Program Description


PTCOG 2020 Online offered three dedicated educational sessions focusing on the core disciplines of clinics, physics, and biology; seven industry symposia from leading particle therapy organizations; six invited keynote sessions; nine oral poster presentation sessions; more than 100 digital posters searchable within an online archive; 15 scientific sessions from 2020 peer reviewed research; virtual exhibitions; and live chat forums.

Objectives for the online conference were to:

  • Address individual needs in compliance with their Continuous Professional Development (CPD) plan.

  • Discuss the latest technological and clinical innovations in particle therapy.

  • Identify educational resources, networks, and other for exchange of knowledge and learning about particle therapy,

  • Discuss current research projects within the Particle Therapy Co-Operative Group (PTCOG) and enhance opportunities for future collaboration between groups of young researchers.

  • Discuss the latest developments about practical clinical application of particle therapy.

  • Describe diagnostics and treatments in the field of particle radiation therapy.

Target Audience

Healthcare professionals who treat cancer patients using radiation therapy/particle therapy and specifically:

  • Radiation Oncologists

  • Medical Physicists

  • Dosimetrists

  • Residents

  • Radiation Therapists

Particle Therapy Cooperative Group Executive and 2020 Online Conference Committees

Members involved in the planning and preparation of the PTCOG 2020 Online Conference:

Jay Flanz, Ph.D., Chairman

Tadashi Kamada, M.D., Vice-Chairman

Marco Durante, Ph.D., Vice-Chairman

Eugen Hug (Past PTCOG Chairman)

Martin Jermann, MSc, Secretary

Anita Mahajan, M.D.

James Metz, M.D., Co-Chairman of Education Subcommittee, Clinics

Niek Schreuder, M.Sc. DABR, Co-Chairman of Education Subcommittee, Physics and Technology

Damien Weber, Chairman of Publication Subcommittee

Anita Mahajan, M.D., Co-Chairman of Scientific Program Subcommittee, Clinics

Michael Story, Ph.D., Co-Chairman of Scientific Program Subcommittee, Biology

Anthony Lomax, Ph.D., Co-Chairman of Scientific Program Subcommittee, Physics

Oral Abstracts

O 01

Automated Patient Specific Knowledge-Based Decision Making for Proton or Photon Treatment Based on NTCP

R. Hytonen1, T.K. Koponen1, R. Vanderstraeten2, C. Smith3, W.F.A.R. Verbakel4, M.R. Vergeer4

1Varian Medical Systems, Palo Alto, CA, United States, Protons, Helsinki, Finland
2Varian Medical Systems, Palo Alto, CA, United States, Protons, Diegem, Belgium
3Varian Medical Systems, Palo Alto, CA, United States, Protons, Milpitas, USA
4Amsterdam University Medical Center - VUmc, Department of Radiation Oncology, Amsterdam, Netherlands

Background and purpose: Selecting patients for proton therapy is laborious, subjective and time consuming. We demonstrate a novel automated solution for creating high-quality knowledge based plans (KBPs) for both proton and photon beams to identify patients for proton treatment based on their normal tissue complication probabilities (NTCP).

Methods and Materials: Two previously validated RapidPlan(PT) model libraries comprising 50 proton and 112 photon plans for head-and-neck cancer (HNC) were used in combination with scripting to create full automatic three-field proton and dual-arc photon KBPs with standard field arrangements for 72 recent HNC patients. The organ-at-risk mean doses were used to calculate the NTCP for each patient, and the patient selection for proton therapy was simulated according to the current Dutch national protocol.

Results: Organ at risk doses for photon/proton predictions were on average 26.3/22.9 Gy and 28.2/21.0 Gy within the final KBP. The total runtime per patient of the automated pipeline was about 20 minutes.

Conclusions: Automated support for decision making using KBP is feasible and fast. The planning solution has potential to speed up the planning and patient-selection process significantly without major compromises to the plan quality.

O 02

Knowledge-based Models for RTOG 1308 Proton Plan Quality Analysis

H. Geng1, Z. Liao2, Q.N. Nguyen2, A.T. Berman3, C. Robinson4, Y. Xiao3

1University of Pennsylvania, Department of Radiation Oncology, Philadelphia, USA
2MD Anderson Cancer Center, Radiation Oncology, Houston, USA
3University of Pennsylvania, Radiation Oncology, Philadelphia, USA
4Washington University - Siteman Cancer Center, Radiation Oncology, St. Louis, USA

Purpose: To build and test a knowledge-based model for proton planning (KBP); And to use the KBP method for RTOG 1308 proton plan quality analysis.

Materials and Method: Fifty proton plans were manually created for 50 cases submitted to RTOG 1308 (Phase III radomized trial comparing overall survival after photon versus proton chemoradiotherapy for inoperable stage II-IIIB NSCLC) using ProBeam clinical Beam data.These plans were used to train a KBP model. The first model was used to re-optimize the 50 plans again. The re-optimized plans were then used to train an updated model. 30 cases not used for the model training were selected from the RTOG 1308 proton cohort for validation. A manual Probeam plan and a model guided Probeam plan were created for these cases. The manual plans were compared with the submitted proton plans (with double scattering beams utilized by participating centers); Comparisons were also made with the model aided plans for model validation.

Results: Re-optimization of the initial 50 plans resulted in both statistically significant better target coverage and organ at risk sparing (Results shown in Table 1). For the 30 QA plans, Probeam plans provide statistically significant sparing of lungs with marginally better target coverage compared with submitted plans. KBP plans show better target coverage and organ at risk sparing consistently, with statistical significance. (Results are shown in Table 2).

Conclusion: The KBP proton model has demonstrated consistent performance for proton plan optimization, showing the potential to be used for proton plan quality analysis in clinical trials.

O 03

GPU-accelerated Treatment Planning System in Intensity Modulated Proton Therapy (IMPT)

J. Shan1, W. Wong1, S. Patel1, M. Fatyga1, M. Bues1, W. Liu1

1Mayo Clinic Arizona, Radiation Oncology, Phoenix, USA

Purpose: GPU suffers from limited memory and opaque programming. To address these problems, we have developed a GPU-accelerated treatment planning system (TPS) and demonstrated its superior performance by generating an IMPT plan with 500 fields.

Method: A clinical plan with two fields for nasopharyngeal carcinoma was re-planned with 500 fields, which is technically impossible with a conventional TPS. We placed 500 fields evenly in the angular space using Fibonacci sequence (Fig.1). In order to minimize memory usage, we employed sparse encoding and dynamic voxel spacing. The influence matrices were calculated using a modified ray-casting analytical dose engine and the dose was optimized using a worst-case robust optimization engine. Both were highly parallelized using CUDA running on a Linux workstation with 2 Xeon E5-2680 CPUs and 4 Tesla K80 GPUs. The GPU-accelerated TPS has been implemented as an EclipseTM plugin, which provides users with real-time interaction during the optimization.

Results: It took 331 seconds to generate such a plan, which has 723,134 spots in total. Of 331 seconds, 61 seconds were used to compute the influence matrices and 100 seconds were used for robust optimization with 16 dose-volume constraints for 13 organs-at-risk. Due to the large number of fields, the resulting plan has dramatically better OAR protection compared to the clinical plan (Fig. 2).

Conclusion: Our GPU-accelerated TPS achieves high efficiency in speed and memory usage. It enables us to pursue advanced studies like beam angle optimization and proton arc treatment, providing an opportunity for significant improvement in IMPT plan quality.

O 04

Developing FLASH Treatment Planning for IBA Proteus System Using the RayStation TPS

E. Traneus1, R. Labarbe2, L. Collignon3

1RaySearch Laboratories AB, Research, Stockholm, Sweden
2IBA, Research, Louvain-La-Neuve, Belgium
3IBA, Product Management, Louvain-La-Neuve, Belgium

Flash irradiation [1] involves irradiating tissue at high dose rate (typically above 40 Gy/s) in order to spare healthy tissue. This high dose rate implies that the TPS must take into account the timing of spot delivery during plan design and evaluation. Treatment plans for FLASH delivery were generated by interfacing a research version of the RayStation TPS (RaySearch Laboratories AB) with a software simulator of the IBA Proteus system (IBA S.A.). The delivery was in PBS mode using a SOBP modulated with a ridge filter and a range compensator for lateral modulation. Using the spot delivery simulator and FLASH dedicated scorers implemented in the RayStation Monte Carlo dose engine, we evaluate and display FLASH metrics derived from voxel wise energy deposition time traces (0.1 ms resolution) such a dose rate vs. volume histograms. The resulting plans are compared to published results [2]. In conclusion, FLASH plans were designed and evaluated in the TPS with dose rate and scanning speed achievable by an IBA Proteus system leading to dose delivery timing compatible with published FLASH conditions. The average dose rate compares favorably to the published literature.

References: [1] Vozeni et al. (2019). Biological Benefits of Ultra-high Dose Rate FLASH Radiotherapy: Sleeping Beauty Awoken FLASH-RT. [2] Montay-Gruel et al. (2017). Irradiation in a flash: Unique sparing of memory in mice after whole brain irradiation with dose rates above 100 Gy/s.

O 05

Potential effects due to variation in linear energy transfer definition and calculation method

E. Smith1, C. Winterhalter1, T. Underwood1, A. Aitkenhead2, J. Richardson2, M. Merchant1, N. Kirkby1, K. Kirkby1, R. Mackay2

1University of Manchester, Division of Cancer Sciences, Manchester, United Kingdom
2The Christie hospital, Medical Physics, Manchester, United Kingdom

Within the PBT community, there is increasing interest in Linear Energy Transfer (LET) calculation for treatment planning. However, the consensus is lacking regarding LET definitions and calculation methods.

In this work, we calculate various LET metrics for clinical PBT plans covering a range of tumour sites. We consider the impact of the scoring method (scoring to water, to medium and to unit density tissue), the use of dose- / track- averaged LET, and the inclusion/exclusion of secondary protons and particles.

Results are shown in Figure 1 for a simple 150 MeV spread-out Bragg peak obtained from Monte Carlo (GATE-RTionV1.0 (GateV8.1/Geant4V10.3.3)) simulations. The LET base settings are dose-averaged LET (LETd) scored to water for primary and secondary protons. When comparing LETd for primary and primary and secondary protons (F1b), there is a difference in the entrance region, up to 20%. Dose-averaged LET (LETd) scoring to medium, to water and to unit density tissue results in differences as large as 10% and 90% for tissue (F1c) and bone (F1d), respectively.

Results are shown in Figure 2 for a representative clinical PBT plan. This shows absorbed dose (F2a), LETd base settings (F2b), LETd primary protons only (F2c), LETd to medium (F2d) and LETt (F2e). For such a clinical case, different methods for LETd lead to differences of up to 68%. LETd and LETt lead to differences greater than 100%.

In conclusion, increasing understanding of how parameters affect LET is an important step towards clinical application and consensus within the PBT community.

O 06

Multi-centric study to harmonize LET-calculations in proton therapy

C. Hahn1, A. Vestergaard2, O. Sokol3, C. Pardi4, A. Leite5, C. Rose6, J. Ödén7,8, L. Grzanka9, A. Lühr1,10

1OncoRay – National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus at Technische Universität Dresden- Helmholtz-Zentrum Dresden - Rossendorf, Dresden, Germany
2Danish Centre for Particle Therapy, Aarhus University Hospital, Aarhus, Denmark
3GSI, Helmholtz Centre for Heavy Ion Research, Darmstadt, Germany
4I-SEE, Internet-Simulation Evaluation Envision, Torino, Italy
5Institut Curie, Centre de protonthérapie d'Orsay, Orsay, France
6Division of Cancer Sciences, School of Medical Sciences and Faculty of Biology Medicine and Health at The University of Manchester, Manchester, United Kingdom
7Stockholm University, Medical Radiation Physics, Stockholm, Sweden
8RaySearch Laboratories AB, RaySearch Laboratories AB, Stockholm, Sweden
9Institute of Nuclear Physics, Polish Academy of Sciences, Krakow, Poland
10Helmholtz-Zentrum Dresden - Rossendorf, Institute of Radiooncology – OncoRay, Dresden, Germany

Emerging clinical evidence for a varying relative biological effectiveness (RBE) in proton therapy poses the urgent need to consider RBE-driving physical parameters such as the linear energy transfer (LET). However, no harmonized concept exists on how to calculate the LET in clinical practice. Therefore, a multi-centric study was set up with the objective to standardize LET-calculations in Europe.

Eight European proton therapy institutions generated non-robust single-field-uniform-dose PBS treatment plans using common strict dose objectives. Multiple treatment field arrangements (single-field SOBP, perpendicular fields, opposing fields) were employed to cover a target cube in a water phantom. The institutions provided their dose and corresponding LET distributions. Here, four different LET-calculation methods (including analytical codes, dedicated LET scripts, Monte Carlo engines) were analysed.

Single-field SOBP ranges (distal R80) and average dose (range: D99 to D1) in the target volume agreed within 2% (Fig.1). In contrast, the corresponding near minimum LET values (LET99), average LET and near maximum LET (LET1) in the target volume differed up to 30%, 19% and 5%, respectively. In the volume 1 cm distal to the target, absolute (relative) LET1 values differed by up to 1.63 keV/μm (17%) in voxels with average physical dose above 40 Gy. Individual institutions included different ions in their LET-calculations partially explaining the observed differences in LET-values and LET-distributions (Fig.2).

Despite comparable dose distributions, substantial LET-differences occurred among the participating institutions. They hamper the consistent analyses of clinical follow-up data and might lead to discrepancies in predicting variable RBE. Therefore, standardized clinical LET-calculations are recommended.

O 07

Proton Beam Therapy for Orbital Rhabdomyosarcomas at the West German Proton Therapy Center Essen (WPE)

D. Geismar1, S. Frisch2, S. Nagaraja2, D. Jazmati2, S. Plaude3, C. Blase4, S. Tippelt5, N. Bechrakis6, B. Timmermann1

1Department of Particle Therapy- University Hospital Essen, West German Proton Therapy Centre Essen WPE- West German Cancer Centre WTZ- German Cancer Consortium DKTK, Essen, Germany
2Department of Particle Therapy- University Hospital Essen, West German Proton Therapy Centre Essen WPE- West German Cancer Centre WTZ, Essen, Germany
3West German Proton Therapy Centre Essen- University Hospital Essen, West German Cancer Center WTZ, Essen, Germany
4Anaesthesia network, Rhine-Ruhr, Bochum, Germany
5Pediatrics III- Pediatric Hematology and Oncology, University Hospital Essen- Germany, Essen, Germany
6Department of Ophtalmology, University Hospital Essen, Essen, Germany

Purpose: Proton beam therapy (PT) is of increasing interest especially in tumors in close proximity of critical structures or in particular sensitive tissues like orbital rhabdomyosarcomas. Early results of PT at WPE are presented.

Methods: Between April 2014 and November 2019, 31 patients (16 male, 15 female, median age 7.2 y (range, 1.8-16.8)) with orbital rhabdomyosarcomas were treated at WPE and were prospectively enrolled in the in-house registry KiProReg. Histopathological subtypes were embryonal (87.1%), alveolar (9.7%), or spindle cell rhabdomyosarcoma (3.2%), respectively. 37.5% had a parameningeal involvement. All patients were treated with chemotherapy (CTx) before radiotherapy and, in 81.2% CTx concomitant to PT was applied. The median PT dose was 50.4 Gy (41.4-55.8Gy) applied with 28 fractions in mean (range, 23-31), using uniform-scanning (38.7%), pencil-beam-scanning (32.2%) or both techniques (28.1%), respectively.

Results: The median follow-up is 20 month (0.4-6,2 y). Twenty-five patients (80.6%) showed disease control. At last follow-up, local recurrence had occurred in five patients, and dissemination in one patient, respectively. No patients had died so far. PT was well-tolerated. New high-grade (CTCAE ≥°3) acute toxicities occurred only for skin (n=2) and blood (n=3). So far, only a low number of new high-grade (CTCAE °3) toxicities was documented. At one year, one patient reported on dry skin and fat atrophy; another one suffered from hematological toxicity. No new grade 4 or grade 5 effects occurred.

Conclusion: Early data suggest good feasibility and local control of PT in orbital rhabdomyosarcomas. However, larger cohorts and long-term follow-up data is desirable to assess long-term outcomes.

O 08

Constraining higher dose volumes for thoracic proton radiotherapy

S. O'reilly1, W. Harris1, C. Cheng1, V. Jain1, B. Burgdorf1, B.K.K. Teo1, L. Dong1, A. Berman1, S. Feigenberg1, W.J. Zou1

1University of Pennsylvania, Radiation Oncology, Philadelphia, USA

Purpose: The dosimetric parameters (V20Gy and mean lung dose) used clinically to reduce radiation pneumonitis (RP) after thoracic radiotherapy for locally-advanced non-small cell lung cancer (LA-NSCLC) have been historically derived from photon therapy. The purpose of this study is to evaluate dosimetric predictors for RP for patients with LA-NSCLC treated with proton therapy and to implement findings into the clinic.

Methods/Materials: One hundred sixty-one (79 photon, 82 proton) LA-NSCLC patients definitively treated with chemoradiotherapy between 2011-2016 were retrospectively identified. 40 (20 photon, 20 proton) patients exhibited grade ≥2 RP after therapy. Statistical analysis (univariate and multivariate logistic regression) using the dose volume histograms for the uninvolved lung was performed. Based on these results, institutional planning protocols were updated to incorporate a three-field technique and a new conformality index was implemented. This new protocol was tested on 10 lung patients.

Results: Multivariate analysis showed V40Gy to best predict RP in our proton therapy cohort. V35Gy-V50Gy were strongly correlated to V40Gy and had the highest AUC compared to other dose levels. A potential dosimetric constraint for RP predictor in proton patients is V40Gy≤23%. The three-field planning technique resulted in statistically significant (p<0.05) reductions to these intermediate dose volumes and mean (as compared to two-field technique), as well as increased conformality by reducing the R40Gy (ratio of 40Gy isodose volume to prescription volume). The V20Gy remained comparable.

Conclusions: The volume receiving higher doses, such as V40Gy, may be a better indicator for RP for LA-NSCLC proton therapy patients and can be reduced through a different planning strategy.

O 10

Evaluation of proton postmastectomy radiation dosimetry using the ESTRO / ACROP breast reconstruction contouring guidelines

N. Depauw1, M. Milligan1, A. Johnson1, R. Jimenez1

1Massachusetts General Hospital, Radiation Oncology, Boston, USA

Proton beam radiation therapy (PBRT) is increasingly utilized for breast cancer patients who undergo mastectomy with reconstruction to ensure adequate target volume and to better spare normal tissue. Recently, ESTRO/ACROP consensus guidelines were updated for postmastectomy radiation therapy (PMRT) with breast reconstruction. In particular, the chest wall (CW) target was redefined to exclude the breast prosthesis and underlying CW tissue. To date, however, the feasibility and dosimetric impact has not been evaluated in patients receiving PBRT.

We performed a treatment planning study of five women with left-sided breast cancer who received proton PMRT after breast reconstruction in 2018 (50.4 GyRBE to the CW and 45 GyRBE to the regional lymphatics). All targets were initially delineated based on conventional RTOG/ESTRO breast atlas guidelines. Comparison treatment plans were then generated using the updated ESTRO/ACROP guidelines. Identical target coverage objectives and normal tissue (OAR) constraints were utilized for both treatment plans, applying the concept of ALARA. A student's t-test was used to compare dosimetric goals between the conventional and updated guideline plans.

Equivalent coverage was achieved for all targets; in particular the mean CW D99/D95 was 47.52/48.65 GyRBE using conventional guidelines vs. 47.60/48.39 GyRBE for the updated guidelines. OARs demonstrated similar to improved dosimetry, with LAD D1 of 6.62 GyRBE vs. 4.60 GyRBE and ipsilateral lung V20 of 5.70 GyRBE vs. 4.70 GyRBE, respectively (all p's = NS).

PBRT for PMRT-reconstruction patients utilizing the updated ESTRO/ACROP consensus guidelines is feasible and provides equivalent target coverage as traditional guidelines, while further sparing cardiopulmonary structures.

O 11

Proton reirradiation of local-regional disease in breast cancer

S. Fattahi1, S.K. Ahmed2, S.S. Park2, I.A. Petersen2, D.A. Shumway2, B.J. Stish2, E.S. Yan2, N.B. Remmes2, R.W. Mutter2, K.S. Corbin2

1Mayo Clinic, Mayo Clinic Alix School of Medicine, Rochester, USA
2Mayo Clinic, Department of Radiation Oncology, Rochester, USA

Purpose: Treatment of local-regional recurrences and/or second primary breast cancers after prior breast RT represents a challenge. This study reports our institutional experience with proton reirradiation.

Methods: Between 2015 and 2019, 21 patients were treated with proton reirradiation. Treatment was delivered with multifield optimized pencil-beam scanning proton therapy (IMPT). Toxicity was assessed with CTCAEv5.0. Freedom from local-regional recurrence and OS were calculated using the Kaplan-Meier Method. Table 1 describes course details.

Results: Median interval between RT courses was 99 months (IQR:43-179mo). With a median follow-up of 22 mo (IQR:5-31mo), freedom from local-regional recurrence among those treated with curative intent was 93.3% at 12 and 24 mo. Overall survival was 100% at 12 mo and 84.4% at 24 mo. Acute (<3 mo) radiation dermatitis: grade 1 (13), grade 2 (5), grade 3 (2). Delayed (>3 mo) grade 2 toxicities: chest wall fibrosis (2), skin infection (2), decreased ROM (1). No delayed grade 3/4 toxicities were present at last follow-up. There was persistence, without worsening, of grade 2 toxicities present before reirradiation: lymphedema (4), decreased ROM (3), chest wall fibrosis (2), brachial plexopathy (1). 2 patients experienced rib fractures. 1 palliative intent patient developed grade 3 skin necrosis at a cumulative dose of 73.55 Gy. Because of diffuse tumor vascular invasion, this may be attributable to tumor or radiation. Table 2 lists dose to organs at risk.

Conclusion: Proton reirradiation is feasible and associated with favorable dosimetry and acute toxicity. Late toxicity and disease control outcomes are promising. Additional follow-up is ongoing.

O 12

Defining dose rate for pencil beam scanning flash radiotherapy

M. Folkerts1, E. Abel1, J. Perez2, V. Krishnamurthi1, C. Ling3

1Varian Medical Systems, Proton Solutions, Palo Alto, USA
2Varian Medical Systems, Proton Solutions, Geneva, Switzerland
3Varian Medical Systems, Advanced Clinical Research, Palo Alto, USA

Purpose: Ultra-high dose rates for electron FLASH (eFLASH) experiments have been reported in the literature to spare healthy tissue while maintaining tumor control. Such deliveries consist of multiple pulses delivering dose to the entire field simultaneously. This is fundamentally different from modern pencil beam scanning (PBS) proton therapy delivery (Figure 1), presenting a challenge when reporting dose rates for PBS fields. The purpose of this work is to develop a method of calculating dose rates for PBS fields that best corresponds to the average eFLASH dose rates reported in the literature.

Methods: Consider the total dose delivered to a point during an entire PBS field irradiation. The irradiation time experienced at a point is calculated as the duration of time starting when the dose delivered to the point rises above a threshold value, and ending when the total dose delivered to that point, minus the threshold value, is reached. This may include elapsed wall time while dose is being scanned elsewhere. PBS dose rate is simply the quotient formed form the total dose delivered to a point, and the irradiation time experienced at that point. This dose rate value can be calculated for points in a region of interest, and statistics reported accordingly.

Results: The PBS dose rate definition can be used to design equivalent pulsed and scanned fields given a sample point (Figure 2).

Conclusion: A PBS dose rate calculation method has been developed to improve correspondence between different modalities of FLASH irradiation and help researchers design PBS FLASH experiments.

O 13

Can proton spot scanning meet the conditions for FLASH?

B. Rothwell1, N. Kirkby1, M. Merchant1, A. Chadwick1, M. Lowe1,2, R. Mackay1,2, K. Kirkby1

1University of Manchester, Division of Cancer Sciences, Manchester, United Kingdom
2The Christie NHS Foundation Trust, Christie Medical Physics and Engineering, Manchester, United Kingdom

The recent resurgence of ultra-high-dose-rate, or ‘FLASH', radiotherapy has attracted widespread interest over the past few years, with promise of improved normal-tissue protection compared to conventional irradiation and no compromise on tumour control. The transient hypoxic state induced by depletion of oxygen at high dose rates provides a well-accepted explanation. Combining this apparent sparing effect with the often-superior dose conformality of proton spot-scanning treatments is an attractive area of research. However, there exists a lack of understanding of the oxygen depletion effects and time characteristics already taking place during conventional spot scanning, where dose rates at locations within the patient from spot delivery can far exceed the supposed FLASH threshold.

Previous work has used cellular automata techniques to solve a model of the complex processes of oxygen diffusion and reaction within tissue, and how these are affected by ultra-high dose rate irradiation. This work has been extended to incorporate clinical parameters to model the delivery of proton spot-scanning at high dose rates, and its spatial and temporal impact on oxygen levels. Results so far suggest that the magnitude of oxygen depletion during a typical proton spot scan can be significant (Fig 1), and strongly depends on the spot-scanning pattern used, as well as a number of other biological and delivery parameters.

Further work will continue to investigate any potential normal-tissue sparing caused by this, and possible methods of harnessing this capability to improve patient outcome. This will further tighten the link between FLASH research and its potential clinical implementation.

O 14

Practical methods to deliver FLASH irradiation with a Proton Therapy machine – from beamline/scattering design to in-vivo experiments

Q. Zhang1, E. Cascio1, Q. Yang1, L. Gerweck1, P. Huang1, B. Gottschalk2, J. Flanz1, J. Schuemann1

1Massachusetts General Hospital and Harvard Medical School, Radiation Oncology, Boston, USA
2Harvard University, Laboratory for Particle Physics and Cosmology, Cambridge, USA

Dr. Schuemann transitioned from particle physics to medical physics in 2010. He is one of the core developers of the TOPAS (Tool for Particle Simulations) Monte Carlo toolkit, a user friendly framework for MC simulations specifically designed for the medical physics community. In 2014, he formed a collaboration to develop TOPAS-nBio, a nanometer scale extension of TOPAS to investigate the underlying mechanisms of cellular responses to radiation damage. To investigate the mechanisms of FLASH therapy, Dr. Schuemann started the proton FLASH project at MGH. \nDr. Schuemann is an Associate Professor at the Massachusetts General Hospital and Harvard Medical School and the head of the Multi-scale Monte Carlo Modeling Lab.

FLASH therapy (irradiations >100 Gy/s) promises to selectively spare healthy tissue. However, most experiments so far have been performed using electron irradiations. To achieve such high dose rates with protons, we designed a new double scattering system for our experimental beamline to deliver a 228 MeV proton beam with a field size of 16 mm diameter. The cyclotron to target efficiency is upwards of 35%. The scattering system and collimation provides a focused beam allowing for the treatment of two mice in a row, reducing the necessary machine time. The scattering system was designed using analytical tools and confirmed with Monte Carlo (TOPAS) simulations (Figure 1a,b). The dose profiles were validated with multiple dosimetric approaches, including thin gap ion chambers, thimble chambers, radiochromic film and Faraday cups. A new scanning system was designed based on a 3D printer to obtain profile measurements of the small fields (see figure 1c). Delivered dose rates of up to 130 Gy/s were achieved.

Mice are positioned perpendicular to the beam and the protons fully pass through the mice (laterally, figure 2a), providing the option of dosimetric assessment for each mouse pair distal to the mice. Here, we present the steps from initial preparations to in-vivo proton FLASH experiments. Results of the first ∼250 mice will be presented, assessing the LD50 for partial gastrointestinal tract irradiations and the importance of lymphocytes. Preliminary results suggest a FLASH normal tissue sparing effect in the order of 20-30% (figure 2b-d).

O 15

Implementation of an online oxygen meter for studying oxygen removal during FLASH irradiation

J. Jansen1,2, J. Knoll1,2, W. Tinganelli3, O. Sokol3, R. Hanley1,2, L. Hehn1,2, G. Echner4, S. Brons5, F. Pagliari1, J. Seco1,2

1German Cancer Research Center DKFZ, Biomedical Physics in Radiation Oncology E041, Heidelberg, Germany
2Ruprecht-Karls-University, Physics and Astronomy, Heidelberg, Germany
3GSI Helmholtzzentrum für Schwerionenforschung GmbH, Biophysics, Darmstadt, Germany
4German Cancer Research Center DKFZ, Medical Engineering, Heidelberg, Germany
5University Clinics Heidelberg, HIT Heidelberg Ion Beam Therapy Center, Heidelberg, Germany

Alternatively to the commonly used conventional radiotherapies (RT) using X-Rays or protons and heavier ions, the treatment with high dose rates, known as FLASH, is of higher interest since it provides a tumor control probability in the range of conventional radiotherapy and a highly improved sparing of the healthy tissue[1]. This phenomenon is not fully understood yet but one common explanation is the oxygen depletion theory due to which tissue enters a hypoxic state and becomes radioresistant. The presented study aims to characterize this effect by measuring the dependence of oxygen solubilized in water from irradiation with different dose rates and different radiation types. Hereby, the initial amount of oxygen in water is set to a physiological range of 0.5-5%. During radiation, the amount of oxygen is measured using an optical sensor in predefined water phantoms of 500μl – 60 ml volumes. The dependence of oxygen is measured as a function of the irradiated volume, the initial concentration of oxygen, the dose rate, the total dose and the particle type. First experiments with an irradiated phantom coupled to an unirradiated part serving as oxygen supply have shown that the percentage of depleted oxygen is independent on the dose rate but the total irradiation time (see Fig 1) and total dose (see Tab 1) needed to reach a saturation point is highly dependend on the dose rate.

References: [1] Favaudon et al (2014) Ultrahigh dose-rate FLASH irradiation increases the differential response between normal and tumor tissue in mice. Sci Transl Med.; 6(245):245ra93.

O 16

Oxygen depletion and ROS production in FLASH radiotherapy with electrons and ions modelled at the chemical track structure level

D. Boscolo1, E. Scifoni2, M. Krämer1, M. Durante1, U. Weber1, C. Schuy1, M. Fuss1

1GSI Helmholtz Centre for Heavy Ion Research, Biophysics, Darmstadt, Germany
2Trento Institute for Fundamental Physics and Applications, Trento Institute for Fundamental Physics and Applications, Trento, Italy

The track structure code TRAX-CHEM allows us to investigate the impact of oxygenation on intra-irradiation pulse oxygen depletion and the related production of superoxide and perhydroxyl radicals in water targets under different irradiation conditions. Based on a mechanistic model for radiolytic yields at 1 μs, we examine the difference between FLASH and conventional irradiation with electrons, protons, and C ions.

For low LET radiation, a significant oxygen consumption in the target at realistic doses is observed only for very hypoxic starting conditions. When accounting for a proportional contribution by carbon-centered radicals to mimic a more realistic chemical environment, the effect is still not compatible with in vivo experimental evidence. According to our model, in physioxic conditions 19.3% of the oxygen is depleted after a pre-clinical experimental dose of 30 Gy, leading to a 1.7% decreased ROS production but showing no effect on OER-weighted dose. In hypoxia representative of radioresistant tumors (<1% pO2), >20% less ROS production and >6% less DOER are predicted, contrarily to the supposed NT sparing.

For high LET radiation expected in target, radical yield and oxygen depletion are considerably reduced due to early intra-track recombination of the chemical species, so that analoguos effects appear at higher doses.

Our results rule out pure oxygen depletion as the main responsible of better normal tissue radioresistence (“FLASH effect”) observed in ultra-high dose rate radiotherapy. However, subsequent stages of the biochemical radiation damage could be affected by the high instantaneous ROS concentrations produced in FLASH, or could entail extra oxygen consumption.

O 17

Feasibility of MR-only dose calculations using deep learning-based synthetic CTs for proton and photon therapy in abdominal pediatric tumors

F. Guerreiro1, M.C. Florkow2, E. Seravalli1, G.O. Janssens3,4, J.H. Maduro4,5, M. van Stralen2, P.R. Seevinck2, A.C. Knopf5, B.W. Raaymakers1

1University Medical Center Utrecht, Department of Radiotherapy, Utrecht, Netherlands
2University Medical Center Utrecht, Image Science Institute, Utrecht, Netherlands
3University Medical Center Utrecht, Department of Radiation Oncology, Utrecht, Netherlands
4Princess Máxima Center, Pediatric Oncology, Utrecht, Netherlands
5University Medical Center Groningen- University of Groningen, Department of Radiation Oncology, Groningen, Netherlands

Objective:To evaluate the dosimetric accuracy of deep learning-based synthetic computed tomography (sCT) images for magnetic resonance (MR)-only proton and photon radiotherapy in pediatric abdominal tumors.

Material and Methods:Planning-CT, T1- and T2-weighted MR images including the whole abdomen of 46 patients (age: 2-7 years) with Wilms' tumor or neuroblastoma were included. Planning-CT and MR acquisitions were separated by less than 45min. MR images were non-rigidly registered to the planning-CT and resampled to the CT resolution. sCTs were generated for each patient using a 3D patch-based model derived from a U-net and taking both MR images as input. The sCT image quality was evaluated by computing the mean absolute error (MAE) between the planning-CT and sCT Hounsfield units (HUs). Intensity-modulated pencil beam scanning (PBS) and volumetric-modulated arc therapy (VMAT) treatment plans (prescribed dose: 10.8-36.0 Gy) were robustly optimized against a patient set-up uncertainty (5mm) using the planning-CT and the internal target volume (ITV). Planning-CT and sCT dose re-calculations were compared using dose-volume histogram (DVH) parameters and 3D gamma analysis (2mm, 2%).

Results:The calculated MAE was 58±12 HU for the body and 185±48 HU for the bone. Average DVH differences between the planning-CT and sCT dose distributions for the ITV/organs-at-risk were ≤|0.2/0.8|% for PBS and ≤|0.3/0.2|% for VMAT (Figure1). Average gamma pass-rates were ≥91.0% for PBS and ≥98.6% for VMAT (Table1).

Conclusion:With the investigated sCT generation method, results show that accurate MR-only dose calculations were feasible for both proton and photon radiotherapy even in a cohort containing children with different ages and tumor locations, sizes and shapes.

O 18

Can the Ocular Option on the pencil-beam-scanning ProBeam nozzle deliver ocular-melanoma treatments of the same quality as a dedicated eyeline?

R. Slopsema1, S. Flampouri1, A. Dhabaan1

1Emory University, Emory Proton Therapy Center, Atlanta, USA

Purpose: To compare the organs-at-risk doses when treating ocular melanoma with an aperture-based, multi-beam technique implemented on the pencil-beam-scanning nozzle of a ProBeam system, or with the traditional, single-field technique of a dedicated eyeline.

Method: The ProBeam Ocular Option (OO) consists of a 3-cm range shifter and aperture holder mounted on a pencil-beam-scanning gantry. RayStation (RS8B) was commissioned allowing 3D spot-map optimization and dose calculation for the OO. For comparison a machine was commissioned in RayStation mimicking the dose distribution delivered by a dedicated eyeline (IBA). A dosimetric study was performed for 10 ocular-melanoma patients. Eye structures were delineated based on a geometrical eye model registered to the patient CT. The eye gazing angle was optimized to limit dose to the OARs. The OO plans consisted of three beams, typically one anterior and two temporal obliques. The aperture of each beam was fit to the beam's-eye-view projection of the target with a 3 mm margin. Inverse planning was used to limit dose to the OARs while maintaining the same coverage as the eyeline plan.

Results: Figure 1 shows how the eyeline fall-off can better limit the dose to the target-abutting optic disc, but the OO allows for better sparing of the anterior segment. Several OAR metrics for all ten patients are summarized in Figure 2.

Conclusion: In cases where critical structures abut the target a dedicated eyeline is superior, but the OO provides more flexibility in prioritizing clinical goals often allowing for better sparing of the anterior critical structures.

O 19

Ultra high-dose distributions (Flash) for lung SBRT using coplanar and non-coplanar scanning proton transmission beams

W. Verbakel1, P. van Marlen1, M. Folkerts2, E. Abel2, B. Slotman1, M. Dahele1

1Amsterdam UMC, radiation oncology, Amsterdam, Netherlands
2Varian Medical Systems, Varian Proton Solutions, Palo Alto, USA

Wilko Verbakel has been medical physicist since 2004. In 2003 he completed his PhD on Boron Neutron Capture Therapy. His research fields are improvement of treatment planning, automation of radiotherapy processes, tumor tracking, deep learning, FLASH treatment planning and lung and spine SBRT. In addition to research, he is working full time in the clinic of Amsterdam UMC.

Background: Pre-clinical research into “FLASH”-radiotherapy (e.g. dose-rate≥40Gy/s) suggests reduced side-effects compared to conventional irradiation, while maintaining tumor control. We propose FLASH using spot scanning proton transmission beams. We investigated dose-rate distributions and delivery times for planar and non-coplanar proton FLASH-plans using stereotactic lung irradiation as the paradigm.

Methods: Stereotactic lung radiotherapy FLASH-plans (3x18Gy), using 10 coplanar or non-coplanar scanning proton transmission beams of 244MeV (Bragg peak behind the body), were made for 3 patients with the intention to achieve OAR dose≤clinical VMAT-plans. For the non-coplanar plans, beams were chosen to avoid heart, spinal cord and contralateral lung. Coplanar plans had 10 equidistant beams. The spot peak dose rate of the FLASH-beams was 360Gy/s, which is currently achievable. We evaluated the OAR doses of both types FLASH-plans compared to VMAT-plans, dose-rate distribution within a beam and irradiation times. A FLASH dose-rate threshold of 40Gy/s was used.

Results: All FLASH-plans had ITV and PTV doses comparable to VMAT-plans, but achieved lower doses for most OARs. Non-coplanar FLASH-plans had no heart, spinal cord and contralateral lung dose, but coplanar FLASH-plans achieved the lowest total lung dose. The FLASH-distributions per beam of both FLASH-plan types are comparable (Fig1; first bin is for low dose (<0.15Gy) in penumbra). Spot irradiation times were mostly 2-4ms and most beams had irradiation times of 0.04-0.11s (total beams<1.1s).

Conclusion: This study shows that proton transmission plans achieve improved or equal plan OAR sparing to VMAT plans. Most high dose receiving volumes of both non-coplanar and coplanar plans achieved >80% dose rates>40Gy/s.

O 21

Robustness analysis including varying relative biological effectiveness for carbon ion therapy

M. Wolf1, C. Graeff1

1GSI Helmholtz Centre for Heavy Ion Research, Biophysics, Darmstadt, Germany

Purpose: In heavy ion treatment planning, RBE is a widely used concept to account for variable radio-sensitivities of tumors and normal tissues. However, RBE modeling has intrinsic uncertainties, which subsequently can lead to dose deterioration. In order to assess the effect of RBE uncertainties, we introduce a robustness analysis (RA) which considers varying LEM-based RBE scenarios.

Methods: The RA is based on 21 scenarios generated by combining nominal, increased and decreased ranges with nominal position and 6 shifted positions. Concerning the RBE scenarios, α/β ratios of 2, 6 (nominal) and 10 are used for the CTVs and α/β ratio of 2 for the OARs. As either α or β can be varied to change the α/β-ratio, there are two RBE tables for α/β ratios of 2 and 10, which yields 210 dose scenarios in total. The RA is performed on a lung cancer patient with a tumor close to the heart (<5mm), comparing three IMPT optimizations: conventional (3mm PTV margins), robust (9 basic scenarios), and robust RBE (9 basic + 5 RBE scenarios).

Results: Compared to the conventional plan, there is a considerable increase in D95% coverage for both robust plans (Tab.1). The conventional plan also shows higher variation of the DVHs (Fig.1). The LET and RBE distribution differed between conventional and both types robust optimized plans.

Conclusion: The new RA reveals decreased dose uncertainty for robust optimization. Robustness against RBE variation was improved regardless of explicitly considering RBE in the optimization.

O 22

A leaf positioning algorithm to optimize conformality and speed with the adaptive aperture proton MLC

J. Wang1, I. Smirnova2, A. Reid2, A. Lajovic3, J. Cooley4

1Mevion Medical Systems, Advanced Development, Littleton- MA, USA
2Mevion Medical Systems, Software Engineering, Littleton, USA
3Cosylab, Software, Ljubljana, Slovenia
4Mevion Medical Systems, Advanced Development, Littleton, USA

The Adaptive Aperture proton MLC is used on the Mevion Hyperscan PBS system to deliver sharp gradients at the edges of treatment fields. The device, which has fourteen leaves that can be translated around the treatment field, is designed to reconfigure its leaves on a spot-by-spot basis throughout the treatment. The leaf positions are not prescribed by the treatment planning optimizer. Rather, the treatment planning system (TPS) optimizes spotmaps by approximating the collimation as continuous aperture curves, as if the treatment were to be delivered with static, energy -dependent apertures. After optimization, an algorithm uses the spotmaps and aperture curves to calculate the placement of the leaves that will reproduce the dose calculated by the optimizer.

We will describe the details of this leaf positioning algorithm (LPA) and how it achieves the optimum set of leaf positions: effective conformality while minimizing delivery time. For each spot in the treatment plan, the LPA generates a set of candidate leaf positions, then scores those based on a “cost” factor that penalizes that candidate for leaked dose outside and blocked dose inside. Every spot is then assigned to a leaf position, with spots grouped so that the total number of leaf positions is minimized at an acceptable total cost.

Through the analysis of several clinical treatment plans, we will show that adjusting the relative importance of the geometric cost factor versus that of the number of leaf positions can have a dramatic influence on the total treatment time with marginal impact on the dose distributions.

O 23

Proton ocular centers with dedicated fixed low-energy beams: Key clinical concepts for new centers

K. Mishra1, A. Kacperek2, I. Daftari1, A. Afshar3, J. Scholey1, B. Damato4, D. Char5, J. Quivey1

1University of California San Francisco, Radiation Oncology, San Francisco, USA
2Clatterbridge Cancer Centre, Proton Beam Radiation, Wirral, United Kingdom
3University of California San Francisco, Ophthalmology, San Francisco, USA
4Moorfields Eye Hospital/University of Oxford, Ophthalmology, London, United Kingdom
5The Tumori Foundation, Ocular Oncology, San Francisco, USA

Purpose: Proton beam treatment for uveal melanoma (UM) is well-established internationally with dedicated beamlines. With the emergence of universal, high-energy, non-fixed beamlines, we conducted a bi-institutional study of key practices to minimize side effects.

Methods: The dedicated, fixed, low-energy, ocular lines at the University of California-San Francisco (UCSF), USA, and the Clatterbridge Cancer Centre (CCC), UK, produce 67.5 and 60.0 MeV proton beams, respectively. 5927 ocular patients have been treated, 93% with UM. Standard UM dose is 56 GyE at UCSF and 57.2 GyE at CCC, in 4 fractions.

Results: Critical beam characteristics at UCSF/CCC include a very sharp lateral and distal dose fall-off, high dose homogeneity, excellent range precision and short treatment time (∼0.5-2 minutes). Doses to critical structures are independent predictors of vision, neovascular glaucoma, and other clinical outcomes (e.g. 28 GyE to macula [P<.0001], optic nerve [P<0.0004], lens [P<.0001], and ciliary body [P<.0001]). A rational tumor/critical structure dose evaluation is used to optimize treatment parameters, i.e. lateral margin, distal range, gaze angle, and aperture shape. Acommon UM dosing regimen is 60 GyE in 4 fractions. Unrandomized early retrospective datashowed lower local tumor control with 48 GyE (P=0.02). Retina, lacrimal gland, cornea, tear ducts, lids, and limbal stem cells are systematically evaluated to minimize side-effects. Eyelid toxicity is minimized by retraction techniques or treatment through closed lids.

Conclusions: The UCSF and CCC dedicated, low-energy, fixed eyelines, provide practical clinical concepts for consideration by new centers to optimize high-energy universal non-fixed beamline designs, particularly to reduce normal tissue toxicity.

O 24

Prospective assessment of optic neuropathy in paraoptic CNS/sinonasal tumors of patients treated with pencil beam scanning proton therapy

J. Thariat1, J. salleron2, P.A. marty3, M. lecornu1, J. bouter1, D. stefan1, A. lecoeur1, J. balosso1, J.L. habrand1, J.C. quintyns3

1Centre baclesse / ARCHADE, Radiation Oncology, caen, France
2institut de lorraine, statistics, nancy, France
3CHU, ophthalmology, caen, France

Introduction: Patients with paraoptic tumors have a ≈7%-risk of radiation-induced optic neuropathy (RION) after PT. Systematic/standardized paraclinical exams are little performed prospectively at baseline/follow-up. We aimed to detect visual dysfunction more systematically with thorough/sensitive exams to better address confounding factors of RION and initiate earlier treatments when possible.

Material and methods: All patients underwent full optic investigations (visual field VF, visual evoked potentials, papillary-optical coherence tomography (OCT), fundus view, angio-OCT) before and 6,12, 18 months after PT.

Results : 60% of patients referred to PT at our department in our first-year (2018) had paraoptic tumors (≤1cm, N=43: pituitary=14%, meningioma=49%, head/neck=14%, other intracranial=23% including 7% reirradiation) vs 5% at our department in 2017. There were 37% male, 72% PS0, 30% with cardiovascular comorbidities. They had hormonal substitution or epileptic treatments in 21% and 12% before PT. On clinical examination, 47% had abnormal baseline neurologic examination, 9% reported clinical-VF defects compared to 46.5% paraclinical-VF defects on standardized measurements (p<0.001, Kappa= 0.21[0.02 ;0.40], i.e. very low agreement. Median total dose was 54 Gy_RBE. At 6 months, no patient had reported new VF defect while paraclinical-VF defects occurred in 60%.

Conclusion: Referral biases should be addressed when reporting non IMRT/SBRT vs PT randomized studies. Paraclinical VF is more sensitive than clinical examination. Full baseline/6-month results and pVF and voxel-based NTCP will be provided for the first 43 patients at PTCOG. Neuronal or vascular origin of RION will be discussed using angio-OCT.

O 25

Randomized phase 2 study of moderately vs standard hypofractionated Proton therapy for large uveal melanomas

J. Thariat1, T. mathis2, C. maschi3, S. baillif3, J.P. caujolle3, J. herault4

1Centre baclesse / ARCHADE, Radiation Oncology, caen, France
2CHU, ophthalmology, lyon, France
3CHU, ophthalmology, nice, France
4centre lacassagne, radation oncology, nice, France

Introduction: Proton therapy is advocated for small to medium size uveal melanomas. Patients with large uveal melanomas are referred for enucleation due to the risk of severe toxicities. Advances in supportive care in recent years have contributed to improved management of toxicities. RBE variations occur in low fraction doses but are not significant in the 4 to 8 Gy range. Large fractions doses delivered to large volumes are associated with inflammatory effects and cytokinic syndromes and might be better managed with smaller fraction doses. In patients with large melanomas who refused upfront enucleation, we evaluated moderately vs standard hypofractionated proton therapy for large uveal melanomas in a randomized phase 2 study.

Material and methods: All patients refusing enucleation for tumors >18mm in diameter or >12mm in thickness (or diameter >15mm and thickness >10mm), or tumor/eyeball volume > 40% were randomized. Patients and referring ophthalmologists were blind to allocated treatment arm. Patients either received 4 fractions and 4 fake ones (standard) or 8 fractions (experimental). Main objective was local control rate without severe toxicity at two years as measured on ultrasound and/or angiography following completion of proton therapy. Patients lost to follow up were considered as local failures. Complications were defined clinically or by any therapeutic interventions (including salvage enucleation) other than topical applications.

Results: The first patient was included in November 2015 and last in April 2017. All treatments were completed.

Conclusion: Local control, complications rates and profiles as well as survival rates will be provided with > 2-year follow-up.

O 26

High preservation of taste and smell after proton beam therapy for nasopharyngeal carcinoma: a prospective longitudinal study

J. Slater1, L. Liu2, D. Sui2, E. Weyman2, A. Chan1

1Massachusetts General Hospital- Harvard Medical School, Radiation Oncology, Boston, USA
2Massachusetts General Hospital, Radiation Oncology, Boston, USA

Purpose/Objectives: To assess chemosensory outcomes and dosimetric predictors of chemosensory dysfunction in nasopharyngeal cancer (NPC) patients treated with proton beam (PBT).

Methods: Twenty-five patients with stage IIB-IVB NPC were enrolled on a prospective, phase II, NCI-funded study. ChemoSensory Questionnaire (CSQ)—a validated tool for head and neck cancer patients—was performed before PBT and at 1.5, 3, 6, 12, and 24 months following completion of PBT. Comprehensive contouring of organs at risk (OARs) was performed.

Results: Twenty-two patients had complete baseline and follow-up CSQ data. Eleven patients had improved or stable CSQ taste scores at 24 months compared to baseline, 5 had a decrease <20%, 6 patients had decrease >20%. Twelve patients had improved or stable CSQ smell scores at 24 months compared to baseline, 2 had decrease <20%, 7 patients had decrease >20%

Combined parotid V5Gy was correlated with severity of decrease in CSQ taste at 24 months (Pearson's coefficient -0.44, p=0.04). The average mean dose to tip of the tongue 1.05 Gy (range: 0.01-4.13 Gy), this did not have any correlation with taste scores. Combined parotid mean dose of >26 Gy was associated with a decrement in taste at 24 months (p=0.009); combined parotid V5Gy of ≥90% had a trend toward association of decrement in taste at 24 months (p=0.063). Change in CSQ smell score was not associated with any OAR dosimetric parameter.

Conclusion: Proton beam therapy results in high rates of chemosensory preservation in NPC patients. Combined parotid V5Gy of >90% was predictive of the severity of late taste decrement.

O 28

Real-time Monte Carlo-based dose calculation algorithm for proton therapy

P. Ibáñez1,2, V. Valladolid1, A. Villa-Abaunza1, P. Galve1, F. Arias1, A. Espinosa1, S. España1,2, D. Sánchez-Parcerisa1,2, L.M. Fraile1,2, J.M. Udías1,2

1University Complutense of Madrid, Grupo de Física Nuclear and IPARCOS, Madrid, Spain
2Hospital Clínico San Carlos IdISSC, Instituto de Investigación Sanitaria, Madrid, Spain

Monte Carlo (MC) treatment planning (MCTP) is increasingly demanded for proton verification, especially when heterogeneous tissues are present as in lung cancer [1]. However, long computation time limits MCTP for daily clinical applications. We developed a GPU-based Hybrid Monte Carlo for protons (GHMCp) incorporating all the physics and particle tracking of realistic MC simulations. GHMCp computes dose distributions, proton fluencies, phase spaces and nuclear activations from proton beams within seconds. It incorporates a precalculated database with interactions of protons, photons and electrons in different materials. The database is easily calculated from any MC code, thus predictions from different MC packages can be obtained. Other particles, such as neutrons or alpha particles, can also be tracked.

Benchmark MC simulations with PENH-NUCL [2,3] and TOPAS [4] were used to test GHMCp dose predictions and activation maps both in homogeneous and heterogeneous media. Figure 1 shows dose distribution in lung from PENH-NUCL and GHMCp. We obtained a 30,000 acceleration factor for the GHMCp compared to MC codes, for similar statistical uncertainty, which makes it possible real time calculations and inverse dose planning.

References: [1] D. Maes et al (2018). Transl. Lung Cancer Res. 2018;7(2):114-121. [2] F. Salvat (2013) Nuc. Instr. Meth. Phys. Res. B, 316, 144-59. [3] E. Sterpin et al. (2013). Med. Phys, 40(11), 111705. [4] J. Perl, et al. (2012). Med. Phys., 39(11), 6818-37.

O 29

Integrating an open source Monte Carlo code, “MCsquare”, for clinical use in intensity-modulated proton therapy

W. Deng1, J. Younkin1, K. Souris2, S. Huang3, K. Augustine1, M. Fatyga1, X. Ding1, M. Cohilis2, M. Bues1, J. Shan1, J. Stoker1, L. Lin4, J. Shen1, W. Liu1

1Mayo Clinic Arizona, Radiation Oncology, Phoenix, USA
2Universite catholique de Louvain, Center for Molecular Imaging and Experimental Radiotherapy, Louvain-la-Neuve, Belgium
3Memorial Sloan Kettering Cancer Center, Medical Physics, Philladelphia, USA
4Emory University, Radiation Oncology, Atlanta, USA

Purpose: To commission an open source Monte Carlo (MC) dose engine, “MCsquare”, for a synchrotron-based proton machine and integrate it into our web-based second check platform for clinical use in intensity-modulated proton therapy (IMPT).

Methods: We commissioned MCsquare using a double Gaussian beam model based on in-air lateral profiles, integrated depth dose of 97 beam energies, and measurements of various spread-out Bragg peaks (SOBPs). Then we integrated MCsquare into our web-based second check platform. We validated the commissioned MCsquare based on twelve different patients and compared the results with another well-benchmarked MC dose engine (gMC). We further improved the MCsquare efficiency by employing the CT resampling approach and expanded its functionality by adding a linear energy transfer (LET)-related model-dependent biological dose calculation.

Results: Differences between MCsquare calculations and SOBP measurements were less than 2.5% (less than 1.5% for ∼85% of measurements) in water. The average 3D gamma analysis (2%/2mm) passing rates comparing MCsquare and gMC calculations in the twelve patient geometries were 98.0±1.0% (Table1). The computation time to calculate one IMPT plan in patients' geometries using an inexpensive CPU workstation (Intel Xeon E5-2680 2.50GHz) was 2.3±1.8 minutes after the CT resampling (Fig.2) was adopted.

Conclusion: MCsquare was successfully commissioned for a synchrotron-based proton machine and integrated into our web-based second check platform. After adopting CT resampling and implementing LET model-dependent biological dose calculation, MCsquare was successfully integrated into clinical workflow for a busy proton clinic and was sufficiently efficient to achieve Monte Carlo-based and LET-guided robust optimization in IMPT.

O 30

Dosimetric impact of rotational set up errors in IMPT and feasibility of six dimensional (6D) robust optimization

D.N. Manthala Padannayil1, D. Sharma1, K. Patro1, R. Jalali2, M. Sawant1, S. Kaushik1

1Apollo proton cancer centre, Medical physics, Chennai, India
2Apollo proton cancer centre, Radiation Oncology, Chennai, India

Purpose: To assess dose delivery errors due to rotational uncertainties in Intensity modulated proton therapy (IMPT) plans and investigate on the proposal of 6D robust optimization to mitigate all uncertainties.

Material and Method: The planning CT and structures of a head and neck anthropomorphic phantom were rotated for ±1° and ±2° separately in three rotational axes {Roll(R), Pitch(P) and Yaw(Y)} using an in-house MatLab analytical program. Robustly optimized nominal IMPT (IMPT-N) plans, which incorporated three translational error and range uncertainty, were projected on to these rotated image datasets to access variation in dosimetric parameters. Subsequently, 6D robust optimization IMPT (6D-IMPT) plans were carried out using the rotational CT datasets. After thorough validation, same investigation was extended to eight previously treated chordoma/chondrosarcoma patients. For each IMPT-N plan, 12 rotational error simulated plans (P+1,P-1,P+2,P-2,R+1,R-1,R+2,R-2,Y+1,Y-1,Y+2,Y-2) were carried out and robustness of each error plans were assessed by voxel worst (VW) increase in D1% for proximal OARs and VW decrease in D95% for targets along with translational (±3mm) and range uncertainties (±3.5%).

Results: The phantom study demonstrate accuracy of our analytical programme for image rotation and feasibility of 6D robust optimization. The presence of patient rotation did not show any appreciable changes in target coverage. However variations up to 20% were observed in D1% depending on degree of rotation and position of OARs (Figure1a-d). 6D-IMPT plans reduces maximum D1% VW case in all OARs (Figure 2).

Conclusion: The proposed 6D robust optimized IMPT plans showed better robustness for both translational and rotational uncertainties.

O 31

Accelerated minimax optimization for intensity-modulated proton therapy

G. Buti1, K. Souris1, A.M. Barragan Montero1, J.A. Lee1, E. Sterpin2

1UC Louvain, Institute of Experimental and Clinical Research, Brussels, Belgium
2KU Leuven, Departement of Oncology, Leuven, Belgium

Purpose: To propose an approximate minimax optimization algorithm, called dynamic minimax, that accelerates IMPT robust optimization.

Material/Methods: Instead of simulating simultaneously the usual 63 scenarios (= 7 setup x 3 range x 3 breathing phases), the proposed dynamic minimax algorithm considers only 5 of the 63 scenarios, at each iteration. These 5 scenarios are updated over time by randomly sampling new scenarios according to a hidden probability acceptance function P. This probability set P is updated by (1) incrementing the value of the worst-case scenario, (2) incrementing the values for yet unconsidered scenarios (giving these the possibility to contribute later on in the optimization) and (3) a normalization step which reduces the values of considered scenarios that are currently not the worst-case.

The proposed algorithm was implemented in the open-source robust optimizer MIROpt and tested for three 4D IMPT lung tumor treatments (prescriptions of 60 Gy). Treatment plans were evaluated by performing robustness tests (simulating range errors, systematic setup errors and breathing motion) using the open-source Monte-Carlo dose engine MCsquare.

Results: An average 83% reduction of optimization time is achieved by the proposed optimization algorithm. In terms of target coverage (see figure for a representative example), the dynamic minimax optimization improves, on average, the worst-case D95 with 0.3 Gy. Moreover, the difference in normal tissue sparing is also comparable (the difference in lung Dmean is only 0.4 Gy).

Conclusion: An approximate worst-case robust optimization algorithm is proposed that achieves an optimization time gain of 83%, without compromising target coverage or normal tissue sparing.

O 32

LETd robustness evaluation in head and neck and breast IMPT

D. Wagenaar1, J. Langendijk1, E. Korevaar1, G. Katgert1, A. Crijns1, A. Knopf1, C. Onyia1, N. Fernandes1, S. Both1

1Department of Radiation Oncology- University Medical Center Groningen- University of Groningen, Radiation Oncology, Groningen, Netherlands

Purpose: The relative biological effectiveness (RBE) of proton therapy varies with the dose-weighted linear energy transfer (LETd) distribution. However, LETd heterogeneity may make it less robust to changes over the treatment course. Therefore, we investigated the robustness of LETd distributions to setup and anatomical variations.

Methods: Ten consecutive head and neck cancer (HNC) and ten breast cancer (BC) patients were included. Patients were treated with robustly optimized intensity-modulated proton therapy (IMPT) treatment plans. All patients were positioned using daily cone-beam CT and a 6D robotic couch. Changes were monitored using weekly verification CTs in treatment position. Dose multiplied with LETd (D·LETd) distributions were calculated on the planning CT using our clinically commissioned Monte Carlo algorithm in the TPS (Raystation v6R). Additionally, the D·LETd distributions were calculated on weekly verification CTs, mapped to the planning CT and summed to calculate the accumulated D·LETd distribution. Target and organ-at-risk (OAR) LETd was determined by dividing the average D·LETd by the average dose.

Results: Target LETd was 3.1-3.2 keV/μm, OAR LETd ranged from 1.8 keV/μm for lung to 7.7 keV/μm for heart and most OARs had higher LETd than the target. The average absolute differences for both targets and OARs between nominally evaluated and accumulated LETd were 1.3% and 0.5% for HNC and BC patients respectively.

Conclusion: The influence of setup and anatomical variations on LETd distributions is small, showing that the LETd calculated on the planning CT for robustly optimized treatment plans will relate to patient outcomes when the relation between RBE and LETd is known.

O 33

A standardized commissioning framework of Monte Carlo dose calculation algorithms for proton pencil beam scanning treatment planning systems

C.W. Chang1, S. Huang2, J. Harms1, J. Zhou1, R. Slopsema1, M. Kang3, T. Liu1, M. McDonald1, K. Langen1, L. Lin1

1Emory University, radiation oncology, Atlanta, USA
2Memorial Sloan Kettering Cancer Center, Radiation Oncology, New York, USA
3New York Proton Center, New York Proton Center, New York, USA

Purpose: Treatment planning systems (TPSs) can involve different implementations of Monte Carlo dose calculation (MCDC) algorithms for pencil beam scanning (PBS) proton therapy. We propose a standardized framework on the use of commissioning data and steps to validate TDS-specific parameters and TPS-specific heterogeneity modeling to potentially reduce these uncertainties.

Methods: A standardized commissioning framework was developed to commission the MCDC algorithms of RayStation 8A and Eclipse AcurosPT v13.7.20 using water and non-water materials. Measurements included Bragg peak depth-dose and lateral spot profiles and scanning field outputs for Varian ProBeam. The phase-space parameters and the number of protons per MU were obtained from measurements. Spot profiles and various PBS field measurements, human tissues in TPS, Gammex phantom materials, and artificial materials were used for the TPS benchmark and validation.

Results: The maximum differences of phase parameters, spot sigma, and divergence between MCDC algorithms are below 4.5 μm and 0.26 mrad in air, respectively. Comparing TPS to measurements at depths, both MC algorithms predict the spot sigma within 0.5 mm uncertainty intervals, the resolution of the measurement device. Beam Configuration in AcurosPT is found to underestimate numbers of protons per MU by ∼2.5% and requires an additional user adjustment, while RayStation is within 1% of measurements using Auto model.

Conclusion: The proposed standardized commissioning framework can detect potential issues during PBS TPS MCDC commissioning processes, and potentially can design efficient measurement and improve dosimetric accuracies. Secondary MCDC can be used to identify the root sources of disagreement between primary MCDC and measurement.

O 34

A novel static beam delivery system for fast proton arc therapy

K.P. Nesteruk1, A.J. Lomax2, S. van de Water2, J.M. Schippers1

1Paul Scherrer Institut, Large Research Facilities GFA, Villigen PSI, Switzerland
2Paul Scherrer Institut, Center for Proton Therapy, Villigen PSI, Switzerland

A new static beam delivery device, foreseen for proton arc therapy, is proposed. This consists of a system of static and fixed magnets, by which the beam is transported to the isocenter from different directions. In contrast to conventional gantries, there is no need to move large magnets to adjust treatment angle at the isocenter (figure 1). Instead, static magnetic field guides the beam along a circular track around the patient in a guiding field region. At a certain radius, the beam enters the region of the enclosed inner bending field, in which a stronger static magnetic field bends the beam to or near the isocenter. The treatment table is located in the field-free central region. The treatment angle at the patient is defined by the azimuthal angle at which the beam enters the bending field region, which is determined by the strengths of the arc-scanning magnets only. Changing the settings of these can be done rapidly, so that a substantial arc span can be covered within a fraction of a second. As such, the device offers the potential to provide extremely rapid dose delivery in proton arc mode. We will report on our studies on beam optics and dose delivery. For a bending field of 2.1 T, the 1 m long device will have a radius of 4.4 m. A superconducting version was also studied, for which a field of 3.1 T results in a radius of 2.6 m.

O 35

Feasibility of proton ultra-high dose rate FLASH irradiation using a clinical synchrocyclotron

A. Darafsheh1, Y. Hao1, T. Zwart2, M. Wagner2, D. Catanzano2, J. Williamson1, N. Knutson1, B. Sun1, S. Mutic1, T. ZHAO1

1Washington University, Radiation Oncology, Saint Louis, USA
2Mevion, Cyclotron, Littleton, USA

Introduction: It has been recently shown that radiation therapy at ultra-high dose rates (>40 Gy/s, FLASH) has a potential advantage in sparing healthy organs compared to that at conventional dose rates. The purpose of this work is to show the feasibility of delivering proton beams in FLASH mode using a clinical synchrocyclotron.

Method: A clinical synchrocyclotron (HYPERSCAN®, Mevion) was modified to deliver ultra-high dose rates. Pulse widths of protons with initial kinetic energy of 230 MeV were changed from 1 μs to 20 μs to deliver in ultra-high dose rate and conventional dose rate. An energy absorber block made of boron carbide was positioned in the beam line to reduce the range of the beam to 4.1 cm. The number of protons per pulse at various dose rates was measured using a Faraday cup in order to establish the number of pulses required to deliver the same number of protons at various dose rates. The dose rate at the entrance and at the Bragg peak was measured using a plane-parallel ionization chamber. Monte Carlo simulation was performed in TOPAS to verify the experimental measurements.

Results: The integral depth dose measured using a Bragg chamber agreed well with the simulation. The average dose rate measured using the ionization chamber showed 101 Gy/s at the entrance and 216 Gy/s at the Bragg peak with a full width at half maximum (FWHM) field size of 1.2 cm. Feasibility of proton FLASH irradiation using a synchrocyclotron was demonstrated.

O 38

Advancing proton minibeam radiation therapy: magnetically focussed proton minibeams for a clinical centre

T. Schneider1, L. De Marzi2, A. Patriarca3, Y. Prezado4

1Imagerie et Modélisation en Neurobiologie et Cancérologie IMNC, CNRS - Univ Paris Sud - Université Paris-Saclay - Université de Paris, Orsay, France
2Institut Curie - University Paris Saclay - PSL Research University, Radiation Oncology Department - Centre de protonthérapie d'Orsay - Inserm U 1021-CNRS UMR 3347, Orsay, France
3Institut Curie - University Paris Saclay, Radiation Oncology Department - Centre de protonthérapie d'Orsay, Orsay, France
4Institut Curie - University Paris Saclay - PSL Research University, Inserm U 1021-CNRS UMR 3347, Orsay, France

Proton minibeam radiation therapy (pMBRT) [1] is a novel therapeutic strategy that has proven to significantly increase the therapeutic index of high-grade gliomas in rodents [2,3]. It uses very narrow proton beams (full width at half maximum, FWHM ≤ 1 mm), roughly one order of magnitude smaller than state-of-the-art pencil beams. The current implementation of pMBRT with mechanical collimators [4,5] is suboptimal due to flux reduction and the production of secondary neutrons.

To overcome these problems and pave the way towards 3D intensity-modulated pMBRT, we investigated how magnetically focussed and scanned minibeams could be integrated at existing clinical centres. For this, Monte Carlo simulations were performed with TOPAS [6] to determine the focussing capabilities of a current pencil beam scanning nozzle. Various modifications of the current nozzle geometry were evaluated which lead to a new, optimised nozzle design suitable for magnetic minibeam generation.

The new design features a shortened focal length and reduced air gap and yields beam sizes (FWHM) as small as 0.66 mm for 100-MeV beams and 0.33 mm for 200-MeV beams. It uses conventional nozzle elements facilitating both manufacturing and integration into existing facilities. Lastly, dose simulations performed in a water phantom suggest an improved dose distribution compared to mechanical collimators.

References: [1] Prezado et al., Med. Phys., 2013. [2] Prezado et al., Scie. Reports, 2017. [3] Prezado et al., Radiat. and Oncology, 2018. [4] Peucelle et al., Med. Phys., 2015. [5] De Marzi et al., Med. Phys., 2018. [6] Perl et al., Med. Phys., 2012.

O 39

Reference dosimetry audit of high-energy pencil beam scanning proton centres with ionisation chamber dosimetry and absorbed-dose calorimetry

A. Lourenco1,2, N. Lee1, I. Patel3, A. Vestergaard4, F. Fiorini5, J. Pandey6, J. Vera7, A. Mazal7, H. Palamans1,8

1National Physical Laboratory, Medical Radiation Science, Teddington, United Kingdom
2University College London, Department of Medical Physics and Biomedical Engineering, London, United Kingdom
3The Christie NHS Foundation Trust, Christie Medical Physics and Engineering, Manchester, United Kingdom
4Aarhus University Hospital, Danish Centre for Particle Therapy, Aarhus, United Kingdom
5Rutherford Cancer Centre Thames Valley, Medical Physics, Reading, United Kingdom
6Rutherford Cancer Centre North East, Medical Physics, Bedlington, United Kingdom
7Centro de Protonterapia Quirónsalud, Medical Physics, Madrid, Spain
8EBG MedAustron GmbH, Medical Physics Group, Wiener Neustadt, Austria

Safe implementation of radiotherapy techniques requires dosimetry audits to ensure best practice and consistency of treatments between centres. In this work, independent reference dosimetry audits were performed in five high-energy scanned proton beam centres.

Measurements were carried out at The Christie, UK and the Danish Centre of Particle Therapy, Denmark in their 250 MeV proton cyclotrons and in three centres equipped with 230 MeV proton synchrocyclotrons at the Rutherford Cancer Centres in Thames Valley and the North East, UK, and the Centro de Protonterapia Quirónsalud, Spain. Measurements were performed using single-energy layers with a field size of 10x10 cm2, a 2.5 mm spot spacing and an equal number of MUs delivered at each spot for 20 representative energies and depths. Measurements were also performed in three box fields of 10×10×10 cm3 homogeneous dose volumes centred at three different depths. PTW Roos ionisation chambers with calibrations carried out at NPL in Cobalt-60 were used and absorbed dose to water was determined according to the recommendations of IAEA TRS-398. The results were also compared to absolute dose measurements made using a graphite calorimeter, measurements which are currently still under analysis.

Absorbed dose values determined in this work and those previously determined by the proton facilities agreed within 2.0%. In the proton synchrocyclotron machines, ion recombination corrections could amount to 3% due to volume recombination. A consistent difference between ionisation chamber measurements and calorimetry was found across the different proton facilities but fell within the uncertainties determined according to the IAEA TRS-398.

O 40

Cancer cell aggressiveness caused by photon vs. proton radiation

M. Vazquez1, A. Bertucci1, H. Wang2, J. Unternaehrer2, J. Slater1, T. Suzuki2

1Loma Linda University Medical Center, Radiation Medicine, Loma Linda, USA
2Loma Linda University, Basic Science, Loma Linda, USA

The majority of ovarian (OC) and glioblastoma (GBM) cancer patients have a poor prognosis despite initial surgery, radiation therapy (RT) and chemotherapy. This high mortality rate is due to tumor recurrence and metastasis, primarily caused by chemo and radioresistant cancer stem like cells (CSC). A major contributor to metastasis and stem cell traits in these tumors is epithelial mesenchymal transition (EMT), which increases the cells' migratory and invasive abilities. Ionizing radiation is known to induce CSC properties. Stemness can be inhibited by let-7, a microRNA that maintains the differentiated state. Over expressing let-7 increases sensitivity to both chemotherapy and RT, disrupts the CSC phenotype, and reduces tumor burden. The EMT factor Snail represses let-7 miRNA transcription. Thus, targeting the Snail/let-7 axis has the potential to prevent radiation induced EMT activation and stemness. We aim to understand mechanisms by which RT results in invasiveness and stemness in cancer cells, and develop strategies to improve outcomes for OC and GBM by inhibiting EMT as well as to determine if protons and photons differ in their potential to modulate cancer cell stemness, proliferation, and invasiveness. Our preliminary results reveal clinically relevant doses of protons (2.5 Gy) were able to induce EMT in LN-18 GBM cells. Stemness markers expressions were increased in LN-18 and OVCAR8 cells at 24h, and in OVSAHO cells at 72h. X-rays were not effective to induce EMT or stemness in all the cell lines tested. These preliminary results suggested that protons and photons had differential biological effects in tumor cells.

O 41

Biological enhanced response (BER) with proton therapy plus PARP-1/-2 inhibitor in head and neck cancer cell mitotic catastrophe and senescence

L. Wang1, J. Cao1, X. Wang2, E. Lin1, X. Lu1, Y. Li2, X. Zhang2, N. Sahoo2, X.R. Zhu2, S.J. Frank3

1The University of Texas MD Anderson Cancer Center, Experimental Radiation Oncology, Houston, USA
2The University of Texas MD Anderson Cancer Center, Radiation Physics, Houston, USA
3The University of Texas MD Anderson Cancer Center, Radiation Oncology, Houston, USA

Purpose: A poly-(ADP-ribose) polymerase (PARP)-1 and PARP-2 inhibitor, Niraparib, can enhance human head and neck cancer (HNSCC) cell response to both photon (XRT) and proton (PRT) radiation with understudied mechanisms. We investigated the influence of Niraparib on mitotic catastrophe and senescence, the two major types of cell death induced by both XRT and PRT in HNSCC cells.

Methods: Human papillomavirus (HPV)-negative HN5, SqCC/Y1, and HPV-positive UPCI-SCC-154, UM-SCC-47 cell lines were used. Cell mitotic catastrophe (co-staining for cytoplasm with γ- tubulin for nucleus with Dapi) and senescence (assessed using SA-β-gal) were determined. Niraparib (1 μM) was given 1 hour (h) before irradiation.

Results: Niraparib significantly increased XRT and PRT induced cellular senescence at 6 days in the two HPV+ cell lines, with higher percentages of senescence in PRT than XRT cells (p all < 0.05); in the two HPV- cell lines, Niraparib significantly raised the percentages of XRT-induced senescence (p all < 0.01) (Figure 1). Niraparib caused significantly more cells undergoing mitotic catastrophe after PRT (in 3 out of 4 cell lines) or XRT (in 2 out of 4 cell lines), with higher mitotic catastrophe rates induced by Niraparib in PRT versus XRT group (p all < 0.05) (Figure 2).

Conclusion: Niraparib induced higher levels of mitotic catastrophe in combination with PRT versus XRT, and Niraparib promoted XRT-induced senescence. Upregulation effects of Niraparib on PRT-induced senescence and PRT- or XRT-induced mitotic catastrophe were cell line dependent. Further studies on personalized therapy of XRT or PRT in combination with Niraparib are warranted.

O 42

Quantifying proton relative biological effectiveness and normal tissue toxicity in the mouse spinal cord

J.M. Denbeigh1, M.H. Howard1, E.K. Debrot1, N.B. Remmes1, C.J. Beltran1

1Mayo Clinic, Radiation Oncology, Rochester, USA

Background: Uncertainty in proton relative biological effectiveness (RBE) may be a cause of radiation-induced normal tissue toxicity. Within late responding tissues such as the spinal cord, risk of radiation necrosis and paralysis resulting from changes in dose delivery and biological effect are of concern. The purpose of this study was to investigate the dependence of RBE on dose and on position along the Bragg curve in a living mouse spinal cord model.

Method and Materials: Cervical spinal cords of female C57BL/6J mice were irradiated (20-80 Gy; lateral opposed beams, A) at low-LET or high-LET (Bragg Peak) positions along the proton curve. Animals were anesthetized and restrained in a custom set up for treatment (B,C). Endpoint was defined as onset of radiation induced myelopathy, while weight and health were recorded weekly. Rotarod tests (D) were used to evaluate motor function. RBE will be calculated as the ratio of the tolerance doses at 50% effect probability (grade II paresis, E) at 365 days post irradiation.

Results: Acute toxicities of weight loss and skin abrasions were observed. Forelimb paralysis was manifest in the highest dose groups and accompanied by performance deficits on the rotarod. Preliminary results suggest Bragg Peak treated mice exhibit a shorter latency time to paralysis.

Conclusions: Robust characterization of proton RBE in our in vivo spinal cord model is crucial for informing clinical treatment strategies that strive to account for the biological variability of proton therapy in order to reduce short and long term side effects to critical organs.

O 43

Examining the differential cellular response of a panel of pancreatic cancer cell lines to 12C vs. photon irradiation

A. Davis1, B. Sishc1, J. Saha1, C. Angelica2, D. Saha1, M. Ciocca2, M. Story1

1University of Texas Southwestern Medical Center, Radiation Oncology, Dallas, USA
2National Centre for Oncological Hadrontherapy, Radiation Oncology, Pavia, Italy

Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal malignancy that is expected by 2020 to become the second most common cause of cancer-related deaths globally. The 5-year overall survival rate for patients with PDAC is <10%, which is due to the majority of pancreatic cancer patients presenting with advanced disease and the inherent resistance of PDAC to conventional chemotherapy and radiotherapy. Clinical trials have suggested that carbon ion radiotherapy (CIRT) used concurrently with gemcitabine is effective against unresectable locally advanced PDAC. We have initiated pre-clinical studies to identify patients who would most likely benefit from CIRT. In particular, we have focused on examining the role of an aberrant DNA damage response (DDR) in PDAC treatment responsiveness. This is supported by exome sequencing data of >150 surgically excised PDACs that showed that >35% of PDAC tumors have an alteration in a DDR gene and that these alterations correlate with a poor outcome. In this study, we utilized a panel of PDAC cell lines, whose survival fraction of 2 Gy (g-rays) ranged from average to radioresistant, to determine if CIRT can overcome photon radioresistance. Furthermore, we will present data examining targeted DNA repair inhibitors to potentiate CIRT-mediated cell killing of g-rays resistant cell lines. Finally, we aimed to identify DDR-related biomarker(s) that predict responsiveness to CIRT.

O 44

Enhancing 12C radiotherapy for head and neck cancer via biomarkers of radioresponse, conditional vulnerabilities, and modeling of relative clinical effectiveness

B. Sishc1, J. Saha1, L. Ding1, F. Angelia2, A. Aroumougame1, P. Arnold1, D. Saha1, M. Ciocca3, A. Davis1, M. Story1

1UT Southwestern Medical Center, Radiation Oncology, Dallas- TX, USA
2National Center of Oncological Hadrontherapy, Biology, Pavia, Italy
3National Center of Oncological Hadrontherapy, Physics, Pavia, Italy

Head and neck squamous cell carcinoma (HNSCC) is the sixth most common malignancy worldwide. 30% of patients will experience recurrence for which median survival is less than one year. Factors contributing to treatment failure include inherent resistance to X-rays and chemotherapy, hypoxia, EMT, and immune suppression. The unique properties of 12C radiotherapy including enhanced cell killing, a decreased oxygen enhancement ratio, generation of complex DNA damage, and ability to overcome immune suppression make its application well suited to the treatment of HNSCC. We examined the 12C radioresponse of six HNSCC cell lines, whose surviving fraction at 3.5 Gy ranged from average to resistant when compared to a larger panel of 49 cell lines. To determine whether 12C irradiation can overcome X-ray radioresistance and whether there may be biomarkers predictive of 12C radioresponse. Cells were irradiated with 12C using a SOBP with an average LET of 75 keV/μm (CNAO: Pavia, Italy). RBE values at 10% survival ranged from 2.13 to 4.2. DNA damage foci resolution suggested that unrepaired, complex double strand breaks contribute to an enhanced RBE relative to photons. A comparative analysis of gene expression post-12C exposure vs g-Rays revealed differential regulation of DNA double strand break repair, differentiation, cell cycle progression, and cell death modulation. These data were used in conjunction with agent based modeling to predict real clinical effectiveness (RCE). Based on the above analysis, we present the framework of a strategy to utilize biological markers to predict which HNSCC patients would benefit the most from 12C radiotherapy.

O 45

Prognostic significance of PD-L1 expression in carbon-ion radiotherapy for adenocarcinoma of the uterine cervix

M. Iijima1,2, N. Okonogi3, K. Banno2, K. Tsuji2, Y. Kobayashi2, E. Tominaga2, S. Yamada1, H. Tsuji3, D. Aoki2, S. Hasegawa1

1National Institutes for Quantum and Radiological Science and Technology, National Institute of Radiological Sciences, Chiba, Japan
2Keio University School of Medicine, Department of Obstetrics and Gynecology, Tokyo, Japan
3National Institutes for Quantum and Radiological Science and Technology, QST Hospital, Chiba, Japan

Objective: Carbon-ion (C-ion) radiotherapy (CIRT) is expected as an effective treatment option for adenocarcinoma of the uterine cervix, which is more radioresistant than squamous cell carcinoma of the uterine cervix. Programmed cell death-ligand 1 (PD-L1) is expressed in tumor cells and has been shown to predict the clinical outcomes of several types of malignancies. However, the prognostic value of PD-L1 expression in cervical cancer is controversial. The aim of this study was to investigate the effects of C-ion beams irradiation on PD-L1 expression in both cell line and clinical samples of adenocarcinoma of the uterine cervix and to identify the prognostic factors for outcomes after CIRT.

Methods: The effects of C-ion irradiation on PD-L1 expression in cells of adenocarcinoma of the uterine cervix was examined by flow cytometry. In biopsy specimens of adenocarcinoma of the uterine cervix from 33 patients collected before CIRT started (pre-CIRT) and after 12 Gy (relative biological effectiveness [RBE]) irradiation of CIRT (post-12Gy-C), PD-L1 expression was examined to investigate the correlation between PD-L1 status and clinical outcomes.

Results: The PD-L1 expression was upregulated by C-ion beams in a dose-dependent manner in cells of adenocarcinoma of the uterine cervix. The post-12Gy-C PD-L1 expression was significantly elevated compared to the pre-CIRT PD-L1 expression. The post-12Gy-C PD-L1 expression had favorable correlation with progression-free survival (PFS).

Conclusion: Our study revealed that CIRT can upregulate PD-L1 expression in adenocarcinoma of the uterine cervix. Moreover, PD-L1 expression in adenocarcinoma of the uterine cervix after C-ion irradiation had favorable correlation with PFS after CIRT.

O 46

Comparison of proton stopping power measurements of animal tissues from proton CT and x-ray CT systems

F. DeJongh1, E. DeJongh1, V. Rykalin1, M. Pankuch2, G. DeFillippo2, J. Welsh3, R. Schulte4, N. Karonis5, C. Ordonez5, G. Coutrakon6

1ProtonVDA LLC, Physics, Naperville, USA
2Northwestern Medicine Chicago Proton Center, Physics, Warrenville, USA
3Hines VA Hospital, Radiation Oncology, Hines, USA
4Loma Linda University, Basic Sciences, Loma Linda, USA
5Northern Illinois University, Computer Science, DeKalb, USA
6Northern Illinois University, Physics, DeKalb, USA

Purpose: Characterize proton CT (pCT) images from a prototype clinical proton imaging system that is simple, lightweight, easily scaled to large field sizes, operates at high speed to maximize throughput, and uses the minimum possible radiation dose for a given resolution. The system fully exploits individual proton path information and utilizes pencil beam scanning systems delivering low fluence beams.

Methodology: Acquire pCT images with the system operating with a scanning pencil beam at the Northwestern Medicine Chicago Proton Center. Compare with relative stopping power (RSP) maps derived from x-ray CT images using stoichiometric calibration.

Results: Figure 1 shows a 1 mm thick pCT slice of a sample of pork shoulder and ribs along with a slice, at the same location, of an RSP map derived from x-ray CT. Figure 2 shows a difference map and a photograph of the setup. The average agreement is better than 1% in uniform regions of muscle, fat and bone. Local discrepancies occur in regions with high density variation. Noise in the pCT slice could be improved by using more protons.

Conclusions: We have characterized the first pCT images of complex animal tissue samples mimicking human patients in size and composition. In comparison, x-ray CT images using stoichiometric calibration produce good agreement in these samples.

O 47

CBCT image correction using a Monte Carlo trained deep convolutional neural network to enable fast and accurate adaptive proton therapy

A. LALONDE1, B. Winey1, J. Verburg1, H. Paganetti1, G. Sharp1

1Massachusetts General Hospital and Harvard Medical School, Radiation Oncology, Boston, USA

Purpose: Daily volumetric imaging is key to enable adaptive proton therapy (APT). While such data can be acquired using cone-beam CT (CBCT), scatter artifacts make uncorrected CBCT images unsuitable for APT dose calculation. Scatter artifacts can be accurately corrected using Monte Carlo (MC) simulations, however MC is too computationally demanding for real-time usage. In this work, we evaluate the performance of a MC trained U-shape deep convolutive neural network (Unet) to correct scatter artifacts in CBCT images and enable fast and accurate APT dose calculation.

Methods: CBCT projections are generated for 24 head and neck patients using a GPU accelerated MC code. A Unet is trained to estimate scatter distributions from the uncorrected CBCT projections. Patients are distributed in training (17), testing (3) and validation (4) sets. The accuracy of the scatter correction is evaluated using CT and CBCT images of an additional anthropomorphic head phantom. The potential of the method for APT is assessed by comparing proton therapy dose distributions calculated on scatter-free, uncorrected and scatter-corrected CBCT images.

Results: The mean error on CT numbers is reduced from 44.5 to 1.4 HU. In a head phantom, the root-mean square difference of proton ranges calculated in CT and corrected CBCT is 1.41 mm. The 3%/3mm Gamma pass rate is >99% for all validation patients and each CBCT volume is corrected in less than 2 seconds.

Conclusion: The potential of MC trained Unet to correct CBCT images for APT is demonstrated for the first time. The method is shown to be fast and accurate.

O 48

First experimental validation of daily adaptive proton therapy – dosimetric evaluation with an anthropomorphic phantom

L. Nenoff1,2, M. Matter1,2, S. Krier1, D.C. Weber1,3,4, A.J. Lomax1,2, F. Albertini1

1Paul Scherrer Institute, Center for Proton Therapy, Villigen PSI, Switzerland
2ETH Zurich, Department of Physics, Zurich, Switzerland
3University Hospital Bern, Department of Radiation Oncology, Bern, Switzerland
4University Hospital Zurich, Department of Radiation Oncology, Zurich, Switzerland

To compensate for anatomical changes during proton therapy, fast plan adaption is necessary. Therefore, online adaptive strategies are increasingly investigated in the proton community (Albertini et al., BJR, 2019). At the Paul-Scherrer-Institute we are currently implementing an online adaptive workflow for daily adaptive proton therapy (DAPT).

To experimentally validate our DAPT workflow, a customized anthropomorphic phantom, sliced in the coronal direction and with exchangeable nasal cavity fillings, was used. A reference IMPT treatment (3 fractions, 6 Gy-RBE) was optimized on a CT of the phantom with empty nasal cavity filling, and then delivered to the phantom with three different nasal cavity filling states. For the same states, online DAPT plans were then re-optimized and delivered based on CT data acquired immediately before each fraction. Accumulated delivered dose for all 3 fractions were measured using films and compared to the reference plan.

3%/3mm gamma rates for the non-DAPT treatment were 87% and 66% when compared to the dose calculated for the reference plan at two different depths. With DAPT, agreement increased to 91% and 93% (Figure 1), despite additional uncertainty of dose accumulation across the 3 fractions. Simulations of the non-adapted treatments showed a CTV V95% reduction of 26%, whilst with DAPT, CTV V95% improved by 9% (Figure 2) compared to the nominal plan.

For the first time we have successfully tested an online DAPT workflow using a customized anthropomorphic phantom. We could show that online plan adaption is feasible and clearly improves the dosimetric quality of the applied treatment dose.

O 50

Performance validation of the CNAO dose delivery system at GSI: Control system for treatment moving targets with scanned ion beams

M. Lis1, M. Donetti2, M. Wolf1, M. Durante1, W. Newhauser3, C. Graeff1

1GSI Helmholtzzentrum fur Swerionenforschung, Biophysics, Darmstadt, Germany
2CNAO, Research and Development, Pavia, Italy
3Louisiana State University, Medical Physics, Baton Rouge, USA

We implemented and tested a conformal motion mitigation technique into the clinical dose delivery system (DDS) of CNAO and in the duplicate system found in the therapy research room of GSI.

The mitigation strategy involves optimizing treatment plans to each motion state of 4DCTs and creating a treatment plan library. These plans are loaded onto the DDS and delivered synchronously to the measured target motion; treatment sequence is directed by the detected motion state.

Validation tests were performed to test compensation of deliveries with range changes and multiple tissue densities. A linear stage was used to create sinusoidal motion patterns for 12 pinpoint ICs in a water tank and Gafchromic films. Various geometric and patient plans for 10 phases of compensation were delivered to investigate the degree of interplay compensation and the extent of residual motion within a single motion phase. Measurements were compared to static reference irradiations.

Geometric cube deliveries demonstrated that effective motion compensation was possible with 10 amplitude-based motion state plans at a 20 mm amplitude. 60x60x60 mm3 cubes were delivered to the pinpoint ICs through a water-equivalent plastic phantom with a bone insert. Average (STD) doses were 9.7 Gy (2.1%) and 9.8 Gy (1.3%) and homogeneity indices were 95.2% and 96.2% for 10 motion phase and static cubes, respectively, showing effective compensation of 17.6 mm range changes.

Motion-synchronized delivery of target conformal plans was shown to be feasible. Further tests for more complex phantoms are necessary to translate this strategy to clinic.

O 51

Extremely hypofractionated proton radiotherapy in the treatment of low and intermediate risk prostate cancer – long-term outcome

J. Kubes1, S. Sláviková2, P. Vítek2, V. Vondráček3, M. Navrátil3, M. Andrlík3, K. Dědečková2, J. Rosina4, S. Vinakurau2, B. Ondrová3

1PTC Czech- University Hospital Motol, Proton therapy dept., Prague, Czech Republic
2PTC Czech- University Hospital Motol, Radiotherapy, Prague, Czech Republic
3PTC Czech, Proton therapy dept., Prague, Czech Republic
43rd Faculty of Medicine- Charles University Prague-, Department of Medical Biophysics and Informatics-, Prague, Czech Republic

Purpose: To analyze 5-years biochemical free survival (bDFS) and late toxicity profile in patients with prostate cancer treated with pencil beam scanning (PBS) proton radiotherapy.

Material and methods: Between January 2013 and July 2015 136 patients with prostate cancer were treated with IMPT (intensity modulated proton therapy), with extremely hypofractionated schedule (36.25 GyE/5 fractions). Five pts were lost from follow up immediately and were excluded from analysis. 131 pts were analyzed with median of follow up time 55 months. Mean age was 64.3 (40.1-85.7) years, mean PSA value was 6.16 μg/l (0.67-17.3 μg/l), 67 (51.1%) and 64 pts (48.9%) had low and intermediate risk cancer, respectively. 20 (15.2%) pts had neoadjuvant hormonal therapy, no patients had adjuvant hormonal therapy. bDFS and late toxicity profile were evaluated.

Results: Median treatment time was 9 (7-13) days. Estimated 5-years bDFS is 95 % and 84% for low and intermediate risk group, respectively. Late toxicity (CTCAE-v.4) was: Gastrointestinal: G1- 28 (21%) pts., G2-8 (6%), G3 1 (0.7%); Genitourinal: G1-26 (19%), G2-10 (7.7%), no G3 toxicity was observed. PSA relaps was observed in 10 pts, lymph node or bone recurrence was detected in 7 of them. No local recurrence was detected. Two pts. died from other reasons than prostate cancer.

Conclusion: Extremely hypofractionated proton beam radiotherapy of prostate cancer is effective with long term bDFS comparable with other fractionation schedules, with minimal serious long term GI and GU toxicity.

Funding: Supported by the European Regional Development Fund-Project: “Engineering Applications of Microworld Physics” (No. CZ.02.1.01/0.0/0.0/16_019/0000766)”.

O 53

Comparative clinical outcomes of proton-beam therapy (PBT) versus intensity-modulated radiotherapy (IMRT) for prostate cancer in the postoperative setting

A. Barsky1, R. Carmona1, P. Santos2, V. Verma3, S. Both4, N. Vapiwala1, C. Deville5

1Hospital of University of Pennsylvania, Department of Radiation Oncology, Philadelphia, USA
2Memorial Sloan-Kettering Cancer Center, Department of Radiation Oncology, New York, USA
3Allegheny General Hospital, Department of Radiation Oncology, Pittsburgh, USA
4University Medical Center Groningen, Department of Radiation Oncology, Groningen, Netherlands
5Johns Hopkins University, Department of Radiation Oncology and Molecular Radiation Sciences, Baltimore, USA

Purpose: Data comparing clinical disease-control outcomes of post-prostatectomy PBT to IMRT for prostate cancer (PC) are limited. Herein, we provide the largest such report.

Methods: We reviewed all patients who underwent post-prostatectomy prostate bed-only PBT or IMRT at our institution between 2009-2017 for biochemical failure (BF) (using institutional, GETUG, and SPPORT definitions), local/regional/distant failure, and mortality. Median radiotherapy doses were 70.2 Gy (RBE) for both groups (PBT, range 66.6-75.6; IMRT, range 70.2-75.6), with 11 (16.9%) PBT and 28 (14.4%) IMRT patients receiving concurrent androgen-deprivation. We performed a case-matched cohort analysis using 3-to-1 nearest-neighbor matching for age-at-diagnosis, pathologic Gleason score, and pT3 vs.<pT3 disease. A multivariable Cox proportional hazards model (MVA) was used to estimate hazard ratios (HR) for disease-control outcomes by treatment group. We identified potential confounders using stepwise-backwards elimination to p<0.1.

Results: We matched 260 out of 295 men (n=65 PBT, 195 IMRT) for analysis. Mean age at diagnosis (63.1±7.3 PBT vs. 60.1±6.1 yrs. IMRT, p<0.01) significantly differed between matched cohorts. At a median follow-up of 59 (range 16-87, PBT) and 59 (3-128, IMRT) months, radiation modality was not significantly associated with BF on MVA (institutional HR 1.16, 95%CI 0.74-1.82, p=0.52; GETUG HR 1.16, 95%CI 0.68-1.97, p=0.58; SPPORT HR 1.03, 95%CI 0.62-1.73, p=0.90). Modality also was not significantly associated with local (p=0.82), regional (p=0.11), or distant failure (p=0.36), or all-cause mortality (p=0.69), though such events were limited (n=6, 19, 24, and 10, respectively).

Conclusions: Disease-control outcomes for PC patients treated with post-prostatectomy radiotherapy do not significantly differ by radiation modality.

O 54

Carbon-ion radiotherapy for pelvic recurrence of colorectal cancer after primary surgery

H. Takiyama1, S. Yamada1, Y. Isozaki1

1National Institute of Radiological Sciences NIRS- QST, Radiation Oncology, Chiba, Japan

Purpose: We investigated the safety and efficacy of carbon-ion radiotherapy (C-ion RT) for pelvic recurrence of colorectal cancer after primary surgery.

Patients and methods: Data from patients with pelvic recurrence of colorectal cancer treated by C-ion RT with 73.6Gy(RBE; relative biological effectiveness weighted absorbed dose) from April 2001 to March 2019 at our institution were retrospectively analyzed. Local control rate (LC) and overall survival rate (OS) were statistically estimated.

Results: A total of 486 patients' data were collected. 386 cases had pelvic recurrences only, and one hundred cases had synchronous distant oligo-metastases. No case had a history of radiation therapy to pelvis. 305 cases had received any kind of chemotherapy before referral to C-ion RT. A median follow-up period from the initiation of C-ion RT was 43 months (range; 3–182 months). The LC rate was 76% (95%CI 73%–79%) at 5 years. The OS rate was 42% (39%–45%) at 5 years. Grade 3 treatment related acute intestinal toxicity was observed in two cases and grade 3 late toxicity was observed in 16 cases: skin reaction (n=2), gastrointestinal toxicity (n=7), neuropathy (n=6), and bladder toxicity (n=1). There was no grade 4 or 5 acute or late toxicity. Besides these toxicities, few patients developed pelvic infection, intestinal perforation, bleeding or ileus related to a polytetrafluoroethylene (PTFE) spacer sheet which was inserted prior to C-ion treatment for the protection of intestines.

Conclusion: C-ion RT may be a safe and effective treatment option for pelvic recurrence of colorectal cancer after primary surgery.

O 55

Pencil beam scanning proton therapy with deep thermal therapy is safe with potential for increased efficacy in advanced abdominopelvic malignancies

J. Molitoris1, D. Rodrigues1, J.W. Snider1, A. Rao1, S. Mossahebi1, M. Zakhary1, B. Nrusingh1, K. Lehman1, Z. Vujaskovic1

1Maryland Proton Treatment Center, Radiation Oncology, Baltimore, USA

Introduction: Little is known about the safety and feasibility of Proton Beam Therapy (PBT) and Deep Thermal Therapy (DTT), a potent radio-sensitizer. Our spot scanning PBT center has been treating patients with locoregional DTT for more than a year. We hypothesize this dual modality treatment is safe and feasible.

Methods: We performed a retrospective IRB-approved single institution review of all patients treated with PBT and concurrent DTT. Patient and disease characteristics were collected along with RT dose and DTT parameters. 40°C for tumor and 39°C for tumor surrogates were used as minimum temperatures for therapeutic treatment time. Cumulative equivalent minutes at 43°C (CEM43) and maximum temperature (Tmax) were recorded thermal parameters

Results: Combination DTT+PBT treatment was delivered to 24 patients with median age 66 (range, 19-78 years). Histologies included colorectal (n=15) and prostate (n=4). 19 (79.2%) were treated for locally recurrent disease and 17 (70.8%) with re-irradiation. Median PBT dose was 48.0Gy(RBE) (30-66Gy[RBE]) and half received BID. Seven (29.2%) received concurrent chemotherapy. DTT median treatments was 6 (1-14) with a therapeutic time of 35min (0-67 min). The TDmax average was 4.1CEM43 (0-25.5 CEM43). There were no grade 4 or higher acute or late toxicities. With a median follow up of 9 months (range, 1-14 months) there have been 4 reported local failures (median 6 months, range 4-11). Updated toxicities and outcomes will be presented.

Conclusions: Initial combination of PBT with DTT is well tolerated with limited acute and subacute toxicities despite a heavily pre-treated population. Continued follow up is required.

O 56

Update on international recommendations for light ion beam therapy

O. Jäkel1,2, L. Burigo1, S. Greilich1

1German Cancer Research Center DKFZ, Medical Physics In Radiation Oncology, Heidelberg, Germany
2On behalf of the ICRU report committee 22, Int commission on Radiation Units and Measurement, Bethesda, USA

By the end of 2019, an ICRU report on Prescribing, Recording and Reporting Light Ion Beam Therapy was finalized and submitted to the Journal of the ICRU. This report is an attempt to harmonize and standardize the clinical use of light ion beam radiotherapy. It follows the concepts previously developed by the ICRU for reporting other therapies but emphazises especially the use and reporting of RBE-weighted quantities. Such harmonization will facilitate the comparison of therapeutic results obtained with ions not only between ion beam therapy centers but also with centers using other modern forms of radiation therapy, such as proton-RT and IMRT with photon beams.

The report outlines the different biological models currently used for calculating RBE weighted doses for light ion beam therapy and attempts to clarify their clinical use in order to enable a common understanding of clinical practice in various facilities. It gives detailed recommendations on how light ion beam therapy should be prescribed, recorded and reported and gives detailled clinical examples. The recommendations on dosimetry were also harmonized and updated according to ICRU report 90 (Key Data for Ionizing-Radiation Dosimetry: Measurement Standards and Applications, 2016) and with the upcoming revision of IAEA's TRS-398 (Absorbed Dose Determination in External Beam Radiotherapy: An International Code of Practice for Dosimetry based on Standards of Absorbed Dose to Water, 2016).

An overview of the recommendations in ICRU 93 and the changes in TRS-398 for light Ions will be presented. The implications of both reports on the clinical use will be outlined.

O 57

Quality assurance and verification of a proton flash lung SBRT plan with a commercial proton therapy system

M. Folkerts1, J. Speth2, E. Lakew3, A. Magliari4, C. Smith1, J. Perez5, E. Abel1, A. Mascia6

1Varian Medical Systems, Proton Solutions, Palo Alto, USA
2University of Cincinnati Medical Center, Cincinnati Children's Hospital, Cincinnati, USA
3University of Cincinnati, Biomedical Engineering, Cincinnati, USA
4Varian Medical Systems, Medical Affairs, Palo Alto, USA
5Varian Medical Systems, Proton Solutions, Geneva, Switzerland
6University of Cincinnati, UC Health, Cincinnati, USA

Purpose: Ultra-high dose rate FLASH therapy promises increased sparing to healthy tissue when compared to conventional dose rates. Currently, proton therapy is the only modality capable of treating non-superficial lesions at FLASH dose rates. Creation and accurate delivery of FLASH plans for human indications is a crucial step towards clinical trials. In this study, we investigate whether a Varian ProBeam® can deliver therapeutic dose at FLASH dose rates (>40 Gy/s) within clinically acceptable tolerances.

Methods: We used 250 MeV transmission fields (Bragg peak behind body) delivered by a ProBeam® in service mode. Varian Eclipse® was used to target a 2 cm lung lesion on an SBRT phantom with three non-coplanar fields (Figure 1). The prescription was 60 Gy of total physical dose delivered in 3 fractions covering 98 % of the target. Film and ion-chamber and measurements were recorded and compared to the treatment planning system using the gamma index test.

Results: FLASH fields were measured to be deliverable at or above a clinically acceptable pass rate of 90 % for 2 %, 2 mm gamma index test (Figure 2). Each field was delivered in 64 ms, depositing 6.7 Gy of physical dose to the isocenter position.

Conclusion: Varian Eclipse® and ProBeam® can create and deliver FLASH lung SBRT plans with transmission fields within clinically acceptable criteria. This positive result has immediate impact on pre-clinical animal studies exploring the nature of the FLASH effect with protons. This result supports proton therapy as a viable modality to deliver FLASH treatments to humans.

O 58

Statistical analysis of patient specific plan verification in ion beam therapy with protons and carbon ions

M. Schafasand1, G. Kragl1, J. Osorio1, M. Stock1, A. Carlino1

1EBG MedAustron GmbH, Medical Physics, Wiener Neustadt, Austria

At MedAustron, patients are treated with protons in two rooms both equipped with fixed beam lines. Since July 2019, carbon ions are also available at one of the fixed horizontal beam lines. Active scanning technology is implemented as beam delivery system and, for superficial tumors, an automatically inserted range shifter (RaShi) can be used. The complexity of this technique request high demands on patient specific quality assurance (PSQA). The aim of this study is to report trend lines for PSQA based on different criteria e.g. dose calculation algorithm (DCA), anatomy, etc.

In the TPS (RayStation V6-V8, RaySearch Laboratories, Stockholm, Sweden) generated plans were recalculated in a virtual water phantom. The DCAs commissioned were the Pencil beam (PB v3.5 and v4.1) and the Monte Carlo (MC v4.0 and v4.2) algorithms for protons and PB for carbon ions. Absorbed dose to water measurements were accomplished with a 3D array of 24 PinPoint ionization chambers (model 31015, PTW Freiburg) mounted on a PTW water phantom.

For all the 2978 proton and 152 carbon ion beams the mean global dose deviations between the measurements and the TPS planned dose for beams without and with RaShi were -0.13±0.66%, 0.24±0.97% and 0.22±0.68%, -1.01±0.95%, respectively. Larger deviations were found for beams with RaShi especially when the PB algorithm was used (see figure 1). A slight over-prediction of the dose for PB algorithms and in presence of RaShi is evident for both particle types. The results revealed that the average dose deviations in all criteria are within ±2%.

O 59

The PAir PRoduction Imaging ChAmber (PAPRICA)

I. Mattei1, G. Battistoni1,2, G. Calvi1,3, Y. Dong1,3, M. De Simoni4,5, A. Embriaco1, M. Fischetti5,6, M. Toppi7, G. Traini5,8, S.M. Valle1

1INFN, Physics, Milan, Italy
2INFN-TIFPA Trento Institute for Fundamental Physics and Applications, Physics, Trento, Italy
3Università degli Studi di Milano “La Statale”, Physics, Milan, Italy
4"Sapienza” Università di Roma, Physics, Rome, Italy
5INFN, Physics, Rome, Italy
6"Sapienza” Università di Roma, Dipartimento di Scienze di Base e Applicate per l'Ingegneria SBAI, Rome, Italy
7INFN-LNF Laboratori Nazionali di Frascati, Physics, Frascati, Italy
8Museo Storico della Fisica e Centro Studi e Ricerche E. Fermi, Physics, Rome, Italy

In particle therapy, safety margins are applied when planning the treatment in order to account for the multiple sources of beam range uncertainty. Reducing safety margins is fundamental in the treatment of tumors close to organs at risk and paediatric patients. Several range monitoring techniques are being investigated, all based on the detection of secondary particles produced in the nuclear interactions of the ion beam with the patient's tissue nuclei. Many efforts are focussed on the prompt photons detection.

In this contribution, a novel range monitoring technique is proposed, based on the exploitation of the prompt photons pair production mechanism as prompt gamma imaging technique. The PAPRICA (PAir PRoduction Imaging ChAmber) project will be discussed: the chamber will reconstruct the prompt photons 3D spatial emission distribution, requiring an energy E>4 MeV to select the event population that is most correlated to the Bragg peak position. The PAPRICA detector will be able to monitor proton and carbon ion treatments, implementing neutrons background reduction strategies and profiting from the e+,- pair clear topological event signature. No collimation nor time of flight information on the detected photons will be needed. The PAPRICA detector design and the expected performances evaluated by means of a Monte Carlo simulation in a real case scenario will be presented.

O 60

Detailed in-vivo mapping of dose and dose-rate distributions in pre-clinical proton pencil beam scanning experiments

P. Poulsen1, E. Kanouta2, G. Kertzscher2, C. Overgaard3, B.S. Sørensen3, J.G. Johansen2

1Aarhus University Hospital, Department of Oncology and Danish Center for Particle Therapy, Aarhus N, Denmark
2Aarhus University Hospital, Danish Center for Particle Therapy, Aarhus N, Denmark
3Aarhus University Hospital, Department of Oncology, Aarhus N, Denmark

Purpose: The current interest in FLASH proton therapy increases the need to monitor proton dose delivery with high spatial and temporal resolution. We developed and applied an in-vivo dosimetry system for detailed mapping of the delivered dose and dose-rate in pre-clinical proton pencil beam scanning (PBS) experiments.

Methods: During proton PBS mouse experiments, the point dose in well-defined positions on mouse leg targets was measured in-vivo with sub-spot temporal resolution using a ZnSe:O scintillator system with 1100Hz sampling rate (Figure 1). Post-treatment, the time-resolved point dose measurements were synchronized with log files containing the time and position of each spot delivery. Triangulation of individual spot signals allowed determination of the detector position relative to each energy layer's spot pattern. The mean and standard deviation (SD) of the triangulated detector position across the layers were determined for each mouse. Combining the detector position with the log files and camera monitoring of the mice gave detailed 3D mapping of the mouse dose as function of time.

Results: In-vivo dosimetry provided 3D dose-rate mapping in nine mice experiments. The accuracy of the triangulated detector position relative to the layer spot pattern (mean SD across nine mice) was 0.21 mm (see example in Figure 2). Figures 2A-B show deduced dose and dose-rate maps overlayed with the mouse leg position.

Conclusion: In-vivo dosimetry with point dose measurements of individual spots was developed and used for mapping of delivered doses and dose-rates in pre-clinical proton PBS experiments. It could prove very useful for dose-rate QA in proton FLASH experiments.

O 61

Timing features of 4D detector and Pencil Beam Scanning Proton Therapy can improve WET map accuracy with deep CNN

C.W. Chang1, S. Zhou2, J. Zhou1, T. Liu1, X. Yang1, T. Zhang2, L. Lin1

1Emory University, radiation oncology, Atlanta, USA
2Washington University, Radiation Oncology, St. Louis, USA

Purpose: Proton radiography using conventional 2D detectors cannot provide accurate water equivalent thickness (WET) map as residual proton energy detection is contaminated by adjacent scatter protons. Using deep learning, this work combines the time features of a newly designed in-house 4D ion chamber array with scanning feature of pencil beam spot scanning proton therapy (PBS) to de-convolve the residual energy contamination from adjacent spots therefore derive a more accurate WET map.

Methods: 4D detector has 2 mm and 0.3 ms spatial-temporal resolutions over 128Xx128Yx64Z channels. Deep convolutional neural networks (CNNs) were used to assist data analysis from the 250k frames of spatial-temporal measured data using 41x41 proton spots per 4 mm spacing (128x128). The anthropomorphic phantom was irradiated with an anterior beam and the exit dose was measured with both 4D detector and conventional MatriXXPT. WET maps were derived by matching measurements with RayStation 8B dose calculation.

Results: Derived WET maps agree within 2 mm for the most parts and large discrepancies of 4-8 mm often happen near the edge of dense bone that involve high heterogeneity. The difference of WET maps increased with the complexity of clinical sites. The results indicate that WET map can be improved with 4D detector by de-convolution of proton scatter involved in MatriXX PT measurements.

Conclusion: Utilize deep CNN and timing features in new 4D detector and PBS, proton radiography can derive more accurate WET maps for more heterogeneous anatomy.

O 62

Incorporation of HiC-measured DNA arrangement in the mechanistic modelling of proton, photon, helium and carbon biological effect

S. Ingram1, N. Henthorn1, J. Warmenhoven1, A. Chadwick1, E. Santina1, N. Kirkby1, N. Burnet1, R. Mackay2, K. Kikrby1, M. Merchant1

1University of Manchetser, Division of Cancer Sciences, Manchetser, United Kingdom
2The Christie NHS Foundation Trust, Christie Medical Physics and Engineering, Manchetser, United Kingdom

Increased use of particles in radiotherapy has made the question of relative biological effect critical for treatment optimisation. We have developed a mechanistic model, DaMaRiS, employing track structure simulation to predict radiation-induced DNA damage and the subsequent repair of DNA double-strand breaks (DSBs), which highlights the sensitivity of biological effect on the initial spatial arrangement of DSBs. We find that an increase in local DSB density leads to an increase in misrepair, resulting in chromosome aberrations. The modelled description of nuclear DNA arrangement is a key parameter. Commonly, models describe this arrangement with distinct chromosome domains. HiC, now offers the possibility to measure the DNA arrangement. In this work, we have implemented a HiC-based geometry model into Geant4-DNA. Contact maps are taken from the literature, for a variety of cell lines, and solved with a polymer model, providing the spatial arrangement of the 46 chromosomes. The polymer beads are constructed as Geant4 geometries within a cell nucleus. The cell is irradiated with mono-energetic protons, Co-60 photons, helium- or carbon-ions and DSBs are scored. This information is then passed to our repair model, where DSB ends diffuse (using sub-diffusion) and recruit DSB repair pathways proteins. Predictions are made for the yields of DSBs and their repair efficacy, scoring residual and mis-repaired DSBs, as a function of linear energy transfer (LET). For the first time a realistic, experimentally informed, DNA geometry is constructed in a track structure simulation. This gives confidence in model predictions and is a step closer to clinical significance.

O 63

Normal tissue response depends on combined temporal and spatial fractionation of proton minibeams

M. Sammer1, J. Schauer1, A. Dombrowsky2, S. Bicher2, S. Rudigkeit1, N. Matejka1, J. Reindl1, S. Bartzsch2, T. Schmid2, G. Dollinger1

1Universität der Bundeswehr München, Institute for Applied Physics and Metrology, Neubiberg, Germany
2Institute of Radiation Medicine, Helmholtz Zentrum München GmbH, Neuherberg, Germany

Proton minibeam radiotherapy (pMBRT) is a spatial fractionation method that widens the therapeutic window. By reducing radiation toxicities in healthy tissue it allows the hypofractionation with high doses. Combining temporal fractions with spatial fractionation schemes raises questions of reirradiation accuracies due to the beam pattern.

In this study, healthy mouse ears were irradiated with four fractions with a mean dose of 30 Gy applying Gaussian proton minibeams of σ=222 μm in three different schemes (FS). Mice were followed for 150 days after course completion to assess acute and late side effects. A pattern of 16 minibeams with center-to-center (ctc) distances of 1.8 mm delivered to the same position in each fraction (FS1) induced the least tissue response. A pattern shift by ctc/2 between the fractions (FS2) led to a stronger inflammation than FS1. The strongest tissue response was measured for FS3 where 64 positions with a ctc of 0.9 mm were irradiated. FS3 also significantly increased fibrosis, a radiation induced late side effect. The total dose patterns of FS2 and FS3 coincided but differed in their temporal fractionation.

The findings show that temporal fractionated proton minibeams have the potential to spare healthy tissue compared to fractionated homogeneous x-ray irradiation. The greatest benefit is achieved by an accurate reirradiation of the minibeam pattern (FS1) with only minor skin toxicities. A shifted reirradiation of the pattern (FS2) significantly reduced skin toxicity compared to a pattern which only differed by the temporal fractionation (FS3) indicating a fast “tissue repair” mechanism between daily fractions.

O 64

First evaluation of temporally fractionated proton minibeam radiation therapy in glioma-bearing rats

Y. Prezado1, A. Patriarca2, M. Juchaux3, D. Labiod4, R. Ortiz-Catalan1, D. Crepin3, R. Dendale2, F. Pouzoulet4, L. De Marzi2

2Institut Curie, Orsay proton therapy center, Orsay, France
3CNRS, Imagerie et Modelisation en Neurobiologie et Cancerologie, Orsay, France
4Institut Curie, Experimental Radiotherapy platform, Orsay, France

Proton minibeam radiation therapy (pMBRT) is a novel therapeutic strategy that uses spatially modulated narrow proton beams [1]. pMBRT has proven to widen the therapeutic window for high-grade gliomas in rodents [2,3]. The experiments performed so far used high doses (> 25 Gy) in one single fraction. In this work we investigated the tumor control effectiveness when a temporal fractionation of the dose is used. With that aim we irradiated two groups of RG2 glioma bearing rats: one group received pMBRT in one fraction (30 Gy mean dose), lateral irradiation, n=9; the second group received a crossfired irradiation: one lateral irradiation the first day (15 Gy mean dose) and a craneo-caudal irradiation 48 h later (15 Gy mean dose), n=7. See figure 1 left. The survival curves are shown in figure 1 right. A significant increase of mean survival time with respect to the controls was observed in the two groups. A percentage of 30 % long-term survivals free of tumor was obtained in the group that received one unique fraction, whereas no long-term survivals were observed in the group receiving two fractions. Analysis are ongoing to assess the impact on normal tissues. Our preliminary results seem to favour the use of a unique fraction. Future perspectives include the evaluation of other temporal schemes and configurations.

References: [1] Prezado et al., Med. Phys., 2013. [2] Prezado et al., Scie. Reports, 2018. [3] Prezado et al., Radiat. Oncology, 2018.

O 65

The TOPAS-nBio project – status and outlook

J. Schuemann1, A. McNamara1, J. Ramos-Mendez2, J. Perl3, K.D. Held1, S. Incerti4, H. Paganetti1, B. Faddegon2

1Massachusetts General Hospital and Harvard Medical School, Rad. Onc., Boston, USA
2University of California San Francisco, Radiation Oncology, San Francisco, USA
3SLAC National Accelerator Laboratory, n/a, Menlo Park, USA
4IN2P3- Cenbg, Cnrs, Gradignam, France

The goal of the TOPAS-nBio project is to provide intuitive nanometer scale Monte Carlo (MC) simulations for radiobiology experiments that do not require programming expertise. TOPAS-nBio thereby aims to foster interdisciplinary work between biologists, chemists and physicists with interest in radiobiology.

Since the open-source release of TOPAS-nBio in 2019, the framework offers to connect energy deposition within irradiated cells (physics) via molecular reactions (chemistry) to cell kill/repair (biology) at the level of sub-cellular targets such as DNA. TOPAS-nBio is an extension to TOPAS and layered on top of the Geant4/Geant4-DNA MC toolkit. To facilitate the setup of simulations we developed a Graphical User Interface (to be released, see figure 1a).

Radiobiological effects at the cellular and sub-cellular scale can be modeled by simulating detailed biological geometries, such as various DNA models, mitochondria and cell shapes (e.g. fibroblasts or neurons) that have been developed as shown in figure 1b. Chemical reactions can be included using two implementations of chemistry, either via step-by-step modeling or using the independent reaction time (IRT) method. We reproduced measured time-dependent G-values in water within 7% for OH and e-aq and 50% for H2O2 as shown in figure 1c. Several biological endpoints were evaluated by including the MEDRAS model showing reasonable agreement with experimental data.

Overall, TOPAS-nBio promises to advance our understanding of the fundamental processes within a cell immediately after radiation induced damages. Further improvements are envisioned to the physics, chemistry, biology and functionality including expanding cell specificity and microenvironments.

O 66

The biological effect of carbon ion irradiation for two different conditions of human dermal fibroblast on various linear energy transfers

M. Nakata1, K. Minami2, T. Tsubouchi3, T. Kanai3, M. Koizumi1, K. Ogawa2

1Osaka University Graduate School of Medicine, Department of Medical Physics and Engineering, Osaka, Japan
2Osaka University Graduate School of Medicine, Department of Radiation Oncology, Osaka, Japan
3Osaka heavy ion Therapy Center, Radiation Physics, Osaka, Japan

Background: Spread-out Bragg peak (SOBP) is created by sequentially irradiating various energies of the incident particle beam with definite weighting factor, and it impacts tumor located in the depth of patient's body. We cannot completely protect normal tissues from particle irradiation so far. In clinic, only late damage reaction sometimes occurs without acute reaction on normal tissues. Therefore, it is essential to investigate the reaction of normal tissues to particle beam. We need firstly clarify the LET dependence of in the LQ model for normal tissues. And also the response of the normal tissue can be estimate in the frame work of the LQ model. In this study, we focused on human dermal fibroblast. We found the unique character of this cell line which was that two cell conditions could be created from this. One was logarithmic growth, another was intermitotic cell condition. We would like to talk about the biological effect of carbon ion irradiation for two different conditions of human dermal fibroblast on various linear energy transfers (LETs).

Methods and Materials: Human dermal fibroblast cell line (HDF) was irradiated by carbon ion beams using different LET (15keV, 60keV, 78keV) at Osaka heavy ion therapy center. Cell surviving fraction was evaluated by colony formation assay.

Results and conclusion: Cell surviving fractions decrease as LET increase for the same physical doses. Cell survivals of intermitotic cells increased more than that of logarithmic growth cells. These results suggested that biological effect of normal tissue cells was different among their conditions.

O 67

Toxic non-homologous end-joining (NHEJ) increases the effectiveness of protons in vivo in homologous recombination deficient triple negative breast cancer (TNBC)

R. Mutter1, Q. Zhou1, M. Howard1, J. Denbeigh1, N. Remmes1, Q. Zhu2, J. Yuan2, C. Beltran1, J. Sarkaria1, Z. Lou2

1Mayo Clinic, Department of Radiation Oncology, Rochester, USA
2Mayo Clinic, Department of Oncology, Rochester, USA

Background: TNBC is enriched for mutations in BRCA1, BRCA2, and other homologous recombination (HR) elements. In vitro studies have suggested that some HR deficient cells may be sensitive to protons. Whether the proton relative biologic effectiveness (RBE) of certain repair deficient tumors is greater in vivo is not known.

Methods: Isogenic TNBC cell lines, BRCA1 mutant (BRCA1mut) or wild-type BRCA1 complemented (BRCA1wt) MDA-MB-436 cells were injected into the neck of nude mice. Following tumor formation, mice were randomized to sham or 5 Gy administered with 6 MV photons, proton pencil-beam scanning (PBS) through beam (PBSTB, linear energy transfer [LETd] 1.1 keV/μm) or proton PBS LET optimized Bragg peak PBS (PBSBP, LETd 3.9 keV/μm). Mechanism of RBE differences were investigated using colony formation, immunoblot, RNA interference and immunofluorescence in MDA-MB-436 and BRCA2 wild-type (V79) and mutant (VC8) Chinese hamster isogenic lines, BT549, and U2OS.

Results: Median time to tumor tripling was significantly greater with PBSBP compared with PBSTB or photons in BRCA1mut, but not BRCA1wt MDA-MB-436 xenograft models (figure). The RBE of PBSBP was also increased in MDA-MB-436 BRCA1mut ( 1.75±0.21) and BRCA2 mutant VC8 ( 1.67±0.11) compared to their wild-type isogenic counterparts, in colony formation assays. Upregulation of pChk1 and BrdU, pRPA, and RAD51 foci was observed following PBSBP, suggesting greater activation of HR-dependent resection machinery. Knockdown of XRCC4, a key NHEJ component, or treatment with a DNAPK inhibitor reduced the RBE of PBSBP to ∼1.

Conclusion: Toxic end-joining increases the RBE of clinically administrable PBSBP in HR deficient TNBC.

O 68

Introducing the periventricular region as a novel organ at risk in proton treatment planning of gliomas

J. Eulitz1,2, F. Raschke1,3, E.G.C. Troost1,2,3,4,5, J. Thiele2, S. Makocki2, S. Menkel2, S. Appold2, W. Enghardt1,2,3,4,5, M. Krause1,2,3,4,5, A. Lühr1,2,3,4,5

1OncoRay - National Center for Radiation Research in Oncology- Faculty of Medicine and University Hospital Carl Gustav Carus- Technische Universität Dresden, Helmholtz-Zentrum Dresden -Rossendorf, Dresden, Germany
2Department of Radiotherapy and Radiation Oncology- Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Dresden, Germany
3Helmholtz-Zentrum Dresden - Rossendorf, Institute of Radiooncology – OncoRay, Dresden, Germany
4German Cancer Consortium DKTK- Partner Site Dresden, and German Cancer Research Center DKFZ- Heidelberg- Germany, Dresden, Germany
5National Center for Tumor Diseases NCT- Partner Site Dresden- Germany: German Cancer Research Center DKFZ- Heidelberg- Germany- Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden- Dresden- Germany- and- Helmholtz Association / Helmholtz-Zentrum Dresden - Rossendorf HZDR, Dresden, Germany

The periventricular region (PVR) has been shown to have an increased susceptibility to dose-dependent brain injury after proton therapy of gliomas (Eulitz 2019, Harrabi 2019). However, the PVR has not been considered in proton treatment planning so far. Here, we present an approach for incorporating the PVR as a novel organ at risk in proton treatment planning.

The PVR was defined as a 4 mm uniformly expanded margin around the ventricular system. The PVR tolerance dose was estimated as the observed near-min dose derived within late post-treatment radiation-associated contrast-enhancements (CE) on T1wCE-follow up images. All grade II/III glioma patients treated between 2014 and 2018 with (adjuvant) proton radio(chemo)therapy to 54/60 Gy(RBE), respectively, were analyzed. Retrospective PVR-sparing treatment planning was performed for 11 (4/7) consecutive glioma patients (Fig.1). Patient-wise comparison to the clinically delivered plans considered the near-max dose D2% and the volume Vx that received more than x% of prescribed dose in PVR tissue outside the target boost volume.

The median distance of the observed CE lesion centers to the cerebral ventricles was 2.5 mm. The median and near-min CE lesion dose was 57.6 Gy(RBE) and 53.0 Gy(RBE), respectively. PVR-sparing treatment planning reduced D2% and V90% by 4.3 (±1.4)% and 53.2 (±22.4)%, respectively, while maintaining clinical goals (Fig.1).

PVR-sparing in treatment planning is a promising approach to reduce late brain injury. In many cases this should be possible without compromising target coverage and field homogeneity. Further experience will improve the arrangement of PVR-sparing in the priority order of planning goals.

O 69

Prospective toxicity analysis of reirradiation with proton therapy for central nervous system tumors: a proton collaborative group study

S. Hasan1, R.H. Press2, J.I. Choi2, L. Halasz3, C. Vargas4, H. Tsai5, M. Mishra6, J. Chang7, W. hartsell8, C. Simone2

1New York Proton Center, Radiation Oncology, New York, USA
2New York Proton center, Radiation Oncology, New York- NY, USA
3University of Washington, Radiation Oncology, Seattle- WA, USA
4Mayo Clinic, Radiation Oncology, Rochester- MN, USA
5ProCure Proton Center New Jersey, Radiation Oncology, Somerset- NJ, USA
6University of Maryland, Radiation ONcology, Baltimore- MD, USA
7ProCure Proton Center Oklahoma City, Radiation Oncology, Oklahoma City- OK, USA
8Northwestern University, Radiation Oncology, Chicago- IL, USA

Introduction: We analyzed the acute toxicities of reirradiation with proton therapy (PT) for central nervous system (CNS) tumors using the Proton Collaborative Group (PCG) registry.

Methods: The multi-institutional, prospectively collected PCG registry was queried for CNS tumors treated with PT reirradiation between 2010-2019. Acute grade 2 (G2) and grade 3 (G3) toxicities were reported, with binomial regression analysis to identify correlates thereof.

Results: Overall, 119 consecutive patients 19-80 years old (median 48) were identified, including 69 males and 50 females with 36 benign tumors, 61 gliomas, and 22 medulloblastoma/ependymomas/neuroendocrine tumors, located in cerebral hemispheres (n=77), infratentorium (n=21), base of the skull (n=14), and spinal cord (n=8). Patients had either a single (n=103), two (n=13), or 3 courses of prior radiation, and median time to PT reirradiation was 63 months. Median PT dose (EQD2) was 50 Gy10 (13 – 66 Gy). Median cumulative EQD2 dose was 103 Gy10 (51-210 Gy10). Chemotherapy was given with PT in 48 patients. Baseline ECOG was 0 (n=40), 1 (n=38), and 2+ (n=32). Median follow-up was 10 months. Acute G2 and G3 toxicities occurred in 52.9% and 9.2% of cases, respectively. Twelve patients had G3 symptoms at baseline, and all but one resolved after PT. Correlates of G3 toxicity include concurrent chemotherapy (HR=8.94, P=0.008), ECOG 2+ (HR=11.9, P=0.03), and cumulative EQD2 > 114 Gy10 (HR=3.5, P=0.05).

Conclusion: Higher cumulative radiation doses and receipt of chemotherapy in the salvage setting correlated with more G3 toxicity, but in appropriately selected patients reirradiation with PT for CNS tumors is well tolerated.

O 70

Outcomes after proton therapy reirradiation for malignant glioma: prospective analysis from the proton collaborative group

R. Press1, S. Hasan1, J.I. Choi1, L.M. Halasz2, C.E. Vargas3, H. Tsai4, M.V. Mishra5, J.H. Chang6, W.H. Hartsell7, C.B. Simone 2nd1

1New York Proton Center, Department of Radiation Oncology, New York, USA
2University of Washington, Department of Radiation Oncology, Seattle, USA
3Mayo Clinic, Department of Radiation Oncology, Phoenix, USA
4ProCure Proton Therapy Center, Department of Radiation Oncology, Somerset, USA
5University of Maryland School of Medicine, Department of Radiation Oncology, Baltimore, USA
6Procure Proton Therapy Center, Department of Radiation Oncology, Oklahoma City, USA
7Northwestern Medicine Chicago Proton Center, Department of Radiation Oncology, Warrenville, USA

Introduction: There are limited data on proton therapy reirradiation (PT-ReRT) for malignant glioma.

Methods: The prospective, multi-institutional Proton Collaborative Group (PCG) registry was queried for patients with malignant glioma who previously received radiotherapy (RT) and underwent PT-ReRT between 7/2011-10/2018. Patient/treatment details were assessed. Kaplan-Meier method was used to estimate overall survival (OS) and local control (LC).

Results: Sixty-one consecutive patients were identified. Median age was 45 years (19-75). Median follow-up was 7.8 months (1-37). Histology included oligodendroglioma (n=27), astrocytoma (n=21), glioblastoma/gliosarcoma (n=4), glioma-NOS (n=5), and pilocytic astrocytoma (n=4). Patients had 1 (n=52) or 2+ (n=9) prior RT courses. Median EQD2 of prior course(s), PT-ReRT, and cumulative RT were 57.0 Gy (38.8-158.4), 48.9 Gy (30.4-60.0), and 105.5 Gy (73.6-210.2), respectively. Median time between most recent RT courses was 60.5 months (1-440). For patients with >3 months follow-up (n=38), median, 1-year, and 2-year OS were 14.6 months, 62%, and 24%, respectively. OS was associated with glioblastoma (Hazard ratio [HR] 1.41, p=0.039) and pilocytic histology (HR 2.02, p=0.017), age (HR 0.96, p=0.045) and time from prior RT (HR 9.41, p<0.001). Median, 6-month, and 1-year LC was 9.2 months, 73%, and 39%, respectively. Acute grade 3+ toxicities occurred in 9 (15%) patients. No acute grade 4+ events occurred. Cumulative EQD2 >114 Gy trended towards increased LC (HR 0.48, p=0.09) and higher acute grade 3 toxicity (HR 6.15, p=0.07).

Conclusion: PT-ReRT is well tolerated with low rates of high-grade acute toxicities. Cumulative dose may be associated with LC and acute toxicity.

O 71

Skull base Chordomas and Chondrosarcomas in adolescents and young adults treated with pencil beam scanning proton therapy

P.S. Lim1, S. Tran2, S.G.C. Kroeze3, A. Pica4, J. Hrbacek4, B. Bachtiary4, M. Walser4, A.J. Lomax4, D.C. Weber4

1University College London Hospitals, Clinical Oncology, London, United Kingdom
2Geneva University Hospital, Radiation Oncology, Geneva, Switzerland
3University Hospital Zürich, Radiation Oncology, Zürich, Switzerland
4Paul Scherrer Institute, Centre for Proton Therapy, Villigen, Switzerland

Background: Skull-base chordomas(Chs) and chondrosarcomas(ChSas) in adolescents and young adults(AYAs) are rare and not well-documented. We aim to assess the outcomes, prognostic factors(PF) and employment status of AYAs with Chs/ChSas treated with pencil-beam-scanning proton-therapy(PT).

Methods: 108 patients with Chs(n=58) and ChSas(n=50) treated with PT between October 1998 and July 2017 at the Paul Scherrer Institute were analysed.

Results: Median age was 30 years(range, 15-39). The median prescribed dose was 74 and 70 GyRBE for Chs and ChSas, respectively. With a median follow-up of 86 months(range, 12-236), the 7-year local control(LC) rate was 73.7%(95%CI 61.1-86.2) for Chs and 93.6%(95%CI 86.5-100) for ChSas. The 7-year distant metastasis-free survival and overall survival(OS) was 93.1%(95%CI 85.3-100) and 87.2%(95%CI 77.6-96.8) for Chs; 100% and 95.5%(95%CI 89.4-100) for ChSas respectively. Negative PF for LC were chordoma histology(p=0.019), recurrent disease(p=0.004) and optic apparatus compression at PT(p=0.002). OS was negatively influenced by recurrent disease(p=0.009), optic apparatus compression at PT(p=0.002) and brainstem compression at PT(p=0.045). The 7-year ≥G3 PT-related toxicity-free-survival was 85.3%(95%CI 77.3-93.3). The unemployment rate was 8.3% at PT, increasing to 31% at last follow-up post-PT.

Conclusions: PT for AYAs with Chs and ChSas result in good tumour local control with acceptable long-term toxicity. Recurrent disease and optic apparatus compression at PT are significant negative PF for both LC and OS. Focus must be given to develop strategies in improving psychosocial support for AYA survivors such as workplace re-engagement, increasing awareness amongst professionals and patient empowerment.

O 72

Carbon ion radiotherapy for sacral chordoma

R. Imai1, H. Tsuji2

1National Institutes for Quantum and Radiological Science and Technology, QST Hospital, Chiba, Japan
2National Institutes for Quantum and Radiological Science and Technology, QST Hospital, 4-9-1 Anagawa Inage-Ku- Chiba, Japan

Objective: The oncologic and functional outcomes of carbon ion radiotherapy for sacral chordoma were evaluated.

Methods: Between 2000 and 2016, 235 sacral chordoma patients treated with carbon ion radiotherapy alone at a single institute were analyzed retrospectively.

Results: Of 235 patients, median age was 66 years old ranging 26 to 89 years old. The study group consisted of 75 females and 160 males. Median overall survival time was 75 months and 5 % of all patients were surviving but followed for less than 5 years. Five-year overall survival rate was 84 %. In 77 deceased patients 32 % died of another disease. The tumor location of the cranial level reached at L5 in 9 patients, S1 in 58, S2 in 87, S3 in 61 and S4 in 20. Median tumor volume was 321 cm3 ranging 31 to 1497 cm3. Five-year local control rate was 74 %. Regarding adverse events 2 % of patients had grade 3 neuropathy and 1 patient had grade 3 skin ulcer. Excluding patients who could not walk at the first visit, in 99 % of patients their ambulatory remained.

Conclusions: Carbon ion radiotherapy alone was a safe and effective treatment for sacral chordoma.

O 73

Long-term outcomes and toxicities of over 100 patients prospectively treated with proton therapy for chordoma: the proton collaborative group experience

S. Rice1, A. Holtzman2, J. Chang3, L. Halasz4, H. Tsai5, L. Rosen6, C. Stevens7, W. Hartsell8, C. Vargas9, C.B. Simone II10

1State University of New York Medical Center, Radiation Oncology, Syracuse, USA
2University of Florida, Proton Therapy Insitute, Jacksonville, USA
3Oklahoma Proton Center, Radiation Oncology, Oklahoma City, USA
4University of Washington, Radiation Oncology, Seattle, USA
5ProCure Proton Therapy Center, Radiation Oncology, Somerset, USA
6Willis Knighton Cancer Center, Radiation Oncology, Shreveport, USA
7Beaumont, Radiation Oncology, Royal Oak, USA
8Proton Therapy Center- Northwestern University, Radiation Oncology, Chicago, USA
9Mayo Clinic Scottsdale, Radiation Oncology, Scottsdale, USA
10New York Proton Center, Radiation Oncology, New York City, USA

Background: We present chordoma outcomes from the Proton Collaborative Group prospective registry.

Materials and Methods: All consecutive chordoma patients treated between 2010-2018 were evaluated. Of the 150 patients identified, 101 had adequate follow-up information for inclusion. Location included base of skull (n=60, 44%), spine (n=23, 17%), and sacrum (n=15, 11%). Median age was 59 years (range 18-92) with a performance status of 0-1 (91%). Sixty-two percent of patients underwent surgical resection. The median proton RT dose was 74 CGE (range 22-86 CGE, IQR=5.45) using passive scatter proton RT (PS-PRT) (59%) and pencil beam scanning proton RT (PBS-PRT) (49%). Rates of local control (LC), progression-free survival (PFS), overall survival (OS) and toxicities were assessed.

Results: 2/5-year LC, PFS, and OS rates are 97%/87%, 90%/47%, and 92%/54%, respectively. LC did not differ based on surgical resection (p=0.61). Twelve patients experienced acute grade 3 toxicities, most commonly pain (n=5), radiation dermatitis (n=3), fatigue (n=1) and dysphagia (n=1). No grade >/=4 acute toxicities were reported. Late toxicities were experienced in 28 patients, most commonly grade 1-2 fatigue (n=13), muscle weakness (n=4), and dysphagia (n=4). Grade 2 CNS necrosis occurred in one patient who received 75 CGE after gross total resection with close margins. No grade >/=3 late toxicities were reported.

Conclusions: Chordoma proton RT shows excellent safety and efficacy outcomes with very low rates of treatment failure. Necrosis is exceedingly low (<1%) despite the high doses of PRT delivered. Further maturation of data and larger patient numbers are necessary to study differences in PS-PRT versus PBS-PRT.

O 74

Imaging quinine sulfate radioluminscence for proton beam dosimetry

M. Rahman1, P. Bruza1, R. Zhang2, Y. Lin3, A. Stanforth3, J.R. Gunn1, D.J. Gladstone1,2, B.W. Pogue1,4

1Dartmouth College, Thayer School of Engineering, Hanover- NH, USA
2Dartmouth Hitchcock Medical Center, Radiation Oncology, Lebanon- NH, USA
3Emory University, Emory Proton Therapy Center, Atlanta- GA, USA
4DoseOptics, Limited Liability Company, Lebanon- NH, USA

Quinine Sulfate (QS) in water can quantify dose accurately for clinical x-ray beams due to isotropic fluorescence emission. The fluorophore was previously investigated for Cherenkov excited fluorescence, but a recent study demonstrated the fluorescence is predominantly x-ray stimulated. Minimal Cherenkov radiation is produced in clinical proton beams, but the scintillation properties of QS motivated investigation of proton radiation stimulated emission for accurate dosimetry. Scintillation yield dependency on concentration (0.5-2g/L) was explored. An intensified CMOS camera imaged 2g/L QS in an acrylic tank phantom, while irradiated by a proton pencil beam. Images were deconvolved with the point spread function of the detector system and corrected for linear energy transfer (LET) dependent quenching. Quenching parameters were determined from Monte Carlo (MC) simulated LET, utilizing Birks' Law. Gamma analysis comparing MC dose and optical corrected image showed imaging QS accurately quantifies dose. Quenching correction of optical images can produce 2D projection dose maps for all clinical proton beam energies. The linearity and increase in scintillation with increasing concentrations (R^2=0.996) justifies further investigation into scintillation yield and quenching parameter with higher concentrations. The quenching corrected optical images can potentially be extended to provide accurate 3D dose of proton beams.

O 75

Absorbed dose to water measurements in the SOBP of a clinical carbon-ion beam using water calorimetry

K.M. Holm1,2, U. Weber3, Y. Simeonov4, O. Jäkel2,5, S. Greilich2, A. Krauss1

1Physikalisch-Technische Bundesanstalt PTB, Department of Dosimetry for Radiation Therapy and Diagnostic Radiology, Braunschweig, Germany
2German Cancer Research Center DKFZ, Division of Medical Physics in Radiation Oncology, Heidelberg, Germany
3GSI Helmholtzzentrum für Schwerionenforschung GmbH, Department for Biophysics, Darmstadt, Germany
4University of Applied Sciences, Institute of Medical Physics and Radiation Protection, Giessen, Germany
5University of Heidelberg, Heidelberg Ion Beam Therapy Center HIT, Heidelberg, Germany

Ionization chamber-based dosimetry for carbon-ion beams shows a significantly higher uncertainty than for high-energy photons. This is mainly caused by the high standard uncertainty of the correction factor for beam quality k_Q, of which the values are based on theoretical calculations due to a lack of experimental data.

To reduce this uncertainty, k_Q factors have been determined experimentally for the entrance channel of a monoenergetic carbon-ion beam [1] by means of water calorimetry and are now being determined for the spread-out Bragg peak (SOBP). Due to heat dispersion effects, water calorimetric measurements are time-critical; a homogeneous dose distribution irradiated in a short time is needed for a low overall uncertainty in the experiment. Therefore, we created a suitable irradiation technique using a 2D range modulator (Fig. 1) to modulate the monoenergetic beam in depth. Thus, only a lateral scanning of the field is needed, which dramatically decreases the irradiation time compared to a fully active scanned field. A scanning time of 90 s for a 6x6x6 cm^3 dose cube of 1.5 Gy and a field homogeneity with a standard uncertainty of measured dose values below 1% was achieved.

For the given irradiation parameters, initial water calorimetric measurements showed good results concerning new k_Q factors. Further measurements will be performed to determine reliable k_Q factors for the SOBP for the first time.

References: [1] OSINGA-BLAETTERMANN et al., Phys. Med. Biol. 62 (2017), 2033–2054

O 76

Quantifying measurement uncertainty in microdosimetry for carbon ion radiotherapy

S. Hartzell1, F. Guan1, O. Vassiliev1, P. Christine2, P. Taylor1, S. Kry1

1UT MD Anderson Cancer Center, Radiation Physics, Houston, USA
2UT MD Anderson Cancer Center, Quantitative Science, Houston, USA

Microdosimetry is an important tool in assessing the mixed radiation field produced in carbon radiotherapy, and serves as a direct input parameter in to at least two common models of relative biological effectiveness (Microdosimetric Kinetic Model and Repair Misrepair Fixation model). Various detectors can be used to make microdosimetric measurements, with the most common being the Tissue-Equivalent Proportional Counter (TEPC). TEPC measurements are subject to various inherent sources of uncertainty resulting from the detector design and beam setup. This work quantifies the effect of physical uncertainty on fundamental microdosimetric quantities made by TEPC measurements in a therapeutic carbon beam.

Microdosimetric spectra and corresponding lineal energy values were calculated using Monte Carlo (GEANT4) for five monoenergetic and three SOBP carbon beams of clinical energies. The impact on the microdosimetric spectra from eight unique sources of uncertainty associated with TEPC measurements were simulated: wall effects, pulse pile-up, electronic uncertainty, gas pressure, W-value, gain instability, low energy cut-off, and counting statistics. These sources were quantified by statistically introducing each into the simulated spectra 200 times and sampling the resultant data.

Both the variance and the bias in the dose-, frequency- and saturation corrected dose mean lineal energies were quantified. The overall uncertainty was typically less than 3% and was, at most, 6% (1-sigma). The individual contribution by each source is displayed below.

This study quantified uncertainty associated with microdosimetric measurements made with a TEPC to establish an error budget for beam measurements. Means of reducing error for clinical use in RBE verification were also identified.

O 77

Dosimetry for advanced radiotherapy approaches using particle beams with ultra-high pulse dose rates in the EMPIR UHDpulse project

A. Subiel1, M. McManus1, F. Romano1,2, N. Lee1, H. Palmans1,3, W. Farabolini4,5, A. Gilardi4, A. Schueller6

1National Physical Laboratory, Medical Physics, Teddington, United Kingdom
2Istituto Nazionale di Fisica Nucleare, Sezione di Catania, Catania, Italy
3MedAustron, Proton Therapy Center, Wiener Neustadt, Austria
4CERN, CLEAR Facility, Geneva, Switzerland
5CEA, IRFU Department, Gif-sur-Yvette, France
6Physikalisch-Technische Bundesanstalt, High-Energy Photon and Electron Radiation, Braunschweig, Germany

The European Metrology Programme for Innovation and Research (EMPIR) UHDpulse project aims to develop the metrological tools needed to establish traceability in absorbed dose measurements of particle beams with ultra-high intra-pulse dose-rates. Advances in technology stimulated novel research ideas exploring new means of delivering radiotherapy. Delivery of doses at ultra-high dose-rate has been of particular interest due to remarkable reduction of normal tissue toxicity (known as the FLASH effect) with respect to conventional treatments. It is not clear which are the optimal beam parameters responsible for enhancing this effect, however there is evidence that beam parameters such as dose-per-pulse, instantaneous dose-rate and the number of intra-pulses delivered can considerably impact therapeutic outcomes. Pulses with dose-rates orders of magnitude higher than in conventional radiotherapy present significant metrological challenges, which need to be addressed to enable the translation of these novel radiotherapy techniques to clinical practice. In dosimetry of high intra-pulse dose-rate beams it is important to understand the effects that influence detector response. Ionization chambers, used routinely in radiotherapy, exhibit significant recombination effects with increased dose-rates. Within the framework of this EMPIR project we investigated the relationship between the ion recombination factor of Roos ionisation chamber and a wide range of dose-per-pulse values ranging from 0.03 to 5.26 Gy/pulse for a 200 MeV electron beam. The chamber measurements (at 200V) were compared to NPL's graphite calorimeter to determine ion recombination correction factors. The corresponding collection efficiencies for the lowest and highest dose-per-pulse were found to be 97% and 10%, respectively.

O 78

Ultra-high dose rate (FLASH) proton beam dosimetry using a high resolution 2D transmission ion chamber

J.W. Zou1, E. Diffenderfer1, K. Ota2, P. Boisseau2, J. Konzer3, C. Koumenis1, K. Cengal1, J. Metz1, L. Dong1, B.K. Teo1

1University of Pennsylvania, Radiation Oncology, Philadelphia, USA
2Pyramid Technical Consultants, Systems Engineering, Boston, USA
3IBA PT- Inc, PT Engineer-Beam Physics, Louvain-La-Neuve, Belgium

Charged particle radiotherapy under FLASH dose rate has been reported to reduce normal tissue toxicity compared to radiation at conventional dose rates. A critical component of flash radiotherapy is accurate dosimetry under very high dose rates. In this study we characterized a newly designed high-resolution position sensing transmission ionization chamber with a fast-response electrometer for flash proton radiotherapy dosimetry.

The 2D detector is a 6x6cm ultra-thin (1mm air gap) ion chamber with 64 channels readout per axis (IC64-6, Pyramid Technical Consultants). The dose and positioning accuracy of the ion chamber were first characterized using conventional PBS system. For FLASH experiment, the cyclotron beam spot current varied from conventional to flash dose rate as high as 350nA. The charges collected by the ion chamber were compared with measurements from a Faraday cup. The ion recombination effect under various bias voltages was characterized for the flash proton beam. The positions as well as beam current of the beam spots were digitized using 100μs sampling interval.

The ion chamber reading demonstrated excellent linearity with the collected charge of the Faraday cup from 5-350nA cyclotron current. At high FLASH dose rates (>200 Gy/s), the correction factor for ion recombination (Pion) was less than 1.002 when the bias voltage was at least 200Volts. The spot position measurement variations was 0.10±0.08mm in two orthogonal directions.

We characterized a 2D transmission ion chamber system with a fast integrated electrometer and demonstrated its capability to measure both proton dose and beam spot position for conventional or FLASH dose rates.

O 79

Calorimetry techniques for absolute dosimetry of proton beams with ultra-high pulse dose rates

F. Romano1,2, S. McCallum3, N. Lee2, A. Subiel2, M. Borghesi3, H. Ahmed3, A. McIlvenny3, G. Milluzzo3, H. Palmans2,4, R. Thomas2

1Istituto Nazionale di Fisica Nucleare, Sezione di Catania, Catania, Italy
2National Physical Laboratory, CMES - Medical Radiation Science, Teddington, United Kingdom
3Queen's University Belfast, Centre for Plasma Physics, Belfast, United Kingdom
4EBG MedAustron GmbH, Medical Physics Group, Wiener Neustadt, Austria

Over the past few years, advances in proton accelerators, such as synchrocyclotrons and laser-based accelerators, have led to the use of irradiation modalities characterized by ultra-high dose-rates. The use of an almost “instantaneous” delivery of a radiation dose has indicated a decrease in undesired effects to healthy tissues whilst still preserving the required tumour control achieved in conventional modalities, thanks to the so-called FLASH effect. Moreover, proton and ion beams accelerated by high-intensity lasers, characterized by even larger dose-rate per pulse, are very promising due to the high fields and compactness of plasma accelerators in comparison with conventional accelerators, with potential advantages in terms of overall cost and compactness.

Accurate dosimetry of high dose-rate particle beams is challenging and requires the development of alternative dosimetric approaches with respect to those used for conventional radiotherapy. A portable graphite calorimeter has been developed by the National Physical Laboratory (NPL) in the UK, with the aim of developing absolute dosimetry for high dose per pulse proton beams, in the framework of the European Metrology Programme for Innovation and Research (EMPIR) UHDpulse project. This prototype calorimeter has been operated for the first time in the ultra-high dose-rate laser-driven proton beam (up to 40 MeV) produced by the PW Vulcan Laser of the Central Laser Facility at the Rutherford Appleton Laboratory in the UK. Despite the harsh experimental environment and the presence of large electromagnetic pulses (EMP), a very good signal-to-noise ratio was achieved, with an average delivered dose per pulse of up to 3 Gy.

O 80

Can prompt-gamma-based verification detect anatomical changes in PT? First systematic clinical investigation

J. Berthold1,2, A. Jost1,3, C. Khamfongkhruea1, J. Petzoldt4, J. Thiele5, T. Hölscher5, P. Wohlfahrt1,2, G. Janssens4, J. Smeets4, C. Richter1,2,5,6

1OncoRay - National Center for Radiation Research in Oncology, High Precision Radiotherapy, Dresden, Germany
2Helmholtz-Zentrum Dresden - Rossendorf, Institute of Radiooncology - OncoRay, Dresden, Germany
3Beuth-Hochschule für Technik, Department of Mathematics-Physics-Chemistry, Berlin, Germany
4Ion Beam Application SA, Louvain-la-Neuve, Belgium
5Department of Radiotherapy and Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus- Technische Universität Dresden, Dresden, Germany
6German Cancer Consortium DKTK, partner site Dresden- Germany and German Cancer Research Center DKFZ, Heidelberg, Germany

Introduction: Anatomical changes during proton therapy can cause severe dosimetric deviation. Treatment verification is thus highly desirable. Here, we present the first systematic evaluation of the sensitivity of a Prompt-Gamma-Imaging (PGI) based range verification system to detect anatomical changes in prostate-cancer treatments.

Materials and Methods: Spot-wise range deviations were monitored with a PGI slit camera during in total 16 fractions of hypo-fractionated Pencil-Beam-Scanning (PBS) prostate-cancer treatments (2 patients, 2 fields, each 1.5GyE). For all monitored fractions, in-room control-CT scans were acquired, serving as ground-truth reference for the identification and scoring of anatomical changes (strong/moderate/light). The sensitivity to detect these changes was determined for both, clinically measured and simulated PGI-data, respectively: For distal PBS spots, expected shifts, determined from line-dose profiles (planning-CT vs. control-CT), were manually compared with PGI-derived spot-wise shifts (Fig.1). Furthermore, a simple two-parametric model was established to classify each monitored field into scenarios of global, local and no-clinically-relevant anatomical changes.

Results: Overall 66% (84%) of the 64 detected anatomical changes were identified from measured (simulated) PGI-data (Fig.2a). All strong changes (14/64) were identified correctly. The first attempt for automated field-wise classification was able to correctly classify most global changes (9/11). However, differentiation between non-relevant from local changes seemed more difficult (4/6 and 7/14 fields classified correctly, respectively); but even ground-truth classification was often borderline in those cases (Fig.2b).

Conclusion: In the first systematic investigation of the sensitivity of clinical PGI-based treatment verification, its capability to detect strong anatomical changes has been clearly demonstrated. Moving towards automated interpretation of PGI-data, a simple two-parametric model already showed encouraging results.

O 81

New prompt gamma-ray detection system to achieve full three-dimensional range verification in pencil beam scanning proton therapy

C. Panaino1,2, R.I. Mackay1,2,3, K.J. Kirkby1,2, M.J. Taylor1,2

1The University of Manchester, Division of Cancer Sciences, Manchester, United Kingdom
2The Christie NHS Foundation Trust, Proton Therapy Research, Manchester, United Kingdom
3The Christie NHS Foundation Trust, Christie Medical Physics and Engineering, Manchester, United Kingdom

We report the first results of a new method to reconstruct the proton beam range in 3D, through the detection of prompt gamma (PG) rays.

The system is comprised of 16 LaBr3(Ce) detectors. The position reconstruction capability of the system is investigated by means of Geant4 simulations. To reconstruct the PG-rays emission positions, the information recorded by each detector is fed into an in-house developed reconstruction algorithm. Both a spectrometer, with realistic energy and temporal resolution, and a water phantom were modelled. The beam stops in the phantom within the spectrometer central volume.

Figure 1a shows the detector/algorithm performance when a 180 MeV clinical pencil beam impinges the phantom. An excellent agreement is observed between the algorithm-reconstructed PG-rays emission positions (blue curve), the phantom-scored PG-rays emission positions (red dashed curve) and the dose distribution (black dot-dashed curve). This analysis was repeated for two additional internal radii (15 and 25 cm). Subsequently a 175 and a 177.5 MeV beams were shot on the phantom; this translates into a 5 / 10 mm range undershoot. Figure 1b depicts the algorithm-reconstructed PG-rays emission positions for the 175 (purple dot-dashed curve) and 177.5 MeV (green dashed curve) beams. The plot for a 180 MeV beam is repeated for comparison. The lateral spread and the centroid of the algorithm-reconstructed PG-ray distribution are reported, for all evaluations, in Table 1.

A new range verification technique is developed. A prototype is under construction. The final goal is a clinically compliant system for on-line, real-time verification.

O 82

Clinical results of in-vivo treatment verification with the INSIDE in-beam PET scanner

E. Fiorina1, M.G. Bisogni2, P. Cerello1, V. Ferrero1, C. Luongo3, E. Malekzadeh4, M. Morrocchi5, F. Pennazio1, G. Sportelli2, V. Vitolo6

1INFN, Sezione Torino, Torino, Italy
2Università di Pisa and INFN, Physics Department, Pisa, Italy
3Università di Pisa, IT Department, Pisa, Italy
4Tarbiat Modares University, Medical Physics Department, Teheran, Iran Islamic Republic of
5INFN, Sezione Pisa, Pisa, Italy
6Fondazione CNAO, Clinical Unit, Pavia, Italy

The INSIDE clinical trial, ongoing at the CNAO centre, aims at performing online in-vivo treatment verification in particle therapy ( NCT03662373).

Since July 2019 patients affected by selected head-and-neck and brain pathologies, have been monitored with the INSIDE bi-modal system (Fig.1) to assess the irradiation quality during the treatment course and to promptly detect possible interfractional morphological changes.

In this contribution, the results on the first set of 20 patients recruited before Spring 2020 will be reported.

The INSIDE system assesses the actual particle range with an in-beam Positron Emission Tomography (PET) scanner, that reconstructs positron emitter distributions, and a tracker for secondary protons emitted during carbon ion treatments, named Dose Profiler. The trial endpoint is to test the clinical performance of this innovative instrumentation as a tool for optimizing particle therapy and improving treatment plan robustness.

The in-beam PET images (Fig.2), acquired during proton and carbon ion irradiations, are analysed to detect possible significant differences in the particle range.

Despite the short acquisition time (about 100-150 s for each treatment field), preliminary results show that PET images referring to proton irradiations have enough statistics to be compared with a reference distribution (other measurement or Monte Carlo simulation) with an accuracy within 2 mm and a sensitivity in particle range difference detection of 6-8 mm FWHM. In the case of carbon ions, the sensitivity is lower (about 1.5 cm FWHM) due to reduced statistics, hence advanced algorithms to improve analysis performance are under development.

O 83

Clinical sensitivity of a prototype proton radiography system

M. Pankuch1, E. DeJongh2, C. Sarosiek3, D. DeJongh2, G. Coutrakon3, J. Welsh4, V. Rykalin2, N. Karonis5, R. Schulte6, W. Hartsell7

1Northwestern Medicine Chicago Proton Center, Medical Physics, Warrenville, USA
2ProtonVDA, Physics, Naperville, USA
3Northern Illinois University, Physics, DeKalb, USA
4Loyola University, Radiation Oncology, Maywood, USA
5Northern Illinois University, Computer Science, DeKalb, USA
6Loma Linda University, Radiation Oncology, Loma Linda, USA
7Northwestern Medicine Chicago Proton Center, Radiation Oncology, Warrenville, USA

Background: Proton Radiography (pRad) images provide information on the patient's position within the beam path in addition to the proton water equivalent pathlegnth through the patient. A quantification of the sensitivity of pRad images for alignment accuracy, WET accuracy and differential detectability was performed.

Methods and Materials: To quantify the alignment accuracy, Monte Carlo pRad images of a brain patient were created from CT image sets offset from a nominal position. Single proton positions and the residual energy were scored in the pRad detectors and processed through an iterative most likely path back-projection algorithm to create pRad DRR's.To quantify the accuracy and sensitivity to pathlegnth changes, actual pRad images were acquired of a cylindrical phantom and head phantom. Various inserts were placed in a cavity of the head phantom and compared to the known WET.

Results: pRad images can be used to align a patient in the Lt/Rt, Ant/Post and Sup/Inf directions by an average of 0.0mm, 0.0mm and -0.1mm, (standard deviation of 0.4mm, 0.4mm and 0.3mm). Average angular accuracy in the Yaw, Pitch and Roll directions were -0.1deg, 0.1deg and 0.1deg, (standard deviation of 0.2deg, 0.2deg and 0.5deg).Accuracy test show that the pRad images predict the WET of several phantom materials to within 0.3% of know values. WET consistency results of pRad images of the head phantom were able to detect pathlegnth differences as small as 0.5mm with statistical certainty.

Conclusion: pRad images in the head region can be used to correct for positional offsets and quantify pre-treatment pathlegnth consistency.

O 84

Towards a clinical application of carbon-ion pencil beam monitoring in depth using prompt secondary-ion tracking

L.M.H. Ghesquiere-Dierickx1,2,3, R. Félix-Bautista1,3,4, T. Gehrke1,3, J. Jakubek5, G. Echner1,3, A. Höss3,6, M. Winter3,6, M. Ellerbrock3,6, O. Jäkel1,3,6, M. Martišíková1,3

1German Cancer Research Center DKFZ, Department of Medical Physics in Radiation Oncology, Heidelberg, Germany
2University of Heidelberg, Heidelberg Medical Faculty, Heidelberg, Germany
3Heidelberg Institute for Radiation Oncology HIRO, National Center for Research in Radiation Oncology NCRO, Heidelberg, Germany
4University of Heidelberg, Department of Physics and Astronomy, Heidelberg, Germany
5ADVACAM, s.r.o, Prague, Czech Republic
6Heidelberg Ion-Beam Therapy Center HIT, Department of Radiation Oncology- Heidelberg University Hospital, Heidelberg, Germany

On-line monitoring of the dose delivery in patients is of great interest for ion-beam therapy to monitor inter-fractional changes, potentially reduce tumor safety margins and irradiate less normal tissue. Carbon-ion beam monitoring methods using secondary-ion tracking have been developed in our group and showed promising results in anthropomorphic head phantoms. In this contribution, such a monitoring method was investigated to detect potential inter-fractional changes in a head phantom. Furthermore, a first clinical test of secondary-ion tracking was performed in a real clinically treated patient.

The monitoring method is based on the detection and tracking of secondary charged nuclear fragments emerging from an irradiated target. Fractions of a 3 Gy (RBE) carbon-ion treatment of a brain tumor were irradiated at the Heidelberg Ion-Beam Therapy Centre (HIT), Germany, in an Alderson head phantom. The directions of the emerging fragments were measured using pixelated silicon detectors (Timepix3, developed at CERN). Thin PMMA plates were placed upstream from the phantom to mimic inter-fractional variations of the whole fraction delivery. Moreover, tracking of secondary ions was performed during a real patient treatment.

During the head phantom's irradiation, significant differences were found in the secondary ions' distribution for each 12C-ion beam energy used. Moreover, inter-fractional variations simulated by 2mm PMMA upstream from the head phantom were detected. Additionally, a first tracking of secondary ions during a clinical patient treatment was successfully performed at HIT.

This demonstrates the potential of using secondary ions for monitoring inter-fractional variations of a carbon-ion treatment and the technical feasibility of its clinical application.

O 85

Multichannel acquisition of thermoacoustic emissions generated during delivery of a clinical plan to an anthropomorphic phantom

S. Patch1, C. Nguyen2, D. Dominguez-Ramirez2, R. Labarbe3, J. Lister4, C. Finch5, D. Cammarano6, J. Pandey7, C. Chirvase7, J. Lambert8

1UW-Milwaukee and Acoustic Range Estimates, Physics, Milwaukee, USA
2UW-Milwaukee, Physics, Milwaukee, USA
3Ion Beam Applications, Research and Development, Louvain-la-Neuve, Belgium
4Ion Beam Applications, Engineering, Louvain-la-Neuve, Belgium
5Ion Beam Applications, Engineering, Bedlington, United Kingdom
6Ion Beam Applications, Engineering, Bedlington, Belgium
7The Rutherford Cancer Center, Physics, Bedlington, United Kingdom
8The Rutherford Cancer Center, Physics, Newport, United Kingdom

Purpose: To confirm that thermoacoustic emissions generated by a pencil beam synchrocyclotron delivering a clinical plan can provide useful clinical information about beam range.

Methods and Materials: A single view plan with posterior-anterior beam orientation treated a 34x55x48 cm3 volume around a liver lesion in a triple-modality anthropomorphic phantom (CIRS FCT057A). Proton pulses were delivered one layer-by-layer at a 1 kHz repetition rate using the physics mode of a ProteusOne. For each spot, the number of proton pulses and maximum charge created in the primary IC were limited to 32 and 9 nC, respectively.

A custom system with 4 thermoacoustic channels at the corners of a 34 × 52 cm rectangle surrounding a wireless ultrasound array (Clarius L7) was positioned to generate an ultrasound image of the lesion and detect thermoacoustic emissions.

Signals were acquired by a remotely controlled oscilloscope (Siglent 1104x-e) and averaged per spot. The ultrasound image was manually co-registered to a multiplanar reformat of the planning CT using 3D Slicer (Fig 1).

Results: Amplitudes and arrival times varied with respect to layer and spot position, but all pulses were bandlimited below 100 kHz. Channels 1-2 located distal to the treatment volume measured characteristic (broadband) “N” shapes; laterally offset channels 3-4 often measured (bandlimited) ringing (Fig 2).

Conclusions: Carefully designed receivers are sensitive to very weak thermoacoustic emissions generated during treatment. However, ringing confounds straightfoward range estimation. A priori information from the planning CT may be required to overlay the Bragg peak location onto the ultrasound image.

Oral Poster Presentations

OP 01

Comprehensive validation of the Local Effect Model with the PIDE database

T. Pfuhl1, T. Friedrich1, M. Scholz1

1GSI Helmholtzzentrum für Schwerionenforschung, Biophysics, Darmstadt, Germany

The increased relative biological effectiveness (RBE) of ions is one of the key benefits of ion radiotherapy compared to conventional radiotherapy with photons. Thus, for treatment planning in ion radiotherapy a RBE model such as the Local Effect Model (LEM) is indispensable to account for the increased RBE of ions, which varies with biological and physical impacting factors. As the model needs to be accurate over a large variety of ion types and energies the Particle Irradiation Data Ensemble (PIDE) is used to extensively validate the LEM. The database covers over 1100 photon and ion survival experiments in the form of their linear-quadratic parameters and raw data (freely available under This makes it an optimal tool to determine systematic dependencies of the LEM. This study confirms that LEM broadly reflects the RBE systematics observed in experiments. A precise quantitative analysis allows investigating also precision limits of the model: E.g. for carbon ions with a linear energy transfer (LET) > 60 keV/μm the predicted RBE is accurate within 10% on average, whereas towards lower LETs the RBE is underestimated by up to about 20%. Similar trends are generally seen for ions heavier than helium; the corresponding deviation increases with the ions' atomic number. The accurate quantification of such systematic deviations enables to identify potential limitations of the model and thus to develop adequate countermeasures to correspondingly further improve the model accuracy, which is now being approached.

OP 03

Analysis of radio-induced DNA damage as a tool for driving Monte Carlo simulations of cell survival

A. Cicchetti1, M. Carante2, M. Ciocca3, T. Rancati1, R. El bezawy4, N. Zaffaroni4, A. Facoetti5, M. carrara6, R. Valdagni1, F. Ballarini2

1Fondazione IRCCS Istituto Tumori, Prostate Cancer Program, Milan, Italy
2National Institute of Nuclear Physics INFN, INFN-sezione di Pavia, Pavia, Italy
3CNAO Foundation, Medical Physics, Pavia, Italy
4Fondazione IRCCS Istituto Tumori, Experimental Oncology and Molecular Medicine, Milan, Italy
5CNAO Foundation, na, Pavia, Italy
6Fondazione IRCCS Istituto Tumori, Medical Physics, Milan, Italy

In this study we aimed at characterizing, for the first time, the radiobiological response of Human Skeletal Muscle Cells(HSKMCs) from healthy donors, by in vitro analysis and Monte Carlo simulations (MCs) driven by experimental data.

HSkMCs were irradiated with photons/carbon ions at different doses. Survival curves were obtained by cell growth assay.To experimentally derive useful data for the implementation of MCs, DNA damage repair kinetics (immunofluorescence analysis: 0-0.5-1-48h) were assessed.

The BIANCA biophysical model of cell death [Carante Phys.Med.Biol.2019] was used , based on 2 radiobiological parameters: the percentage of un-rejoined chromosome break-ends (f), which is cell-line dependent but radiation-quality independent, and the mean number of DNA “critical lesions” (CL) /Gy/cell, which depends both on cell type and radiation quality. The code was adapt to deal with HSkMCs. Moreover, based on a database of CL-yields derived in a previous study [Carante2019], full predictions of cell survival for the cell-line of interest can be performed.

Model parameter f was experimentally derived from the g-H2AX foci analysis upon 4Gy photon-irradiation. The calculated percentage was defined by the use of kinetics curve of DNA repair (Fig.1). We adjusted the CLs parameter in order to describe the survival curve with photons. Analogous simulation was performed for carbon ions keeping fixed the f and deriving the CL yields from reference database developed in previous study (Fig.2).

A radiobiological characterization of HSkMCs was determined. The method suggests that immunofluorescence assay of g-H2AX foci can be used to estimate un-rejoined chromosome break-ends and also to perform full prediction of cell survival.

OP 06

UNIVERSE: Unifying the modelling of radiosensitization for photons and protons

H. Liew1,2,3,4,5,6, C. Klein3,4,5,6, A. Abdollahi3,4,5,6, J. Debus1,2,3,4,5,6, I. Dokic3,4,5,6, A. Mairani3,4,5,6

1German Cancer Research Center, Clinical Cooperation Unit Radiation Oncology, Heidelberg, Germany
2Heidelberg University, Faculty of Physics and Astronomy, Heidelberg, Germany
3Heidelberg Ion-Beam Therapy Center HIT, Radiation Oncology, Heidelberg, Germany
4German Cancer Consortium, German Cancer Research Center, Heidelberg, Germany
5German Cancer Research Center, Heidelberg Institute of Radiation Oncology HIRO, Heidelberg, Germany
6National Center for Tumor Diseases NCT, Division of Molecular and Translational Radiation Oncology, Heidelberg, Germany

Mechanistic approaches to modelling the effects of ionizing radiation on cells are on the rise, promising a better understanding of predictions and higher flexibility concerning therapeutical conditions to be accounted for. Previously, we showed that we could modify and extend a published mechanistic model of cell survival after photon irradiation to account for radiosensitization by genetic or pharmacological suppression of the DNA repair machinery. It was also shown, that the parametrization was invariant under change of the oxygenations status and could be introduced independently from our implementation of the oxygen effect into the model for photons. In our current work we have extended our model to predict cell survival of different cell lines after proton irradiation over a wide range of LET. Furthermore, we could show that our parametrization of radiosensitization is also applicable to the case of proton irradiation. The Figure shows survival data of the HT1080 cells and its RAD51 deficient mutants (taken from Bright et al. 2019) after photon and proton (7.3 keV/μm) irradiation. The model parameters, including the radiosensitization factor (RSF) of the repair deficiency, were derived solely from the photon data. The RSF was directly introduced to our model of survival after proton irradiation and predicted the survival of the mutant cell line very well. An extended summary of the concept, development and benchmarking of our model well be presented at the conference.

OP 07

Concurrent pencil beam scanning proton therapy and superficial hyperthermia: a growing clinical experience

J. Snider- III1, J. Molitoris1, A. Koroulakis2, C. Decesaris2, O. Siddiqui2, E. Kowalski2, S. Samanta2, D. Rodrigues1, W. Regine1, Z. Vujaskovic1

1University of Maryland School of Medicine, Radiation Oncology, Baltimore, USA
2University of Maryland Medical Center, Department of Radiation Oncology, Baltimore, USA

Purpose: Hyperthermia (HT) acts as an excellent adjuvant to radiotherapy. However, there remains a general lack of experience in conjunction with particle therapies due to the lack of centers with paired capabilities. Herein, we update the largest experience utilizing concurrent superficial-HT and pencil beam scanning proton therapy (PBS-PT) from a major academic center.

Methods: Over 2,000 patients have undergone PBS-PT at our institution, of which 46 (50 courses) have received concurrent superficial-HT. Histologies include sarcoma (n=21), breast cancer (n=17), squamous cell carcinoma (n=3), adenocarcinoma (n=2), ureteral (n=1), ovarian (n=1), vulvar (n=1), mesothelioma (n=1), desmoid (n=1), esophageal cancer (n=1), and multiple myeloma (n=1). PBS-PT was delivered to a median dose of 50.4 Gy(RBE) (range,30-72 Gy(RBE); mean,53.8 Gy(RBE)). The BSD-500 device was utilized for superficial HT treatment (median,10 applications; range,4-28).

Results: With a median follow-up of 8 months (range,1-31 months), concurrent PBS-PT and superficial-HT has been well-tolerated. Thirty six (72%) of the courses were delivered in the setting of a recurrent malignancy. There have been no grade 4 or 5 toxicities. Grade 3 toxicities included radiation dermatitis (n=7), chronic lymphedema (n= 2), and non-healing wounds (n=2). Grade 1-2 toxicities included pain, dermatitis, hyperpigmentation, and GI disturbance. Thirty eight (83%) patients remain alive with 28 (61%) free of disease, and 39 (78%) tumors locally controlled.

Conclusions: Concurrent PBS-PT and superficial-HT remains a safe and generally well-tolerated regimen. Long-term and prospective clinical trial data should be sought to clarify the role of this approach in individual disease entities.

OP 08

Improvements in left-sided breast cancer proton DIBH treatments with 3D SGRT

W. Xiong1, H. Lin2, M. Kang2, L. Hu2, P. Tsai2, A. Zhai2, C. Apinorasethkul2, A. Shim2, C. Simone2, I. Choi2

1New York Proton Center, Medical Physics, roslyn heihts, USA
2New York Proton Center, Medical Physics, New York, USA

Background: This study aimed at improving left-sided breast cancer proton DIBH treatment with 3D SGRT. We initiated DIBH treatment for our left-sided breast cancer patients to achieve better target dose coverage as well as significantly dose reduction to hearts and lungs. However, the robust planning for the large setup uncertainty in breast treatment compromised dose reduction to OARs. We assessed the potential benefits of applying SGRT to monitor patient motion and reduce setup uncertainty and further reduce doses to hearts and lungs.

Method and Materials: A video-based imaging system has been installed in our proton facility that generates high density 3D models of skin surface without markers. This SGRT, applied to our breast/chestwall proton DIBH treatment, can help improve patient positioning accuracy and efficiency. Six consecutive patients who had left breast/chestwall proton treatment with DIBH were selected for this study. We replaced the large setup uncertainty of 5mm with smaller setup uncertainty 3mm enabled with SGRT in robust planning and compared doses to patient's lungs and heart.

Results: By reducing uncertainty in our robustness planning from 5mm to 3mm, the maximum dose Dmax to left Lung dropped by 1.6% for these six patients, at the same time the average lung mean dose dropped by 1.5%. The heart average Dmax reduced by 1.0[CS3] % and average mean dose reduced by 1.2%.

Conclusions: The use of SGRT allows proton breast/chestwall patients to be positioned accurately and efficiently. The robust planing with reduced setup uncertainty further leads to dose reduction to heart and lungs.

OP 09

Reirradiation with proton therapy for recurrent esophageal and gastroesophageal junction malignancies: Results of the Proton Collaborative Group multi-Institutional registry trial

J.I. Choi1, C. DeCesaris2, J. Molitoris3, S. Hasan1, R. Press1, N. Mohammed4, H. Tsai5, C. Vargas6, M. Chuong7, C. Simone1

1New York Proton Center, Radiation Oncology, New York, USA
2University of Maryland Medical Center, Radiation Oncology, Baltimore, USA
3University of Maryland School of Medicine, Radiation Oncology, Baltimore, USA
4Northwestern Medicine Chicago Proton Center, Radiation Oncology, Chicago, USA
5Procure New Jersey, Radiation Oncology, Somerset, USA
6Mayo Clinic, Radiation Oncology, Scottsdale, USA
7Miami Cancer Institute, Radiation Oncology, Miami, USA

Background: Photon reirradiation for isolated locoregional recurrences of esophageal/gastroesophageal junction (GEJ) cancers is associated with significant toxicity. Proton therapy improves normal tissue sparing and may more safely allow reirradiation dose escalation.

Materials/Methods: The Proton Collaborative Group prospective, multi-institutional registry was queried for recurrent esophageal/GEJ cancers treated with a second course of radiotherapy using protons. Outcomes and toxicities (CTCAEv4.0) were evaluated.

Results: Twenty-five consecutive patients retreated from 7/2012-11/2019 were analyzed. Patients were a median of 70 years (52-82), predominantly male (88%) and non-Hispanic Caucasian (88%) with adenocarcinoma (68%). Initial stages: T1-4%, T2-8%, T3-76%, unknown-8%; N0-24%, N1-48%, N2-20%, unknown-8%; stage I-4%, IIA-4%, IIB-24%, IIIA-40%, IIIB-16%, IV-4%, unknown-4%. During their initial treatment (median 50.4/1.8Gy), all received concurrent chemotherapy, and five patients (20%) underwent resection. Median time to recurrence was 14.0 months (4.1–134.0). Proton reirradiation was delivered using uniform scanning/passive scattering (n=21) or pencil beam scanning (n=4) to a median 45.9Gy(RBE) (30.1–60.5) in 12-42 fractions. Eleven patients (44%) received concurrent chemotherapy. Median follow-up from reirradiation completion was 23.1 months among living patients. 6/18-month survivals were 48%/24%. Only two patients (8%) developed locoregional recurrences (9.2, 11.1 months following reirradiation) both salvaged with resection. Grade 3 toxicities occurred in 20% (anemia=1; anorexia=2; dysphagia=2; esophagitis=2); grade 2 in 44% (most commonly, fatigue=6). Grade ≥2 esophagitis and pneumonitis occurred in 16% and 0%, respectively. No esophageal fistula/stricture/necrosis and no grade 4-5 events occurred.

Conclusions: Proton reirradiation for locoregionally recurrent esophageal/GEJ cancers is feasible and achieves durable local control with limited toxicity. Additional prospective investigations and cumulative dose constraint analyses are warranted.

OP 10

A dosimetric study of optimization target volumes for CBCT-guided spot-scanning stereotactic body proton therapy for locally advanced pancreatic cancer patients

D. Han1, C.C. Chen1, H. Hooshangnejad2, K. Ding1

1Johns Hopkins University, Radiation Oncology, Baltimore, USA
2Johns Hopkins University, Biomedical Engineering, Baltimore, USA

Purpose: To evaluate dosimetric impacts of different optimization target volume (OTV) schemes for spot-scanning stereotactic body proton therapy (SBPT) with CBCT of unresectable, locally advanced pancreatic cancer (LAPC).

Methods and Materials: We carried out different expansions of clinical target volume (CTV) to OTV of six LAPC patients for SBPT plans with prescribed dose 33 Gy(RBE)/5 Fx including: 1) an isotropic expansion of 2 mm (OTV2mm), for setup uncertainty; 2) a beam-specific expansion (OTVWET) based on ray-tracing technique for proximal and distal expansion along the beam direction accounting for the range uncertainties and 2 mm expansion laterally in beam's eye view (Fig. 1). All treatment plans are generated in Raystation TPS (9A), and dose coverage for CTV (D98, Dmax and D50) and OAR dose sparing (V33Gy(RBE), V20Gy(RBE) and V15Gy(RBE)) are compared against robust optimization on GTV (GTVrobust). Robustness evaluations under scenarios of 2 mm setup and ±3.5% range uncertainties are analyzed.

Results: All three optimization schemes achieved decent target coverage with no statistically significant difference (Fig. 1).While there is no significant difference in dose distributions between OTVWET and GTVrobust, OTV2mm shows superior OAR sparing, especially for proximal duodenum; however, OTV2mm plans demonstrate severe susceptibility to range uncertainties (Tab.1).

Conclusion: Although a setup uncertainty of 2 mm can be achieved by CBCT in SBPT, the geometric expansion of OTV2mm fails to account for range uncertainty. The use of OTVwet mimics a union volume for all scenarios in robust optimization but save the plan optimization time (1 vs. 21 scenarios) noticeably.

OP 11

Carbon-ions / protons for hepatocellular carcinoma: a meta-analysis

Z. Yang1, Z. Qiuning2, W. Xiaohu3, L. Hongtao4, S. Lihua5

1Lanzhou University, Basic Medical College, Lanzhou City, China
2Lanzhou Heavy Ion Hospotial, Department of Radiation Oncology, Lanzhou, China
3Lanzhou Heavy Ion Hospotial, Department of Radiation Oncology, Lanzhou City, China
4Lanzhou University, Department of Radiotherapy, Lanzhou, China
5The First Clinical Medical College of Lanzhou University, Department of Radiation Oncology, Lanzhou, China

Objective: To evaluate the efficacy and safety of carbon ion/proton therapy for hepatocellular carcinoma (HCC) by Meta-analysis.

Methods: PubMed, The Cochrane Library, EMBASE, Chinese Journal Full-text, Chinese Biomedical Literature and Wanfang Database were searched to collect relevant clinical studies on carbon ion and proton therapy for HCC.Two reviewers independently screened the literature and extracted data based on inclusion and exclusion criteria. Meta-analysis was carried out by using Stata 12.0.

Results: A total of 7 carbon ion and 23 proton therapy studies were included. According to the combined results, the 1-, 2-, 3-, and 5-year survival rates of carbon ion/proton therapy for HCC were 94.2%,82.3%,65.4%,29.2% and 80.2%,65.2%,57.4%,36.9% respectively, the 1-, 2-, 3-, and 5-year local control rates were96.0%,88.4%,83.0%,90.9% and 97.7%,90.2%,87.8%,85.6%, respectively. Subgroup analysis showed that carbon ion and proton therapy seemd to achieve similar survival rates compared to surgery for HCC with good prognosis like better liver function, single or smaller lesion, and for HCC with poor prognosis such as poor liver function, recurrence, and locally advanced disease also showed good efficacy.

Conclusion: Both carbon ion and proton therapy for HCC can achieve good local control and low toxicity, and it seems to be able to achieve a survival rate comparable to surgery for HCC patients with good prognosis, and it also shows good efficacy for HCC patients with poor prognosis. Therefore, carbon ion and proton for HCC has broad clinical application prospect; however, it needs to be confirmed by prospective clinical studies.

OP 12

Completion time in prostate cancer patients treated with proton beam therapy: Do interruptions matter?

S. Hasan1, J.I. Choi1, J. Chang2, R. Lane3, C. Rossi4, W. Hartsell5, H. Tsai6, G. Laramore7, C. Vargas8, C.B. Simone 2nd1, J. Han2

1New York Proton Center, Radiation Oncology, New York, USA
2Oklahoma Proton Center, Radiation Oncology, Oklahoma City- Oklahoma, USA
3Willis Knighton Cancer Center, Radiation Oncology, Shrieveport, USA
4California Protons USCD, Radiation Oncology, San Diego, USA
5Northwestern University, Radiation Oncology, Chicago, USA
6Procure Proton Therapy Center New Jersey, Radiation Oncology, Somerset, USA
7University of Washington, Radiation Oncology, Seattle, USA
8Mayo Clinic, Radiation Oncology, Rochester, USA

Introduction: The association between completion time of proton therapy (PT) in prostate cancer and biochemical control currently is unknown.

Methods: We queried the multi-institutional, prospectively collected Proton Collaborative Group registry for prostate cases treated definitively with PT. Kaplan Meier methodology was used for biochemical failure free (bFF) rates and multivariable regression analyses (MVA) were used to identify correlates of treatment interruptions (TI) and bF.

Results: After exclusion, 2,794 consecutive men with low (n=693), favorable intermediate (n=869), unfavorable intermediate (n=627), and high (n=605) risk prostate cancers had available data. The median age was 68 years (40-92), 90% were white and 8% were black. Androgen deprivation therapy (ADT) was given to 676 patients, 312 courses were hypofractionated, and the median EQD2 dose was 75 (74-86) GyE1.5. Kaplan-Meier median follow-up was 79 months. In total, 900 patients (32%) had at least one TI. Shorter treatments (HR=0.95 per day, P<0.01) and high risk (HR=0.72, P<0.04) cases were less likely to have TIs on MVA. There was no difference in 5-year bFF rates with (92.7%) and without (93.1%) TIs. In a subset of high risk patients treated with ADT (n=385), the 5-year bFF was 83% without TIs and 75% with TIs (HR=2.10, P=0.06). This discrepancy was significant with multivariable binomial regression (HR=2.36, P=0.03), and the bF difference was greatest when >5 treatment days were missed.

Conclusion: There was no correlation between TIs and bF in low/intermediate risk prostate cancer treated with PT; however, completion time may play a role in high risk disease.

OP 14

Ultra-fast Monte-Carlo simulation for in-beam proton range verification

L. Ma1, Y. Zhong1, M. Chen1, X. Jia1, X. Gu1, Y. Shao1, W. Lu1

1University of Texas Southwestern Medical Center, Radiation Oncology, Dallas, USA

Range verification is critical in proton therapy to ensure that delivery matches plan. In-beam positron emission tomography (PET), combining with mid-range probing strategy, is promising to detect particle-activated positron activity with high accuracy. An ultra-fast activity and dose simulation tool is demanded to translate in-beam range verification from benchtop to bedside.

In order to verify the range of invisible dose through visible PET signal, the spatial relationship between dose and activity need to be studied. We developed a GPU-based Monte Carlo simulation package (gPMC) for simulating positron emitter production of proton beam. The inputs of gPMC packages are physics data (cross section of proton in materials, energy and angular distribution of each 2nd isotope), voxelized patient geometry data and proton treatment plan. Simulation results include the phase space of each positron-emitting isotope (C10, C11, N13, O15) and physical dose distribution.

We benchmarked gPMC and validated it with GATE v8.2, both with 10^7 incident protons on water and soft tissue phantom. Both gPMC and GATE simulation run on a work station with 4core i7 9700K 3.6GHz CPU and RTX2080i graphical card. The depth profiles of dose and O15 activity in water phantom by gPMC is presented in Figure1. Figure2 shows the energy dependency of the distance between R50 of dose and R50 of O15. For 10M protons, simulation time of gPMC is 1.2 seconds while GATE takes 27000 seconds. The 20000X speedup provided by gPMC makes it feasible for in-beam range verification of a clinical proton plan associated with patient-specific geometry.

OP 15

In-vivo monitoring of 12C ions treatments with the DoseProfiler: preliminary results from a clinical trial at the CNAO centre

A. Sarti1,2,3, M. De Simoni3,4, Y. Dong5,6, A. Embriaco5, M. Fischetti3,7, F. Galante1,3, R. Mirabelli3,4, S.M. Valle5,6, G. Traini3, V. Vitolo8

1Università di Roma “La Sapienza”, Scienze di Base e Applicate per l'Ingegneria, Roma, Italy
2Museo Storico della Fisica e Centro Studi e Ricerche “E. Fermi”, Sezione di Roma, Roma, Italy
3Istituto Nazionale di Fisica Nucleare, Sezione di Roma 1, Roma, Italy
4Università di Roma “La Sapienza”, Fisica, Roma, Italy
5Istituto Nazionale di Fisica Nucleare, Sezione di Milano, Milano, Italy
6Università degli Studi di Milano, Fisica, Milano, Italy
7Università di Roma “La Sapienza”, Scienze di Base e Applicate per l'Ingegneria, Roma, Italy
8Fondazione CNAO, Staff Medico, Pavia, Italy

The production of secondary fragments occurring in 12C ions particle therapy treatments depends on the densities of the tissues crossed by the beam. By detecting the particles that have enough energy to exit from the patient it could be possible to detect inter-fractional morphological changes that could lead to a wrong dose release pattern if not accounted for in the treatment plan. As no online monitoring of 12C ions treatments is nowadays available in the clinical practice, whenever the clinical practice suggests it a re-evaluation CT is performed after the first half of fractions has been delivered.

The DoseProfiler (DP) is a charged particles tracker that is operational as an online monitoring device of 12C ions since August 2019 (the picture shows the first patient monitored) in a clinical trial at the CNAO centre as part of the INSIDE bi-modal monitoring system. The DP collects the charged fragments produced at 60° wrt the incoming beam direction that exit from the patient and is able, during the treatment, to reconstruct their production point. The obtained spatial emission maps can be compared in different fractions to identify the occurrence of morphological changes. The preliminary results obtained with a first set (10) of patients included in a clinical trial involving different head and neck pathologies will be reviewed and the technique potential in assessing the insurgence of morphological changes in the patient that require a treatment plan re-evaluation will be discussed.

OP 16

Dosimetry quality assurance in particle therapy: first results from a European survey conducted by the EPTN-ESTRO task force

M. Togno1, P. Trnková2,3, K. Henkner4, O. Nørrevang5, M. Stock6, S. Menkel7, O. Jäkel4,8, D.C. Weber1,9,10, T. Lomax1,11, S. Safai1

1Paul Scherrer Institut, Center for Proton Therapy, Villigen PSI, Switzerland
2Holland PTC, Department of Medical Physics and Informatics, Delft, Netherlands
3Erasmus MC, Department of Radiation Oncology, Rotterdam, Netherlands
4Heidelberg Ion-Beam Therapy Center HIT, Department of Radiation Oncology, Heidelberg, Germany
5Danish Centre for Particle Therapy DCPT, Department of Medical Physics, Aarhus, Denmark
6MedAustron Ion Therapy Center, Department of Medical Physics, Wiener Neustadt, Austria
7Universitätsklinikum Carl Gustav Carus, Proton Therapy Center, Dresden, Germany
8Deutsches Krebsforschungszentrum, Medical Physics in Radiation Oncology, Heidelberg, Germany
9University Hospital of Bern, Department of Radiation Oncology, Bern, Switzerland
10University Hospital of Zürich, Department of Radiation Oncology, Zürich, Switzerland
11ETH Zürich, Department of Physics, Zürich, Switzerland

Comprehensive quality assurance (QA) programs and guidelines are needed in order to safely treat patients and to make efficient use of available resources for particle therapy. The lack of dedicated QA standards for this treatment modality in recent years has led to the implementation of local institutional procedures mostly based on those executed by experienced centers or based on commissioning programs, resulting in a heterogeneous number of QA protocols.

In the framework of the European Particle Therapy Network (EPTN), the second Working Party is promoting a comprehensive survey of dosimetry QA performed by the particle centers across the network. The aim of the survey is to gather information on the current methods, tools, tolerances and resources (manpower) and to verify if there is a consensus among the European centers with respect to the QA programs. The survey focuses on PBS delivery technique, and it consists of both specific and general questions on the QA methods. It was first distributed in 2018 and updated end of 2019.

As of December 2019, 18 centers (proton n=15; proton and carbon ion n=3) returned a completed survey (completion rate 78%). The FTE ratio for QA is detailed in Fig. 1. Fig. 2 displays the percentage of centers using a specific action level for QA tests. The analysis provides an overview of the spectrum of approaches to dosimetry QA, showing that currently there is large variety among European centers, not only in the methods used but also in the tolerances and type of tests.

OP 17

Evaluations of a flat-panel imager for proton PBS patient specific QA

Z. Su1, S. Rossomme2, W. Hsi1

1University of Florida Proton Therapy Institute, Department of Radiation Oncology, Jacksonville, USA
2IBA Dosimetry, IBA Dosimetry, Schwarzenbruck, Germany

Purpose: To evaluate a commercial flat-panel imager signal responses and LET quenching effect for proton PBS plan QA.

Methods and Material: The flat-panel signal linearity of different proton energies were obtained for all 6 gains settings (0.25, 0.5, 1, 2, 4, 8pF). The flat-panel spatial responses were measured by irradiation of 100MeV proton 10x10cm2 beam through 7cm solid water then an aluminum plate with 1cm spaced matrix of holes. The distances between the centers of measured adjacent spots were compared to the plate mechanical specifications. PDDs of 100, 150 and 200MeV protons were measured by the flat-panel with different slabs of solid water and compared to ones in water. The panel LET quenching effects were then quantified. A prostate PBS plan was measured at various depths using the flat-panel. The obtained images were converted to planar dose through counts-to-dose calibration and compared to corresponding planar doses from treatment plan using gamma tests.

Results: The signal responses are linear in all gain settings. The measured inter-hole distances are in excellent agreement with the plate mechanical design. The distal R80 of measured PDDs are 7 to 8mm shorter than those in water indicating flat-panel detector LET quenching effect (Figure1).The gamma test pass rates of prostate PBS plan at depths of 0cm to 15cm are from 96.6% to 100% (3mm, 3%) (Figure2).

Conclusion: The flat-panel signal response is linear and spatial response is accurate. Detector has LET quenching effect and is similar across energies. The PBS patient-specific QAs are acceptable at the measured depths.

OP 18

Introducing multi-detector prompt gamma timing with an in-beam PET device

F. Pennazio1, P. Cerello1, V. Ferrero1, E. Fiorina1, S. Garbolino1, V. Monaco2, B. Sharifi3, R. Wheadon1, M. Rafecas4

1INFN, sez. Torino, Torino, Italy
2University of Torino and INFN, Physics dept., Torino, Italy
3Amirkabir University Of Technology- was with University of Torino, Physics dept., Torino, Italy
4University of Lübeck, Institute of Medical Engineering, Lübeck, Germany

We propose an innovative multi-detector layout and a data analysis algorithm exploiting the emission of prompt gamma photons (PG) during proton therapy to monitor the beam range. In this work, this technique is implemented with in-beam PET detectors, allowing us to design a scanner able to simultaneously monitor the target (patient) activation and the PG emission.

The Prompt Gamma Timing (PGT) method is based on assessing the range variation in the patient by measuring the time of flight between the primary proton irradiation and the PG detection. We propose to expand it by exploiting multiple PET detectors arranged at different azimuthal angles along the beam direction. We introduce then a Maximum Likelihood Expectation Maximization algorithm (MLEM) to reconstruct the emission of prompt photons along the beam path in both spatial and time domains.

This opens the perspective of directly assessing the primary particles speed along their path, a quantity closely related to their energy loss.

We performed a FLUKA-based Monte Carlo simulation of a proton beam impinging on a PMMA phantom with cavities of different thickness, with PG detected by nine PET pixelated detectors (250 ps FWHM time resolution). Results show a millimetric PG emission agreement along the beam range.

This work is part of the I3PET INFN project, which is presently testing a demonstrator of combined in-beam PET and Multi-Detector PGT scanner, based on Silicon Photomultipliers-based PET detectors, and monitoring the beam with innovative Ultra-Fast Silicon Detectors, which can detect single protons with tens of ps time resolution at therapeutic rates.

OP 19

Measurement based estimation of relative biological effectiveness in carbon ion radiotherapy

S. Hartzell1, F. Guan1, O. Vassiliev1, P. Taylor1, C. Peterson2, S. Kry1

1UT MD Anderson Cancer Center, Radiation Physics, Houston, USA
2UT MD Anderson Cancer Center, Quantitative Sciences, Houston, USA

Inconsistencies in dose delivered during carbon radiotherapy is due largely to uncertainties in relative biological effectiveness (RBE), which is calculated using one of several models. Each model requires unique input parameters. While RBE by the Microdosimetric Kinetic Model (MKM) can be measured using a microdosimeter, there exist no direct means of measuring RBE by other common models, including the Repair Misrepair Fixation (RMF) and Local Effect Model I (LEM). This study investigates a means of estimating RBE with a uniform, microdosimetric measurement for each model, to allow both measurement-based validation of model implementation and comparison of RBE across models and institutions.

Monte Carlo (GEANT4) was used to simulated monoenergetic and SOBP carbon beams (>160 energies). Input parameters for each RBE model (microdosimetric spectra, double strand break yield, kinetic energy spectra, dose fragment contributions) were calculated and used to determine RBE by each model. The saturation corrected dose mean lineal energy (i.e., measured values) were then used to fit each RBE model. The RBE calculated using the full input parameters versus that calculated using the measurement-based fit were compared to quantify estimation error.

RBE was estimated by microdosimetric parameters to within 5% accuracy in 100%, 97%, and 93% of data points assessed for MKM, RMF, and LEM models, respectively. The figure below displays a histogram of estimation error by model for various beams.

While true RBE has extensive associated uncertainty, modeled RBE can be estimated with reasonable accuracy in a common, measurement-based framework, to enhance model intercomparisons and verify model implementation by institutions.

OP 20

Microdosimetry at the 62 MeV proton beam line of CATANA

A. Bianchi1,2,3, V. Conte3, A. Selva3, D. Bortot4,5, D. Mazzucconi4,5, A. Pola4,5, B. Reniers2, A. Parisi1, G.A.P. Cirrone6, G. Petringa6

1Belgian Nuclear Research Centre SCK-CEN, Environment- Health and Safety, Mol, Belgium
2UHasselt, Faculty of Engineering Technology, Diepenbeek, Belgium
3Istituto Nazionale di Fisica Nucleare, Laboratori Nazionali di Legnaro, Legnaro, Italy
4Istituto Nazionale di Fisica Nucleare, Sezione di Milano, Milano, Italy
5Politecnico di Milano, Dipartimento di Energia, Milano, Italy
6Istituto Nazionale di Fisica Nucleare, Laboratori Nazionali del Sud, Catania, Italy

In proton therapy an RBE of 1.1 is mostly used [1] but an increasing RBE was found in several biological assays [2]. Microdosimetry measures stochastic features of the radiation interaction at the micrometer level. The dose distribution of the lineal energy d(y) and the dose-weighted mean lineal energy are useful to estimate the radiation quality of therapeutic beams [3]. The response of three detectors was studied at several depths at the 62MeV proton SOBP of CATANA.

The TEPC developed by LNL-INFN has an internal diameter of 0.9mm and works without gas flow [3]. The silicon telescope of the Politecnico di Milano consists in E and ΔE (1.9μm) stages fabricated on a single silicon substrate [4]. The MicroPlus probe, of the University of Wollongong [5] is based on silicon-on-insulator wafers (10μm).

Measurements with the three detectors were performed at the same depths in water. The lineal energy spectra were analyzed and compared (fig.1). The dose-mean lineal energy was compared to the calculated total dose-averaged LET [6] (fig.2).

The three detectors show different limits and potentialities that will be discussed in view of their potential use in clinics for optimization of the therapeutic treatment plans.

References: 1. ICRU, Report 78 (2007). 2. Paganetti, Phys.Med.Biol. 59(22), R419–72 (2014). 3. Conte et al., Phys.Med. 64, 114-122 (2019). 4. Agosteo et al., Radiat.Prot.Dosim. 122(1-4), 382-386 (2006). 5. Rosenfeld, Nucl.Instrum.Methods.Phys.Res.A. 809,156–170, (2016). 6. Cirrone et al., Nucl.Phys. B 150, 54–57 (2005).

OP 22

Evaluation of radiation-induced second malignant neoplasms risk from state-of-art radiotherapy treatments in paediatric abdominal neuroblastoma

C. Veiga1, P. Lim2, A. Alhadi1, S. Taylor1, W. Harris1, V. Rompokos3, M. Gaze2, J. Gains2

1University College London, Department of Medical Physics and Biomedical Engineering, London, United Kingdom
2University College London Hospital, Department of Oncology and Radiotherapy Department, London, United Kingdom
3University College London Hospital, Radiotherapy Physics, London, United Kingdom

Purpose: To investigate the theoretical benefits of pencil beam scanning proton therapy (PBS-PT) versus intensity modulated arc therapy (IMAT) in paediatric abdominal neuroblastoma using risk models of radiation-induced second malignant neoplasms (SMN).

Methods and Materials: Twenty patients (median 3y; range, 1-9y) were retrospectively planned for IMAT and PBS-PT (21Gy[RBE] in 14 fractions). The risk ratio (RR) between radiotherapy modalities was estimated with the concept of organ equivalent dose (OED) using doses exported from the treatment planning system and site-specific dose-response relationships (Fig1).

Results: The risk of SMN due to primary radiation was overall reduced in PBS-PT against IMAT (OEDPBS-PT/OEDIMAT<1, Fig2). PBS-PT consistently decreased the mean liver dose in all subjects (8.4±4.0Gy[RBE] vs 4.4±3.5Gy[RBE]) but the RR was variable across the population (0.52±0.33). In the skin, the RR is favourable for PBS-PT (0.33±0.12) but at the cost of higher V10Gy. For soft tissues and bone, the benefits are moderate (0.89±0.05 and 0.95±0.03, respectively) since V10Gy are similar for PBS-PT and IMAT. PBS-PT spared the liver, lung and bladder in 10%, 40% and 100% of the patients. For out-of-field organs there are larger uncertainties in RR due to not accounting for secondary radiation; however, the absolute risk is small.

Conclusion: We quantitatively describe the theoretical extent of the benefits of PBS-PT in the treatment paediatric neuroblastoma in terms of late radiation-induced SMN. Lower risk was estimated for PBS-PT in the lungs, liver, and skin. Further work is needed to evaluate the risks in the gastrointestinal tract and impact of secondary radiation.

OP 23

Effect of pencil beam scanning on cardiac implantable electronic devices – an in vitro study

H. Bjerre1,2, C.J.S. Kronborg1, M.B. Kronborg2, M.F. Jensen1, H. Nyström3, C.S. Søndergaard1, E. Almhagen3, M. Høyer1, J.C. Nielsen2

1Aarhus University Hospital, Danish Centre for Particle Therapy, Aarhus, Denmark
2Aarhus University Hospital, Department of Cardiology, Aarhus, Denmark
3The Scandion Clinic, The Scandion Clinic, Uppsala, Sweden

Background: Cancer patients with cardiac implantable electronic devices (CIED) who undergo radiotherapy, present a complex challenge, because CIEDs are sensitive to ionizing radiation. The harmful effects of radiotherapy on CIEDs consist mainly of stochastic effects from interactions with particles, especially neutrons. With proton therapy, little is known about radiation effects on CIEDs, but risk of neutron exposure is generally higher than with conventional photon radiotherapy. This experimental in vitro study investigates CIED malfunctioning risk during pencil beam scanning.

Methods: Twenty-four explanted and fully functional CIEDs (12 pacemakers and 12 implantable cardioverter defibrillators) were used. Each CIED was placed in a polymethyl-methacrylate phantom, 0.5 cm outside the 10x10x10 cm spread-out Bragg peak, to best simulate clinical conditions. The CIEDs were subjected to 72 Gy in 2 Gy fractions, five days a week, to allow for daily interrogation. All CIEDs were interrogated, using manufacturer-specific standard telemetry equipment. Secondary neutron exposure was estimated using Monte Carlo simulations.

Results: After 50 Gy, 26 reset-to-backup-mode malfunctions occurred in six CIEDs, corresponding to a risk of 4.3% (2.7;6.0%) per fraction, in 25% (7.7;42.3%) of devices. All resets occurred in CIEDs from the same manufacturer. All devices were successfully reprogrammed and returned to normal function.

Conclusion: Risk of CIED malfunction during pencil beam scanning is non-negligible. Malfunctions so far have been clinically manageable, and no permanent damage has been recorded. Results and conclusions from the full 72 Gy course will be presented at PTCOG 2020.

OP 25

3D PBS dose rate calculation and analysis for flash lung treatment planning

J. Perez1, M. Folkerts2, A. Magliari3, E. Abel4

1Varian Medical Systems, Proton Solutions, Geneva, Switzerland
2Varian Medical Systems, Proton Solutions, Dallas, USA
3Varian Medical Systems, Medical Affairs, Chicago, USA
4Varian Medical Systems, Proton Solutions, Palo Alto, USA

Introduction: FLASH therapy has been shown to spare normal tissues while maintaining tumor control in a variety of preclinical models. Protons can reach deep-seated tumors and achieve FLASH dose rates on clinical systems. The definition of dose rate becomes complex in pencil beam scanning (PBS) as each point receives dose contributions from its surroundings in a unique scanning-pattern dependent fashion.

Materials and Methods: A FLASH plan was designed in transmission mode to follow standard lung SBRT constraints. PBS dose rate is defined as the dose to a point divided by the irradiation time experienced at that point. The irradiation time starts once a point is first irradiated and stops when the total dose to the point is delivered. Based on the plan and the scanning dynamics of the system, 3D PBS dose rate was calculated for each field and displayed on the CT for analysis.

Results: Figure 1 shows dose (left) and PBS dose rate (right) on the CT for one field with the 40 Gy/s isodose rate line (white). A line (red line) through healthy tissue shows both dose (blue) and dose rate (red) profiles in Figure 2. At the 40 Gy/s intercept, the corresponding dose is 2 Gy. Regions of lower dose rate tend to correspond to low dose regions. Additionally, variations in dose rate dependent on scanning direction can be observed.

Conclusion: Dose rate calculation studies such as this one will allow researchers to quickly and easily evaluate dose rate levels in different OARs and study the potential benefits of FLASH.

OP 26

A clinical decision tool for optimal treatment delivery of brain SRS: VMAT, intensity modulated proton therapy or spot-scanning proton arc?

G. Liu1, X. Li1, L. Zhao1, C. Stevens2, D. Yan2, P. Chinnaiyan1, P. Kabolizadeh1, X. Ding1

1Beaumont Health, Radiation Oncology Proton Therapy Center, Royal Oak, USA
2Beaumont Health, Radiation Oncology, Royal Oak, USA

Purpose: To provide a reference and guideline in the management of single brain SRS target with different sizes and locations among three different treatment modalities: Intensity Modulated Proton Therapy(IMPT), Spot-scanning Proton Arc(SPArc) and Volumetric Modulated Arc Therapy(VMAT).

Materials and Methods: To simulate the different target locations, a GTV(0.3cc) was inserted in the deep central and peripheral region of a head CT set. To simulate the different target sizes, the GTV was expanded with a uniform margin(2mm increments), corresponding to a different target volume (from 0.3cc; to 26.05cc)(Fig. 1). Three planning groups: IMPT, SPArc and VMAT were generated using the same planning objective functions and robust optimization parameters. Prescription was 18Gy(RBE) in 1 fx(Fig. 2). Multiple dosimetric metrics were analyzed to assess the plan quality such as dose Conformity Index(CI); R50; V12Gy and mean dose of brain.

Results: IMPT showed advantage over VMAT in large peripheral target(26.05cc) in terms of normal brain tissue dose sparing where single fraction SRS might not be considered clinical feasible via VMAT. However, VMAT showed its advantage in the CI and R50 in all the other targets where a sharp dose fall-off is clinically desired. In comparison with VMAT, SPArc has an equivalent or better CI in any size of peripheral targets and deep centrally located targets which were bigger than 4.74cc while significantly reducing the normal brain tissue dose(Fig. 3).

Conclusion: Different treatment modalities offered different dosimetric advantage in brain SRS. Location and size of the target could be used as a reference for choosing the optimal treatment solution.

OP 27

Comprehensive and quantitative evaluation of Monte Carlo (Geant4) physics settings for clinical proton dose simulations

C. Winterhalter1,2, D. Boersma3, L. Grevillot4, S. Guatelli5, L. Maigne6, A. Resch7, P. Sitch2, M. Taylor1,2, K. Kirkby1,2, A. Aitkenhead1,2

1University of Manchester, Division of Cancer Sciences, Manchester, United Kingdom
2The Christie NHS Foundation Trust, Christie Medical Physics and Engineering, Manchester, United Kingdom
3ACMIT Gmbh, ACMIT Gmbh, Wiener Neustadt, Austria
4EBG MedAustron GmbH, EBG MedAustron GmbH, Wiener Neustadt, Austria
5University of Wollongong, Centre For Medical Radiation Physics, Wollongong, Australia
6Campus Universitaire des Cézeaux, Laboratoire de Physique de Clermont, Aubière Cedex, France
7Medical University of Vienna, Department of Radiation Oncology and Christian Doppler Laboratory for Medical Radiation Research for Radiation Oncology, Vienna, Austria

Goal: Monte-Carlo simulations are crucial for independent proton therapy dose calculations. This study aims to quantify the influence of the underlying Monte-Carlo settings on dose and calculation time.

Methods: Using GATE-RTionV1.0(GateV8.1/Geant4V10.3.3), dose distributions are simulated with a range of physics lists which are most commonly used in proton therapy (QGSP_BIC/QGSP_BIC_EMY/QGSP_BIC_EMZ/QGSP_BIC_HP_EMZ). Additionally, various thresholds on the production of secondary particles (range-cuts,0.1-10mm) are investigated. Simulation results are compared to a vast amount of measurements consisting of commissioning and patient specific quality assurance data, and within patient-CT.

Results:- Physics-lists: The use of High Precision neutron data libraries does not substantially influence dose in the target region (Figure1b/Figure2b,2f), but increases calculation times by up to 15%. Due to different scattering models, doses simulated with different electromagnetic settings differ by up to 1% (Figure1c/Figure2c,2g). 2%/2mm gamma-agreement to patient specific quality assurance measurements is not substantially different (140 measurements, average 96±4% for both QGSP_BIC_EMZ and QGSP_BIC_EMY). The EMY-constructor (corresponding to G4EMStandardPhysicsOption3) decreases calculation times by up to 22% compared to EMZ (G4EMStandardPhysicsOption4). -Range-cuts: (1mm/10mm) versus (0.1mm/1mm) in (phantom/world) range-cuts do not substantially influence dosimetric calculations in solid-water (Figure1d), but lead to differences up to 2% in the presence of heterogeneities (Figure2d,2h). Calculation times are up to 79% lower when using larger range-cuts.

Conclusion/Outlook: Calculation times are substantially reduced by optimising Geant4 settings for proton therapy: Changing physics-list and range-cuts decreases calculation times by factors of 6 (solid-water) and 3 (patient-CT-based modelling) while dosimetric impact is within clinical tolerances. Future work will investigate effects on LET and biological models.

OP 28

Impact of spot positional error in robustly optimized Intensity modulated proton therapy plan of cranio-spinal Irradiation (CSI)

D.N. Manthala Padannayil1, D. Sharma1, K. Patro1, M. Arjunan1, G. Krishnan1, R. Thyagarajan1, C. Srinivas2, R. Jalali2

1Apollo proton cancer centre, Medical physics, Chennai, India
2Apollo proton cancer centre, Radiation Oncology, Chennai, India

Purpose: To investigate the influence of random spot positioning error (SPE) on the dosimetric outcome of robustly optimized IMPT (RB-IMPT) plans of CSI

Methods: Six patients were planned and treated for CSI using RB-IMPT technique. Random SPE of 1 mm were simulated in each of the nominal plans in positive, negative and both direction for 25%, 50% and 75% of total spot positions (SPs) using an in-house Matlab code. The percentage dose variation (ΔD%) amongst the 6 nominal and 54 error-introduced plans were evaluated using standard dose volume indices, line dose difference and 3D gamma analysis method.

Results: The introduction of random SPE of 1mm resulted in reduction of D99%, D98%, D95% of both CTVs and PTVs by <1%, as compared to corresponding nominal plans. However, it leads to increase in D1% of OARs located outside the target, such as lens up to 16.9%. Line dose in the junction region showed in Fig 1. The 3D gamma values for 3% at 3mm and 2% at 2 mm were above 99% and 95% in all 54 SPE introduced plans. The worse decrease in gamma values were observed for 1% at 1mm with values ranging from 64-78% for SPE of 75 to 25%. (Fig 2)

Conclusion: RB IMPT plan for in this study is robust enough for target coverage including cribriform plate even if there is random spot positioning error of 1 mm. However, spot positioning error introduced plans lead to increase in OARs dose located close proximity to irregular shaped target.

OP 29

Proton therapy in Argentina: an opportunity for Latin America

G. Santa Cruz1, P. Menendez2

1NATIONAL ATOMIC ENERGY COMMISSION, Research and Development in Nuclear Applications for Human Health, Buenos Aires, Argentina
2UNIVERSITY OF BUENOS AIRES- Angel H Roffo Oncology Institute, Radiotherapy Department, Buenos Aires, Argentina

Latin America is made up of 20 countries, with approximately 630 million inhabitants. It is a highly heterogeneous region in terms of development, access to health, child mortality rates, life expectancy at birth and other parameters that determine each country's qualification as developed or developing. Argentina is a developing country of 45 million people, a leader in Latin America in the field of nuclear technology for peaceful purposes lead by the activities of the National Atomic Energy Commission (CNEA), from basic sciences and engineering to radiation applications to human health, including hadron therapy with thermal neutrons.

Within the framework of an extensive national plan for nuclear medicine and radiotherapy initiated several years ago, and in a joint effort between the University of Buenos Aires and CNEA, we recently started the construction of the “Centro Argentino de Protonterapia” (CEARP), the first proton therapy center in Central and South America.

In this presentation we will show the significance that this center has for our country and the region, its characteristics, infrastructure, organizational concepts and clinical and scientific prospects, the work done between CNEA, the University of Buenos Aires and its teaching oncology hospital, the Angel H. Roffo Oncology Institute.

It will be the first time in Latin America that a state-of-the-art proton technology will be available, providing cutting edge research and offering health services to the community, promoting public and private initiatives to reinforce radiation therapy options and encouraging Latin American countries to plan the possibility of installing additional centers.

OP 32

Adaptive proton therpy planning – indications, frequency, dos and don'ts, Apollo Proton Cancer Center (APCC) experience

U. Gaikwad1, S. Chilukuri1, N. Sapna1, D. Sharma2, P. Ch Katikeshwar1, M. Naufal2, R. Thiyagarajan2, M. Wakde1, S. Karthekeyan1, R. Jalali1

1Apollo Proton Cancer Center Chennai- India, Radiation Oncology, Chennai, India
2Apollo Proton Cancer Center Chennai- India, Medical Physics, Chennai, India

Introduction– Retrospective audit of 110 consecutive patients, to analyse the incidence and need for adaptive re-planning was performed.

Patients and Methods–Initial 110 consecutive patients treated with PBS technique with daily image guidance were included. All plans were generated on Raystation v.7 using Monte Carlo algorithm 4.1. Patients underwent imaging either with CT or MRI after every 5-10 fractions or earlier based on CBCT. The plans were overlaid or recalculated on QA (quality assurance) scans to assess dose perturbations and need for adaptive re-planning.

Results– At least 3 QA images were acquired for every patient, and a total of 382 QA images were analysed. 37 patients (33.6 %) required adaptive re-planning, of which 28(75.7%) underwent re-planning once, 6(16.2 %) twice and 3(7.1%) thrice. Of the patients undergoing re-planning, 16 (43.2%) were of head neck, 7(18.9%) brain, 5 (13.5%) pelvic, 3 (8.1%) skull base, 3 (8.1%) CSI, 2 (5.4%) breast and 1 (2.7 %) was of thoracic subsite. Deformation of either skin, subcutaneous tissues, sinuses or body cavities accounted for 35.1% of re-plans followed by target deformation in 24.3%, OAR deformation in 13.5%, large setup errors in 5.4% and combination of either in 21.6%. 59% of plans were adapted due to overdose in OARs while 29.7 % were due to undercoverage of target and 10.8 % had both.

Conclusion – Despite careful immobilisation, avoidance of certain beam directions and robust optimisation, frequent QA imaging is required to assess need for adaptive re-planning owing to high sensitivity of PBS technique to tissue deformation.

OP 33

Demographic analysis of first 100 patients treated with image-guided IMPT at South-Asia's first proton therapy centre in India

P.K. Panda1, S. Chilukuri2, S. Nangia2, D. Shamurailatpam3, R. Jalali2

1Apollo Proton Cancer Centre, Clinical Research, Chennai, India
2Apollo Proton Cancer Centre, Radiation Oncology, Chennai, India
3Apollo Proton Cancer Centre, Medical Physics, Chennai, India

Background: This is a demographic analysis of the first 100 patients treated using image-guided intensity modulated proton therapy (IMPT) at Apollo Proton Cancer Centre, Chennai, India which is South-Asia's first proton beam therapy (PBT) centre.

Methods: Data was collected from a prospectively maintained institutional database. Quality of life (QOL) and treatment related toxicity assessments were done at pre-defined regular intervals. Descriptive statistical analysis and Chi-square test and Fisher's exact test was used for comparisons.

Results: Between January-October 2019, 100 patients were treated with IMPT using Ray-Station treatment planning system and Monte-Carlo algorithms. The first gantry clinically operational for 8 hours per day, 5 days a week, subsequently the second gantry being operational since September 2019. The machine up-time during this time had been consistently 98% and 75% during March-April. Twenty-three (23%) were paediatric patients. Seven paediatric patients have been treated under anaesthesia. As per site specific cancer management teams (CMT), 48(48%) were treated under neuro-oncology CMT including 11 skull base tumours and 7 patients with cranio-spinal irradiation, 16 (16%) under the head and neck CMT, 21 (21%) under bone and soft tissue (BST) CMT, 6 (6%) under thoracic CMT, 7(7%) under genito-urinary CMT which includes prostate cancers, 1(1%) under gastro-intestinal CMT and 1(1%) under breast CMT.

Conclusion: Our initial experience of treating our first 100 patients with image-guided IMPT has shown that central nervous system tumours, head and neck tumours along with skull-base tumours are the major indications for IMPT in this part of the world.

OP 34

Efficient proton arc therapy plan with minimal energy layer switches

M. Chen1, X. Gu1, W. Lu1

1UT Southwestern Medical Center, Radiation Oncology, Dallas, USA

Proton arc therapy (PAT) has been demonstrated for its potential benefit of improved dose homogeneity and normal tissue sparing. However, the energy layer switching is still a major component that affects the total treatment delivery time. The determination of spot parameters (energy, position, and intensities) for proton arc beams can become a very large optimization problem and thus difficult to solve and inefficient to deliver. In this work, we proposed a novel energy layer optimization algorithm to achieve quality plans and efficient delivery for PAT.

We formulated the optimization as finding single energies, one energy for each arc, for multiple arcs, such that deviation of delivered dose from desired dose is minimized. We solved the optimization using column generation that progressively improves plan quality.

For the simulated head-and-neck cases, the proposed energy layer optimization yielded proton arc plans with better plan quality compared to those of static fields with full energy layers. With up to four energy layers it is sufficient to achieve comparable plan qualities (measured by homogeneity index and maximum OAR dose) as PAT with full energy layers, while the delivery time is only fractional.

Proton arc therapy is a type of rotational radiotherapy, known for improved plan quality because of cross-firing radiation. With the novel energy layer optimization algorithm, we demonstrated the feasibility of proton arc plans with tractable solutions that can be delivered in a minimal number of energy layers to achieve efficient quality plans. The proposed method may pave the road to clinical implementation of PAT.

OP 35

Beam model for the GPU-accelerated Monte Carlo FRED for clinical routine support in a compact spot scanning proton therapy system

J. Gajewski1, A. Schiavi2, N. Krah3, V. Patera2, G. Vilches-Freixas4, J. Martens4, A. Rucinski1, B. Nijsten4, G. Bosmans4, I. Rinaldi4

1Institute of Nuclear Physics, Pan, Krakow, Poland
2University of Rome, Sapienza, Rome, Italy
3University of Lyon/CNRS CREATIS UMR5220, Centre Léon Bérard- Creatis UMR5220, Lyon, France
4MAASTRO Clinic, Proton therapy center, Maastricht, Netherlands

The purpose of this work was to implement and validate a proton beam model for a compact MEVION machine in the GPU-accelerated Monte Carlo (MC) code FRED.

The beam model was implemented and validated against commissioning measurements and compared to clinical Treatment Planning System (TPS). This step required significant changes in the core of FRED because of the peculiarities of the machine: The beam exits the accelerator with a pristine energy of around 230 MeV and travels through a dynamically extendable nozzle, which contains the beam monitor system, the Range Modulation System (RMS), and the Adaptive Aperture (AA). In certain machine configurations, the nozzle overlaps with the patient CT voxel geometry (figure 1). We extended the MC code to properly handle such situations. We successfully implemented and validated the AA system which moves the leaves partially into the beam path. The AA module trims the spots laterally over the 20x20 cm2 area at the isocenter plane. We also implemented an alternative conversion method from CT-number to relative stopping power based on a lookup table between CT-number and mass density. Patient dose distributions recalculated with FRED are comparable with TPS, but calculation time is significantly shorter (figure 2).

FRED is now ready as tool for plan verification based on machine log files and daily (on-the-fly) dose recalculations. As a next step, we will implement an automatic pipeline in our facility which exploits the high computational efficiency of FRED.

OP 36

A minimalist GPU-based Monte Carlo proton dose and dose-averaged LET algorithm

N. Depauw1, M. Williams1, T. Madden1, B. Clasie1, H. Kooy1

1Massachusetts General Hospital, Radiation Oncology, Boston, USA

The use of Monte Carlo (MC) calculations is now standard radiotherapy practice. A clinical MC should model the minimal set of physical effects of clinical relevance to improve computational performance. We here present GMC, a minimalist MC that only models the Coulomb scattered proton tracks in heterogeneous medium in a voxel grid. At each interaction point, the MC uses a constant – medium and energy dependent – scattering angle (Gottschalk) and uses a single random number to compute the azimuthal angle around the initial proton direction. This track calculation correctly models range straggling and scatter in heterogeneous medium. Energy loss is derived from the measured, halo-corrected, depth dose as a function of medium and radiological path-length. The use of measured depth doses eases commissioning and ensures dosimetric accuracy. Dose-averaged LET is accumulated from a functional form described by Wilkens and Oelfke. We accumulate in each voxel: dose, squared dose (for error estimation), and a depth-dose function that is the product of TOPAS calculated dose-averaged LET and depth dose. These values, after the simulation at sufficient statistics, yield the dose and dose-averaged LET per voxel. GMC uses an efficient spot-to-dose-point indexing to compute dose data structures for optimization algorithms without loss of efficiency compared to the bulk transport of protons in MCs. We present the design of the MC, the physical models to compute dose and LET, and clinical results. GMC is deployed on a GPU and performance is proportional to the number of available cores.

OP 37

The accuracy of different dose calculation algorithms for pencil beam with range shifter in heterogeneous medium interface

Y.H. Lin1, H.Q. Tan1, J.H. Phua1, L.K.R. Tan1, Z. Master1, K.W. Ang1, J.C.L. Lee1, S.Y. Park1

1National Cancer Center Singapore, Department of Radiation Oncology, Singapore, Singapore

Purpose: To evaluate the accuracy of Proton Convolution Superposition and Acuros PT for pencil beam with range shifter in heterogeneous medium interface.

Methods: Monte Carlo (MC) algorithm with GEANT4 code is used to simulate the Hitachi proton beam with range shifter at different positions from isocenter. The simulated integrated depth doses and beam-spots are configured in Eclipse treatment planning system (version 15.6) using proton convolution superposition (PCS) and Acuros PT (APT) dose models. The effects of range shifter at different positions from isocenter are studied based on the single spot characteristics in heterogenous medium interface and compared against an independent MC calculation.

Results: The effect of beam-spot broadening is observed in range shifter for both PCS and APT. This effect increased when the range shifter is positioned far from the isocenter. The difference in spot size is larger for more distal depths of higher energy proton. APT modelled better than PCS in the heterogeneous medium interface when compared against an independent MC calculation. The difference of R80%distal end is less in APT but more in PCS when compared against an independent MC calculation.

Conclusion: The potential inaccuracies of dose calculation models for range shifter in heterogeneous medium interface using Eclipse treatment planning system have been identified. However the differences are within the tolerance recommended by AAPM Task Group.

OP 38

First experimental validation of daily adaptive proton therapy – workflow implementation and performance

M. Matter1,2, L. Nenoff1,2, M. Charmillot2, D.C. Weber1,3,4, A.J. Lomax1,2, F. Albertini1

1Paul Scherrer Institut, Center for Proton Therapy, Villigen, Switzerland
2ETH Zürich, Depatment of Physics, Zürich, Switzerland
3University Hospital Zurich, Department of Radiation Oncology, Zurich, Switzerland
4University Hospital Bern, Department of Radiation Oncology, Bern, Switzerland

Proton therapy treatment quality can be improved by the use of Daily Adapted Proton Therapy (DAPT). This however requires a highly efficient workflow that minimizes the time between imaging and delivery. The performance of such a workflow, in terms of total treatment time, has been experimentally investigated using an in-house developed DAPT-workflow.

Our DAPT workflow is shown in figure 1. With a CT on rails, a daily image is first obtained. Structures are then rigidly propagated to this CT and the daily plan fully re-optimized using DVH constraints of the nominal plan. After plan QA, this plan is then delivered. To minimize the time between imaging and delivery, clinically integrated software for efficient execution of all online adaption steps has been developed, as well as tools for comprehensive and automated QA checks.

Total treatment time for DAPT was investigated for a 3 field treatment delivered in 3 fractions to an anthropomorphic phantom with interchangeable nasal cavity fillings. For each delivered fraction, the timing of each step of the workflow was recorded (figure 1), and compared to a non-DAPT delivery to the same case. Film measurements performed in the phantom demonstrated significantly improved target coverage using the DAPT approach (figure 2). Total treatment time however increased by only 10 minutes/fraction for DAPT vs. non-DAPT treatments (25 vs. 15 minutes/fraction, respectively).

With our DAPT workflow, we have experimentally demonstrated significantly improved delivery accuracy, whilst treatment time per fraction was increased by 10 minutes compared to the non-adaptive approach.

OP 39

Inter-fraction dosimetric effects on dominant intraprostatic lesions dose escalation with intensity-modulated proton therapy

J. Zhou1, S. Tian1, T. Wang1, P. Patel1, S. Patel1, M. McDonald1, J. Bradley1, K. Langen1, T. Liu1, X. Yang1

1Emory University, Radiation Oncology, Atlanta, USA

Purpose: To investigate the inter-fraction anatomy changes and the corresponding dosimetric effects in the dominant intraprostatic lesion (DIL) boost protocol using intensity-modulated proton therapy (IMPT).

Methods: Twenty-five previously treated prostate patients with the first-week daily CBCT images (122 total) were used to investigate such dosimetric effects. Treatment plans using robust optimization (5mm setup, 3.5% range uncertainties) were applied to the deformed planning CT for dosimetric evaluation. DILs were contoured based on co-registered multiparametric MRI images. Simultaneous integrated boost (SIB) plans (70Gy in 28 fractions) were created so that a) the prostate CTV V100%>99%, b) DIL V98Gy>95% when possible, c) DIL V140Gy as much as possible, and d) all OARs meet the clinical goals. The distance from DIL to the bladder (Dist_DIL_Blad), rectum (Dist_DIL_Rectum), and urethra (Dist_DIL_Urethra) in both CT and CBCT were calculated. The femur deformation vectors (Femur_Deform) were evaluated. The predictors for dosimetric effects to DIL V98Gy, bladder, and rectum Dmax were calculated.

Results: Average bladder and rectum volume changes between CT and CBCT images are 100.6±39.5% and 112.8±27.2% (p=0.02), respectively. The Femur_Deform was 3.16±2.52mm. The DIL V98Gy in the initial and evaluation plans were 89.3±19.9% vs. 81.8±26.3% (p=0.18). The corresponding rectum and bladder Dmax were 71.96±0.63 vs. 74.53±5.54Gy (p=0.02) and 70.12±2.39 vs. 74.88±11.58Gy (p=0.04), respectively. The predictors for DIL V98Gy reductions are Dist_DIL_Rectum change (p<0.01), Femur_Deform (p=0.03), and initial Dist_DIL_Blad (p<0.05).

Conclusions: Dosimetric parameters could be subjected to Inter-fraction anatomy changes. An online dosimetric prediction tool is necessary for SIB of the DIL using IMPT.

OP 40

Motion mitigation in hypofractionated scanning proton therapy for lung tumors: Potential of 4D robust optimization in comparison to gating

E. Mastella1, S. Molinelli1, A. Pella2, A. Vai1, D. Maestri1, V. Vitolo3, G. Baroni2, F. Valvo3, M. Ciocca1

1CNAO Foundation – National Center for Oncological Hadrontherapy, Medical Physics Unit, Pavia, Italy
2CNAO Foundation – National Center for Oncological Hadrontherapy, Clinical Bioengineering Unit, Pavia, Italy
3CNAO Foundation – National Center for Oncological Hadrontherapy, Clinical Radiotherapy Unit, Pavia, Italy

Purpose: To evaluate the potential of four-dimensional robust optimization (4DRO) to reduce interplay effects in hypofractionated pencil beam scanning proton therapy for lung tumors.

Methods and Materials: Full 4DRO was compared to a restricted approach combined with gating, in which only 3 breathing phases of the reconstructed 4DCT (corresponding to the gate window) were included in the optimization problem instead of the whole breathing cycle. Plans were optimized for 10 patients: target coverage(D98%) and dose to healthy tissues were evaluated using Wilcoxon signed-rank test.

Plans were recalculated on re-evaluation 4DCTs to compare the robustness to interfractional respiratory variations.

Treatment delivery accuracies of the two strategies were compared with ionization chambers measurements on a heterogeneous moving phantom(see Figure 1). Robustly optimized plans with a prescription of 6 Gy(RBE) were delivered in different dynamic conditions.

Results: The same target coverage was reached with both 4D robustly optimized plans(p=0.42), while statistically significant decreases of the dose to healthy tissues and of the total number of particles were found using the restricted approach(p=0.002). The same robustness to interfractional respiratory variations was shown(p=0.18 for D98%). Phantom measurements confirmed the delivery accuracies of the two strategies(mean dose deviations<5%). Higher deviations were found for ungated full 4DRO and large motion amplitude(see Figure 2).

Conclusion: Both 4D strategies showed to be feasible for hypofractionated treatments of lung tumors. The restricted approach combined with gating improved normal tissue sparing and was shown to be more robust to single fraction deliveries. Interplay became more severe with amplitude for ungated full 4DRO plans.

OP 41

Delivery uncertainty estimation using daily breath-hold cone-beam CTs for liver proton stereotactic body radiotherapy

H. Chen1, J. Meyer2, A. Narang2, S. Han-Oh2, K. Ding2, J. Wong2, C. Tsien1, H. Li1

1Johns Hopkins Unverisity, The Johns Hopkins Proton Therapy Center, Washington DC, USA
2Johns Hopkins Unverisity, Radiation Oncology and Molecular Radiation Sciences, Baltimore, USA

Purpose: To estimate for proton stereotactic body radiotherapy (SBRT) plan delivery uncertainty with daily cone-beam CT (CBCT) imaging for liver patients under breath-hold with active breath coordinator (ABC).

Methods: Images from one liver patient previously treated with photon SBRT with ABC at our institution are used for this study. The prescribed dose for the proton SBRT plan is 60 Gy (RBE) in 5 fractions. There is a total of 21 setup CBCT images collected during the patient's previous photon SBRT treatment. The planning CT and all CBCT images are imported into Raystation 9A. A CT relative linear stopping power (RLSP) curve is calibrated for CBCT and used for proton dose calculation. A four beams proton SBRT plan is optimized with the single field optimization (SFO) technique.

All 21 CBCT images are registered to planning CT with bony anatomy (rib cage and vertebrae) or liver to evaluate the delivery uncertainty. The delivery dose is estimated from the proton dose calculation on each CBCT image with the same proton plan. The dose coverages for targets are evaluated for: gross tumor volume (GTV), gtv_multABC_2mmRad_5mmSI, and ptv_5mmRad_7mmSI.

Results: Liver positions under DIBH with ABC are not consistent under bony alignment as measured (2.2 mm SI, 0.4 mm AP and 1.1 mm LR). Proton plan delivery uncertainty can dramatically reduce under liver alignment (5% prescription dose).

Conclusion: Daily CBCT images could be potentially used to estimate proton plan delivery uncertainty for liver patients.

OP 42

Impact of radiosensitivity and sex on paediatric cranial tumours following Intensity-modulated photon and proton radiotherapy- a radiobiological response modelling

M. Dell'oro1,2, M. Short1, P. Wilson2,3, C.H. Hua4, M. Gargone4, T.E. Merchant4, E. Bezak1,5

1University of South Australia, Cancer Research Institute and School of Health Sciences, Adelaide, Australia
2Royal Adelaide Hospital, Radiation Oncology, Adelaide, Australia
3University of South Australia, School of Engineering, Adelaide, Australia
4St. Jude Children's Research Hospital, Radiation Oncology, Memphis, USA
5University of Adelaide, Department of Physics, Adelaide, Australia

Objectives: Proton therapy has superior dose distribution compared to photon therapy, potentially reducing normal tissue complication probability for organs at risk (OARs). Previously published 3D-conformal comparative planning studies have been superseded by intensity-modulated proton and photon therapy (IMPT and IMRT). The aim of this study was to compare proton and photon plans with respect to target size and location for paediatric intracranial tumours.

Methods: Six gender-matched paediatric cranial datasets (5, 9 and 12 years) were planned in Varian Eclipse treatment planning system (version 13.7). Up to 108 IMPT were robustly optimized and 108 IMRT plans retrospectively planned to treat supratentorial (ependymoma) and infratentorial (medulloblastoma) target volumes, including simulated variations in size (diameter ranging from 1 to 3 cm) and position (central, 1 and 2 cm shifts). Dose and volume data were extracted for the comparative plans to assess the impact of target volume size/position on OAR outlined by radiation oncologists.

Results: A total of 216 plans were created by a single planner with IMRT/IMPT planning training and experience. Planning objectives were achieved for all plan pairs as per clinical protocols (Figure 1). Preliminary results on average mean dose to selected OARs are shown in Table 1.

Conclusion: The study compared latest clinically relevant proton and photon treatment techniques across a large range of simulated clinical scenarios demonstrating a dose reduction to normal tissues, particularly significant for centralised tumour, using IMPT across both tumour sites. These findings inform the next phase of research related to modelling intrinsic radiosensivity of OARs.

OP 44

IMRT or proton therapy in the treatment of early stage mediastinal Hodgkin Lymphoma (HL): Choosing technique – an Institut Curie experience

Y. Kirova1, F. Goudjil2, O. Blot1, A. Boileve1, M. Amessis1, B. Deau3, R. Dendale4

1Institut Curie, Department of Radiation Oncology, PARIS, France
2Institut Curie, Department of Radiation Oncology, Orsay, France
3APHP Cochin-, Hematology Department, Paris, France
4Institut Curie Centre de Protontherapy, Department of Radiation Oncology, Orsay, France

Background and Purpose: Proton therapy (PT) may help to reduce the radiation dose to the organs at risk (OARs) and reduce toxicities. Because PT may have disadvantages, identifying candidates with real benefice is important. We present the evolution in our institutional selection criteria.

Material and Methods: We studied all early stage mediastinal HL patients treated in our Department for radiation therapy (RT) between January 2018 and July 2019. All patients had diagnosed histologically HL and had PET scan. All patients received chemotherapy and RT in respect of ESMO and ILROG Guidelines. The selection of RT technique: Intensity modulated RT (IMRT) or PT were selected using the ILROG Guidelines. For the dosimetry PBS, Eclipse and Tomotherapy TPS were used.

Results: We studied 23 HL patients. There were 10 females and 13 males, aged between 18-66 (median, 29). All were irradiated using IS (involved site) RT. The median follow-up was 6 months (3-16). Of 23, 11 were presented at the Proton weekly meeting, of them 5 were not accepted for protons. Six were accepted and underwent 4D-CT scan free breathing and with spirometry control. Of them, 3 female patients, aged 23-24-29, were treated by protons because the benefice in terms of dose to heart, lungs and breasts. They were treated with spirometry breath hold. The early tolerance was excellent with grade I radiodermatitis in 1 case and 1 grade I dysphagia. No late toxicity is observed.

Conclusion: Careful selection of patients-candidates for PT for HL is needed as well as evidence-based guidelines on how to select them.

OP 45

Head and neck malignancies involving skull base and ipsilateral neck: Comparison of proton therapy vs volumetric modulated arc therapy

M. Kharouta1, R. Pidikiti1, N. Damico1, D. Hawkins1, S. Choi1, J. Dorth1, D. Mansur1, M. Machtay1, M. Yao1, A. Bhatt1

1University Hospitals/Seidman Cancer Center at Case Western Reserve University, Radiation Oncology, Cleveland, USA

Background: Head and Neck location is surrounded by several critical organs at risk (OAR's) and thus minimizing the integral dose while treating tumors in this location is advantageous. Our aim is to evaluate the dosimetric differences of volumetric modulated arc therapy (VMAT) as compared to 3-D proton therapy (PBT).

Methods: Thirty-two patients with head and neck malignancies who were treated with PBT and had a comparison VMAT plan available were evaluated. When neck was treated, only the ipsilateral neck was radiated. The average mean and maximum doses to the brain avoidance, bilateral optic structures, cochlea, parotids, oral cavity, larynx and esophagus were compared across treatment modalities using Wilcoxon test, with use of the Benjamini–Hochberg correction for multiple comparisons.

Results: M: F = 22: 10; Median age was 69 years; Median dose was 60 CGE (55.8-70). Target volume coverage was comparable in both PBT and VMAT plans. Compared to VMAT, PBT plans showed a significant reduction in several mean and maximum doses to the OAR's as shown in Table 1, especially in the eyes, parotid glands, larynx, and oral cavity.

Conclusions: PBT as compared to VMAT resulted in meaningful dose reductions to OAR's while maintaining comparable target coverage. PBT should be considered as the optimal modality in head and neck malignancies requiring treatment to skull base and/or ipsilateral neck. Refinements in proton therapy including intensity modulation may have the potential to further minimize dose to critical structures in this location.

OP 46

Proton radiotherapy for locally advanced nasopharyngeal carcinoma: early clinical outcomes from a single institution

V. Williams1, B. Sasidharan2, S. Aljabab3, U. Parvathaneni1, G. Laramore1, T. Wong4, J. Liao1

1University of Washington, Radiation Oncology, Seattle, USA
2Christian Medical College, Radiation Oncology, Tamil Nadu, India
3Roswell Park Comprehensive Cancer Center, Radiation Oncology, Buffalo, USA
4Seattle Cancer Care Alliance, SCCA Proton Therapy Center, Seattle, USA

Purpose/Objective(s): Advances in radiotherapy have improved tumor control and reduced toxicity in nasopharyngeal carcinoma (NPC). Local failure remains a problem as well as acute/late toxicities. Proton therapy (PT) offers dosimetric advantage over IMRT, which can further improve the therapeutic ratio. We report our early clinical outcomes with PT for locally advanced NPC.

Materials/Methods: We reviewed patients enrolled on a prospective IRB-approved clinical registry study who received PT for NPC. Demographics, dosimetry, disease control outcomes, and acute/late toxicities were reviewed (CTCAE v.4).

Results: 21 patients treated from 2015-2018. 6 female, 15 male, median age 57. T stage: T1, 4; T2, 4; T3, 1; T4, 12. N stage: N0, 2; N1, 7; N2, 10; N3, 2. 95% stage III-IV. WHO classification type 1 in 6, type 2/3 in 14, unknown in 1. 71% EBV positive. Dose-painted PBS approach with 2-5 beams encompassing primary/bilateral neck. Majority were treated to 69.96 CGE, in 33 fractions once daily. All received concurrent CDDP; 5 also received induction chemotherapy. With median follow up 16 mos, 18 patients NED, 1 local failure and 2 distant metastases. Locoregional control 95%, DM free rate 90%, and OS 90%. Acute toxicities: grade 3 mucositis 14, grade 3 dermatitis 9. Late toxicities: grade 2 xerostomia 2, hearing loss 3, and only 1 patient remains PEG-dependent.

Conclusion: Proton therapy is feasible in locally advanced NPC with early outcomes demonstrating excellent locoregional control and favorable toxicity profile. Longer follow up and additional comparative studies are needed to evaluate the relative advantages compared to IMRT.

OP 47

MR-imaging of Uveal Melanoma for Proton Therapy

J.W. Beenakker1, T. Ferreira2, M. Marinkovic3, M. Jaarsma-Coes1, C. Rasch4, G. Luyten3

1Leiden University Medical Center, Ophthalmology and Radiology, Leiden, Netherlands
2Leiden University Medical Center, Radiology, Leiden, Netherlands
3Leiden University Medical Center, Ophthalmology, Leiden, Netherlands
4Leiden University Medical Center, Radiotherapy, Leiden, Netherlands

Introduction: MRI is increasingly being used in the diagnosis and treatment planning of Uveal Melanoma (UM), the most common primary malignant ocular tumor. As the superior soft-tissue contrast and 3D imaging capabilities of MR, compared to the conventional ultrasound, are also beneficial for Proton Therapy (PT) planning, multiple efforts are undertaken for MRI-based PT planning of ocular tumors. However MR-imaging of the eye is challenging due to motion artefacts and its small size. To overcome these challenges, we developed a dedicated MRI-protocol for UM and evaluated its use for PT planning.

Methods: Thirty UM patients were scanned with a dedicated protocol at a 3Tesla MRI, using an eye-coil. The protocol consisted of 3D isotropic sequences (resolution <(0.9mm)3), to assess the 3D tumor geometry, 2D sequences (resolution <(0.5mm)2), to screen for invasion of the sclera or ciliary body, and functional scans [Ferreira, Cancers 2019]. The image quality to assess the tumor dimensions and invasion was scored by two experienced readers.

Results: To determine the tumor dimensions: In 20/30 patients the quality of the 3D MR-images was good to perfect, in 3/30 patients insufficient, and for the remaining 7 patients sufficient. The quality of the 2D MR-images was sufficient in all patients to assess tumor invasion of nearby tissues, with a quality reaching good to perfect in 26/30 patients.

Conclusion: With a dedicated protocol, high resolution MR-imaging of the eye is possible for UM patients, enabling its use for the treatment planning of and follow-up after PT.

OP 48

Can supinely acquired MR-images be used to plan ocular proton beam therapy in sitting position?

M. Jaarsma-Coes1, M. Hassan2, E. Astreinidou3, M. Marinkovic4, F. Peters3, C. Rasch3, J.W. Beenakker1

1Leiden University Medical Center, Radiology and Ophthalmology, Leiden, Netherlands
2Leiden University Medical Center, Radiology- C.J. Gorter Centre for High Field MRI, Leiden, Netherlands
3Leiden University Medical Center, Radiotherapy, Leiden, Netherlands
4Leiden University Medical Center, Ophthalmology, Leiden, Netherlands

Introduction: Proton beam therapy (PBT) is often the therapy of choice for large uveal melanoma (UM). MRI is increasingly being used for the clinical target volume definition and PBT-planning. However, PBT is performed in sitting position, while the acquisition of the MR-images is performed in supine position. We therefore assessed the effect of different patient positions on the eye- and tumour- shape.

Methods: Seven volunteers and five UM-patients were scanned in two positions on a 3T Philips MRI scanner. One set of images was acquired in supine position, while a second set was acquired mimicking the patient sitting for PBT (flexed). Additionally, two volunteers and one patient were scanned twice to assess reproducibility. After registration and segmentation, the distances between the tumor and sclera in both positions were calculated.

Results: The median distance between eye-shape in supine and flexed position was 0.1mm (95thpercentile: 0.3mm), which is in the order of the reproducibility of the method (95thpercentile: 0.4mm). The median difference in tumour shape was 0.2mm (95thpercentile: 0.4mm). Both differences are lower than the acquisition voxel size (<1.0mm3).

Conclusion: Change in gravity direction produces no substantial changes in sclera and tumour shape. This result indicates that supinely acquired MR images can be used to plan ocular PBT, which is performed in sitting position.

OP 51

Development of real-time neutron detectors at whole body position in BNCT

N. Matsubayashi1, T. Takata2, M. Sato1, Y. Sakurai2, H. Tanaka2

1Kyoto university, Graduate school of engineering, Kyoto, Japan
2Institute for Integrated Radiation and Nuclear Science- Kyoto University, Particle Radiation Oncology Research Center, Osaka, Japan

In the irradiation of BNCT, a few neutrons with various energies pass through the collimator components. It is desired to individually measure dose of thermal, epithermal, and fast neutron in real-time, because RBE is different in each energy region. The purpose of this study is the development of real-time detectors that can measure neutron dose discriminated in three energy regions at whole body position. We selected the combination of small LiCAF scintillator and quartz fiber to measure thermal neutrons. Polyethylene is used as moderator for the detection of epithermal and fast neutron, because LiCAF scintillator has a low sensitivity in epithermal and fast energy regions. When the radius of moderator is 10, 51 mm, the contribution of epithermal and fast neutron in total events is 90, 15 %, respectively. Thermal neutrons are shielded by enriched 6LiF ceramic. The irradiation test using water phantom simulated brain tumor irradiation. The count rates of thermal, epithermal, and fast neutron detectors were measured in each position of neck, chest, and abdomen. When the measured results are compared with simulated results of neutron transport code of PHITS, each dose of thermal, epithermal, and fast neutron can be derived. We developed real-time neutron detectors for individually discriminating three energy regions. Performance test was carried out and it was confirmed that the designed system was able to measure thermal, epithermal, and fast neutron dose in real-time. In the future, we plan to develop a system that can measure the components of neutrons more accurately.

OP 52

Tissue imprints in nuclear track detectors to enhance spatial resolution of neutron autoradiography

A. Portu1,2, S. Thorp3, P. Curotto4, E. Pozzi4, G. Saint Martin1

1CNEA, Radiobiology, Buenos Aires, Argentina
2CONICET, conicet, Ciudad Autónoma de Buenos Aires, Argentina
3CNEA, Instrumentation and Control, Buenos Aires, Argentina
4CNEA, RA-3 reactor, Buenos Aires, Argentina

Boron microdistribution can be addressed through neutron autoradiography. Briefly, it consists on registering the impacts of charged particles on nuclear track detectors (NTD) and correlating this information with the microstructure of the biological sample that contains the particles' emitter (e.g. 10B). When small histological regions are analyzed, high spatial resolution is required.

We have reported a methodology to produce cell imprints on polycarbonate, through UV-C sensitization. As tissue structure largely differs from cultured cells, several aspects must be considered in order to extend the methodology to tissue sections.

We studied polycarbonate (Lexan™) and polyallyldiglycolcarbonate (PADC, CR-39) as potential NTDs for the enhanced resolution neutron autoradiography. As a fading effect of UV-C on nuclear tracks had been observed in polycarbonate, it was analyzed for CR-39.

Tissue sections obtained by cryosectioning were mounted on the detectors. Neutron irradiations were performed at different fluences. Several staining agents were tested as potential promoters of the imprint formation. Different UV-C exposure times were assayed. Etching with adequate solutions was performed to reveal both imprints and nuclear tracks. An image segmentation method based on artificial intelligence was applied to quantify nuclear tracks.

While well-defined imprints of samples mounted on Lexan were obtained with only 5 min UV-C exposure, irradiations of about 6 h were necessary to yield imprints on CR-39, regardless of previous staining. Conversely, fading of nuclear tracks was only observed for Lexan. Both effects (imprint formation and fading) are a consequence of the polymers' surface photodegradation.

OP 53

The effects of duocarmycin SA, TMZ and BCNU in combination with proton irradiation on glioblastoma cell line LN-18

A. Bertucci1, A. Camacho-Santos2, R. Fuller3, N. Jankeel4, K. Boyle4, J. Slater1, M. Vazquez1

1Loma Linda University medical center, Radiation Medicine, Loma Linda, USA
2Loma Linda University, School of Medicine, Loma Linda, USA
3Loma Linda University, Basic Sciences, Loma Linda, USA
4Loma Linda University, School of Pharmacy, Loma Linda, USA

Glioblastoma multiforme (GBM) is the most common and the most aggressive form of primary brain cancer, withpatient median survival of 15–18 months. Unfortunately, because GBM tumor cells are highly infiltrative throughout the brain at diagnosis, complete removal of all tumor cells by surgery is not possible and disease recurrence following tumor resection results. New treatment procedures for glioblastoma multiforme (GBM) are urgently needed. Proton therapy is considered the most effective form of radiation therapy for GBM. Also, the most frequentlyused drugs in the treatment of GBM are members of the DNA-alkylator class, such as carmustine (BCNU) with temozolomide (TMZ). While clinically useful, TMZ and BCNU are fairly ineffective compounds. The duocarmycin class of antitumor antibiotics, exemplified by duocarmycin SA (DSA), is an exceptionally potent group of agents capable to induce a selective alkylation of duplex DNA. In order to establish the scientific rationale for future animal studies, we evaluated the comparative efficiency of DSA, TMZ and BCNU in combination with proton irradiation The experiment was performed using the human GBM cell lines U-138 MG and LN-18. Cell toxicity was measured by cell proliferation, colony formation, apoptosis induction and DNA damage. Results confirmed the extraordinary potency of DSA (IC50: 0.124 nM) in comparison with BCNU (IC50: 72 μM) and TMZ (IC50: 446 μM) for cell proliferation in combination with proton irradiation (3 Gy). Preliminary results suggest that DSA is an very interesting candidate for combined therapy approaches if toxicities to normal tissues and penetration to the blood-rain-barrier are resolved.

OP 54

Skin and mucosal reaction of reactor based BNCT

K. Nakai1, T. Aihara2, T. Yamamoto3, K. Hiroaki1, Y. Matsumoto1, A. Matsumura4, H. Sakurai1

1University of Tsukuba, Department of Radiation Oncology- Faculty of Medicine, Tsukuba, Japan
2Osaka Medical College, Kansai BNCT Medical Center, Takatsuki, Japan
3Yokohama City University School of Medicine, Department of Neurosurgery, Yokohama, Japan
4University of Tsukuba, Department of Neurosurgery- Faculty of Medicine, Tsukuba, Japan

Background: We had treated glioblastoma patients, and recurrent head and neck cancers as a non-randomized clinical study. Brain Tumor patients were treated using JRR-4, Japan Atomic Energy Agency, Tokai, Japan. HandN cancer patients were treated using Kyoto University Reactor. JRR-4 was decided to decommissioning. In this reports, focusing on skin and mucosal damage of normal surrounding tissue on the surface.

Material and Methods: We had treated 17 cases of glioblastoma from 1999 to 2010. 4 case of recurrent head and neck cancer on 2014. Retrospective analysis from medical chart and summarized documentation.

Result and Discussion: In the brain tumor BNCT, the most common acute adverse event was mild erythema (CTC-AE Grade 1), and alopecia was prolonged. 3 patients received frontotemporal irradiation without ideal neutron shielding due to skin reflection. In the latter protocol, a patient tolerated additional 40 Gy external X-ray irradiation, while a recurrent patient with needed surgery immediately before BNCT, developed skin ulcer. In the HandN BNCT, all cases had previous radiotherapy including proton beam therapy, BNCT, X-ray. 2 squamous cell carcinoma. 1 rhabdomyosarcoma, 1 adenoid cystic carcinoma. Acute radiation dermatitis was observed in all cases. 3 cases ware Grade 1, 1 case developed Grade 2. Mucositis in 3 cases. 2 cases were grade 2, 1 case was grade 3. Conjunctivitis in 2 cases. Late radiation toxicity was observed including 3 radiation necrosis.

Conclusion: Previous reactor based BNCT cases were retrospective analyzed. Acute radiation dermatitis were generally tolerable but careful planning and observation are required.

OP 55

A practical KV Target Tracking trigger for Proton FLASH treatment

W. Xiong1, H. lin2, M. Kang2, S. Huang3, F. Yu2, Q. Chen1, A. Zhai1, I. Choi2, C. Simone2

1New York Proton Center, Medical Physics, roslyn heihts, USA
2New York Proton Center, Medical Physics, New York, USA
3MSKCC, Medical Physicis, New York, USA

Background: This study aimed at building a near real time KV target tracking trigger for proton FLASH treatment. FLASH therapy delivers at least 40Gy in a second, and a single FLASH treatment field is delivered in much less than a second. It is necessary to build an image tracking system to verify the target is optimally aligned before FLASH treatment, even with DIBH.

Method and Materials: We analyzed previous intra-fraction target motion of eight consecutive patients treated with DIBH using IMR imaging technique, and we developed target residue motion model in the S-I direction. We applied the model in proton DIBH treatment to predict target residue motion and derived target motion time pattern after setup CBCT/KV imaging.

Results: Even though the external markers indicated that the respiratory motion was within 3 mm in DIBH treatment for all patients, significant residual internal target motion was observed for some patients. The range of average motion was from 0.1 to 7.9 mm, with standard deviation ranging from 1.2 to 3.5 mm. For all patients, the target residual motions seemed to be random, with mean positions around their initial setup positions and the maximum residue motion speed is 0.8mm/s.

Conclusions: We will apply DIBH to all FLASH treatments and use KV/KV imaging to verify target position just 1 to 3 seconds before treatment delivery. According to our model, this KV imaging is acting as a trigger and can best ensure the internal organ motion is within 3mm.

OP 56

Nuclear fragmentation model optimization for TPS basic data calculation in helium ion therapy

F. Horst1,2, G. Aricò3,4, S. Brons5, M. Durante1,6, T. Haberer5, A. Mairani5, U. Weber1, K. Zink2,7,8, C. Schuy1

1GSI Helmholtzzentrum für Schwerionenforschung, Biophysics, Darmstadt, Germany
2THM University of Applied Sciences, Institute of Medical Physics and Radiation Protection, Gießen, Germany
3CERN European Organization for Nuclear Research, STI-EN Engineering Department, Geneva, Switzerland
4Present Adress:, San Raffaele Scientific Institute, Milano, Italy
5University Hospital Heidelberg, Heidelberg Ion-Beam Therapy Center, Heidelberg, Germany
6TU Darmstadt, Institut für Festkörperphysik, Darmstadt, Germany
7FIAS, Frankfurt Institute for Advanced Studies, Frankfurt, Germany
8University Hospital Giessen-Marburg, Department of Radiotherapy, Marburg, Germany

Helium ions will be added to the treatment spectrum of the Heidelberg Ion-Beam Therapy Center (HIT) supplementary to protons and carbon ions in the near future. If radiotherapy with helium ions turns out to be beneficial it can be expected that other facilities will follow.

An essential part of the commissioning of the HIT facility for helium ions was the calculation and validation of basic data for the clinical treatment planning system (laterally integrated depth dose profiles, parametrized lateral dose profiles and fragment distributions in water). At HIT these basic datasets are calculated using the FLUKA code which is already well-established and benchmarked for protons and carbon ions. However, for helium ions the initial FLUKA depth dose calculations deviated from measurements by some percent in the Bragg peak region [Tessonnier et al., PMB 62 (2017)]. Those deviations were assumed to be associated with the nuclear fragmentation models in FLUKA.

In order to support the development of the nuclear reaction models in FLUKA, a series of fragmentation experiments with helium projectiles was performed at HIT [Horst et al., Physical Review C 99 (2019)]. The measured cross sections have been used to optimize the nuclear reaction parametrization in FLUKA which resulted in considerable improvements of its dose calculation accuracy.

In this contribution, the helium fragmentation experiments will be described, their implications for radiation transport modeling will be pointed out, the validation against dose measurements will be presented and recommendations for calculation of helium ion basic data will be provided.

OP 57

A parametric approach towards new proton facility shielding considerations in the era of proton FLASH therapy

N. Gupta1, J. DeFilippi2, D. DiCostanzo1, A. Ayan1, J. Woollard1, J. Meissner3, K. Hintenlang1, J. Sussi1, A. Chakravarti1

1The Oho State University, Radiation Oncology, Columbus, USA
2DeFilippi Consulting LLC, DeFilippi Consulting LLC, Arlington, USA
3Meissner Consulting GmbH, Meissner Consulting GmbH, Neubiberg, Germany

With the advent of the modern era of proton therapy and the inclusion of hypo-fractionated/FLASH treatments, the traditional paradigms/assumptions of barrier thicknesses calculation can lead to wall thicknesses that are considerably larger than existing designs.

Our institution is currently evaluating facility shielding for our recently acquired proton equipment, with considerations towards future implementation for FLASH treatments. We have developed a parametric approach to help develop shielding assumptions that are directly connected to our specific clinical experience, rather than general norms. This method uses the following input data: a) target volumes by disease sites extracted from our treatment planning; b) prescribed doses for different disease sites; c) expected disease site breakdown; d) best estimates of conventional and hypo-fractionated/FLASH case mix inclusive of the respective prescribed dose for each; and e) treatment time slots allocated for different complexity of treatments.

Our model allows us to appropriately weight all the parameters, allowing for different levels of conservativeness. The outputs of our model are weighted averages for target volumes, dose per fraction and the patients per hour as assumptions, which are typical inputs for empirical shielding calculations. Additionally, these parameters can be inputs for Monte Carlo calculations for different scenarios to parameterize the source terms directly into the weighting, which allows for the most realistic calculation of barrier thicknesses – an effort currently underway.

We feel that our approach will allow us to adequately account for FLASH and other proton treatments realistically and serve as a systematic approach for future facility design.

OP 58

Producing an accurate beam model of the varian probeam system using TOPAS Monte Carlo toolkit

M. Rahman1, P. Bruza1, Y. Lin2, D.J. Gladstone1,3, B.W. Pogue1,4, R. Zhang3

1Dartmouth College, Thayer School of Engineering, Hanover- NH, USA
2Emory University, Emory Proton Therapy Center, Atlanta- GA, USA
3Dartmouth Hitchcock Medical Center, Radiation Oncology, Lebanon- NH, USA
4DoseOptics, Limited Liability Company, Lebanon- NH, USA

Modern treatment planning system (TPS), such as the one at the Emory Proton Therapy Center (Raystation v9) simulate Varian ProBeam delivered treatment plans via Monte Carlo (MC) methods. Current commercialized TPS does not support the calculation/optimization of linear energy transfer (LET) distribution, vital to radiobiological effectiveness and scintillation dosimetry studies. The Geant4 based TOPAS MC simulation toolkit proved to accurately model other radiotherapy systems. We sought to report the first complete beam model of the ProBeam using TOPAS and simulate LET. The model was based on the commissioning data, which included integral depth dose (IDD's) in water and spot profiles in air measured at varying depths (for energies of 70 to 240MeV in increments of 10MeV, and 242MeV). Emittance was defined based on Courant-Snyder's particle transport theory and depth dependent spot profiles. Simulated and commissioned spot sizes agreed within 3%. Energy spectrum were Gaussian distributions that best matched range (within 0.1mm) and maximum dose (generally within 0.5%) of the commissioned IDD's. The simulated model is exceptionally accurate, capable of computing LET, and may improve dose profile accuracy due to inline(Y) and crossline(X) freedom in beam shaping (TPS assumes cylindrical symmetry). It will be validated further with absolute dosimetry of spread out Bragg peak and patient plans.

OP 59

Optical fibers for proton therapy dosimetry

C. Hoehr1, A. Morana2, C. Penner3, M. Trinczek4, C. Duzenli5, S. O"Keeffe6, M. Bouazaou7, S. Girard8

1TRIUMF, Life Science, Vancouver, Canada
2University of St. Etienne, Physics, St. Etienne, France
3TRIUMF, Life Sciences, Vancouver, Canada
4TRIUMF, Protn Irradiations, Vancovuer, Canada
5BC Cancer, Medical Physcis, Vancouver, Canada
6University of Limerick, Physcis, Limerick, Ireland
7University of Lille, Physcis, Lille, France
8University of St. Etienne, Pysics, St. Etienne, France

In order to measure doses in small-field proton therapy, as for example in the treatment of choroidal melanomas, small detectors with superior spatial resolution are needed. Optical fibers are ideal except for the quenching in the high LET region of the Bragg peak, which causes the deposited dose to be severely underestimated. To overcome this problem, we manufactured several different fibers, organic and inorganic, and tested them at the Proton Irradiation Facility at TRIUMF (PIF, 63 MeV), Canada. To estimate their behavior in the secondary neutron fields routinely present at proton therapy facilities, we also tested the fibers' response with neutrons at TRIUMF's neutron facility (TNF, 400 MeV to thermal).

Silica-based fibers of 0.5 mm diameter were manufactured via the sol-gel technique and doped with copper, cerium or gadolinium. Organic fibers were manufactured by drilling a small hole in the end (0.5 mm thick and 1-2 mm deep) and scintillation powder was added, Y2O2S:Eu, Gd2O2S:Eu, Gd2O2S:Tb and YVO4:Eu.

It was found that the fiber with the least amount of quenching was the inorganic Gd-doped silica fiber, which showed almost no quenching in the proton irradiation, making it an ideal detector for small-field proton therapy dosimetry. The organic fibers do show different response in neutron and proton fields, potentially making them ideal to distinguish between dose deposition in the fiber caused from different particles.

OP 61

Beam monitor calibration for scanned proton and carbon ion beams in a synchrotron-based facility

J. Osorio1, R. Dreindl1, P. kuess2, L. Grevillot1, V. Letellier1, A. Carlino1, A. Elia1, M. Stock1, S. Vatnitsky1, H. Palmans3

1MedAustron, Medical Physics, Wiener Neustadt, Austria
2Medizinische Universität Wien Medical University of Vienna, Medical physics, Wien, Austria
3MedAustron and National Physical Laboratory NPL, Medical Physics, Wiener Neustadt, Austria

Introduction: This work presents the implementation and comparison of two independent methods of beam monitor calibration in terms of number of particles (Np) per spot for scanned proton and carbon ion beams at MedAustron.

Materials and Methods: For all ionization chambers used (Farmer, Roos and Bragg Peak (BPC)) readings were corrected for ion recombination and polarity-effects. Roos and BPC were cross-calibrated against a calibrated Farmer chamber in 179.2 MeV proton and 346.6 MeV/n carbon ion beams. Single-layer scanned fields (SLSF) with identical number of particles in each beamlet and equidistant lateral spacing were used to determine dose-area-product (DAPw) as a reference method for the beam intensity monitor calibration in the clinically relevant proton and carbon energy ranges (E). As alternative method, DAPw of static beamlets was measured with the BPC. Np was determined as:
where (S/ρ)w(E,zref) is the mean mass stopping power of water per incident particle, calculated for both particle species by Monte Carlo simulations.

Results: Both calibration methods agreed on average within 1.1% for protons and within 2.8% for carbon ions (Figure-1). The uncertainty using the SLSF method is 2.5% with a major component of uncertainty in determination of kQ. The uncertainty using DAPw method is 2.6% with a major component of uncertainty in the non-uniformity BPC response, which can lead to a correction of up-to 3.2%.

Conclusion: The agreement between both dosimetry methods enhances confidence in the beam monitor calibration and the estimated uncertainty.

OP 62

A clinical quality assurance concept for conformal motion-synchronized dose delivery systems used for four-dimensional ion therapy

M. Lis1, M. Donetti2, T. Steinsberger1, M. Wolf1, A. Paz1, W. Newhauser3, C. Graeff1

1GSI Helmholtzzentrum fur Swerionenforschung, Biophysics, Darmstadt, Germany
2Fondazione Centro Nazionale Adroterapia Oncologica, Research and development department, Pavia, Italy
3Louisiana State University, Medical Physics, Baton Rouge, USA

Technical quality assurance essential to ensuring the consistency and safety of radiotherapy. There is a lack of standardized QA procedures for moving tumor treatments in ion therapy.

We have developed a comprehensive series of QA tests which can be performed with dosimetry equipment available in most clinics. The proposed tests for daily, monthly, annual and patient-specific QA (PSQA) were experimentally verified for the conformal motion-synchronized dose delivery system in development.

In order to verify the QA concept, a linear stage, with programmable motion patterns, was set to move QA phantoms with 20mm amplitude sinusoidal motion. Motion amplitudes detected with an optical distance laser sensor were converted into discrete motion states. Experimental set-ups were designed to modify existing QA procedures at CNAO to simplified patient respiratory scenarios using water phantoms and for regular and irregular respiration patterns. Test plans (cubes, squares, patient plans) were designed for these set-ups. Existing PSQA protocols, using 12 pinpoint ICs in a water phantom or 2D IC array detectors, were extended for motion mitigation tests.

The motion QA protocol was executed in conjunction with standard QA and resulted in an estimated 12-minute increase to daily QA time. Homogeneity, calculated on Gafchromic film, was 98.1% for static and 95.0% for 10 phase squares. Annual QA dose distributions were 103.7% +/- 1.2% and 103.8 +/- 3.5% of the planned dose, respectively, for cubes delivered to the pinpoint ICs.

We have demonstrated that the proposed QA concept for conformal, motion-synchronized dose delivery is feasible for clinical use.

Poster Presentations

P 001

Carbon ion radiotherapy in the treatment of gliomas: a systematic review

T. Malouff1, J. Peterson1, A. Mahajan2, D. Trifiletti1

1Mayo Clinic, Radiation Oncology, Jacksonville, USA
2Mayo Clinic, Radiation Oncology, Rochester, USA

Introduction: Gliomas are among the most common primary brain malignancies, with a poor prognosis for high grade gliomas despite aggressive therapy. Carbon ions, which exhibit favorable biological and physical characteristics, have recently been studied in intracranial malignancies as a way to escalate dose to the tumor while minimizing dose to normal tissue.

Methods: Pubmed/Medline, SCOPUS, EMBASE, CINAHL and the Cochrane database were systematically reviewed using the search terms “carbon ion” and “glioma” or “glioblastoma” in August 2019. Out of 332 articles screened, 46 were included in this analysis.

Results: This comprehensive review describes the pertinent physics and radiation biology studies relevant to the treatment of gliomas with carbon ions and summarizes the important clinical studies for both high and low grade gliomas. Studies investigating carbon ions as both definitive radiotherapy and as a boost to traditional radiotherapy are reviewed. The use of carbon ion radiotherapy in the setting of recurrent disease is also described.

Conclusions: Carbon ion radiotherapy is both efficacious and safe based on early clinical studies. Current trials, including the CLEOPATRA and CINDERLLA trials, hope to define the role of carbon ion radiotherapy in the treatment of gliomas.

P 002

The efficacy and safety of proton modulated scanning radiotherapy for low grade gliomas: the Maria Skłodowska-Curie Institute–Oncology Centre experience

D. Wojton Dziewonska1, O. Urbanowicz1, E. Pluta1, K. Urbanek1, A. Patla1, K. Kisielewicz2, E. Góra2, R. Kopeć3, T. Skóra1

1Maria Skłodowska-Curie Institute-Oncology Centre, Department of Radiotherapy, Kraków, Poland
2Maria Skłodowska-Curie Institute-Oncology Centre, Department of Medical Physics, Kraków, Poland
3Institute of Nuclear Physics Polish Academy of Sciences, Bronowice Cyclotron Centre, Kraków, Poland

Purpose: The aim of the study is to describe early clinical outcomes in patients treated with proton modulated scanning radiotherapy for low grade gliomas.

Methods and Materials: We analyzed the outcomes of proton modulated scanning radiotherapy of 58 patients with low grade gliomas treated at our facility from December 2016 to December 2018. Mean age at diagnosis was 41 (range 21-76). 53% of patients were female. Median dose delivered was 54 Gy RBE in 30 fractions. Prior to irradiation 69% of the patients underwent at least partial resection and in 31% of patients stereotactic biopsy was performed. Most common histopathologic subtype was diffuse astrocytoma (45% of patients). For each patient, we recorded tumor location, extent of resection, dose-volume comparison with alternative photon plan and use of adjuvant chemotherapy. Acute and late toxicity scores were recorded using CTCAE version 4.0 at weekly on treatment visits. Maximum grade of fatigue, headache, insomnia, nausea, vomiting, alopecia, dermatitis and need for steroids over the radiation therapy treatment course were recorded, and rates of acute toxicity were calculated. Imaging findings following proton beam radiation therapy including radiation necrosis and pseudoprogression were recorded.

Results: After a mean follow-up of 17,6 months 5 failures were observed. The mean time to recurrence was 12 months .There were no G3 toxicities due to radiotherapy.

Conclusion: Proton radiotherapy is an effective and safety treatment for low grade gliomas. Despite the high volumes of irradiated brain, the toxicity of treatment is low. The results are encouraging and further follow-up is pending.

P 003

Proton radiation therapy in patients with chordoma and chondrosarcoma of the skull base: Maria Skłodowska-Curie Institute – Oncology Centre experience

T. Skóra1, E. Pluta2, D. Wojton-Dziewońska1, A. Patla1, K. Urbanek1, K. Kisielewicz3, E. Góra3, R. Kopeć4, D. Kabat3

1Maria Skłodowska-Curie Institute - Oncology Center, Department of Radiotherapy, Kraków, Poland
2Maria Skłodowska-Curie Institute – Oncology Center, Department of Radiotherapy, Kraków, Poland
3Maria Skłodowska-Curie Institute - Oncology Center, Department of Medical Physics, Kraków, Poland
4Institute of Nuclear Physics Polish Academy of Sciences, Bronowice Cyclotron Center, Kraków, Poland

Purpose: The aim of the study is to describe early clinical outcomes (tumor control and treatment toxicity) in patients treated with proton radiation therapy for chordoma or chondrosarcoma oft he skull base.

Methods and Materials: Between November 2016 and December 2018, 50 patients (median age, 43.5 years; range, 22-77 years) with chordoma (n=32) and chondrosarcoma (n=18) of the skull base were treated with proton radiation therapy at our institution. Fifty-eight percent of patients were women. The median chordoma and chondrosarcoma dose was 74.0 GyRBE and 70.0 GyRBE, respectively. Surgical tumour resection before proton radiotherapy was performed in 49(98%) cases. Macroscopically residual disease was found in 44(88%) patients. Disease-free survival and overall survival were calculated. Treatment toxicity was evaluated using CTCAE (version 5.0) grading system.

Results: After a mean follow-up of 18.5 months (range, 1-33 months), 2 local and 1 distant failure were observed. 2-year disease-free survival and overall survival rates were 90.3% and 98.0%, respectively. The mean time to disease reccurence was 12.7 months (range, 6.5-23.8 months). High grade (G3) toxicities were rare in both the acute (n=2) and late (n=3) settings.

Conclusion: Proton therapy is an effective treatment for skull base tumours. Despite the use of high doses of radiation therapy, the toxicity of treatment is highly acceptable. These preliminary results are encouraging but should be confirmed during a longer follow-up.

P 005

Pilot study on critical structure dosing for ocular melanoma radiotherapy: Analysis of brachytherapy plaque and dedicated proton eye beamline dosimetry

J. Scholey1, V. Weinberg1, I. Daftari1, K. Mishra1

1The University of California San Francisco, Radiation Oncology, San Francisco, USA

Purpose: To analyze critical structure dose distributions using I-125 brachytherapy plaques and proton plans generated with a dedicated eye beamline for ocular melanoma treatment.

Methods: Nine ocular melanoma tumor cases were treated at our institution with I-125 brachytherapy plans (prescribed to 85Gy at tumor apex delivered over 7 days) during proton cyclotron maintenance. Clinically comparable proton plans were generated for the same tumors on the 67.5 MeV eye beamline (prescribed to 56GyE in 4 daily fractions). Maximum absolute dose (Dmax) and maximum dose as a percentage of prescription (Dmax/%Rx) were extracted for optic disc (OD), macula, and lens, with the latter metric accounting for differences in prescription values between regimens. Due to the small sample size, distributions were compared using a nonparametric Wilcoxon matched pairs test with statistical significance at p-value<0.05.

Results: Tumors characteristics were as follows [range (median)]: (a) maximum clinical basal diameter [6.7-15.8mm (9.7mm)]; (b) tumor height [1.1-10.75mm (3.4mm)]; and (c) OD-tumor distance [1.2-14mm (4.6mm)]. Dosimetric results are shown in Table 1. Dmax and Dmax/%Rx to OD was significantly different for protons versus plaques, with 7 proton plans sparing the disc. Dmax and Dmax/%Rx to macula was significantly different between protons and plaques. Dmax to lens was significantly different for protons versus plaques; Dmax/%Rx to lens did not differ significantly.

Conclusion: Dedicated eye proton beam plans demonstrated significantly different dose distributions compared with I-125 plaque plans with increased critical structure sparing and lower median doses for the OD and macula for both Dmax and Dmax/%Rx, and Dmax to lens.

P 006

Dose-volume statistics comparison of pediatric intensity-modulated proton therapy sparing the inner ear and parotid and X-ray therapy for whole brain

T. Hashimoto1, T. Mori2, S. Takao3, Y. Matsuo3, M. Tamura3, T. Matsuura3, R. Onimaru4, H. Shirato5, S. Shimizu1, H. Aoyama4

1Hokkaido University, Radiation Medical Science and Engineering, Sapporo, Japan
2Hokkaido University, Oral Radiology, Sapporo, Japan
3Hokkaido University, Medical Physics, Sapporo, Japan
4Hokkaido University, Therapeutic Radiology, Sapporo, Japan
5Hokkaido University, Proton Beam Therapy, Sapporo, Japan

Purpose: Dose-volume statistics for whole brain (WB) irradiation with intensity-modulated proton therapy (IMPT) sparing the inner ear and parotid, volumetric modulated arc therapy (VMAT), and conventional X-ray therapy in pediatric and young adult patients with central nervous system (CNS) tumors were analyzed.

Methods and Materials: Seven pediatric and young adult patients with CNS tumors treated with IMPT were selected. The median age was 9 (5-29) years. The prescribed doses to the CTV for the craniospinal region/ primary site were 23.4-36.0/ 50.4-56.0 Gy(RBE). Simulation of WB treatment plans with VMAT, and two laterally opposed fields using X-rays were generated. Dose-volume histograms were used to evaluate the dosimetric parameters, and the normal-tissue complication probability (NTCP) for xerostomia and hearing loss was calculated and compared.

Results: The mean of the average parotid and inner ear doses were 8.1 (95%CI: 3.3-12.9) Gy(RBE)/ 8.3 (-4.4-21.0) Gy/ 14.1 (7.6-20.5) Gy (IMPT<X-ray (P=.0013)), and 34.2 (95%CI: 30.1-38.1) Gy(RBE)/ 40.0 (36.0-44.1) Gy/ 38.6 (35.0-42.1) Gy (IMPT<X-ray (P<.0001), IMPT<VMAT (P=.0003)) in the IMPT/ VMAT/ X-ray plans. The mean NTCPs for xerostomia and hearing loss were 25.9 (95%CI: 21.5-30.3) %/ 26.0 (95%CI: 22.0-30.1) %/ 31.6 (24.9-38.4) % (IMPT<X-ray (P=.0024), VMAT<X-ray (P=.0167)), and 55.9 (95%CI: 37.8-74.0) %/ 74.8 (57.9-91.6) %/ 67.0 (49.6-84.5) % (IMPT<X-ray (P<.0001), IMPT<VMAT (P=.0275)), in the IMPT/ VMAT/ X-ray plans.

Conclusions: IMPT enabled significantly lower dose exposure to the parotid gland and inner ear than conventional X-ray therapy of the WB. Furthermore, IMPT enabled significantly lower dose exposure to the inner ear than VMAT.

P 007

Pencil beam scanning proton therapy maximizes cardiac substructure sparing in breast cancer

N. Burela1, S. Nangia1, K. Patro1, N. Mp1, D. Shamurailatpam sharma1, S. Chilukuri1, R. Jalali1

1Apollo Proton Cancer Centre, Radiation Oncology, Chennai, India

Purpose: To report dosimetric details of specific cardiac and coronary artery segments (cardiac substructures)in patients treated for breast cancer using hypofractionated intensity modulated proton therapy, with simultaneous infield boost.

Material and methods: IMPT was planned for 4 patients, with Stage I-III breast cancer, post breast conserving surgery, from June to November 2019. Two patients had right sided and 2, left sided breast cancer; 3 of the 4 planned patients underwent proton therapy. The whole breast, IMN, SCF, 5 left ventricular segments, 10 coronary artery segments and rest OARs were delineated. The axilla was delineated and treated in one patient. Plans were generated using 2 field, SingleFieldOptimization technique. The dose prescribed was 45CGE, 57CGE and 42CGE in 20 fractions to whole breast, tumor bed (SIB) and regional nodal sites. All patients underwent daily image guidance with kV imaging and CBCT. Surface guidance was used to monitor intra-fraction motion.

Results: The mean heart dose, V5 and V25 were 0.61CGE, 2.2% and 0.4%respectively. For individual left ventricular segments, the mean doses were; septal 0.09CGE, anterior 0.5CGE, lateral 0.05CGE, apex 0.52CGE and inferior 0CGE. The mean of D0.02cc for ten coronary arterial segments were LCA 0.06, LADproximal 1.25, LADmid 1.9, LADdistal 2.34, LCxAproximal 0.03, LCxAdistal 0.01, RCAproximal 3.7, RCAmid 4.5, RCAdistal 0.7 and PDA 0.02CGE. Ipsilateral lung mean dose, V5 andV20 were 6CGE, 31% and9.8%respectively. The mean contralateral breast dose achieved was 0.35CGE.

Conclusion: Modern proton therapy maximally reduces doses to coronary and cardiac substructures, thus minimizing risk of cardiac injury.

P 008

Is Proton Therapy better than Tomotherapy or VMAT intensity-modulated radiotherapy for breast cancer regional nodal irradiation? A dosimetric study

T.Y.A. Chang1, B. Yang2, A. Mui3, M. Poon3, W.W. Lam2, W.W.K. Fung3, S.K.B. Yu2, G. Chiu3, W. Wong4

1Hong Kong Sanatorium and Hospital, Comprehensive Oncology Centre, Hong kong, Hong Kong
2Hong Kong Sanatorium and Hospital, Department of Medical Physics and Research, Hong Kong, Hong Kong
3Hong Kong Sanatorium and Hospital, Department of Radiotherapy, Hong Kong, Hong Kong
4Mayo Clinic, Department of Radiation Oncology, Arizona, USA

Introduction: For patients receiving adjuvant irradiation to whole breast/chest wall (WBRT/CWRT) and regional lymph nodes (RNI) for high risk breast cancer, volumetric modulated arc therapy (VMAT) or helical tomotherapy (HT) may provide improved coverage of internal mammary chain (IMC) and sparing of normal tissues compared to 3D-conformal radiation therapy (3DCRT). Intensity modulated proton therapy (IMPT) potentially further reduce heart and lung doses. This study aims to compare dosimetry of WBRT/CWRT and RNI using 3DCRT, VMAT, HT, and IMPT plans.

Methods: Five patients were included. Clinical target volumes include chest wall, whole or reconstructed breast, supraclavicular fossa (SCF), IMC, and level III axillary chain. 5mm margin was added to form PTVs. Target volumes were treated with 40.05Gy in 15 fractions. Plans were re-optimized with partial arcs VMAT, HT and IMPT. Doses for optimal target coverage and organs at risk (OARs) were compared among different modalities.

Results: VMAT, HT and IMPT achieved at least 90% prescribed doses to PTV except 3DCRT plans (V38Gy92.34%, 90.82%, 90.54% and 81.52% respectively; p=0.02). IMPT achieved significant reduction in mean and V20Gy heart dose, and contralateral breast and lung sparing. Mean dose to heart were 4.15Gy (VMAT), 4.72Gy (HT) and 0.77Gy (IMPT; p<0.01). Compared to VMAT plans, HT delivered higher V8Gy and V16Gy to heart for left-sided breast cancers; overall OAR constraints were similar for both modalities (Table 1).

Conclusions: 3DCRT is suboptimal for regional nodal irradiation. VMAT and HT can achieve acceptable and comparable dosimetry. IMPT achieves the most significant heart, lung, and contralateral breast sparing.

P 009

Explore the feasibility of using single-iso Spot-scanning Proton Arc therapy(SPArc) for large breast and bilateral chestwall cancer treatment

X. Li1, L. Zhao1, G. Liu1, J. Dilworth1, S. Jawad1, D. Yan1, P. Chen1, C. Stevens1, P. Kabolizadeh1, X. Ding2

1Beaumont Health, Radiation Oncology, Royal Oak, USA
2Beaumont Health, Radiation Oncology Proton Therapy Center, Royal Oak, USA

Purpose: Due to the limited field size, multiple-iso plans were normally used in some large breast or bilateral chestwall cancer patients. This study explored the feasibility of using a single-iso Spot-scanning Proton Arc therapy(SPArc) technique to eliminate the iso-shifts and simplify the clinical treatment workflow.

Methods and Materials: A bilateral chestwall case and a left breast case previously treated via Intensity Modulated Proton Therapy (IMPT) were selected in this study. The IMPT plans used three-iso and two-iso for bilateral chestwall and left breast cases, respectively. Single-iso SPArc plans were generated using the same clinical plan objectives and robust optimization parameters. Machine log files were used to analyze the beam delivery time and the additional time that therapists spent in the 2nd and 3rd iso setup/shift during the IMPT. SPArc beam delivery time was calculated using same proton system with 1-second energy layer switching time.

Results: SPArc provided an equivalent or better dosimetric quality compared to the multi-iso IMPT. Both SPArc and IMPT's beam delivery time was similar (10.3 mins (IMPT) vs 10.1min (SPArc) and 24.2mins (IMPT) vs 24.1mins (SPArc) for Lt Breast and bilateral chestwall cases, respectively. The study found that an additional 13 mins and 16 mins were used in IMPT treatment for Lt breast and bilateral chestwall cancer.

Conclusion: SPArc plans could be used to treat large breast or bilateral chestwall cancer patients without additional iso shift. Such a simplified clinical workflow could not only reduce the chance of a patient's intra-fraction motion but also increase the proton therapy daily treatment throughput.

P 010

Dosimetric comparison to cardiac substructure between proton therapy and breath-hold volumetric modulated arc therapy for left-sided breast cancer

P. Loap1, F. Goudjil1, M. Ribeiro1, B. Baron1, A. Fourquet1, Y. Kirova1

1Institut Curie, Department of Radiation Oncology, Paris, France

Purpose: Deep-inspiration breath-hold (DIBH) combined with volumetric modulated arc therapy (VMAT) substantially reduces mean heart dose while treating left-sided breast cancers with photons. This study compares dosimetry to cardiac substructure between DIBH-VMAT and pencil-beam scattering (PBS) proton therapy.

Material and Methods: Seven left-sided breast patients who underwent breast conserving surgery and DIBH-VMAT irradiation were included. Patients received 51.8 Gy to the left breast, 50.4 Gy to the left axillary and internal mammary lymph nodes, and 63 Gy to the tumor bed. PBS plans were generated on simulation computed tomography scans with similar PTV dose coverage. The endpoint was the mean doses delivered to the heart, to the cardiac cavities, to the left ventricular walls and to the coronary artery segments (Figure 1).

Results: Compared with DIBH-VMAT, PBS significantly reduces mean doses delivered to all cardiac substructures and coronary artery segments (Table 1). Mean heart dose was reduced from 4.0 Gy with DIBH-VMAT to 0.3 Gy with PBS (p<0.01). PBS substantially decreases coronary artery exposure : mean dose to the left main artery was reduced from 2.9 Gy with DIBH-VMAT to < 0.01 Gy with PBS (p=0.016) and mean dose to the left anterior descending coronary artery was reduced from 9.8 Gy with DIBH-VMAT to 1.0 Gy with PBS (p=0.033).

Conclusion: Pencil beam scanning proton therapy for left-sided breast cancer significantly reduces mean doses delivered to all cardiac substructures and coronary artery segments compared with deep inspiration breath-hold volumetric modulated arc therapy.

P 011

The impact of concurrent chemoradiotherapy on survival in patients with stage IV squamous non-small-cell lung cancer

Y. Geng1,2,3,4, X. Wang1,4,5, B. Lu2,3, Q. Zhang4,5, L. Wang1,4, S. Feng1,4, H. Luo1,4,5, R. Liu1,4,5, C. Li1,4, X. Zhao1,4

1Lanzhou University, The First Clinical Medical College, Lanzhou, China
2Guizhou Provincial Cancer Hospital, Department of Thoracic Oncology, Guiyang, China
3Affiliated Hospital of Guizhou Medical University, Department of Thoracic Oncology, Guiyang, China
4Lanzhou heavy ions hospital, Department of Radiation Oncology, Lanzhou, China
5Gansu Provincial tumor hospital, Department of Radiation Oncology, Lanzhou, China

Objective: To analyse the impact of survival with three-dimensional radiotherapy for stage IV squamous non-small cell lung cancer (NSCLC).

Methods: Data for 629 eligible patients who received three-dimensional radiotherapy (X-ray) between 2002 and 2016 were retrospectively analyzed.161 of 183 cases were included pre-protocol. Patients received platinum-doublet chemotherapy with concurrent irradiation of the primary tumour. Primary endpoints were overall survival (OS) and progress-free survival (PFS).

Results: Of 161 patients, the 1-, 2-, 3- and 5-year OS rates and median survival time (MST) were 45.7%, 14.1%, 11.2%, 2.2% and 11months, respectively. Using contrastive analysis PTV dose ≥63 Gy and <63 Gy, the 1-, 2-,3- and 5-year overall survival (OS) rates and median survival time (MST)were 48.9% vs 43.3%, 21.8% vs 8.2%, 18.4% vs 4.4%, 5.1% vs 0%, and 12 months vs 11months (χ2=7.222, P=0.007). Contrastive analysis patients received radical concurrent chemoradiation therapy, and the 1-, 2-, 3- and 5-year overall survival (OS) rates and median survival time (MST) were 54.3% vs. 37.2%, 27.2% vs. 7.5%, 24.9% vs. 4.8%, 8.3% vs. 0%, and 14 months vs. 10 months (χ2=13.180, P=0.000). Multivariate analysis showed that PTV ≥63 Gy was an independent favourable factor for survival.

Conclusion: Concurrent chemotherapy and three-dimensional radiotherapy to the primary tumour in stage IV squamous NSCLC could prolong survival, and with increasing intensity of comprehensive treatment, OS gradually improved. PTV ≥63 Gy is the independent prognostic factors for OS.

P 012

Induction chemotherapy followed by neoadjuvant chemoradiotherapy for patients with locally advanced rectal cancer: a systematic review and meta-analysis

S.W. Feng1,2, Q. Zhang2,3, Z. Li2,4, C. Li1,2, Y. Geng1,2, L. Wang1,2, X. Zhao1,2, Z. Yang2,5, X. Wang1,2,3

1Lanzhou University, The First School of Clinical Medicine, Lanzhou, China
2Lanzhou Heavy Ions Hospital, Department of Radiation Oncology, Lanzhou, China
3Gansu Provincial Cancer Hospital, Department of Radiation Oncology, Lanzhou, China
4Chinese Academy of Sciences, Institute of Modern Physics, Lanzhou, China
5Lanzhou University, Basic Medical College, Lanzhou, China

Background: Additional induction chemotherapy (ICT) before neoadjuvant chemoradiation (NCRT) and surgery(S) has been proposed as a means of eradicating subclinical micrometastases and improving long-term survival, but controversy persists about whether this approach yields improved oncological outcomes.

Methods: We searched the PubMed, EMBASE, Cochrane Library, and China Biology Medicine (CBM) databases. The endpoints were long-term oncologic outcomes, and safety measures.

Results: We identified 9 relevant trials that enrolled 1538 patients. We detected no significant difference in the 5-year OS (OR 1.50; 95 % CI 0.48–4.64; P = 0.486), DFS (OR 1.03; 95 % CI 0.73–1.46; P = 0.847), LR (OR 0.80, 95 % CI 0.45–1.43; P= 0.457) and DM rates (OR 1.03; 95 % CI 0.55–1.93; P=0.922) between patients who did and did not receive ICT. Our findings suggest that between the ICT+NCRT+S and NCRT+S groups, ICT improved the incidence of grade 3 to 4 toxicity effects (OR 4.81; 95 % CI 2.38–9.37; P=0.000), but between the ICT+NCRT+S and NCRT+S+ACT groups, ICT might achieve more favorable toxicity profiles (OR 0.19; 95 % CI 0.08–0.50; P=0.01). ICT had no significant impact on surgical complications (OR 0.97; 95 % CI 0.63–1.51; P=0.906).

Conclusion: Although there were no differences in the long-term outcomes between the groups with and without ICT, our findings suggest that between the ICT+NCRT+S and NCRT+S groups, ICT improved the incidence of grade 3 to 4 toxicity effects, and between the ICT+NCRT+S and NCRT+S+ACT(adjuvant chemotherapy) groups, ICT might achieve more favorable toxicity profiles.

P 013

Rectal D8cm3 and D10cm3 as a predictor of side effects during prostate radiotherapy using proton pencil beam scanning

M. Navratil1, V. Vondracek1, M. Andrlik1, J. Kubes2, J. Rosina3, A. Grebenyuk4

1Proton Therapy Center Czech- s.r.o., Medical Physics Department, Praha 8, Czech Republic
2Proton Therapy Center Czech- s.r.o., Department of Health care - Proton Therapy, Praha 8, Czech Republic
3Faculty of Biomedical Engineering. Czech Technical University in Prague, Department of Health Care Disciplines and Population Protection, Kladno, Czech Republic
4Pavlov First Saint Petersburg State Medical University, Department of Health Protection and Disaster Medicine, Saint Petersburg, Russian Federation

Purpose: Dose in relative rectal volume and mean rectal dose are widely used as main constraints for prostate irradiation. But these criteria are very sensitive to changes in rectal filling and also to used contouring standard. Using absolute volumes avoids these issues.

Material and Methods: 300 patients with early-stage prostate cancer were treated using IMPT (intensity modulated proton therapy) - 36.25 GyE in 5 fractions between February 2013 and July 2016. During this time was same number of patients treated with advanced prostate carcinoma with IMPT using SIB (simultaneous integrated boost) - 63 GyE in 21 fractions for prostate resp. 48.3 GyE for lymph nodes. D5cm3 – D20cm3 for rectum were recorded for all these patients. Toxicity was scored according to Common Terminology Criteria for Adverse Events (CTCAE) v 4.0. These data were analysed together.

Results: The median of follow-up time in both groups was 46 months. D8cm3 and D10cm3 were discovered as statistically significant parameters related to occurrence of GIT sideffects with specific grade.

Conclusion: Rectal D8cm3 and D10cm3 are feasible parameters for GIT toxicity prediction. Search for critical values of D8cm3 and D10cm3 for different kinds of side effects is in progress. Stability of these parameters against the quality of rectal contouring and changes in rectal content during treatment course was confirmed.

P 014

Pencil beam scanning (PBS) proton radiotherapy (PRT) for esophageal cancer (EC) – pattern of treatment failure supports dose escalation

P. Vitek1, J. Kubes2, V. Vondracek3, A. Pazdro4

1Proton Therapy Center Czech, Proton radiotherapy, Praha 8, Czech Republic
2Proton Therapy Center Czecg, Radiation Oncology, Prague, Czech Republic
3Proton Therapy Center Czech, Radiation Oncology, Prague, Czech Republic
4University Hospital Motol, Clinic of Surgery, Prague, Czech Republic

EC belongs to less frequent diagnoses in Czech republic. Stages III, IV comprise more than 70% and mortality-to-incidence ratio remains high. Radiotherapy is indicated within trimodal therapy or as a definitive therapy. PRT was employed to increase treatment efficacy assuming good tolerance.

PBS-PBT, was administered in 43 EC patients, squamous-cell cancer (38) or adenocarcinoma (5), stage T2-T4, N0-2, M0. Dose 50 GyE/25 fractions in preoperative setting or 70 GyE/35 fractions for definitive radiotherapy. The treatment volume included mediastinal and supraclavicular or coeliac nodes accordingly to primary tumor location. Concomitant chemotherapy CBDCA+paclitaxel weekly or CDDP weekly was added.

Complete and partial regression was confirmed at thoracotomy or CT plus endoscopy in 16 (37,2%) and 17 (39,5%) pts. respectively, stable and progressive disease in 2 (4,7%) and 8 (18,6%) pts. Median survival time 26 months. 10 pts.(23,3%) died due to progression with a median onset time 3 months. Acute toxicity > gr.3: Mucosal in 2 pts., leukopenia in 2 pts., dermatitis in 3 pts. No effusions were observed. Weight loss >10% in 2 pts. Chronic toxicity: Radiation pneumonitis, oesophageal bleeding gr. 4, oesophago-mediastinal fistulation, each in 1 pt. Median PTV volume 1130 ccm, median of lung Dmean 5,8 Gy (0,07- 14,65), median of heart Dmean 4,9 Gy (0-16,5). No correlation between neutrophil-lymphocyte ratio and prognosis was observed.

PBS PRT was administered with favourable dosimetry and favourable tolerance.Treatment failure consists predominantly in early local progression, presumably related to more advanced stages diagnosed in Europe. Tolerance and dosimetry of PBS support dose escalation even above 70GyE.

P 015

Preliminary clinical results of carbon ion therapy for bone and soft tissue sarcoma of trunk and limbs

R.F. Liu1, Q.N. Zhang1, H.T. Luo1, Q. Li2, Y.C. Geng3, X.S. Zhao3, Z.Q. Liu1, X.H. Wang1

1Gansu provincial cancer hospital, Radiation oncology center, Lanzhou/ Gansu province, China
2Institute of Modern Physics- Chinese Academy of Sciences, Key Laboratory of Heavy Ion Radiation Biology, Lanzhou/ Gansu province, China
3Lanzhou University, The first clinical medical college, Lanzhou/ Gansu province, China

Objective: To verify the safety and effectiveness of a carbon ion therapy system made by China.

Methods: A clinical trial was conducted in Wuwei Heavy Ion Center of China from November 2018 to February 2019. A total of 47 subjects were recruited, including 5 cases of soft tissue sarcoma. All enrolled patients signed the informed consent. The treatment process including position fixation, CT simulation, target delineation, dose verification, position verification and treatment implementation. The adverse effects and short-time response rate were evaluated according to CTCAE 5.0 and RECIST1.0 respectively.

Results: Among the 5 patients, 1 case of malignant fibrohistiocytoma of upper limb, 1 case of rhabdomyosarcoma in paraspinal, 1 case of liposarcoma of left leg, 1 case of osteosarcoma of left tibia, and 1 case of postoperative residual of alveolar soft tissue sarcoma of pelvic wall, among which 3 cases were male and 2 cases were female, the average age was 46 years and ECOG score was 0-1. Follow-up was conducted every month for 3 months, including physical examination, laboratory indicators and enhanced MR. In terms of short-term efficacy, CR in 1 case, PR in 2 cases, SD in 2 cases, ORR (CR+PR) was 60%, and DCR (CR+PR +SD) was 100%. In terms of toxic and side effects, the main radiotherapy related toxicity was skin reaction, including 2 cases of grade 1 and 2 cases of grade 2, there was no other treatment-related toxicity.

Conclusions: Carbon ion radiotherapy for bone and soft tissue sarcoma resulted in a good local effect with minor toxicities.

P 016

Early experience with high-dose proton therapy for chordomas and chondrosarcomas of the spine

T. Skóra1, E. Pluta1, D. Wojton-Dziewońska1, A. Patla1, K. Urbanek1, E. Góra2, K. Kisielewicz2, T. Kajdrowicz3, D. Kabat2

1Maria Skłodowska-Curie Institute - Oncology Center, Department of Radiotherapy, Kraków, Poland
2Maria Skłodowska-Curie Institute - Oncology Center, Department of Medical Physics, Kraków, Poland
3Institute of Nuclear Physics Polish Academy of Sciencec, Bronowice Cyclotron Center, Kraków, Poland

Purpose: The aim of the study is to evaluate the short-term outcomes in terms of tumor control and toxicity of patients with spine chordoma and chondrosarcoma treated with proton radiation therapy.

Methods and Materials: From November 2016 through December 2018, 22 patients with chordoma (n=17) and chondrosarcoma (n=5) of the spine were treated with proton radiation therapy at our institution. Median patient age was 54.5 years (range, 23-82 years); 54.5% were male, and 45.5% were female. Primary tumour location: the cervical spine (n=6), the thoracic spine (n=2), the lumbar spine (n=3), the sacrum (n=11). The median radiation therapy dose for chordomas and chondrosarcomas was 74.0 GyRBE and 70.0 GyRBE, respectively. Eighty-two percent of patients (n=18) had undergone surgery before radiotherapy. In 8 cases (36.4%) surgical stabilization with titanium hardware was performed. Disease-free survival and overall survival were calculated. The early and the late treatment toxicity was evaluated according to CTCAE (version 5.0).

Results: With a median follow-up of 17.1 months (range, 4.3-34.3 months), 2-year disease-free survival and overall survival rates were 72.3% and 79.9%, respectively. Five patients (22.7%) experienced disease recurrence: local failures (n=2) and distant metastases (n=3). The mean time to disease reccurence was 7.5 months (range, 2.9-19.8 months). Acute grade ≥3 toxicities were observed in 3 patients. Late toxicities higher than G2 were not noticed so far.

Conclusion:For spinal chordomas and chondrosarcomas high dose proton therapy remains a highly effective therapeutic method with acceptable toxicity. Nevertheless, a longer follow-up period is needed to confirm these results.

P 017

Comparative analysis of scanning beam proton therapy, intensity modulated radiation therapy, and 3D conformal radiation therapy in soft tissue sarcoma

R. Thomas1, H. Chen2, Z. Fellows2, C.C. Chen2, H. Li2, C. Deville- Jr.2

1George Washington University, School of Medicine and Health Sciences, Washington- DC, USA
2Johns Hopkins University, Department of Radiation Oncology and Molecular Radiation Sciences, Washington- DC, USA

Background: Soft tissue sarcoma (STS) radiotherapy reduces recurrence after limb salvage surgery, while increasing wound healing complication and late fibrosis risks. Proton therapy (PT) spares excess dose to normal tissues, potentially reducing adverse effects.

Purpose: To evaluate dosimetric differences in modern scanning beam PT versus photon therapy for neoadjuvant STS treatment.

Methods: Existing IMRT and 3DCRT data of 15 STS patients treated preoperatively were used to create 2-3 field, single field optimization PT plans using Raystation TPSv8.A. Target volumes were robustly optimized for 100% CTV coverage with 3.5% range uncertainty and 3mm setup uncertainty. The prescribed dose was 50.4 GyRBE in 28 fractions. Constraints for organs at risk (OARs) were: joint and bone V50<50%. Mean dose to each OAR and integral dose were compared by student T-test with P<0.05 significance.

Results: A minimum 99% CTV coverage and OAR constraints were achieved for all PT plans. Mean dose to the proximal joint (477.35 ± 521.13 cGyRBE for PT and 1706.91+1311.52 cGyRBE for photon) and bone (1442.93+1068.33 cGyRBE for PT and 2588.39+1250.17 cGyRBE for photon) were significantly reduced for PT, ps<0.05. Mean dose to Body-CTV was 298+192 cGyRBE for PT and 614+229 cGyRBE for photons (p <0.001). Integral dose was significantly reduced by an average of 139% (range 28%-243%, p <0.001) (Figures 1 and 2).

Conclusions: PT maintained target coverage while significantly reducing dose to adjacent OARs, normal tissues, and integral dose compared to photons. Further clinical investigation is warranted to determine whether these dosimetric benefits translate to reduced treatment toxicities.

P 018

IMRT or proton therapy in the treatment of early stage mediastinal Hodgkin Lymphoma (HL): Dosimetry selection – an Institut Curie experience

O. Blot1, F. Goudjil2, M. Amessis1, R. Dendale2, Y. Kirova1

1Institut Curie, Department of Radiation Oncology, Paris, France
2Institut Curie Centre de Protontherapy, Department of Radiation Oncology, Orsay, France

Purpose: We present our institutional selection criteria using the modern ILROG guidelines to identify adult mediastinal HL patients who may have an advantage of proton irradiation and describe the used procedure.

Material and Methods: We studied all early mediastinal HL patients who were treated in our Department for radiation therapy (RT) after chemotherapy between 2018 and 2019. All patients received chemotherapy and radiotherapy in respect of ESMO and ILROG Guidelines. The selection of RT technique: intensity modulated radiotherapy (IMRT) or protons were selected using the ILROG Guidelines for the use of proton therapy (PT) for adults with mediastinal lymphomas.

Results: After images registration between the pretreatment PET scan and the CT scan in treatments position, delineation of the target volumes (IS) and OAR was realized on the free breathing as well as at the CT with spirometry control. After optimization of the dosimetries, the committee of 2 radiation oncologists, 1 physicist specialist in protons and 1 dosimetrist, specialized in photon treatment compared the dosimetries and selected the most appropriated patients for PT. During this period, 3 female patients, aged 23-24-29, were accepted and treated by protons because there was a significant benefice in terms of dose reduction to heart, lungs and breasts in all 3 cases.

Conclusion: Careful selection of patients- candidates for PT for HL is needed, especially in countries with limited number of proton facilities. The next step is to write the evidence-based guidelines for how to select this population of patients.

P 019

Patterns of care and outcomes of adjuvant proton therapy for oropharyngeal cancer: Analysis of the multi-institutional proton collaborative group registry

A. Koroulakis1, G. Alexander1, N. Kline2, C. Vargas3, J. Chang4, H. Tsai5, K. Hatten6, M. Witek1, J. Molitoris7, J. Snider7

1University of Maryland Medical Center, Radiation Oncology, Baltimore, USA
2University of Maryland School of Medicine, University of Maryland School of Medicine, Baltimore, USA
3Mayo Clinic, Radiation Oncology, Phoenix, USA
4Oklahoma Proton Center, Radiation Oncology, Oklahoma City, USA
5Procure Proton Therapy Center, Radiation Oncology, Somerset, USA
6University of Maryland School of Medicine, Otolaryngology, Baltimore, USA
7University of Maryland School of Medicine, Radiation Oncology, Baltimore, USA

Background: The combination of treatments for oropharyngeal cancers (OPXC) for best functional outcomes remains uncertain. Transoral robotic surgery (TORS) for OPXC is often followed by adjuvant radiation therapy (RT) +/- chemotherapy (CHT). Experiences describing adjuvant proton therapy (PT) are limited. We hypothesized that PT may be ideal adjuvant therapy for reducing long-term morbidity.

Materials and Methods: Prospectively collected data for patients undergoing surgery and adjuvant PT enrolled on PCG-REG001 were analyzed. Toxicities were scored according CTCAE 4.0. Descriptive statistics were used for patient, disease, and treatment characteristics. The Kaplan-Meier (KM) method was used for locoregional control (LRC), progression-free survival (PFS), and overall survival (OS).

Results: Between 2012-2019, 70 patients underwent adjuvant PT at 7 proton centers. Sixty were analyzable [median follow-up 13 months, median age 63 years (41-88)]. Fifty (83.3%) had AJCC 7th Stage III-IV disease, 46 (76.7%) p16+. Median PT dose was 60 Gy (30-74 Gy). Twenty-four patients (40%) received concurrent CHT. Grade 3 acute toxicities were reported in 12 (20%) [8 (66.7%) with concurrent CHT], with no higher-grade events. Gastrostomy tubes were placed in 6 (10%) during treatment. Grade 3 late toxicities were reported in 4 (6.7%). One patient suffering recurrence underwent re-irradiation with Grade 5 toxicity attributable to re-irradiation. One-year LRC, PFS, and OS were 94.3% (+/- 3.2%), 90.1% (+/- 4.2%), and 98.1 % (+/- 1.9%), respectively.

Conclusion: Adjuvant PT is safe, with low rates of acute and delayed toxicity and none greater than Grade 3 in this database study. Further follow-up and accrual will clarify toxicity outcomes.

P 020

Efficacy and safety of carbon-ion radiotherapy for the malignant melanoma in head and neck: a systematic review

C. Li1, Q. Zhang2, Z. Li3, S. Feng1, H. Luo2, R. Liu2, L. Wang1, Y. Geng1, X. Zhao1, Z. Yang4

1Lanzhou University, The First School of Clinical Medicine, Lanzhou, China
2Gansu Provincial Cancer Hospital, Radiotherapy department, Lanzhou, China
3Institute of Modern Physics- Chinese Academy of Sciences, no, Lanzhou, China
4Lanzhou University, Basic Medical College, Lanzhou, China

The purpose of this study is to systematically review the safety and efficacy of carbon-ion radiotherapy (CIRT) for malignant melanoma in head and neck. The review protocol is available in PROSPERO (CRD42019141495). Eight studies were eligible. Data showed that melanoma patients obtained better local control with low recurrence and mild toxicities by CIRT. For overall survival, uveal melanoma had a slight advantage than mucosal melanoma. Concurrent CIRT with chemotherapy exhibited superior PFS than CIRT alone. Although available data demonstrated that malignant melanoma is effectively and safely treated with CIRT, it is suitable to organically combine systemic therapy to obtain the best prognosis for melanoma patients.

P 021

A Magnetic Beam splitter for the 250 MeV Proton Beam therapy transportation line

S. Alshehri1, A. Outif2

1King Abdulaziz City for Science and Technology KACST, National Center for Radiation Detectors Technology NCRDT, Riyadh, Saudi Arabia
2Saudi Particle Therapy Center, Saudi Particle Therapy Center, Riaydh, Saudi Arabia

To utilize proton beam therapy facility efficiently with low cost the proton beam should be provided to different treatment rooms simultaneously. The current technique to split the beam is either using electrostatic beam splitter or using pulsing mode to deflect the beam at specified intervals. These techniques could cause significant beam loss. The current work presents a new method to split the proton beam to different lines simultaneously and continuously using magnetic dipole. This method based on redesign the magnetic dipole in a way to create two areas with high magnetic flux density to bend part of the energetic beam to a transportation line at a certain angle and area with low magnetic flux density to allow the rest of the proton beam to pass. The proton beam (800nC) with energy of about 250 MeV was splitted to a line at 90 degree angle branched from the main line. The numerical simulation was carried out by using COMSOL Multiphysics simulation software to validate the function of the new design of magnetic dipole. It was found that approximately 60 % of the beam current deflected to the 90 degree.

P 022

An investigation to find out an agile positioning solution for fixed beam in Protontherapy

F. Bourhaleb1, R. Prisco2, M. Cretoni1, M. Mazzone2, C. Pardi1

1I-See Computing Ltd, I-See Computing Ltd, Turin, Italy
2LinearBeam snc, particle accelerator department, Ruvo di Puglia-Bari, Italy

Background: Proton therapy has provided many evidences on quality and accuracy in planning a treatment. Pencil-beam scanning technology with protons enables clinicians to precisely manipulate and steer the beam to conform dose to the tumor. The actual main issue concerning proton therapy has been technology behind it and the huge expanses compared to conventional radiotherapy. To take full advantage of this relatively new treatment technique, we defined new configurations that would lower costs. Indeed, we do a full study on the use of fixed source instead of a gantry.

Methods: This work explore the feasibility and the advantages in the use of fixed particles beam and a positioning of two robotic arms, each with six degree of freedom respect to the traditional gantry-couch system under the TPS point of view. The first robotic arm is handling the couch while the second one is endowed with an imaging system that uses a Cone Beam CT installed on a C-arc ring. The positioning system is controlled and interfaced with a dedicated software translating and converting complex motions to equivalent once in conventional treatments.

Results/Conclusion: Full Monte Carlo investigations have been performed for all configurations. We see then advantages of combining such solutions in offering new smart way of taking full advantage of the hardware in our disposal nowadays. The full solution in well integrated within the whole treatment workflow from the accelerator control to the planning system. The study provides a cost and benefits report and additional non-standard configurations that propose planning improvements.

P 024

Exploring beamline acceptance for fast energy regulation with an upstream degrader in pencil beam scanning (PBS) proton therapy

A.C. Giovannelli1, D. Meer1, S. Safai1, S. Psoroulas1, M. Togno1, C. Bula1, D.C. Weber1, A.J. Lomax1, G. Fattori1

1Paul Scherrer Institut, Center for proton therapy, Villigen, Switzerland

At PSI, the beam energy is defined using a fast upstream degrader, together with associated tuning of the beamline to ensure that position and range are accurate at treatment isocenter. Indeed, it is the latter that currently limits energy layer changes to 100 milliseconds. Striving for efficiency, we have therefore evaluated whether energy changes within the momentum of the gantry acceptance, thereby obviating the need to re-tune the gantry beamline, can be exploited for even faster layer changes over small magnitudes (∼2mm WER).

Integral depth-dose curves have been measured in water with a range-scanner at four non-nominal momentum around 100, 150 and 200 MeV to obtain changes in ranges of ±0.5mm and ±1mm. Experimental data show negligible distortion in shape with overall gamma pass rates of 1%/1mm above 94%. The maximal variation in range were respectively 0.15mm, 0.5mm and 2mm due to beam loss in the beamline at low energies. Under the same settings, beam position error at isocenter has been verified with strip-chamber measurements to be as low as 0.0895mm (IQR:0.1135mm) for the average 2D in-plane distance. Even without ad-hoc changes of the clinical beamline, the upstream degrader could be controlled with a latency of 44.06ms (IQR:15.46ms) in repeated measurements.

Very fast energy changes can be realized within the beamline acceptance while preserving clinical level beam quality. However, these preliminary results are for a limited portion of the nominal gantry acceptance, and potentially larger changes are possible, albeit with increased potential losses in beam current.

P 025

Nozzle design for FLASH irradiation of single mice eyes

J. HEUFELDER1, J. Bundesmann2, A. Denker2, V.H. Ehrhardt3, J. Gollrad4, G. Kourkafas2, A. Weber1

1Charité - Universitätsmedizin Berlin, BerlinProtonen am HZB, Berlin, Germany
2Helmholtz-Zentrum Berlin für Materialien und Energie, Protonentherapie, Berlin, Germany
3Charité - Universitätsmedizin Berlin, Klinik für Strahlentherapie und Radioonkologie - Campus Virchow Klinikum, Berlin, Germany
4Charité - Universitätsmedizin Berlin, Klinik für Strahlentherapie und Radioonkologie - Campus Benjamin Franklin, Berlin, Germany

FLASH irradiation is a novel radiotherapy technique using high dose rates and has been shown to have a great potential to treat tumors isoeffectively to conventional radiotherapy at lower toxicity. The Charité – Universitätsmedizin Berlin is treating about 220 to 270 patients with intraocular tumors per year at the proton facility of the Helmholtz-Zentrum Berlin (HZB).

Within our already established experimental mice program we therefore designed a specific nozzle to study ocular adverse effects in animal studies under FLASH conditions. In order to meet the specific anatomical requirements we optimized our nozzle for a field size of 9 mm diameter with a penetration depth of 5 mm. The contralateral eye is used as an internal control group.

In detail, we used a small range shifter to degrade the 68 MeV proton beam to the desired water equivalent range of 5 mm. The range shifter works simultaneously as scattering device to create the necessary small field, covering the mice eye. A small rotating modulator wheel or a ridge filter design will be used to produce the spread-out Bragg peak. The goal is to apply 15 Gy within 200 ms resulting in a dose rate of 75 Gy/s.

First experiments showed that dose rates up to 200 Gy/s are feasible.

P 026

Next-generation dose delivery system for Particle Therapy

D. Perusko1, M. Schokker2, J. Bobnar1

1Cosylab, Products, Ljubljana, Slovenia
2MedAustron, Medical Frontend, Wiener Neustadt, Austria

The final piece of equipment of a PT system before the patient is the scanning or the dose delivery system (DDS). There are high performance requirements on the DDS, while it must at the same time adhere to all the relevant standards to ensure patient safety.

Cosylab and MedAustron have been involved in the co-development of a novel DDS. It is being developed according to all the relevant medical safety standards, such as the IEC 60601-2-64. In addition, the DDS has strict performance requirements, such as delivering high beam intensities and fast layer switching, to increase patient throughput. Multiple operation modes are foreseen, such as a clinical mode, with strict safety constraints; a research mode, with dedicated unique non-clinical research functionality; and a QA mode, where calibration of the system, based on external and internal measurements, can be performed. The DDS is being designed with future functionalities in mind, such as FLASH, adaptive proton therapy, tracking, proton imaging, range verification and arc therapy. The modular design of the DDS, as presented in the figure, allows to adapt it to various PT facilities, regardless of accelerator type or particle species used. This design strategy enables the DDS to incoporate future updates in an easy manner thus permitting introduction of revolutionary techniques in beam generation, while minimizing efforts for recertification.

A next-generation DDS is being designed by Cosylab and MedAustron with strict performance and safety specifications in order to enable fast integration, different operating modes and upgrade capabilities.

P 027

New beam extraction mode on Protom synchrotrons for proton tomography

A. Pryanichnikov1,2,3, P. Zhogolev1,3, A. Shemyakov1,3, M. Belikhin2,3, V. Rykalin4

1Protom Ltd., Research and Development, Protvino, Russian Federation
2Lomonosov Moscow State University, Accelerator Physics and Radiation Medicine, Moscow, Russian Federation
3Lebedev Physical Institute RAS, Physical-Technical Center, Protvino, Russian Federation
4ProtonVDA, Research and Development, Chicago, USA

Protom synchrotron was specially designed to accelerate protons to 330 MeV which energy is enough to have proton visualization of any part of human body without any restrictions . Moreover, the use of a synchrotron significantly simplifies the design of the proton tomography scanner, and makes it cheaper. Also, one of the main advantages of proton tomography system is a lower, compared with similar X-ray imaging systems, equivalent dose that is received by the patient. However, proton tomography systems can't work with beam intensity that uses in standard proton therapy. The purpose of this work is to show the possibility of using the proton tomography system as part of the proton therapy complexes that are based on Protom synchrotron. A special tomographic accelerator operation mode was developed and tested. It will increase the percentage of useful proton events registration that are used for further proton track reconstruction. This mode will increase the efficiency of the tomography setup and reducing the dose. The report presents the results of the first testing of the tomographic beam extraction mode at the Protom synchrotron and new concept how to increase number of useful for proton imagine events. This work shows the possibility of the PTC «Prometheus» (as well as other facilities based on the Protom synchrotron) to operate in a special proton beam extraction mode, in which single protons are released for each revolution or even less than one revolution, that allows such facilities to work effectively in tomographic mode.

P 028

Modifying the ProBeam pencil-beam-scanning nozzle to treat ocular melanoma

R. Slopsema1, A. Dhabaan1

1Emory University, Radiation Oncology, Atlanta, USA

Purpose: To evaluate the impact of design parameters - the location and thickness of beam-modifying devices - on both the dosimetric properties of the delivered dose (penumbrae, irradiation time) as well as on the quality of ocular-melanoma treatment plans.

Method: The Ocular Option (OO) is an add-on for the ProBeam pencil-beam-scanning nozzle. It consists of a range shifter and a field-specific collimator, and uses standard spot-scanning dose delivery. The range shifter position and thickness, aperture air gap, and spot-scanning map were varied. Per configuration the lateral penumbra and distal fall-off were calculated using the RayStation Monte Carlo algorithm. For a subset of configurations the calculation was verified against measurement. Treatment plans were developed for 10 ocular melanoma cases using the various OO designs as well as a dedicated eyeline. Dose to critical organs was assessed to determine the difference in treatment quality between the different designs.

Results: The lateral penumbra was improved by increasing the isocenter-to-range-shifter distance, decreasing the aperture air gap, and decreasing the spot-map size (Fig 1). A design in which a 3.5 cm Lucite range shifter is placed just downstream of the nozzle's ionization chambers, the aperture air gap is 7 cm, and the spot map size is 3x3 cm2, the lateral penumbra in air is found to be 2.4 mm.

Conclusion: With limited modifications to the existing ProBeam system, an Ocular Option can be installed whose dosimetric properties are not as good as a dedicated eyeline but that can deliver eye treatments of acceptable quality. (Fig 2)

P 030

Innovative silicon detectors for qualification and monitoring of clinical proton beams

A. Vignati1,2, O. Hammad Alì1,2,3, F. Fausti2, S. Giordanengo2, O.A. Martì Villarreal1,2, F. Mas Milian1,2,4, V. Monaco1,2, R. Sacchi1,2, Z. Shakarami1,2, R. Cirio1,2

1Università degli Studi di Torino, Physics Department, Turin, Italy
2INFN, National Institute for Nuclear Physics, Turin, Italy
3FBK, Fondazione Bruno Kessler, Trento, Italy
4Universidade Estadual de Santa Cruz, Ciencias Exactas e Tecnologicas, Ilheus, Brazil

The University of Torino and the National Institute for Nuclear Physics are investigating Ultra Fast Silicon Detectors (UFSD) for proton beam monitoring. A single particle counter for up to 100 MHz/cm2 beam fluence and a beam energy measurement device, using time-of-flight (TOF) techniques, are being developed. Dedicated strip sensors and a 24 channels front-end ASIC prototype (discrimination capability up to 100 MHz/channel instantaneous rate, with 4-150 fC expected charge range) were produced. Tests were performed at CNAO (Pavia, Italy) and at the Trento Proton Therapy Center (Italy), with 60-250 MeV clinical proton beams (106–1010 p/s fluxes). Varying the flux at different energies, the particle rate was measured and compared with the one estimated by a pin-hole ionization chamber, while a telescope of two UFSDs was built for TOF measurements. Fig. 1 shows the achieved efficiency of the counter prototype (>98% up to 108 p/s cm2,i.e. a rate of 2 MHz per 2 mm2 channel), while Fig. 2 shows the coincident signals in the two telescope sensors (CNAO, 97 cm distance between sensors, 228 MeV beam energy). Few hundreds of keV deviations from nominal energies were achieved for all beam energies at 67 and 97 cm distances between the sensors, corresponding to < 1 mm range, as clinically required. These results demonstrate that UFSD could improve the sensitivity and the response time of conventional monitors, allowing as well the online beam energy measurement.

P 031

Commissioning strategy and results of proton pulsed pencil-beam-scanning synchrocyclotron based compact system

M. Artz1, W. Hsi1, Z. Su1, M.W. Ho1, J. Park1, Y. Zhang1, X. Liang1, S. Huh1, M.M. Hunter1, L. Zuofeng1

1UFHPTI, Radiation Oncology, Jacksonville, USA

Results and strategy are presented for commissioning an IBA Proteus ® ONE (P1) proton pulsed pencil-beam-scanning (PBS) compact system; including a S2C2 super-conducting synchrocyclotron, a 220° partial-rotation gantry, with cone-beam CT and stereoscopic imaging capabilities, and a 6D robotic couch.

The commissioning of pulsed PBS proton system was expedited by using the data acquired during the validation and verification (VnV) procedure performed by the vendor during the tuning of beam optics.

Clinical Acceptance-Test (AT) only started once the data acquired during VnV was properly reviewed by both the vendor and medical-physics teams. During the VnV data review, the stability and reliability of the proton system was established. The IBA P1 system utilizes partial layer repainting, averaging three proton “bursts” per spot with dose reproducibility measured to be within 2% and energy layer switching below 0.5 seconds.

During acceptance testing, partial data was acquired for use in the treatment planning system (TPS), such as spot size variation with beam energy, and pre-set clinical protocols for cone-beam CT.

The commissioning of the P1 system, including the acquisition of beam data for RayStation TPS, clinical protocols for IGRT, and the training program for both therapist and physicists to perform full clinical treatment workflows for the disease sites of head/neck, thorax, breast, and pelvis, was completed in 8 weeks. Three patients began treatment on 12/9/19, disease sites included, head and neck, chordoma and prostate.

P 032

A single beam data set for a multi-room synchrotron-based proton therapy system: Beam matching during commissioning of a new facility

C.C. Chen1, E. Gelover1, Y. Kushida2, H. Satake2, D. Han3, T. Hrinivich13, K. Sheikh3, H. Chen1, J. Wong3, L. Heng1

1Sibley Memorial Hospital, The Johns Hopkins Proton Therapy Center, Washington- DC, USA
2Hitachi America- Ltd., Particle Therapy, Tarrytown- NY, USA
3The Johns Hopkins University School of Medicine, Radiation Oncology and Molecular Radiation Sciences, Baltimore- MD, USA

Purpose: A 3-room proton therapy center equipped with the first Hitachi PROBEAT-CR system in the U.S. was commissioned with matched beam characteristics provided by the manufacturer.

Methods: A single beam data set was generated in the RayStation TPS with integral depth-dose curves (IDDCs), spot profiles, and energy-dependent absolute output factors obtained from the first treatment room (G3). The second room (G2) was delivered by the manufacturer with matched beam characteristics, i.e. R80 offsets <0.2 mm, spot size differences <10%, and 1 cGy/MU for 1-liter homogeneous dose at 15 cm water depth. The TPS data was then validated in G2 with measurements of IDDCs, spot profiles, 3-D dose distributions of various target shapes and volumes, as well as daily, monthly and patient-specific QA (PSQA) procedures.

Results: The R80s measured with a 0.2 mm high resolution MLIC in G2 found to be <0.05 mm different compared to G3 measurements and <1.0 mm compared to TPS data. The difference in spot sizes was <10% in comparison to G3 measurements and TPS calculations. The absolute depth-dose measurements showed <1% discrepancies at treatment depths. 2D-Gamma passing rates >95% (3%/3mm) were found in planar dose measurements.

Conclusions: G3 was commissioned in 16 weeks including beam data collection, modeling, and validation. The two consecutive rooms delivered with matched beam characteristics could be commissioned in 8 weeks only. Patient treatments and PSQA could be transferred and performed in all 3 rooms interchangeably. Our data could be used as a reference for other PROBEAT-CR centers around the world.

P 033

Commissioning a Monte Carlo model for the mevion S250I hyperscan pencil beam scanning proton nozzle

B.H. Chiang1, S. Ahmad1, H. Jin1, Y. Chen1

1University of Oklahoma Health Sciences Center, Radiation Oncology, Oklahoma City, USA

The main objective of the present study has been to provide an overview of the design and commissioning process for the first Monte Carlo (MC) model of MEVION S250i Hyperscan Pencil beam scanning (PBS) proton system. S250i is equipped with a 190-degree rotating, a gantry-mounted synchrocyclotron and a compact delivery nozzle including range shifter plates for energy modulation and Adaptive Aperture (AA) for spot sharpening. It is capable of delivering fast volumetric PBS proton treatment with a 6D robotic couch system. In this study, the treatment nozzle of Mevion S250i system was modeled in detail using TOPAS (TOolkit for PArticle Simulation) MC code and was validated with commissioning data and patient treatment plans. All nozzle elements were modeled with actual dimensions and materials following the detailed information provided by the vendor. The unique designs of Energy Modulation System (EMS) and AA were modeled using an in-house geometry component class while other components were modeled using TOPAS build-in geometry. Scanning beam characteristics including integral depth doses (IDDs) of pristine Bragg peaks, in-air beam spot profiles and sizes were simulated and compared with the corresponding commissioning data. In-water penumbrae and FWHMs with and without AA were also simulated and compared with measurement. Complete end to end simulations with actual delivery was compared with treatment plan and measurement. With the treatment head of Mevion S250i PBS proton system was modeled and validated, it can be served as a viable tool for complicated research and patient treatment verification in the future.

P 035

Contributions to commissioning and operational radiation protection in proton centers with new delivery techniques (PMAT)\r

G. Garcia1, E. Gallego1, J.M. Gómez-Ros2, A. Carabe-Fernández3, H.R. Vega-Carrillo4, R. García-Baonza1, A. Bertolet-Reina3, L.E. Cevallos-Robalino5, K.A. García-Guzmán4

1Polytechnic University of Madrid UPM, Industrial Engineers ETSII - Department of Energy Engineering, Madrid, Spain
2Ciemat, Radiation Dosimetry Unit, Madrid, Spain
3Hospital of the University of Pennsylvania, Department of Radiation Oncology, Philadelphia, USA
4Autonomus University of Zacatecas, Academic Unit of Nuclear Studies, Zacatecas, Mexico
5Polytechnic University Salesian, Radiation Protection and Nuclear Systems, Guayaquil, Ecuador

Proton therapy is in continuous ever-evolving and improvement to obtain more precise and beneficial treatments for patients. Some prominent current trends involve cutting-edge delivery techniques or building compact proton centers incorporating the most advanced technologies, reducing their size while achieving more affordable facilities. This work is framed into Contributions to operational radiation protection and neutron dosimetry in compact proton therapy centers (CPTC) project, which is focused on assessing the impact of these innovations on the operational radiation protection and commissioning of the facilities.

Thus, several tasks related to such project have been carried out over the last three years in fields as: studying wide range Rem-meters to characterize neutron fields in CPTC, checking shielding and comparing ambient dose yielded by neutrons in CPTC, or analyzing activation in shielding of CPTC with different types of concrete, among others. Proton monoenergetic arc therapy (PMAT) is a new delivery modality, currently in the research and development stages, which aims to take advantage of irradiation of the tumour volume under fields with a full 360° angle, using monoenergetic protons and optimizing the LET inside the target.

The aim of this work was to carry out experimental measurements, using Prescilla detector, of neutronic fields yielded with proton monoenergetic arc therapy (PMAT) comparing by those generated with the conventional intensity-modulated proton therapy (IMPT) treatment, at different distances and angles of the circular phantom used in the tests. The measurements have been carried out at the Roberts Proton Therapy Center (RPTC) of the Hospital of the University of Pennsylvania (UPenn).

P 037

Current status of east Japan heavy ion center, faculty of medicine, Yamagata University

T. Iwai1, H. Souda1, T. Kanai1, Y. Miyasaka1, S. Kawashiro2, K. Nemoto2, H. Yamashita3, T. Kayama4

1Yamagata University, Graduate School of Medical Science, Yamagata, Japan
2Yamagata University, Department of Radiation Oncology- Faculty of Medicine, Yamagata, Japan
3Yamagata University, Department of Ophthalmology and Visual Science- Faculty of Medicine, Yamagata, Japan
4Yamagata University, Faculty of Medicine, Yamagata, Japan

Carbon-ion therapy project of Yamagata University called “East Japan Heavy Ion Center” is now within a year until first treatment. The cubic-designed building construction (Fig. 1) was completed in May 2019. An injector and a main accelerator (synchrotron) are located on the basement floor in the building. 12C beam extracted from the synchrotron is bent upward and guided to the 2nd floor where two treatment rooms are located. The treatment rooms are one fixed beam room (horizontal) and one gantry beam room. This vertical configuration of the treatment system reduces the construction area to ∼ 2,000 m2, almost 2/3 of the current minimum carbon ion therapy facility. Beam delivery is active scanning only and energy scanning method is applied without any physical range shifter. A rotating gantry with superconducting magnets is even smaller than NIRS gantry thanks to the new short-length scanning delivery system developed by NIRS and TOSHIBA Energy Systems. System installation works has almost been completed. Beam test, radiation inspection, acceptance test and clinical commissioning will follow until winter 2021. Fixed beam room will open in early 2021 by horizontal beam only for prostate cancer patients. Rotating gantry will begin treatment in spring 2021 for all other cancer types. Updated information on the project overview, schedule and current status will be presented.

P 038

Beam matching for the latest commissioned ProBeam multiple-room facility at the New York Proton Center

M. Kang1, S. Huang2, Q. Chen1, H. Zhai1, W. Xiong1, F. Yu1, J.I. Choi1, C. B. Simone- II1, H. Lin1

1New York Proton Center, Radiation Oncology, New York, USA
2Memorial Sloan Kettering Cancer Center, Radiation Oncology, New York, USA

Background: Beam matching allows for optimal clinical efficiency and is becoming more standard in multiple-room proton centers. This study investigated the beam characteristics of the latest commissioned 4-room ProBeam system at the New York Proton Center. A single double-Gaussian model was implemented in Eclipse TPS for the match treatment rooms.

Method: The integrated depth dose curves (IDDs), the in-air spot profile and machine output were fine tuned to match the previous two commissioned rooms. Only the data acquired for the first two rooms were utilized to feed into a double-Gaussian beam model using an efficient commissioning procedure. Dosimetric verifications and validations were carried out in water and non-water phantom to assess the quality of the beam modeling and room matching.

Results: The range differences between rooms were < 0.5 mm. There was no observable difference on IDD curves between all treatment rooms. The difference on outputs were within 0.5% for all the compared single-energy beams. The differences on in-air spot sigma were within 10% without RS. The spot sigma differences for 2, 3 and 5 cm RSs were less than 6%, 5% and 4%, respectively. The dose differences between TPS calculation and measurement were within 3% for all the SOBP beams for all the rooms. The validations and verifications showed all rooms were < 2% for dose measurements with and without RS, and all passed IROC TLD within 98-99% agreement.

Conclusion: All four treatment rooms are well matched. A single double-Gaussian TPS model successfully satisfied the clinical application.

P 039

An efficient commissioning workflow to commission a multiple-room ProBeam system using proton PCS algorithms

M. Kang1, H. Sheng2, Q. Chen1, F. Yu1, H. Zhai1, W. Xiong1, H. Lei1, J.I. Choi1, C.B. Simone- II1, L. Haibo1

1New York Proton Center, Radiation Oncology, New York, USA
2Memorial Sloan Kettering Cancer Center, Radiation Oncology, New York, USA

Purpose: This study aimed to propose a standardized and efficient commissioning procedure to commission a multiple-room proton system using Pencil-Beam-Convolution-Superposition(PCS) TPS.

Method: A novel standardized commissioning procedure has been developed to commission the Eclipse PCS v15.6(Fig1). The IDDs were measured using a 12cm large-diameter ionization chamber in a 3D-water-tank. A 2D-array scintillation detector Lynx was used to acquire the in-air spot profile at 5-location along the central beam axis. An IBA PPC40 parallel-plate chamber was used to calibrate the output and to scale the relative IDDs curve to Gy*mm2/MU. The Double-Gaussian(DG) beam model was used to compensate for the deficiency calculation phantom scattering of PCS, and the primary-Gaussian sigmas were directly derived from in-air spot profile, and the 2nd Gaussian sigmas and weighting factor were optimized based on TOPAS Monte Carlo simulation. A group of SOBP plans with various field sizes from 4×4 to 20×20 cm2, ranges and modulations were generated to perform dosimetric verifications to assess the quality of the beam modeling.

Results: All agreements of the range between TPS and measurement were better than 1mm. Among all the 200+ SOPB plans generated, 98% of plans have a dose difference smaller 2%, and the remaining 2% of plans have a dose difference of 2% to 3%(Fig2).

Conclusion: The proposed novel standardized commissioning procedure can expedite the commissioning process to be as short as 4 weeks per room. The DS model can achieve dose accuracy within 2% and enable new centers to pass QA measurement without re-scaling any MUs.

P 040

Estimation of competing beam requests for a multi-room particle treatment facility from conventional linear accelerator utilisation

T. Kron1, C. Fox1, N. Anderson2, A. Yeo1, G. Hanna3

1Peter MacCallum Cancer Centre, Physical Sciences, Melbourne, Australia
2Peter MacCallum Cancer Centre, Radiation Therapy, Melbourne, Australia
3Peter MacCallum Cancer Centre, Radiation Oncology, Melbourne, Australia

In many facilities one particle accelerator supplies several treatment rooms. We seek to estimate the incidence of conflicting beam time between different rooms based on utilisation of four identical medical linear accelerators (linacs) at our institution.

Data pertaining clinical use from four Varian Truebeam linacs were collected over one week in October 2019. Approximately 30 patients were treated per linac daily. The beam start time for photon beams (3DCRT, 51%; IMRT, 16%; VMAT, 33%) was recorded and an estimate of the beam-on time obtained from the total number of monitor units (MUs) delivered. For 3DCRT, each field was recorded separately. Auto-field sequencing made this impossible for IMRT. Beam demand overlap was determined from the initiation of beam and the beam-on time at three nominal dose rates (200, 500 and 1000MU/min) using an Excel spreadsheet. A variable ‘switching time' was introduced to mimic the operation of a single accelerator multiroom particle treatment facility.

Based on our utilisation of four linacs, approximately 15% of fractions would have had some beam on-time overlap assuming medium dose rate.

This doubles if the dose rate was reduced. Utilising three or two linacs reduced the overlap to 9 or 4.5%, respectively. The introduction of even a small ‘switching time' of 30s significantly increased the risk of overlap except for the lowest dose rate.

A simple spreadsheet model allows estimation of beam demand overlap in a multiroom treatment facility based on data reflecting local practice.

P 041

Effective validation of proton PBS commissioning: cross-center dosimetric comparison

C.S. Lin1, M. Kang2, Y. Wu3, S. Huang4, J. Zhou5, S. Tang6, Q. Chen2, F. Yu2, C.B. Simone2, H. Lin2

1National Taiwan University Cancer Center, Radiation Science and Proton Therapy, Taipei, Taiwan Province of China
2New York Proton Therapy Center, Proton Therapy, New York, USA
3National Taiwan University Hospital, Radiation Oncology, Taipei, Taiwan Province of China
4Memorial Sloan Kettering Cancer Center, Medical Physics, New York, USA
5Emory Proton Therapy Center, Radiation Oncology, Atlanta, USA
6Texas Center for Proton Therapy, Proton Therapy, Dallas, USA

Background: High precision on machine calibration and high consistency between the treatment planning system(TPS) and machine performance are the key factors for commissioning a new proton system. This study proposes a novel method to independently validate the machine calibration and TPS configuration through cross-center comparison.

Materials and Methods: Cross-validation between established-center A and new-center B is carried out through two steps. Machine calibration: Two sets of TLDs for patient dose measurement were ordered from the Accredited Dosimetry Calibration Laboratory (UWADCL) and irradiated separately at A and B. The TLD-reading from B is based on reading of A as A is submitted as the calibration. TPS validation: It is well known that Varian ProBeam system is globally matched on beam data. Plans of different treatment sites from A are forward calculated at B using a different TPS system. Doses are compared to cross-check the TPS configurations.

Results: For machine calibration, all the TLDs dose differences were <2% between two centers. For TPS validation, the dose differences were within 3% for all the cases of prostate, head-and-neck, and lung at the middle of SOBP(mid-SOBP), and the gamma passing-rate (3%/3mm) for all cases at the mid-SOBP was >95%.

Conclusions: A novel cross-center evaluation of beam data, configuration, and PCS algorithms is feasible. High agreement of absolute dose and gamma-comparison between the two centers was obtained. This workflow can improve the quality of treatment and patient safety, and it can be used as an alternative for new proton centers, especially in the area that no external audit is available.

P 042

First year's experience operating a network of single-room proton therapy centres across the UK

J. Pettingell1, J. Lambert2

1Rutherford Cancer Centres, Physics and Radiotherapy, London, United Kingdom
2Rutherford Cancer Centres, Physics and Radiotherapy, Newport, United Kingdom

Our company has built, commissioned and now operates a network of proton therapy centres across the UK. The first three centres started proton treatments in April 2018, June 2019, August 2019, and a fourth centre in under construction. Each centre is equipped with IBA Proteus ONE single-room PBT, as well as linac, CT and MR scanners. All sites are on the same IT network with centralised OMS server (Elekta Mosaiq) and centralised TPS (Pinnacle and RayStation).

Previously we presented commissioning results (J.Lambert et al PTCOG58) demonstrating the beam data from our PBT machines match well enough that we could transfer patients between sites for a proportion of treatment fractions without the need to re-plan, and without data transfer since all sites use the same Mosaiq server.

Now we'll present our first year's experience of running a network of three proton therapy centres. So far we transferred patients on four occasions and will discuss: 1) Why and when we decide to transfer patients instead of linac backup plan or treating at weekends etc.; 2) Reasons patients do or don't want to transfer; 3) Practicalities and logistics of transferring patients between sites, moving staff and immobilisation equipment etc.; 4) Patient QA results for the same plans delivered on different sites; and 5) Downtime on each machine and the network as a whole (three cyclotrons and three gantries). For example clinical up-time for the network of centres for first three months was 99.5%, because only 0.5% of the time were all three sites down at the same time.

P 043

Proton center staffing: A multicenter survey in the United States

L. Zhao1

1St. Jude Children's Research Hospital, Radiation Oncology, Memphis, USA

Introduction: The number of proton centers in the US continues to increase. Appropriate staffing at these centers is critical for patient safety and quality care. In this study, we review staffing information from operational proton therapy centers in the US and propose a modified ASTRO worksheet for calculating physicist staffing.

Methods: Seventeen operational proton therapy centers in the US with multiple treatment rooms were emailed during October and November 2018. Two separate questionnaires were created for centers offering proton therapy only (Group A survey: 14 questions) and centers providing both proton therapy and photon therapy (Group B survey: 22 questions). The survey included questions on the FTE for therapists, dosimetrists, physicists, radiation oncologists; number of patients undergoing procedures; and number of treatment machines and imaging equipment.

Results: Fourteen centers, with 7 providing proton therapy only, completed the questionnaire (response rate 82.3%). Average daily patient per staffing ratios for therapists, dosimetrists, physicists, and radiation oncologists were 4.2, 8.8, 11.9, and 20.6 for Group A and 4.5, 12.4, 12.0, and 12.8 for Group B. When relative FTE for physicists was increased from 0.25 to 0.35 per proton treatment machine, as a modification to the ASTRO worksheet, predicted FTE for physicists compared well with survey data.

Conclusions: There is noticeable difference in the patient per staffing ratio between two groups for dosimetrists (p=0.05) and radiation oncologists (p =0.09) however negligible difference was found for therapists and physicists. A small adjustment to the physicist FTE factor in the ASTRO staffing model improves predictions.

P 044

A multi-institute dosimetry audit of small animal irradiation units using an anthropomorphic murine phantom

E. Biglin1, A. Aitkenhead2, G. Price3, K. Williams4, A. Chadwick1, K. Kirkby1

1University of Manchester, Division of Molecular Clinical Cancer Studies- Faculty of Biology- Medicine and Health, Manchester, United Kingdom
2The Christie NHS Foundation Trust, Christie Medical Physics and Engineering, Manchester, United Kingdom
3University of Manchester, Manchester Academic Health Science Centre, Manchester, United Kingdom
4University of Manchester, Division of Pharmacy and Optometry, Manchester, United Kingdom

Preclinical radiotherapy studies using small animals are an indispensable step in the pathway from in vitro experiments to clinical implementation. As radiotherapy techniques advance in the clinic it is important preclinical models evolve to keep in line with these developments. So far this includes the use of orthotopic tumour sites, small animal image-guided radiotherapy platforms that mimic clinical treatment delivery and the development of tissue equivalent phantoms. One significant issue with preclinical radiation research is the lack of traceable standards to a primary calibration source, affecting accuracy and reproducibility.

Utilising 3D printing our group have produced a phantom of varying density, based on a CT scan, to create an anthropomorphic phantom reflecting the size, physiological features and soft tissue and bone densities of a real mouse. 3D printing also allows the unique incorporation of various detectors in specifically designed orientations.

Our existing work provides the basis for a dosimetry audit with protons and heavier ions. Through the PTCOG radiobiology sub-committee a questionnaire was sent to multiple institutions which provided detailed information regarding the techniques implemented in preclinical studies. Murine phantoms holding Gafchromic EBT3 film and Alanine pellets within them will then be distributed across centres enabling performance of a dosimetry audit across multiple institutions.

P 045

Reference dose measurement of proton beams with solid water phantom

M.L. Kang1, L. Hu1, Y.T. Wu2, W.J. Xiong1, F. Yu1, Q. Chen1, C.S. Lin3, C. Ackerman1, J.Q. Zhang1, H.B. Lin1

1New York Proton Center, New York Proton Center, New York, USA
2National Taiwan University Hospital, Radiation Oncology, Taipei, Taiwan Province of China
3National Taiwan University Cancer Center, Radiation Oncology, Taipei, Taiwan Province of China

Background: Reference dosimetry of proton beams is typically measured in water for its ubiquity and uniformity. The depth of measurement for single energy beams is usually placed at 2 cm, where the depth dose curve is relatively flat and the measurement is less subject to uncertainty in low energy ranges. However, this optimal configuration might not be achievable in practical situations such as for fixed horizontal beams. Here we propose a workflow of reference dose measurements using Solid Water (SW) Phantom.

Materials and Methods: The reference doses of 18 different energy beams for a fixed beam room (FBR) in a cyclotron proton center was measured with a PPC05 chamber (IBA) and SWs. The chamber was placed at 2cm SW in SAD setup with sufficient backscatter. The water equivalent thicknesses (WET) of the chamber and SWs were measured with Zebra (IBA). The doses of FBR determined with SW were compared to that of gantry rooms measured in water and SW, respectively. A reproducibility study was then conducted with three representative energies (70, 150, and 240 MeV).

Results: The WETs of the PPC05 and SPs were 1.6mm and 20.4mm, respectively. The absolute doses measured in the FBR were all within 1.0% difference compared to the gantry rooms (Fig.1). Daily dose variation of the FBR was within 0.6% (Fig.2).

Conclusion: Our result suggests that SW can be used as an alternative to water for reference dose measurements so as to provide flexibility in special clinical situations. Similar suggestion was also provided in TRS-398.

P 047

Dosimetric measurements of synchrotron-based ultra-high dose rate proton beam using a float glass

T. Toshito1, C. Omachi1, Y. Kibe1, M. Kimura1, H. Iwata2, H. Ogino2, Y. Shinozawa3, T. Katayose3, M. Umezawa3

1Nagoya Proton Therapy Center, Department of Proton Therapy Physics, Nagoya, Japan
2Nagoya Proton Therapy Center, Department of Proton Therapy, Nagoya, Japan
3Hitachi- Ltd., Smart Therapy Division, Kashiwa, Japan

We have developed a dose monitoring system for a biological experiment using ultra-high dose rate proton beam so called FLASH. A system consists of a 0.5-mm thick float glass plate, surface mirror, and a cooled CCD camera. The scintillation image on the glass plate is reflected by the surface mirror and recorded by the CCD camera. The system was designed to measure dose distribution at the bottom of a petri dish put on the float glass across a 2-mm thick opaque acrylic plate and assembled in a light shielded black box. The glass plate was reported to have enough light yield, undetectably small quenching effect, i.e. LET dependence, and expected to have negligible dose rate effect. We have carried out beam tests at a treatment room equipped with spot-scanning irradiation nozzle in Nagoya Proton therapy Center. A coefficient to convert CCD signal to absolute dose was obtained by calibration against an Advanced Markus chamber calibrated under the JSMP 2012 protocol with conventional dose rate. Measured depth-dose profile around Bragg-peak agreed well with PDD measured by the Advanced Markus chamber.

P 051

Benefits of functional programming in treatment planning systems

N. Depauw1, T. Madden1, H. Kooy1

1Massachusetts General Hospital, Radiation Oncology, Boston, USA

Functional programming is a paradigm where deterministic functions produce and consume immutable data. Immutable data is deeply embedded in radiotherapy through predictable DICOM data representations and transfers. When applied to a treatment planning system (TPS/FP), the paradigm has a profound impact on plan data handling and on system testing and validation. In a TPS/FP, a treatment plan is a structured set of unique immutable data and function IDs. The set forms a computational graph that records the actions that produce planning results. The graph is a lightweight representation of the whole or parts of the treatment plan. We present our experience with the Astroid™ TPS/FP which uses the FP paradigm. Figure 1 shows a simple plan based on its immutable functions and data. One benefit applies to clinical validation of TPS releases. New releases require validation against the release statement as well as previous behavior. This requires the recomputation of plans and comparison to previous plans or measurements. With a conventional TPS, the user cannot “measure” the internal changes of a system and must per-force recompute all. With a TPS/FP, however, each plan has a unique signature graph that records all functions and data. Therefore, only those plans where a change in a function or data input occurs are recomputed and only the results of those recomputed plans need comparison. This greatly reduces effort and allows a rigorous recording of changes and effects. Other benefits include version management and automation of plan changes such as occur in adaptive radiotherapy.

P 052

Dictionary-based protoacoustic imaging for range verification

C. Freijo1, J.L. Herraiz1,2, D. Sanchez-Parcerisa1,2,3, J.M. Udias2

1Universidad Complutense de Madrid, Nuclear Physics Group- EMFTEL and IPARCOS, Madrid, Spain
2Hospital Clínico San Carlos, Instituto de Investigación Sanitaria IdISSC, Madrid, Spain
3Sedecal, Molecular Imaging, Algete- Madrid, Spain

The detection of acoustic pressure waves generated by a proton beam can be used to reconstruct the dose distribution. Nevertheless, the low signal-to-noise ratio of these measurements and the finite number of detector locations limit the accuracy of the achievable dose-maps using standard methods, i.e. the ones used in photoacoustic tomography.

We propose to use the a priori knowledge of the shape of the proton dose distribution to create a dictionary of deposited dose for variations from the canonical case (X-Y displacements, range/density variations), together with their expected ultrasonic signals at the detector locations. This library is then used to detect errors in treatment delivery by choosing the combination of dictionary dose distribution that better agrees with the detected signals.

Dose and generated waves were simulated with TOPAS and k-Wave using information of the patient-CT. We evaluated the capability of this method to determine the location of the proton beam in several anatomical locations, such as liver and prostate. In all cases, our method was able to locate the beam maximum within <1mm in depth and transverse directions. This approach could also be used for online verification, as detected and dictionary signals can be compared in real-time.

P 053

Log files-based fraction dose recalculation in a treatment planning system

T. Henry1, E. Almhagen1,2, C. Vallhagen Dahlgren1

1Skandionkliniken, Skandionkliniken, Uppsala, Sweden
2Uppsala University, Department of medical radiation sciences, Uppsala, Sweden

Purpose: To recalculate in a treatment planning system the 3D dose distribution delivered on a daily fraction based on log files recorded during irradiation.

Methods: Varian Eclipse v15.6 was used. A proton treatment plan was created, and a fraction was subsequently delivered at the Skandion Clinic in Uppsala, Sweden. During irradiation, the delivery was interrupted and resumed at a different point, resulting in an energy layer in one of the treatment fields being excluded from the delivery. This created a discrepancy between the original planned dose and the delivered dose. Log files from this irradiation containing delivered “real” spot weights and positions were exported and a new proton plan, containing the delivered spot maps, was created and imported in Eclipse. Finally, the dose was recalculated based on the new spot maps from the log files.

Results: The new log files-based plan was successfully imported in the TPS. The recalculated 3D dose distribution showed a reduced dose at the depth where the omitted energy layer would be expected. The discrepancy was 0.2 Gy of the planned 1.8 Gy at the maximum point, which lowered the PTV dose by about 0.1 Gy.

Conclusion: This new tool allows for the recalculation of daily fraction dose based on spot parameters obtained from the log files created after the fraction has been delivered. This will help evaluating the impact on the total dose of a wrongly delivered fraction, in case of machine interruption for example, or compare planned and delivered dose for a normal treatment.

P 054

Preventing mechanical collisions at a proton therapy center using 3D modelling

T. Henry1, C. Vallhagen Dahlgren1, A. Dasu1

1Skandionkliniken, Skandionkliniken, Uppsala, Sweden

Purpose: Treatment plans for patients treated at the Skandion Clinic are optimized at seven university hospitals around Sweden in a distributed model. A tool is needed to easily and rapidly check for mechanical limitations of a plan with prediction of potential collisions through gantry angle/snout position/table rotation combinations.

Material and Methods: The treatment rooms of The Skandion Clinic were 3D-modelled in a CAD software (Fusion 360, Figure 1). The included mechanical parts were: gantry, nozzle, snout, range shifter, couch, robotic arms and imaging panels. Full motion was simulated with the corresponding mechanical rotational/translational constraints of the different parts. Testing/validation of the model's accuracy was performed by comparing the distance between two nearly colliding parts in the model and in the room for a number of different scenarios that included general tests and geometries used for patient treatments.

Results: The maximum distance difference between room and model measurements was 5.4 mm, with a mean value of 2.5 mm. Implementation of the model in the treatment-planning workflow has prevented several collision scenarios and hence potential replans. Solutions are under evaluation to distribute the model to all University Hospitals in the near future. Other applications of the model are also under development.

Conclusion: A 3D model of the treatment rooms at the Skandion Clinic has been developed and has proven to accurately predict potential collisions. Its efficiency during the treatment-planning process has been demonstrated.

P 058

Beam stability as a function of gantry angle in a synchrocyclotron based protontherapy facility

A. Mazal1, J.A. Vera1, J.M. Perez1, I. Lorenzo1, J. Goffe2, J. Cal Gonzalez2, F. Dessy2, S. Rossomme3, F. Bisello3, J. Castro1

1Quironsalud, Centro de Protonterapia, Pozuelo- Madrid, Spain
2Ion Beam Applications, Particle Therapy, Louvain-la-Neuve, Belgium
3IBA Dosimetry, Particle Therapy, Schwarzenbruck, Germany

The stability of a PBS proton beam has been measured as a function of the gantry angle (each 10 degrees) and energy (105, 155, 170 and 225 MeV), including off axis values, for a facility based on a cryogenic synchrocyclotron and an open 220 degrees rotational gantry (Proteus One, IBA, Belgium). We used as main measuring methods ion chambers (cylindrical Razor and plane-parallel PP05) as well as a 2D scintillator detector (Lynx) combined with a dedicated PBS phantom (Sphinx), all detectors from IBA Dosimetry (Germany), either fixed in the room coordinates or rotating with the gantry with a specific holder. The measured range variations for all energies, off axis positions and gantry angles were 0.024 mm +/- 0.26 mm (1std.) including the measuring errors, that are under evaluation. The variations of the distal fall-off were at the level of the detector resolution (+/- 0.1 mm). Spot sizes have variations in their sigma of 0.05 mm +/- 0.09 mm (1std). Values are also referred to industrial validation and verification methods e.g. localization of isocenter with theodolites and coincidence with imaging and beam systems. Other values will be presented: e.g. peak width, spot positioning, skewness, and intensities. This approach can be used at acceptance, commissioning and/or systematic quality controls. Acknowledgments: participating staff from the institutions CPTQ, IBA and IBA dosimetry with special thanks to S.Ruiz, S.Lopes, L.Oliveira, S.Marcelis, C.Ares and R.Miralbell.

P 060

Simulated proton tomography of a thoracic patient in a half-gantry delivery system

M. Pankuch1, D. Robertson2, E. DeJongh3, M. Bues2, D. DeJongh3, A. Mahajan4, S. Keole2

1Northwestern Medicine Chicago Proton Center, Medical Physics, Warrenville, USA
2Mayo Clinic, Radiation Oncology, Phoenix, USA
3ProtonVDA, Physics, Warrenville, USA
4Mayo Clinic, Radiation Oncology, Rochester, USA

Background: Proton computed tomography (pCT) can produce 3-dimensional patient images at isocenter. Consequently, pCT could provide image guidance, particularly in half-gantries that cannot accommodate gantry-mounted cone-beam CT imagers. Unfortunately, pCT is limited by the ∼230MeV maximum energy in most proton centers. Because pCT requires the imaging proton beam to traverse the patient, some projections lack proton range for imaging. Additionally, the limited rotational range of half-gantry systems may limit image quality. Despite these challenges, the iterative nature of proton CT reconstruction may enable adequate imaging for patient alignment.

Methods and Materials: A planning CT of an adult thorax patient was used to simulate the image quality of a prototype pCT detector system (Figure 1). Images were reconstructed using 320MeV (all projections present) and 230MeV (some projections missing) proton beams and 190 degrees of gantry rotation. To simulate the effects of a larger set of missing projections, another reconstruction omitted a complete section of 30 degrees in the lateral sector.

Results: The 190 degree tomographic images at 320MeV and 230MeV demonstrate good contrast. The omission of 30 degrees from the lateral sector introduces imaging artifacts(Figure 2), but all images are of sufficient quality for volumetric alignment.

Conclusion: pCT in a half gantry may produce isocentric images of sufficient quality for volumetric alignment even with lateral sectors omitted in the reconstruction.

P 061

Towards FLASH proton therapy validation: imaging dose rates achieved by clinical scanned pencil beam

M. Rahman1, P. Bruza1, K. Langen2, D.J. Gladstone1,3, X. Cao1, B.W. Pogue1,4, R. Zhang3

1Dartmouth College, Thayer School of Engineering, Hanover- NH, USA
2Emory University, Emory Proton Therapy Center, Atlanta- GA, USA
3Dartmouth Hitchcock Medical Center, Radiation Oncology, Lebanon- NH, USA
4DoseOptics, Limited Liability Company, Lebanon- NH, USA

Interest in measuring dose rates in proton FLASH-RT, e.g. proton pencil-beam scanning (PBS) with ultra-high dose rate, has been heightened by the superior normal tissue sparing observed in FLASH-RT. We developed a technique that resolves clinical PBS' fast dynamics and dose rate using a fast-intensified CMOS camera and scintillation screen. The captured images were background subtracted, perspective transformed, and corrected for uniformity of the camera and scintillation screen response. Linearity in scintillation was verified for varying dose, dose rates, field sizes, and beam energies. Dose-rate temporal profiles were resolved at the isocenter for mono-energetic beam treatments plans of square fields. Maximum dose rate was mapped spatially for different spot-spacing. Cumulative dose-rate histogram was introduced for evaluation of dose rate to treated areas and showed the camera resolved a maximum dose rate of 25.6Gy/s. The imaging technique revealed PBS dose rates (1mm spatial and 10ms temporal resolution) and is available for quality assurance of proton FLASH-RT (with higher frame rate). The temporal profiles and histograms provide metrics for evaluation and potentially optimization of proton FLASH-RT plans.

P 062

Comparison of 2D, 2.5D, and 3D gamma analysis for pencil beam scanning patient-specific quality assurance

A. Stanforth1,2, R. Slopsema1, K. Stiles1,2

1Emory University, Department of Radiation Oncology and Winship Cancer Institute, Atlanta, USA
2Georgia Institute of Technology, Nuclear and Radiological Engineering and Medical Physics, Atlanta, USA

The purpose of this study was to compare the accuracy of 2D, 2.5D, and 3D gamma analysis for patient specific QA in pencil beam scanning proton therapy.

All treatment plans were calculated using a clinical Monte Carlo algorithm on a 2mm dose grid. A total of 849 2D dose distributions were measured using a 32x32 grid ion chamber array in water. For the 2D γ-analysis, the plane corresponding to the nominal depth was chosen. For the 2.5D γ-analysis, planes within ±3mm were analyzed, choosing the maximum γ passing value. All γ-analysis was done using in-house software written in Matlab. The geometric interpretation of DTA was used.[1] This breaks the 2D dose distribution into triangles and the 3D distribution into tetrahedral and calculates the closest distance between points using matrix multiplication and inversion.

The average 3%/3mm 2D, 2.5D, and 3D passing rates were 99.0%±3.38, 99.7%±1.04, and 99.4%±2.14 respectively. The 2%/2mm passing rates were 95.1%±8.61, 97.4%±5.99, and 96.9%±6.47. The percent of measurements above a 95% passing rate with the 3%/3mm criteria were 93.9%, 98.0%, and 96.2%, and 73.2%, 85.0%, and 81.8% for 2%/2mm.

The 2.5D and 3D passing rates are both larger than the 2D. The 2.5D criteria does not change for different planes, giving a looser overall criteria for errors in depth measurement whereas the 3D has a true criteria quantifying errors in depth, giving a higher passing rate.

References: 1. Ju T, Simpson T, Deasy JO, Low DA. Geometric interpretation of the γ dose distribution comparison technique: Interpolation-free calculation. Med. Phys. 2008;35(3);879-887.

P 063

Log data based patient-specific quality assurance at Osaka-HIMAK

M. Takashina1, T. Nakaji2, A. Komatsu3, N. Hamatani1, T. Tsubouchi1, Y. Wakisaka4, M. Yagi5, T. Kanai1

1Osaka Heavy Ion Therapy Center, Medical Physics, Osaka, Japan
2QST Hospital, Quality Control Section, Chiba, Japan
3Osaka University Graduate School of Medicine, Health Science, Ibaraki, Japan
4Osaka Heavy Ion Therapy Center, Radiation Technology, Osaka, Japan
5Osaka University Graduate School of Medicine, Carbon Ion Radiotherapy, Ibaraki, Japan

Purpose: To verify the treatment beam, we are developing a method to construct the dose distribution from the irradiation log data as Patient-specific Quality Assurance (PSQA) instead of the measurements.

Methods: We take the following three steps.

[0th step] (1) Make a treatment plan by VQA manufactured by Hitachi, Ltd., and calculate physical dose distribution in water for PSQA. (2) Obtain the monitor unit (MU) value and spot position (SP) from logfile of the treatment plan irradiation. (3) Input the MU and SP to VQA and calculate physical dose in water. (4) Perform gamma analysis with the dose distributions of (1) and (3).

[1st step] (1)(2) Same as the 0th step. (3) Prepare one spot physical dose distributions in water (DDw) for all nominal energies by VQA or Monte Carlo simulation with Geant4, and calculate dose distribution with MU, SP and DDw using in-house software. (4) Same as the 0th step.

[2nd step] (1) Make a treatment plan by VQA and calculate clinical dose distribution. (2) Same as the 0th step. (3) Calculate the clinical dose distribution on the planning CT image with MU and SP using the Monte Carlo simulation code Geant4. (4) Same as the 0th step.

Results: The 0th step has been successfully curried out, and its feasibility was confirmed. We next moved on the 1st step, and the in-house software has been developed.

Conclusions: The verification of the treatment beam using log data without dose distribution measurement bring the simplification of PSQA.

P 064

Measurement of resonant ionoacoustic waves in fixed-field alternating gradient accelerator for particle beam range verification

T. Takayanagi1,2, T. Uesaka1, Y. Nakamura3, M.B. Unlu4,5, Y. Kuriyama6, T. Uesugi6, Y. Ishi6, N. Kudo7, K. Umegaki5,8,9, T. Matsuura5,8,9

1Hokkaido University, Graduate School of Biomedical Science and Engineering, Sapporo, Japan
2Hitachi Ltd, Research and development group, Hitachi, Japan
3Hokkaido University, Graduate School of Engineering, Sapporo, Japan
4Bogazici University, Department of Physics, Istanbul, Turkey
5Hokkaido University, Global Station for Quantum Medical Science and Engineering- Global Institution for Collaborative Research and Education GI-CoRE, Sapporo, Japan
6Kyoto University, Institute for Integrated Radiation and Nuclear Science, Kumatori, Japan
7Hokkaido University, Faculty of Information Science and Technology, Sapporo, Japan
8Hokkaido University, Faculty of Engineering, Sapporo, Japan
9Hokkaido University Hospital, Proton Beam Therapy Center, Sapporo, Japan

Metal spherical markers used for patient positioning act as a strong pressure source when irradiated with proton beams; the marker briefly acts as an acoustic transmitter. To explore the possible use of this phenomenon as a range verification tool, an experiment was done in the fixed-field alternating gradient accelerator at Kyoto University, Japan.

Figure 1 shows the experimental setup. A hydrophone and an amplifier specialized for the measurement of spherical waves with the resonance frequency were developed. The beam intensity was about 1.2×108 particles/pulse. This is equivalent to the clinical dose per spot in the spot scanning irradiation method. Figure 2 shows the measured waveform in one pulse irradiation. Specific high-frequency waves originating from the marker were observed. Their frequency was 1.54 MHz, and this almost agreed with the value (1.62 MHz) simulated by Matlab k-Wave. The cause of the difference is currently under investigation.

The resonant waves were observable in clinical beam conditions. Moreover, the amplitude was linearly correlated with the distance between the marker and the Bragg peak. If this correlation coefficient is estimated before treatment, the residual beam range at the marker can be obtained in real time from the in-situ acoustic measurement. It is concluded that the measurement of resonant ionoacoustic waves generated from metal markers is useful for particle beam range verification.

P 065

A comprehensive and efficient daily QA for proton pencil beam scanning

S. Tang1

1Texas Center for Proton Therapy, Physics, Irving, USA

Purpose: No commercial or standardized daily QA product is available for proton radiotherapy. A comprehensive and efficient daily quality assurance (QA) program for pencil beam scanning (PBS) proton radiotherapy is established using the MatriXX PT device.

Materials and Methods: A single field with customized buildup was designed for pencil beam scanning machine to measure the proton beam range, output, the size and location of individual spots at multiple energies. The proton energy is measured at the distal falloff of a single energy pertained to the customized buildup. The spot size is determined by fitting the dose distribution. A web-based analysis program was implemented to perform the data analysis and show the trend of the results. The streamlined procedure can let therapists perform the proton beam QA in less than 10 minutes.

Results: The daily QA program was successfully run for more than three years at our center. An accuracy of 0.1 mm in range variation and 0.2 mm in spot size variation was demonstrated to be achievable.The trend of variation is observed and can be used to identify the degradation of the proton beam system as confirmed independently by the vendor.

Conclusions: The comprehensive and efficient daily QA program is proved to be useful to monitor and ensure the complicate proton beam system to meet clinical requirements.

P 067

Enable large-diameter multilayer ionization chambers for quality assurance of proton central axis pencil-beam spread-out Bragg peak

P. Wang1, M. Zhu2, L. Katja3

1Inova Health System, Inova Schar Cancer Institute, Fairfax, USA
2University of Maryland School of Medicine, Radiation Oncology, Baltimore, USA
3Emory University, Radiation Oncology, Atlanta, USA

Purpose: To introduce a method by which a large-diameter (12-cm) multilayer ionization chamber (LD-MLIC) can measure spread-out Bragg peak (SOBP) signals in proton treatment plans, including mixed energies, for quality assurance (QA).

Methods: Two stages using two types of contours are needed in treatment planning(Fig.1). The nominal plan is composed of proton pencil-beam spots with mixed energies close to the central axis. The integrated depth dose (IDD) curve contains a flat SOBP region. The LD-MLIC–measured IDD was compared to the IDD curve exported from the treatment planning system (TPS). In addition, three plans with intentionally modified energy layers were created using TPS and measured by the LD-MLIC. The water-equivalent thickness (WET) difference between the inserted and replaced energies was 0.2 cm. Six weeks of measurements were analyzed for QA purposes. A low-pass filter was introduced to mitigate the high-frequency noise in the IDD signal ratios. The filtered IDD signal ratios between the modified plans in different weeks and the baseline were used to check the energy accuracy.

Results: The differences between the LD-MLIC–measured and TPS-exported IDDs of the nominal plan were within 2% in most parts of the curve. Bumps/dips (∼1%) were noted in the filtered IDD ratio between the modified plans and the baseline(Fig.2).

Conclusion: The LD-MLIC can be used to measure IDDs with mixed energy layers for QA purposes. The LD-MLIC was sensitive in identifying plans in which an erroneous energy (0.2 cm in WET) layer replaced the correct one.

P 071

Handling high minimum spot weight constraints in scanned particle therapy of moving organs

C. Graeff1, M. Wolf1

1GSI, Biophysics, Darmstadt, Germany

Several strategies for the handling of moving targets in scanned particle therapy require splitting the delivered dose, such as rescanning but also conformal 4D-optimization. While some rescanning strategies explicitly perform less rescans for smaller spots, a high minimum spot weight (MSW) is beneficial in such plans, permitting more flexibility in the rescanning strategy and a faster overall delivery. Here, an approach for achieving good plan quality with very high MSW is presented for proton, helium and carbon ion plans.

After each optimizer iteration, the original implementation of GSI's TRiP98 zeroed spots which are below the MSW. Here, to promote higher MSW, too small spots are ignored with a probability decreasing with the number of iterations, and are rounded up if >0.5 MSW. The deleted spot weight is added to the neighbor with the highest optimizer gradient. In this way, spot weights can accumulate and pass the MSW gradually over several iterations. The method was tested in a single field for a lung cancer patient for physical and LEM-based RBE-weighted doses of 3Gy, and dose coverage was evaluated.

Dose coverage was consistently higher for the new method, and the achievable MSW with D95>95% increased by a factor of 5 for all ions, except RBE-weighted carbon with a factor of 2. The MSW exceeded the smallest deliverable spot weight at GSI by a factor of 15 to 40, sufficient for fast rescanning. Plans were also more regular with more spots at the same MSW, potentially increasing robustness.

P 072

Evaluation of pencil beam scanning proton therapy planning with 3D and 4D robust optimization versus photon IMRT for lung cancer

M. Huang1, J. Torok1, H. Chen2, L. Yuting3, L. Ren1, M. Blakey4, A. Kassaee5, N. Schreuder4, C. Kelsey1, F.F. Yin1

1Duke University Cancer Insitute, Radiation Oncology, Durham, USA
2Johns Hopkins Sibley National Proton Center, Radiation Oncology, Washington. DC, USA
3Emory University Proton Center, Radiation Oncology, Atlanta, USA
4Provision Proton Center, Radiation Oncology, Knoxville, USA
5University of Pennsylvania, Radiation Oncology Proton Center, Philadelphia, USA

The study aims to investigate the effects of lung tumor size on photon and proton planning comparison, and the robustness of 3D and 4D optimization for proton plans. Three categories of ten lung tumor patients were evaluated: A) Prescription 50Gy (1250cGyx4) early-stage SBRT, with peripheral lung tumor diameter 2cm-4.5cm; B) Prescription 60Gy (750cGyx8) hypofractionated treatment (HIGRT), with central lung tumor 2.5-6cm; C) Prescription 60Gy (200cGyx30) advanced-stage lung tumors.

Previously treated photon IMRT/VMAT plans cover Rx>95%PTV. For category A) and B), all planning target volumes (PTV) included 5mm expansion to the internal target volume (ITV). Proton planning IGTVs (union of 10 phases) were the same as photon ITVs. For C) PTV included the CTV plus 5mm margin expansion.

Three planning strategies were implemented to generate proton plans based on Monte Carlo calculation with RayStation(V8bSP1) proton planning pencil beam scanning system (PronovaSC360), with/without Robustness Optimization (RO). The 3D optimization was done on ITV for average CT, and 4D optimization (4% range uncertainty, 5mm setup uncertainty) was implemented for GTV(or CTV) on each individual breathing phase. Each phase-GTV and all plans were examined by physicians to ensure optimal target coverage and OAR constraints. The comparison among photon and proton plans for three tumor types were evaluated, the proton plans (no-RO/3D-RO/4D-RO) were compared. Selected results were presented in Figure1 and Table1-2.

For tumor size larger than 4cm diameter, the optimized proton plans demonstrated the superior OAR sparing compared to photon plans; higher target coverage was achieved with the 4D robust optimization proton planning method.

P 073

Comparative analysis of radiation treatment planning (RTP) systems using dose calculation algorithms for primary ocular melanoma radiation therapy

D. Jung1, J.G. Baek1, J.H. Cho1, K.C. Keum1

1Yonsei Cancer Center, Radiation Oncology, Seoul, Korea Republic of

This study aims to establish treatment plans for radiation therapy in primary ocular melanoma patients by analyzing and comparing treatment results and present the optimal radiotherapy option.

We developed plans for Brachytherapy (CCA and prescribed 85 Gy), which uses target volumes of 2 and 5.3 cc, CyberKnife (Fixed and MLC and prescribed 64 Gy/4 fx) and Proton (pencil beam and prescribed 64 Gy/4 fx) and compared their results.

For Brachytherapy, Durations were 291 and 372 hrs. OARs were the sclera, optic disk, and lens , which showed 1366, 0, and 15.8 Gy and 1448, 72.67, and 0.153 Gy. For the CyberKnife, Durations were 26 and 25 mins and MUs were 5801.1 and 3922.2. The maximum dose of the OARs were found to be 90.14, 24.19, and 39.06 Gy, and 81.67, 80.83, and 43.21 Gy. The PTV coverage was 94% for all. For Proton, Durations were not provided but, MUs were 633.15 and 2870.49. The maximum dose of OARs were 74.74, 10.19 and 53.28 Gy, and 73.56, 72.77, and 56.48 Gy, while the PTV coverage was 96% for all.

Brachytherapy demonstrated excellent dose reduction for the OARs, but it also resulted in patient discomfort due to long duration. CyberKnife and Proton planning resolved the discomfort due to their shorter durations, while maintaining the doses to OARs at reasonable levels. This Plan showed approximately 27% more dose reduction than CyberKnife. In conclusion, the plans for CyberKnife and Proton can serve as alternatives to Brachytherapy in patients with primary ocular melanoma.

P 074

The effect of titanium implants on proton spine SBRT

M. Kang1, L. Hu1, C. Ackerman1, C. Apinorasethkul1, P. Park1, J. Moreau1, J.I. Choi1, C.B. Simone- II1, J. Yamada2, H. Lin1

1New York Proton Center, Radiation Oncology, New York, USA
2Memorial Sloan Kettering Cancer Center, Radiation Oncology, New York, USA

Purpose: Metal implants are a challenge in radiation therapy and particularly in proton therapy. However, contouring uncertainties, artifact handling and the dosimetric impact of spinal implants to date have not been studied for proton SBRT. This study aimed to investigate those uncertainties for spinal titanium alloy implants placed in vertebrae treated with proton beams.

Methods: Titanium alloy rods and screws with known geometry and density were placed in a mini-water phantom and scanned using clinical protocols(Fig. 1). The contouring uncertainties were studied based on the delineations by different dosimetrists/physicists on iMAR and iMAR extended CT images using different window levels. Plan delivering 8Gy/fraction with field sizes of 2x2cm2 were generated in Eclipse to irradiate the implants. The dosimetry effects were studied by placing EBT3 films beyond the implants.

Results: The contouring uncertainties as the (standard deviation/mean volume) for implants, high- and low-density artifacts were quantified as 15%, 19% and 18%, respectively. The scattered proton after the implants can enhance field dose by ∼10%; however, Eclipse PCS cannot predict the scattered dose accurately(Fig. 2). The delineation uncertainties can introduce a range deviation of ∼6mm and more significant dose errors beyond the implants.

Conclusion: Both the extended and non-extended CT images can provide sufficient quality to contour the implants precisely by selecting the proper window level. The contouring and CT override can introduce large errors in the proton range and dose calculation that can be clinically significant for spine SBRT. The scattered protons can boost the field dose by ∼10% beyond the implants.

P 075

Dosimetric comparison of VMAT and IMPT in Liver plans

M. Keawsamur1, S. Suriyapee2, N. Amornwichet2, C. Khoprasert2

1King Chulalongkorn Memorial Hospital, Division of Therapeutic Radiation and Oncology, Bangkok, Thailand
2Chulalongkorn university, Division of Therapeutic Radiation and Oncology, Bangkok, Thailand

Purpose: The aim of the present study was to compare the dose distribution generated from photon volumetric modulated arc therapy (VMAT) and intensity-modulated proton therapy (IMPT) in the delivery of Liver cancer

Materials and Methods: Ten selected patients who underwent liver radiotherapy therapy between 2013 and 2018 were re-planned for a relative biological effectiveness (RBE) weighted dose of 30-50 GyE delivered with VMAT and IMPT with the same optimization criteria. Treatment plans were then compared.

Results: There were no significant differences in target volume dose coverage or dose conformity. IMPT led to more prominent normal liver sparing at low doses. There was a significant decrease of the mean doses delivered to the nontargeted part of the liver (18.3 vs.11.4 Gy, p < 0.05) and the kidneys (11.8 vs. 7.6 Gy, p < 0.05) and decrease of the maximum dose to the spinal cord (18.8 vs. 0.6 Gy, p < 0.05). The V10Gy of the liver was significantly decreased with the IMPT plans.

Conclusion: IMPT trend to improve sparing normal liver and major sparing organ at risk while the dose to target volume between VMAT and IMPT is comparable.

P 076

Margin design for breast cancers for a new proton center: a retrospective study

S.H. Lu1, C.W. Wang1, S.H. Kuo1, Y.C. Tsai1, C.S. Lin1

1National Taiwan University Hospital, Division of Radiation Oncology- Department of Oncology, Taipei, Taiwan Province of China

Background To establish the baseline of setup error and target delineation of breast cancer, a retrospective study was carried on to obtain the PTV margin based on the workflow of the National Taiwan University Hospital (NTUH). The data provides references for new proton therapy facilities.

Materials and Methods The breast cancer with lymph node-positive patients (including supraclavicular lymph, axilla lymph, and breast tissue) were selected. In order to minimize the reposition uncertainty, the patient was set at supine orientation, positioned with mask in a vacuum bag with a posture of both hands raising overhead . In the treatment room, the patient was initially positioned to align markers using lasers and followed by a weekly CBCT-scanning and image-registration was used to find the optimal position. Thirty patients were selected to find the statistics of a setup error. The mean values (M) of organ-motion and setup-error, besides the systematic (Σ) and random (σ) setup errors, were calculated.

Results The systematic setup-errors were 1.45, 1.72, and 1.03 mm in the left-right (LR), antero-posterior (AP), and supero-inferior (SI) directions, respectively. The random setup-error was 1.42, 1.73, and 1.87mm in the LR, AP, and SI directions, respectively.

Conclusions The statistics of setup errors are analyzed. This is important information to the new proton center to define the margin. More research such as BMI and target volume will continue.

P 078

Precision for automatic segmentation of pelvic anatomies on CT images using deep neutral networks with multi-slices image inputs

S. Otsubo1, Y. Takahashi1, Y. Maeda2, K. Yamamoto2

1University of Fukui, Dept. of Human and Artificial Intelligent Systems- Graduate School of Engineering / Faculty of Engineering, Fukui, Japan
2Fukui Prefectural Hospital, Proton Therapy Center- Proton Therapy Research Institute, Fukui, Japan

Purpose: To evaluate the precision for automatic segmentation of the prostate, rectum, bladder, and seminal vesicles (SVs) using deep neutral networks (DNN) with multi-slice inputs

Methods: We used planning CT-images of 190 patients undergoing proton beam treatment for prostate cancer and separated them into 170 for training and 20 for testing. Label data of all anatomies with the pixel size of 1.07x1.07 mm2 were created for each of CT-images (2 mm slice thickness) by radiation oncologists with the guide of MRI-images and used for training of DNNs developed based on SegNet and FCN, where the multi-images of adjacent slices were used for inputs to segment all anatomies on individual images simultaneously and the number of the input varied from a single to 9 channels. For each of input channels, the anatomies segmented by the best model of the validation were compared with the label data for testing by means of the dice similarity coefficient and the maximum Hausdorff distance.

Results: Table 1 shows means and standard deviations inside parentheses for dice coefficients and Hausdorff distances with each of the best number of channels and the corresponding statistical significance (p-value) from the single input model. The segmentation by the multi-image inputs tends to shows better precision than the one by the single one. The Hausdorff distance for the bladder shows the minimum on the best channel, as the prostate does with less significance (Fig.1). The development of DNNs models will continue to obtain better segmentation precision for these pelvic anatomies

P 079

Towards clinical application of ultra-high-dose-rate proton therapy: Monte Carlo calculations demonstrate the robustness of FLASH transmission plans for lung cancer

T. Pfeiler1, P. van Marlen2, M. Folkerts3, I. Huth1, J. Perez4, E. Abel5, W.F.A.R. Verbakel2

1Varian Medical Systems Particle Therapy, Clinical applications, Troisdorf, Germany
2Amsterdam UMC, Radiation Oncology, Amsterdam, Netherlands
3Varian Medical Systems Particle Therapy, PT Marketing Flash Program, Houston, USA
4Varian Medical Systems Particle Therapy, PT Flash Program Management, Steinhausen, Switzerland
5Varian Medical Systems Particle Therapy, PT Marketing Flash Program, Palo Alto, USA

Purpose: Ultra-high-dose-rate “FLASH” radiotherapy has been shown to reduce normal tissue toxicity without compromising tumor control in preclinical animal studies. A number of publications have explored possible planning paradigms to enable FLASH delivery in patients. The purpose of this study is to investigate the quality and robustness of FLASH transmission proton plans using Monte-Carlo dose calculations.

Methods: For five lung SBRT patients, ten-field non-coplanar 244 MeV transmission beam FLASH plans were created, optimized on PTV and calculated using a pencil-beam algorithm. The plans were designed to be delivered with existing ProBeamTM systems and demonstrated equal or better normal tissue sparing to clinical VMAT plans. Each plan was re-calculated using the Eclipse proton Monte-Carlo algorithm and analyzed based on dose metrics and robustness against setup and range uncertainties (±2 mm, ±3%).

Results: The highest deviations between the Monte-Carlo and pencil-beam algorithms were observed for the target maximum dose (up to 6.6%). The sparing of organs at risk was equal or better than VMAT plans except for V5Gy (total lung - PTV) (Table 1). In all uncertainty scenarios, 100% of the ITV volume was covered by at least 95% of the prescribed dose. Most of the uncertainty DVH curves resembled the nominal curve closely (Figure 1).

Conclusion: Monte-Carlo calculations verified robustness and dosimetrically favorable results of FLASH transmission plans compared to VMAT plans for the examined lung cases. This study confirms the viability of FLASH transmission proton therapy in the clinic.

P 082

Evaluation of margin design and immobilization strategies of lung cancers in NTUH for a new proton therapy center

Y.C. Tsai1

1National Taiwan University Hospital, Division of Radiation Oncology- Department of Oncology, Taipei, Taiwan Province of China

Background A ProBeam (Varian) system will be installed in the National Taiwan University Cancer Center in mid-2020. To establish the baseline of target delineation and review the efficacy of immobilization strategies for lung cases currently, a retrospective review was conducted based on the workflow of the National Taiwan University Hospital (NTUH).

Materials and Methods In NTUH, the patient was initially positioned based on markers and lasers and followed by a procedure consisting of CBCT-scanning and image-registration to find the optimal-position. The setup-error is defined as the position difference between the initial-position and that after image-registration. The organ-motion was obtained through comparing the position shifts between registrations based on bones and soft-tissues. Twenty patients were selected to find the statistics. In general, combinations of immobilization devices were applied(Table-1). The mean values(M) of organ-motion and setup-error were calculated. Moreover, the systematic(Σ) and random(σ) setup errors were calculated.

Results The systematic setup-errors were 1.95, 1.44, and 1.84 mm in the left-right(LR), antero-posterior (AP), and supero-inferior (SI) directions, respectively. The random setup-error was 1.88, 1.24, and 1.66 mm in the LR, AP, and SI directions, respectively. Moreover, the organ-motion was 0.75, 0.85, and 1.25 mm in the three direction, respectively.

Conclusions The statistics(Table-2) of setup errors and organ motion are analyzed. This is an important information to the new proton center to define its new immobilization strategies and delineation of CTVs. In addition, the accuracy of positioning of the new center can be evaluated based on the statistics.

P 083

Clinical benefit of range uncertainty reduction enabled by direct stopping-power prediction from dual-energy CT

M. Tschiche1,2, N. Peters2,3,4, P. Wohlfahrt3,4,5, C. Hofmann6, C. Möhler6, S. Makocki1, C. Richter1,3,4,7

1Department of Radiotherapy and Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus- Technische Universität Dresden, Dresden, Germany
2both authors share first authorship, Germany
3OncoRay - National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus Dresden- Technische Universität Dresden- Helmholtz-Zentrum Dresden - Rossendorf, Dresden, Germany
4Helmholtz-Zentrum Dresden - Rossendorf, Institute of Radiooncology - OncoRay, Dresden, Germany
5Now with Massachusetts General Hospital and Harvard Medical School, Department of Radiation Oncology, Boston, USA
6Siemens Healthineers, Forchheim, Germany
7German Cancer Consortium DKTK- Partner Site Dresden, and German Cancer Research Center DKFZ, Heidelberg, Germany

Purpose: Together with the first clinical implementation of direct stopping-power prediction (DirectSPR) from dual-energy CT (DECT), the resulting range uncertainty could be reduced substantially compared to that of the current state-of-the-art heuristic CT-to-SPR conversion (HLUT). The dosimetric impact was evaluated for the first patients treated with this new approach.

Material/Methods: Range uncertainty in our institution was reduced from 3.5%+2mm (HLUT approach) to 1.7%+2mm for the head and 2.0%+2mm for the pelvic region (DirectSPR approach). The dosimetric impact was analysed for 10 prostate- as well as 10 brain-tumour patients (CTV in close proximity to the brainstem). For this, clinical treatment plans, using DirectSPR with the reassessed uncertainty, were reoptimised with the HLUT approach and its respective range uncertainty. The two approaches were compared concerning the change in integral normal tissue dose, DVH parameters of relevant organs at risk and the conformity index.

Results: For DirectSPR, integral normal tissue dose was reduced on average by 4.4% for prostate- and 0.8% for brain-tumour patients based on our dosimetric evaluation. Mean dose to bladder and rectum was decreased by 1.6% on average for prostate-cancer treatments, with bladder ΔV65Gy/V65Gy=3.5% and rectum ΔV60Gy/V60Gy=2.3%. In the brainstem, the mean dose was reduced by 3.3%. The conformity index was increased on average by approximately 8.6% for prostate-cancer- and 1.9% for brain-tumour patients.

Conclusion: Enabled by the first clinical implementation of DECT-based DirectSPR, range prediction accuracy was substantially increased in routine clinical practice. For the first patients treated using DirectSPR, the clinical benefits regarding dosimetric outcome and normal tissue sparing have been demonstrated.

P 084

Unique non-coplanar irradiation making the most of 45-degree oblique port in carbon ion radiotherapy for head and neck cancer

Y. Wakisaka1, A. Hasegawa2, T. Tsubouchi3, N. Hamatani3, M. Yagi4, M. Takashina3, H. Maruo1, T. Ishii1, T. Kanai3, J. Mizoe2

1Osaka Heavy Ion Therapy Center, Department of Radiation Technology, Osaka, Japan
2Osaka Heavy Ion Therapy Center, Medical Office, Osaka, Japan
3Osaka Heavy Ion Therapy Center, Department of Medical Physics, Osaka, Japan
4Osaka University Graduate School of Medicine, Department of Carbon Ion Radiotherapy, Suita, Japan

Purpose: In the treatment planning of heavy ion radiotherapy, the arrangement of beam directions is important to be robust against setup errors or changes of the tumor size, and not to locate the distal end of spread-out Bragg peak which has high relative biological effectiveness (RBE) on normal tissues. By taking advantages of 45-degree oblique irradiation port and 6-axis movable robotic couch in our facility, a unique irradiation is often carried out which uses non-coplanar oblique (NCO) beams on the 45-degree angle from superior-anterior side instead of coplanar vertical (CV) beams just from anterior side. This study was performed to assess the expected influences using NCO beams.

Methods: The dose distributions of 6 patients with head and neck cancer were analyzed for NCO and CV beams using the same dose constraints. In addition, these two beams were compared according to the following three indices:1) the dose deviation by moving the isocenter position to simulate the setup error, 2) the dose distribution calculated on computed tomography (CT) image retaken during the irradiation, 3) the RBE of surrounding normal tissue.

Results: Considering setup errors, NCO plan hardly decreased the dose to the target and increased the dose to the chiasma and the brain stem as compared with CV plan (Fig.1). The transformation of dose distributions on retaken CT and RBE of normal tissues behind the target were smaller in NCO plan (Fig.2).

Conclusion: Heavy ion radiotherapy using non-coplanar beam with 45-degree oblique port showed advantages of improving robustness and decreasing RBE of normal tissue.

P 085

Intensity Modulated Proton Therapy: Optimization objectives and planning strategies to minimize effects of RBE/LET in normal tissues for anal cancer

T. Williamson1, X. Zhang1, M. Chen1,2, F. Poenisch1, R. Hunter1, P. Diagaradjane3, E. Holliday4, R. Zhu1, N. Sahoo1, G. Sawakuchi3

1The University of Texas MD Anderson Proton Therapy Center, Radiation Physics, Houston, USA
2Ruijin Hospital- Shanghai Jiao Tong University School of Medicine, Department of Radiation Oncology, Shanghai, China
3The University of Texas MD Anderson Cancer Center, Radiation Physics, Houston, USA
4The University of Texas MD Anderson Cancer Center, Radiation Oncology, Houston, USA

Purpose: Perform a dosimetric study to analyze treatment planning system (TPS) objectives, planning strategies, and their impact on variable RBE (vRBE) using the McNamara model and LET times dose (LETxD) to normal tissues.

Methods: Thirteen patients were analyzed who had been treated for anal squamous cell carcinoma to simultaneous integrated boost: 43-45 Gy and 50-54 Gy in 25 or 27 fractions. Patients were retrospectively planned using Eclipse TPS version 13.7. Two planning strategies utilized were 3 field (AP/LPO/RPO), and 5 field (AP/LPO/RPO/RtLat/LtLat). Beam angle variation was 5-15 degrees. Standard treatment plan objectives included aggressive low dose constraints to the bowel. Alternative plan objectives reduce or removed low dose constraints to the bowel. All plans were multi-field robustly optimized (0.5cm isocenter shift and 3.5% range uncertainty). Field specific targets were consistent for each beam across the different optimization approaches.

Results: Significant reduction in bowel bag vRBE max point dose (3.7– 5Gy), v54Gy (79% - 98%), max dose and 1cc for LETxD (19% - 31%) were observed in plans with alternative objects for both 3 field and 5 field plans. All bowel bag low dose constraints were within tolerance regardless of approach.

Conclusion: TPS objectives and planning strategies play an important role in minimizing the effects of LET/RBE in normal tissues. Five field plans with the alternative objectives had the best dosimetric approach in minimizing the effect of vRBE/LETxD to the bowel, without significant increase to bone marrow V10Gy.

P 087

Multiple-CT anatomic robust optimization for intensity-modulated proton therapy in head and neck patients

M. Zhu1, N. Onyeuku2, J. Snider1, K. Langen3

1University of Maryland School of Medicine, Radiation Oncology, Baltimore MD, USA
2Wellstar Cobb Hospital, Radiation Oncology, Austell GA, USA
3Emory University School of Medicine, Radiation Oncology, Atlanta GA, USA

Weight loss in head and neck (HN) cancer patients treated with intensity-modulated proton therapy (IMPT) often causes dose distribution degradation, due to the sensitivity of proton dose to anatomic changes. We hypothesized that robust optimization (RO) with multiple CT (MCT) could improve robustness and reduce the need of re-planning.

Five HN patients treated to primary site and bilateral cervical lymph nodes were selected. Four RO plans were created for each patient: 4-field (4F) (LAO,RAO,LPO,RPO), 3-field (3F) (RAO,LAO,PA), 4F-MCT, and 3F-MCT. All plans accounted for 5mm isocenter position offset and 3.5% proton range uncertainty. MCT RO also included a modified planning CT: superficial 5mm of tissue was overridden to air density. Plans were optimized to nearly identical target coverage and organ-at-risk doses.

CTV D95 and V95, Dmax to brainstem and spinal cord, and Dmean to parotids and oral cavity on re-scan CT were compared. There was no statistically significant difference between the 4-field (4F and 4F-MCT) and 3-field (3F and 3F-MCT) plans, except for oral cavity Dmean (2116cGy vs. 1952cGy, p=0.02). Compared to the single CT plans, MCT RO reduced the brainstem Dmax from 1581cGy to 1426cGy (p=0.01) and the spinal cord Dmax from 2258cGy to 1964cGy (p=0.01); no difference were observed for CTV D95, CTV V95, parotid and oral cavity Dmean.

Simulating weight loss by modifying planning CT can reduce dose to brainstem and spinal cord, while maintaining the target coverage and doses to parotid and oral cavity. MCT RO can reduce the re-planning frequency in treating HN cancer patients with IMPT.

P 090

Proton range verification with PET using an 18O-enriched water phantom

S. España1,2, A. Espinosa Rodríguez1,2, D. Sánchez-Parcerisa1,2,3, V. Valladolid-Onecha1,2, J.M. Udías1,2, C. Bäumer4,5,6,7, C.M. Bäcker4,5,7, B. Timmermann4,5,6,7,8, P. Fragoso Costa5,8, L.M. Fraile1,2

1Instituto de Investigación Sanitaria San Carlos, IdISSC, Madrid, Spain
2Complutense University of Madrid, Grupo de Física Nuclear- IPARCOS, Madrid, Spain
3Sedecal Molecular Imaging, Molecular Imaging, Madrid, Spain
4West German Proton Therapy Center, West German Proton Therapy Center, Essen, Germany
5West German Cancer Center WTZ, University Hospital Essen, Essen, Germany
6German Cancer Consortium DKTK, German Cancer Consortium, Heidelberg, Germany
7TU Dortmund University, Faculty of Physics, Dortmund, Germany
8University Hospital Essen, Clinic for Nuclear Medicine, Essen, Germany

Several techniques have been proposed for in-vivo proton range verification, including the use of PET imaging to measure the β+ isotopes produced by the proton beam through inelastic collisions inside the patient. However, minimal activation occurs near the distal end of the proton beam and radioisotopes created tend to diffuse and perfuse away from the proton interaction point, which causes PET images to be distorted away from the proton activation region.

We tested the capability of 18O-enriched water for in-vivo range verification in proton therapy. A cylindrical phantom was built including 7 cylindrical inserts with 6 mm diameter. Three of those cylinders were filled with 1% agarose gels containing 10% 18O-enriched water. A wedge-shaped HDPE block was placed proximal to the phantom in order to produce a continuous range variation transverse to the phantom and a monoenergetic proton field of 100 MeV was delivered to the phantom.

Later, the phantom was taken to a PET-scanner and scanned in list-mode for 4 hours. PET data was reconstructed in consecutive frames. A Monte Carlo simulation was performed to obtain the corresponding depth dose distribution in the phantom. The decay curves along reconstructed frames were fitted to a sum of exponentials including the decay for 11C, 13N, 15O and 18F in order to obtain the individual contribution for each isotope. It can be observed in the attached figure that falloff region of the activation overlays the falloff of the proton dose 2 hours after irradiation, when most of the signal corresponds to 18F.

P 091

Time-dependent dose calculation for Flash treatment planning

T. Evans1, T. Zwart1, J. Cooley2

1Mevion Medical Systems, Advanced Development, Littleton- MA, USA
2Mevion Medical Systems, Advanced Development, Littleton, USA

Recent well-publicized research into the “Flash” effect suggests that a radiotherapy treatment may be significantly impacted by the timescale over which it is delivered. Accurate time-dependent treatment planning will therefore be critical for clinical Flash applications. We take the first steps toward time-optimized treatment planning by developing time-dependent dose calculation in which the time history of dose delivery to an individual voxel can be examined. As an application we consider the case of dose-delivery to a 3D volume by the Mevion S250i scanning proton therapy system. Monte Carlo simulations of individual spots in intensity-modulated proton therapy (IMPT) plans are performed in the radiotherapy modeling package Topas. The outputs are then combined in a time sequence based on a benchmarked model of the delivery machine. This time sequence produces time-dependent dose histories which can be compared for different treatment plans, such that different sequence strategies and the impact of changing machine parameters can be examined. We use the time between set dose thresholds (e.g. 20% and 80% of maximum) as an initial metric. The dose-averaged dose-rate is also considered for different treatment plans. Of particular interest is the effect of reordering spots within the same treatment volume, potentially allowing for the consolidation in time of dose delivery. We find significant differences in characteristic voxel delivery times for different spot orderings.

P 092

A literature study on LET-definitions in radiation biology

F. Kalholm1, B. Singers Sørensen2, G. Leszek3, N. Bassler1

1Stockholm University, Medical Radiation Physics, Stockholm, Sweden
2Aauhur University, Experimental Clinical Oncology, Aarhus, Denmark
3Institute of Nuclear Physics, Proton Radiotherapy Group NZ62, Kraków, Poland

Introduction: Linear energy transfer (LET) is commonly used for expressing radiation quality of ion beams, where a higher LET typically is associated with an increased relative biological effectiveness (RBE). There are, however, multiple ways of calculating the LET of mixed radiation fields, which in practice is always present, even for quasi-monoenergetic beams, which may lead to varying results. The colloquially used term LET may thus be ill defined and a source of difficulties when results from different publications are compared. In this extensive literature study, we review various aspects on how the LET is calculated and reported, and speculate on the consequences of this.

Methods: Different methods of calculating the LET were evaluated for 84 different papers, involving dose averaged LET vs track averaged LET, which particles were included (e.g. primaries only, primaries + light secondaries or all particles), and whether an energy threshold was adopted or not. Attention was also given to density normalization.

Results: Thirty-one percent of papers used dose averaging, 5% track averaging and 64% did not specify the averaging method. 1% of papers calculated LET of the primary particle, 23% included all particles, and 75% did not specify which particles were included 23% of articles used no threshold, 2% specified a threshold and 75% did not specify a threshold.

Discussion/Conclusion: The multiple ways of calculating the LET may introduce an additional source of error when RBE values from various publications are compared. We suggest that a consensus should be pursued for future reporting of LET of mixed radiation fields.

P 093

Monte Carlo simulation of a double scattering proton therapy beamline using TOPAS

A. Maia Leite1, L. De Marzi1, P. Lansonneur1, A. Da Fonseca1

1Institut Curie- Radiation Oncology Department- Orsay- France, Centre de Protonthérapie d'Orsay, Orsay, France

Purpose: A variable RBE treatment planning has the potential to optimize proton beam therapy, while reducing normal tissue toxicity. Monte Carlo simulations can be employed to model RBE by accurately calculating dose and LET distributions. We report on the Monte Carlo simulations of a complex double scattered proton beamline at Institut Curie.

Methods: A TOPAS based simulation of a universal nozzle was developed to model the double scattering proton therapy. The simulation has been comprehensively validated against measurements of depth-dose and SOBP in water, for several ranges and modulation widths. The Monte Carlo simulations were also validated by recalculating patient treatment plans with TOPAS and benchmarked with the TPS.

Results: Figure 1 (a) shows that the depth-range for different energy options is well within the clinical tolerance, and that the modulation width closely matches the measurements. Figures 2 (a) and (b) present an example of the dose distribution calculated with the TPS and with TOPAS. A 2%/2mm 3D gamma analysis revealed a passing rate of 99%. The DVHs of selected organs at risk are presented in Figure 2. There is a good agreement between the analytical calculation and the MC-simulation, with a maximum difference of 4% evaluated in the PTV.

Conclusion: Our results demonstrate that the Monte Carlo method can be used to independently validate the TPS calculation, and furthermore to estimate the LET distributions as shown in Figure 2 (c). This tool will then be employed to correlate LET and RBE-weighted dose distributions with the incidence of radiation induced toxicity following proton beam therapy.

P 094

Study of secondary neutron production in PT treatments using MONDO, an innovative ultra-fast neutrons tracker

M. MARAFINI1, A. Bochetti2, M. De Simoni2, Y. Dong3, M. Fischetti4, G. Franciosini2, I. Mattei3, R. Mirabelli5, M. Toppi4, G. Traini5

1Centro Fermi - Museo Storico della Fisica e Centro Studi e Ricerche E.Fermi, Physics, Rome, Italy
2Sapienza Università di Roma, Physics, Rome, Italy
3INFN, Sezione di Milano, Milan, Italy
4Sapienza Università di Roma, Scienze di Base ed Applicate per l'Ingegneria, Rome, Italy
5INFN, Sezione di Roma, Rome, Italy

The secondary neutrons produced in Particle Therapy (PT) can deposit a non-negligible energy in- and out-of-field. To account for this contribution when optimising the treatment plans and to minimise the additional dose due to neutrons it is necessary to perform a complete secondary neutron spectra characterisation. In this context it is crucial to separate the secondary neutrons from the ternary neutral component generated in the iterative interactions of fragmentation products with the treatment room and the patient itself. To this aim a tracking detector is needed. The MONDO tracker is composed by a compact matrix of thin plastic scintillating fibres (∼250μm), assembled in orthogonal oriented planes (total size 16x16x20cm3), optimised for the ultrafast neutron tracking using single and double elastic scattering interactions. The tracking of both recoil protons allows for a complete neutron four-momentum reconstruction. An innovative SPAD based system with integrated electronics has been designed for the fibres readout (SBAM sensor). The detector performance has been studied with a MC simulation study that implements both the detector geometry and the trigger strategy. An energy resolution of ∼ 5-8 % is expected for neutrons in the ultrafast energy range while the expected back-pointing resolution is < 5 mm (neutron source @20 cm). The secondary neutron spectra expected from a Carbon ion beam impinging on a PMMA target has been convoluted with the detector neutron tracking capability. The expected results and efficiency will be presented as well as a preliminary SBAM readout characterisation with cosmic rays and protons.

P 095

Evaluating an aperture-based approximation to model a dynamic collimation system

N. Nelson1, B. Smith1, W. Culberson1, D. Hyer2, S. Rana3, P. Hill4

1University of Wisconsin-Madison- School of Medicine and Public Health, Department of Medical Physics, Madison, USA
2University of Iowa, Department of Radiation Oncology, Iowa City, USA
3Miami Cancer Institute- Baptist Health South Florida, Department of Radiation Oncology, Miami, USA
4University of Wisconsin-Madison- School of Medicine and Public Health, Department of Human Oncology, Madison, USA

Aim: The Dynamic Collimation System (DCS) is a collimation system capable of providing energy layer-specific collimation in pencil beam scanning proton therapy. Substantial normal tissue and increased target conformity can be achieved through four nickel collimators that rapidly and independently move to intercept the scanning beam. To enable treatment planning with this device, current analytical dose calculations using the Astroid treatment planning system approximate the DCS as an infinitesimally thin beam-limiting aperture. This work evaluates the validity and limitations of an aperture-based approximation to model the DCS.

Methods: A TOPAS Monte Carlo model of the IBA dedicated nozzle at the Miami Cancer Institute was created and validated through commissioning measurements. The DCS and aperture were then separately incorporated to simulate DCS- and aperture-collimated 6cm x 6cm uniform treatment fields in the presence and absence of an external range shifter with an air gap of 24cm. Simulated treatment fields consisted of equally weighted 100 MeV spot-scanned proton beams with 2.5mm spacing.

Results: With a range shifter in place, dose differences between DCS- and aperture-collimated fields were observed in the penumbra, leading to discrepancies up to 10%. In the absence of a range shifter, the observed differences were within individual simulation uncertainties.

Conclusions: For non-range shifted fields, the single aperture-based approximation is sufficiently accurate for modeling the DCS, but for range shifted fields, a full simulation of the DCS is recommended due to the additional scatter induced in the range shifter that cannot be collimated with an infinitesimally thin aperture.

P 096

RTSandT microdosimetry models for proton and carbon ions RBE calculation

A. Pryanichnikov1,2,3, A. Simakov1,3, M. Belikhin1,2,3, F. Novoskoltsev4, I. Degtyarev4, Y. Altukhov4, E. Altukhova4, R. Sunyukov4

1Protom Ltd., Research and Development, Protvino, Russian Federation
2Lomonosov Moscow State University, Accelerator Physics and Radiation Medicine, Moscow, Russian Federation
3Lebedev Physical Institute RAS, Physical-Technical Center, Protvino, Russian Federation
4NRC “Kurchtov Institute”, Institute for High Energy Physics named by A.A. Logunov, Protvino, Russian Federation

This work contains descriptions of the features of the implementation of the Microdosimetric Kinetic Model (MKM) and Local Effect Model (LEM) as models that are included in the RTSandT software package. The results of theoretical and experimental studies of the main microdosimetric characteristics for cellular structures placed in homogeneous water phantoms irradiated with 454 MeV/u 12C6+ ions are presented.

The report compares the calculated (RTSandT using different hadron generators and several methods for calculation of ionization losses and the standard version of the FLUKA code) with experimental data for the depth-dose distribution, flux- and dose-averaged linear energy transfer from the primary carbon ion beam. There is also a comparison of the depth-RBE distributions at different levels of survival and biological dose distributions, calculated for 9 types of cellular structures using the RTSandT and FLUKA codes. In conclusion, the results of modeling the parameters of modified Bragg curve obtained in the framework of RTSandT for the conditions of irradiation a homogeneous water phantom with a carbon ion beam using a ridge filter are given in comparison with the experimental results

P 097

Experimental validation of a 4D dose calculation algorithm for irregular motion in carbon ion therapy

T. Steinsberger1, M. Lis1, G. LeDoeuff1, M. Wolf1, C. Graeff1

1GSI Helmholtzzentrum für Schwerionenforschung GmbH, Biophysics, Darmstadt, Germany

Current simulations of scanned particle therapy are typically based on regular repetitions of the same 4DCT, failing to account for irregular breathing. We developed a LEM-based method to compute non-linear RBE-weighted dose of carbon ions for sequential CT series. One CT is loaded sequentially for each motion phase and the contributions of the entire particle spectrum of each motion phase are added up in a CT-sized array. Here we present an experimental validation of this algorithm for physical doses.

Scanned carbon beam interplay patterns have been measured with a time-resolving ionization chamber array detector placed on a non-periodically moving linear stage. We applied several motion scenarios including variations of the motion amplitude and baseline drifts. In a modified version of the CNAO, Pavia, Italy dose delivery system, time-resolved motion traces and beam data were recorded. Each detector frame constituted one motion phase for 4D-dose calculation. This way, dose calculations with more than 1000 states where performed. The reconstructed dose was compared to the measurement using a (3 mm, 3%) Gamma analysis for the integral dose and each frame.

Passing rates 94.3(1.2)% where achieved.

The now validated dose calculation is going to be used for simulation studies on realistic motion scenarios for lung cancer patients using phantom CTs created by the XCAT tool and on MR-based motion sequences of real patients. The algorithm will provide a better understanding of interplay mechanics and the effects of irregular motion.

P 098

Radiobiological impact of target and projectile fragmentation respectively in proton and carbon ion therapy: the FOOT experiment

G. Traini1, on behalf of the FOOT collaboration

1INFN - Section of Rome, -, Rome, Italy

In the proton therapy the proton RBE is at present assumed as constant and equal to 1.1. Such an assumption needs to be overcome to account for the known, significant enhancement of the proton RBE above 1.1. One of the possible causes of the reported RBE increase is the fragmentation of patient nuclei that can produce a large amount of highly ionising secondary fragments traveling only tens of μm from their production point. The fragments contribution to the dose due to this process is more relevant in the entrance channel, where the beam crosses the healthy tissues, affecting the Normal Tissues Complication Probability (NTCP). The target fragmentation is not accounted for in the current clinical TPS implementation as the underlying physics processes are scarcely known, and no measures are currently available. The FOOT collaboration, made by 100 researchers from France, Germany, Japan and Italy started in 2017 a measurement campaign of the double differential fragmentation cross sections of interest in proton and carbon therapy. The use of two independent setups will provide a wide angular coverage (up to 70° with respect to the beam direction). The experimental apparatus are respectively based on an emulsion spectrometer and a fixed target detector with magnetic spectrometer, silicon tracking detectors, Time Of Flight system and a BGO calorimeter. In this contribution the status and the measurement program of the FOOT experiment will be presented and and evaluation of the eventual the impact of the fragments on the proton and carbon therapy will be discussed.

P 099

A proton treatment planning platform to facilitate FLASH small animal pre-clinical research

E. Traneus1, R. Nilsson1

1RaySearch Laboratories AB, Research, Stockholm, Sweden

Introduction: We present a set of tools available in the μ-RayStation treatment planning system to facilitate research on proton FLASH planning and delivery using sub-millimeter voxel size and proton ranges down to a few millimeters where FLASH relevant parameters such as the combination of dose, dose-rate within pulses, and overall time of irradiation can be studied.

Materials and Methods: The μ-RayStation system utilizes RayStation's proton Monte Carlo dose engine extended with FLASH relevant scorers and beam models. It is possible to create single energy layer PBS plans with SOBP style fields or high energy overlapping punch through fields. For PBS delivery a dedicated scorer can accumulate a time trace of energy depositions per voxel and report dose-rate versus dose histograms per voxel and other metrics such as average dose rate, LET, “Dirty Dose”, and RBE.

Results: We report FLASH metrics (scorer time resolution 100 μs) for plans where air gap, field size, spot size, spot separation were varied. In one example case, by varying the air gap from 2 to 15 cm, the dose rate vs. dose histograms for an OAR indicates that 75%, 50% and 25% of the dose is delivered at dose rates (unit Gy/s) above 190, 430, 640 for the 2 cm air gap and above 90, 220, 370 for the 15 cm air gap.

Conclusion: We have demonstrated that the μ-RayStation platform is suitable for retrospective plan analysis as well as for design of irradiation experiments to study aspects of FLASH delivery and radio biological effects.

P 101

Feasibility of cardiac and breast-sparing whole lung irradiation using intensity modulated proton therapy: a dose calculation comparison study

R.X. Wong1, J. Faught2, W. Myers2, M. Krasin2, A. Faught2, S. Acharya2

1National Cancer Centre Singapore, Radiation Oncology, Singapore, Singapore
2St Jude Children's Research Hospital, Radiation Oncology, Memphis, USA

Aim: Whole lung irradiation (WLI) is indicated for certain pediatric patients with lung metastases. WLI is conventionally performed with AP/PA photon fields that include the entire heart and bilateral breasts. This study investigates whether intensity modulated proton therapy (IMPT) WLI can significantly spare the heart and breasts as compared to conventional WLI.

Methods: Conventional and IMPT plans were generated for five patients (age 5-22 years). Prescription dose was 16.5GyRBE in 1.5GyRBE fractions. Conventional plans used 6MV photons prescribed to mid-line and field-in-field technique to cover the PTV (ITV+1cm). IMPT plans used scenario-based optimization with 5%/5mm range/positional uncertainty to robustly cover the ITV. Monte Carlo dose calculation was used for all IMPT plans. Doses were compared using paired Student's t test.

Results: ITV Dmean was similar between IMPT and conventional plans, but IMPT plans had lower Dmin and higher Dmax at tissue interfaces (Dmean ratio 0.96 p=n.s, Dmin ratio=0.9 p<0.001, Dmax ratio=1.1 p=0.014). Dmean of the heart (ratio 0.63, p=0.008), left ventricle (0.61, p=0.004), right ventricle (0.45, p=0.003), left atrium (0.79, p=0.004), right atrium (0.81, p=0.008), breasts (0.40 left, 0.46 right, p<0.05) and body Dmean (0.62, p=0.002) were reduced with IMPT.

Conclusions: IMPT resulted in comparable ITV coverage and statistically significant lower cardiac and breast mean doses. WLI IMPT is a feasible option for pediatric patients with presumed long-term survival to reduce risk of cardiac morbidity and breast secondary cancers.

P 103

Is daily CBCT necessary for proton breast treatment?

E. Batin1, E. Blokzijl1, D. Wagenaar1, J.A. Langendijk1, S. Both1, A.P.G. Crijns1, J.H. Maduro1

1University Medical Center Groningen, Radiation Oncology, Groningen, Netherlands

This study aims to investigate whether daily CBCT is necessary after surface image positioning.

Five breast cancer patients with nodal involvement were treated using robustly optimized intensity modulated proton treatment plans with a 5 mm/3% setup/range uncertainty setting. Initial daily positioning was performed with AlignRTTN surface image followed by CBCT from which 6D corrections were applied. These patients also received weekly verification CTs for which they were positioned using AlignRT.

Two methods were used to evaluate the theoretical dose that would have been delivered if CBCT corrections would have not been applied before treatment. For the first method (DCBCT), the 6D CBCT correction vectors applied before treatment delivery were inversely applied to the initial planning CT for each treatment fraction, and the dose per fraction was computed and summed. For the second method (DCT), the weekly dose was calculated on unregistered control CTs with a patient position equivalent to the position at treatment before CBCT. Target coverage and organ-at–risks (OAR) doses were evaluated and compared to those from the treatment plan.

For the DCBCT method, the clinical target criteria were met (D98 > 95% prescribed dose) for all cases except one target (D98 = 94.4%). Similar targets doses were obtained with the DCT method and one target at D98 = 92.6%. For both methods, some OARs doses were slightly higher on DCBCT/CT and one case presented a mean heart dose superior to the treatment one by 0.3 Gy(RBE).

Surface image seems sufficient to obtain proton breast treatment patient positioning fidelity.

P 106

Implementation of a patient-specific CBCT dose calculation workflow in TOPAS

T. Henry1, A. Dasu1

1Skandionkliniken, Skandionkliniken, Uppsala, Sweden

Purpose: To accurately determine patient and organ-specific total dose received from CBCT imaging during the course of a treatment.

Material and Methods: The proton-mounted CBCT system at the Skandion Clinic in Uppsala, Sweden, was implemented in the Monte Carlo code TOPAS. Validation of the model was performed by comparing depth-dose curves in water and lateral profiles in air between TOPAS and measurements. Three clinically used protocols (head, pelvis and thorax), with three different beam output parameters, were used for this validation. Subsequently, the possibility to import in TOPAS CT data and RT structures for an anthropomorphic phantom and to perform CBCT simulations on them was evaluated. The end-goal was to be able to perform patient-specific CBCT dose calculations in individual organs to assess the magnitude of the imaging dose received by each patient during a proton treatment course.

Results: Results after validation showed that the implemented CBCT model agreed within 4% with the measurements for both depth-dose curves and lateral profiles. CT datasets and RT structure volumes exported from a treatment planning system were also successfully implemented in TOPAS. The CT data was automatically correctly placed both in orientation and position. Preliminary simulations have been performed using the TOPAS embedded volume-based dose filters to determine organ doses based on RT structures volumes.

Conclusion: A workflow for individual determination of proton CBCT doses to specific organs has been developed. The obtained doses from individual patients will be evaluated with respect to diagnosis and age groups.

P 107

Dual-energy CBCT on a proton gantry-mounted system: exploring the potential for improved soft-tissue contrast

T. Henry1, B. Ivanić2, A. Dasu1

1Skandionkliniken, Skandionkliniken, Uppsala, Sweden
2Ion Beam Applications, s.a., Louvain-la-Neuve, Belgiu

Purpose: To evaluate the possibility to perform dual-energy CBCT (DECBCT) on the CBCT system available in one of the gantry rooms at the Skandion Clinic in Uppsala, Sweden. The aim was to improve image quality by increasing soft-tissue contrast.

Material and Methods: The head of an anthropomorphic phantom containing different soft-tissue-like materials for various brain structures was used for imaging. First, a CBCT scan with the default protocol (E=100 kV) currently used for clinical head scans was performed. Then, the same scan was repeated with a lower energy (E=70 kV) and the highest energy possible with the current generator system (E=120 kV). The two low- and high-energy CBCT reconstructions were then blended in Matlab, testing both linear- and non-linear blending. Different degrees of blending were explored. The blended images were then visually evaluated in comparison with the regular head protocol CBCT images.

Results: A clear contrast improvement was observed on the DECBCT images. New soft-tissue structures that were indiscernible in the regular CBCT images could be identified thanks to the dual-energy image improvements. While certain regions of the regular images showed a constant Hounsfield unit, the same region in the DECBCT imaged showed two different structures with a difference of approximately 20 HU.

Conclusion: The potential of dual-energy CBCT to improve soft-tissue contrast was explored. Enhancements in the soft-tissue differences were observed. An optimization process, including spectrum filtering, artifact mitigation and dose lowering will follow.

P 108

Stopping-power calculation based on dual-energy CT

G. Paiva Fonseca1, I. Rinaldi1, C. Hofmann2, I. Almeida3, B. van der Heyden1, W. van Elmpt1, G. Bosmans1, F. Verhaegen1, G. Vilches Freixas1

1Maastricht University Medical Centre+, Department of Radiation Oncology MAASTRO, Maastricht, Netherlands
2Siemens Healthcare GmbH, Computed Tomography, Forchheim, Germany
3Technical University of Lisbon, Center of Nuclear Sciences and Technologies, Lisbon, Portugal

Purpose: Particle therapy requires treatment margins due to range uncertainties compromising its physical advantages. Stopping-power ratio to water (SPR) accounts for a large part of range uncertainties with SPR assignment currently using look-up tables based on pre-defined materials. Dual-energy CT (DECT) allows SPR calculation on a voxel level leading to a patient-specific SPR map that can reduce range uncertainties. A Siemens prototype was used for direct SPR calculation based on DE images (DirectSPR) was evaluated.

Material/Methods: SPR for Gammex tissue-mimicking inserts were measured in our proton beam. Measurements were compared against calculations based on DECT images acquired for five different phantoms (Figure 1). DECT images were used to calculate relative electron density (RED) and effective atomic number (Zeff), which is empirically related to the mean excitation energy. Finally, SPR was calculated using Bethe's equation.

Results: Phantom size affects significantly (>10%) RED and Zeff calibration parameters. This difference is accounted for by using the water equivalent diameter per slice and its corresponding calibration. Table 1 shows the results obtained experimentally and calculated for the evaluated phantoms with deviation <2% for most of the inserts. Larger differences (>6%) were observed for lung inserts most likely due to its heterogeneities and composition as reported in other studies.

Conclusion: The prototype evaluated in this work provides accurate SPR values using its default calibration. However, the manufacturer recommends a CT specific calibration, which we are investigating, for optimal results. SPR is a promising alternative to integrate DECT in the clinical routine potentially reducing range uncertainties.

P 109

Evaluation of dual energy computed tomography for predicting proton stopping-power-ratio

H. Sheen1, H.B. Shin2

1Samsung Medical Center, Radiation Oncology, Seoul, Korea Republic of
2Department of Radiation Oncology, Yonsei Cancer Center, Seoul, Korea

Purpose: The effects of Computed tomography Hounsfield unit (CT HU)and effective atomic number (Zeff) obtained using a dual energy CT (DECT) were investigated in Direct relative stopping power ratio prediction (rSPR) via the Bethe formula (Direct CT No.-to-rSPR).

Materials and Methods: Virtual monochromatic images and polychromatic images of tissue characterization phantom Gammex RMI 467 (Gammex, Middleton, WI, USA) were acquired in a fast-kVp-switching single-source dual energy CT. Direct CT No.-to-rSPR was calculated using single energy CT (SECT) data at 120 kVp and DECT data at 77 keV and 140 keV. Their accuracy was then estimated by means of the relative error percentage of rSPR and root mean square (RMS) percentage.

Results: The relative error percentages (r%Error) between measured rSPR and calculated rSPR were obtained at 120 kVp, 77 keV, and 140 kVp. The lung density (<0.5 g/cm3) were excluded since their Zeff cannot be achieved. In soft tissue range (0.94 g/cm3 < ρ < 1.1 g/cm3), r%Error values were less than ±10% at 120 kVp, 77 keV and 140 keV. In the bone region (ρ >1.15 g/cm3), the relative error percentage was more than 5% at 120 kVp and 77 keV and less than 4% at 140 keV.

Conclusion: The rSPR which was obtained by CT HU and Zeff from CT data at virtual monochromatic 140 keV showed less error than the results in polychromatic 120 kVp. The virtual monochromatic energy CT data at 140 keV and Zeff are the unique functions of DECT in comparison with SECT, and the crucial factors for the accuracy of dose calibration. In this study, the results derived from 140 keV provide feasible evidences for clinical application in proton therapy.

P 110

Status update on the PTCOG-funded project “Proof of Concept Evaluation of Helium imaging for range uncertainty reduction in particle therapy”

L. Volz1,2, C.A. Collins-Fekete3,4, E. Bär3, S. Brons5, V.A. Bashkirov6, R.P. Johnson7, C. Sarosiek8, R. Schulte6, J. Seco1,2

1German Cancer Research Center, Biomdeical Physics in Radiation Oncology, Heidelberg, Germany
2Heidelberg University, Department of Physics and Astronomy, Heidelberg, Germany
3University College London, Department of Medical Physics and Biomedical Engineering, London, United Kingdom
4National Physics Laboratory, Chemical- Medical and Environmental Science, London, United Kingdom
5Universitätsklinikum Heidelberg, Heidelberg Ion-Beam Therapy Center, Heidelberg, Germany
6Loma Linda University, Department of Basic Sciences- Division of Biomedical Engineering Sciences, Loma Linda, USA
7University of California at Santa Cruz, Santa Cruz Institute for Particle Physics, Santa Cruz, USA
8Northern Illinois University, Department of Physics, Chicago, USA

Single-event particle imaging in the form of particle computed tomography (pCT) and radiography (pRad) is a promising tool for particle therapy planning and verification. Initial research in the field has focused on protons, but recently helium ions have shown great potential, especially in terms of spatial resolution due to the helium ions' lower multiple Coulomb scattering (MCS). In this work, we present a status update on our PTCOG funded project to investigate the use of helium imaging for pre-treatment optimization of the patient relative stopping power (RSP) information.

The project utilizes a state-of-the-art pCT scanner prototype developed by the U.S. pCT collaboration. The system was shipped to the Heidelberg Ion-Beam Therapy Center (HIT) and a series of beam tests have been conducted. To date, measurements were taken to benchmark the detector response for helium ions in comparison to protons. These measurements will further be used to investigate image artifacts observed for pCT with the scanner in previous experiments. Additionally, helium and proton radiographs and full CT scans of various plastic phantoms were acquired in order to compare the achievable image quality. A preliminary helium radiograph is shown in Figure 1. In future experiments, the feasibility of combining a few particle radiographs with a planning x-ray CT for personalized treatment optimization of the patient RSP values will be investigated.

P 111

Theoretical investigation of the spatial resolution in particle radiography

L. Volz1,2, C.A. Collins-Fekete3, J. Rambo Soelie4,5, J. Seco2,6

1German Cancer Research Center, Biomdeical Physics in Radiation Oncology, Heidelberg, Germany
2Heidelberg University, Department of Physics and Astronomy, Heidelberg, Germany
3National Physics Laboratory, Chemical- Medical and Environmental Science, Teddington, United Kingdom
4Western Norway University of Applied Sciences, Department of Computer Science- Electrical Engineering and Mathematical Science, Bergen, Norway
5University of Bergen, Department of Physics and Technology, Bergen, Norway
6German Cancer Research Center, Biomedical Physics in Radiation Oncology, Heidelberg, Germany

Particle radiography (pRad) has received increasing interest over the past years as promising tool for pre-treatment verification in particle therapy. However, the spatial resolution of pRad is limited due to the multiple Coulomb scattering (MCS) of the particles inside the phantom/patient. In state-of-the-art single-event pRad, the most likely path (MLP) of each particle through the phantom/patient is computed during image reconstruction in order to more correctly distribute the particles information along the deflected particle trajectories. The accuracy of the MLP is limited and has so far been used to derive the spatial resolution of particle imaging.

In this work, we investigate the theoretical spatial resolution limit of pRad for four different contemporary image reconstruction algroithms. Using simulated pRad data, we show that the spatial resolution of pRad is limited if the particles information is distributed evenly along the particle path, even if the ground truth particle trajectory would be used during image reconstruction (see Figure 1). In order to explain this observation, we developed theoretical models that enable to predict the spatial resolution of pRad reconstructions with the investigated image reconstruction algorithms. We validate the theoretical predictions through Monte Carlo simulations. The limitation of the spatial resolution is explained from the projection of scattered particle trajectories onto a single reconstruction plane. The developed formalisms enable investigation of the effect of different beam parameters on the spatial resolution. This work will be helpful in the development of improved image reconstruction techniques in the future.

P 112

Evaluation of patient specific range uncertainty with single energy CT for proton treatment

J. Zhu1, B. Liu1, X. Wang1, Y. Zhang1, R. Parikh1, R. Davis2, N. Yue1, K. Nie1

1Rutgers-Cancer Institute of New Jersey, Radation oncology, New Brunswick, USA
2Rutgers-Robert Wood Johnson University Hospital, Radiation Oncology, New Brunswick, USA

Introduction: Proton therapy uses a generic planning margin recipe such as, 3.5%±3mm, which may not be optimal in each individual case. In this study, we propose a scheme for which patient-specific range uncertainty (PSRU) can be quantified, thus to effectively improve proton dose distribution in terms of target coverage and OAR sparing at the distal and proximal end.

Methods: Due to the potential differences of range uncertainty (RU) for different tissues, PSRU method focuses on the extraction of RU for each tissue along with the beam path, then combines to be a patient's RU. To verify the proposed concept, a Gammex phantom with known chemical compositions of insert was scanned to analyze RU of CT imaging. RU from the stoichiometric calibration curve was also assessed. A proton beam plan was re-generated by considering PSRU as safety margin. The difference to the initial plan with 3.5% generic margin as used clinically was evaluated.

Results: As a proof of concept, a single beam brain case with 2.2% as PSRU (Table 1) was calculated with 5400cGy as prescription dose. The dose difference compared to the plan with generic margin is shown in Figure 1. At distal edge of the beam, a clear dose difference is observed with max dose difference of 1187.7 cGy, indicating that the proposed method could be potentially useful in patient specific margin design.

Conclusion: We have designed a model to compute PSRU for proton treatment and the results highlight the importance of robust planning and analysis in proton therapy.

P 113

Extraction of two effective atomic numbers and the improvement of stopping power ratio calculation with triple energy CT

J. Zhu1, B. Liu1, X. Wang1, Y. Zhou2, K. Nie1, R. Parikh1, R. Davis3, N. Yue1, Y. Zhang1

1Rutgers-Cancer Institute of New Jersey, Radiation Oncology, New Brunswick, USA
2Fudan University Zhongshan Hospital, Radiation Oncology, Shanghai, China
3Robert Wood Johnson University Hospital, Radiation Oncology, New Brunswick, USA

Introduction: This study is to extract two atomic numbers (Z ˆ and Z ˜, defined in Figure 1) of tissues with triple energy CT (TECT) and evaluate the corresponding improvement on stopping power ratio (SPR) calculation for proton therapy.

Method: In the derivation of tissue stopping powers, the relative electron-density and effective-atomic-number can be derived to degenerate the subtle tissue differences using dual-energy CT (DECT) while assuming Z ˆ=Z ˜=Zeff. By adding a third energy, and can be derived mathematically (Figure 1) to potentially improve the CT based tissue SPR derivation. To verify this method, a CIRS phantom was scanned three times using 80kVp, 120kVp and 140kVp on a GE LightSpeed scanner, simulating a TECT scan. A 3D-monotonous conversion between Z ˆ, Z ˜ and mean excitation energy (Im) was generated (Figure 2(a)). The accuracy of SPR calculated by TECT was compared to DECT approach using 80/140kVp and single energy CT (SECT) approach (120kVp).

Result: It is evident that the Z ˆ can be different from Z ˜ for certain tissues, but also that Z ˆ and Z ˜ can be used for deriving Im. The comparison of SPRs shows TECT and DECT have better performances in SPR calculation than SECT (Figure 2(b)). Moreover, the use of TECT approach led to more accurate SPR calculation and differentiation for most of the soft tissues evaluated in this study than the DECT approach.

Conclusion: The TECT based stopping power calculation approach can be used to more accurately calculate SPR than the DECT and more commonly used SECT approach.

P 114

Fundamental verification of new respiratory gating device for clinical implementation in proton beam therapy wobbler method

I. Maeshima1

1Niigata University of Heaith and Welfare, Radilogical Technology, niigata, Japan

Purpose and Background: Respiratory synchronized irradiation is used in Japanese proton therapyBy using AbchesET, respiratory control by the patient himself can be easily and accurately reproduced by directly seeing indication of rotation angle of pointer displayed on a small monitor. In this study, we compared the conventional respiration synchronizer (ANZAI) as a comparative subject and made various investigations on the usefulness of respiratory synchronous irradiation of proton beam wobbler method (Sumitomo) using a new respiratory synchronizer (AbchesET).

Methods: Verification of proton beam characteristics and delay time at respiratory synchronized irradiation was performed for each of ANZAI and AbchesET. With regard to verification of beam characteristics, verification of synchronism with no synchronization (stopped state) (Gate width: 12%, 25% when maximum expiration is 0%) is performed using a moving body phantom and a two dimensional detector for a rectangular radiation field. We evaluated each beam characteristics (Flatness, Symmetry, Penumbra, Field size and Dose) of the irradiation field. We also evaluated the dose distribution by gamma analysis. The breathing waveform of the moving body phantom was evaluated with respect to three types of respiration rates (Breath per Minute, BPM), two kinds of amplitudes (± 1 cm, ± 2 cm) with reference to the sinusoidal wave, and the breathing waveform of the respiratory synchronizer.

Conclusion: Although respiratory synchronous irradiation of proton beam wobbler method using AbchesET has some use restrictions, it was confirmed that it can be sufficiently used clinically. We conclude that respiration synchronized irradiation of proton beam wobbler method with higher accuracy is achieved by using AbchesET.

P 116

Assessment of Time-Volume-Dependencies in enhanced-DIBH-based lung cancer treatment with PBS proton therapy performed at CPT/PSI

F. Emert1, A. Bolsi1, A. Mayor1, A. Lomax1, D.C. Weber1

1Paul Scherrer Institute, Center for Proton Therapy, Villigen PSI, Switzerland

Introduction: Minimizing motion during proton therapy (PT) for lung cancer is of critical importance. The clinical pilot trial demonstrated that deep-inspiration-breath-hold, enhanced using prior inhalation of combined 100% O2 and hyperventilation (eDIBH), prolongs breath-hold duration by a factor of 3.5±1.1 in healthy subjects [teDIBH=(206±66)sec], whilst assuring maximum intra-fractional lung vessel displacement reproducibility of less than 4mm. In this work, we analyze target volume sizes that could be delivered, for different re-scanning scenarios, within patient-related breath-hold durations given the delivery characteristics of Gantry 2 at PSI.

Materials and Methods: Projecting eDIBH results on lung patients predicts breath-holds of (μ±σ)[teDIBH]=(70±22)sec. Using scaled, volumetric re-scanning calculated for Gantry 2 at PSI, and for a representative 4D-treatment (3-fields-beam arrangement), 6 synthetic CTs were generated shrinking the right lung volume to virtual PTVs from ∼1100cc to ∼100cc (4mm steps as compensation of intra-fractional motion). By calculating 19 identical plans on these PTVs, and applying 0-8 re-scans, we estimated if a single field of a given lung tumor could be irradiated within one eDIBH.

Results: Figure 1 displays calculated treatment (time/target volume) pairs including their linear fits for different re-scanning factors. The introduction of different eDIBH levels reveals that lung tumor volumes up to ∼800cc should be treatable in a single field without re-scanning. The upper limits for 2 and 4 rescans are about 400cc and 230cc (ellipsoids), respectively.

Conclusions: The eDIBH approach in combination with existing re-scanning at PSI's Gantry-2 offers a scalable possibility for safe, effective, and efficient lung tumor treatments. A more extended analysis is foreseen in the future.

P 117

Accuracy and robustness of 4D logfile-based dose reconstruction and start of clinical application

S. Spautz1, M. Arturs2, A. Jakobi1,3,4,5, N. Peters1,3, A.C. Knopf2, E.G.C. Troost1,3,4,6,7, C. Richter1,3,4,6, K. Stützer1,3

1OncoRay - National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus- Technische Universität Dresden- Helmholtz-Zentrum Dresden - Rossendorf, Dresden, Germany
2Department of Radiation Oncology, University Medical Center Groningen- University of Groningen, Groningen, Netherlands
3Helmholtz-Zentrum Dresden - Rossendorf, Institute of Radiooncology – OncoRay, Dresden, Germany
4Department of Radiotherapy and Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus- Technische Universität Dresden, Dresden, Germany
5now with: Strahlentherapie RheinMainNahe, Rüsselsheim, Germany
6German Cancer Consortium DKTK- Partner Site Dresden, and German Cancer Research Center DKFZ, Dresden, Germany
7National Center for Tumor Diseases NCT- Partner Site Dresden- Germany:, German Cancer Research Center DKFZ- Heidelberg- Germany- Faculty of Medicine and University Hospital Carl Gustav Carus- Technische Universität Dresden- Dresden- Germany- and- Helmholtz Association / Helmholtz-Zentrum Dresden - Rossendorf HZDR, Dresden, Germany

Introduction: We established a 4D logfile-based dose reconstruction for monitoring and potential intervention during intensity-modulated proton therapy (IMPT) of moving tumors. Before clinical application, we assessed the validity of reconstructed doses and the sensitivity against changes of selected input parameters by phantom experiments.

Material/Methods: A dynamic thorax phantom (CIRS, USA) with a soft-tissue target and radiochromic film insert was imaged by 4DCT and irradiated with either quasi-monoenergetic fields or 4D optimized proton plans. The surrogate signal (ANZAI, Japan) of the regular motion was recorded in synchronization with the machine logfiles. Reconstructions were performed with different dose grid resolutions (1mm/3mm), deformable image registrations (DIR; manually defined or automatically generated vector-fields) and artificial asynchronies between machine and motion logfiles.

Results: Characteristic dose patterns on radiochromic films were well reconstructed (Fig.1A). Gamma pass rates (2mm, 2%) for extracted characteristic profiles of the reconstructed and measured doses were >98% under static conditions, ranged between 99% and 86% for 5mm motion depending on applied reconstruction parameters, especially the DIR, and were about 80% for 30mm motion due to the predominant residual motion in the 4DCT (Fig.1B). Fig.1C demonstrates the robustness against potential minor asynchronies (≈5ms) between machine and motion logfiles. A workflow test during a pancreatic cancer IMPT treatment (Fig.2) revealed a data processing time of approximately 20min/fraction.

Conclusions: Due to satisfying accuracy and robustness for clinically aimed motion amplitudes (≤5mm), IMPT treatment of non-small cell lung cancer accompanied by daily 4D logfile-based dose monitoring will start in our institute within the first months of 2020.

P 118

Starting phase independency of conformal motion-synchronized dose delivery for carbon ions

M. Wolf1, M. Lis1, A. Paz1, T. Steinsberger1, M. Donetti2, C. Graeff1

1GSI Helmholtz Centre for Heavy Ion Research, Biophysics, Darmstadt, Germany
2CNAO, Research and development, Pavia, Italy

It is known that 4D-dose distributions are susceptible to interplay which complicates treatment plan QA as several repeated measurements are necessary to assess the impact of interplay. We hypothesize that conformal 4D-synchronized dose delivery eliminates interplay and, therefore, a single measurement is representative of the 4D-dose distribution, which will enable reconstruction of 3D dose distributions.

For the measurements, a PTW Octavius 1500 array detector was mounted on linear axis, which provided target motion perpendicular to the beam. A 60mm x 60mm square was then irradiated with a motion-synchronized plan with 10 amplitude-based motion-states. Two motion amplitudes (20mm and 40mm) were investigated with a 4sec Lujan-period. Each measurement was repeated 10 times with a shifted starting phase. Correlation parameters were calculated for each of 45 possible combinations of the 10 dose measurements per motion amplitude.

The resulting average (STD) correlation coefficient R between the dose measurements was 0.9992 (0.0002) and 0.9979 (0.0004) for the 20mm and 40mm amplitude, respectively (Fig.1).

As the repeated dose measurements only showed small variations and high correlation, we conclude that it is valid to reconstruct 3D doses from several measurements a different depths for repeated irradiations of the same motion-synchronized treatment plan.

P 119

The impact of minimal deliverable spot weight on rescanning effectiveness for different 4D planning scenarios

J. Zhang1, S. Safai1, D.C. Weber1, A.J. Lomax1, Y. Zhang1

1Paul Scherrer Institut, Center for Proton Therapy, Villigen-PSI, Switzerland

The effectiveness of rescanning to mitigate interplay effect typically improves with increasing number of rescans. However, rescanning is limited by the minimal deliverable monitor units (MU) of the treatment machine. Consequences of this are investigated in detail here by comparing “realistic” (taking minimal deliverable MU into account) with “ideal” rescanning (no restriction) for various fractionation and planning scenarios.

For two 4DCT-MRI liver datasets (peak-to-peak motion amplitudes of 15(10-20)mm and periods of 6.3(5.2-7.2)s), single-field SFUD plans were optimized for 6 different fraction doses (1/2/4/6/8/10/12Gy) and 2 plan scenarios: 1) 4mm lateral-spacing, 2.5mm water-equivalent energy-intervals; 2) 8mm lateral-spacing, 3.5MeV energy-intervals. 4D doses were calculated for 6 layered-rescanning scenarios (rnom=1:3:15) with 2 different layer switching times (TdE=80/900ms) and 5 different minimal MU conditions (Fig1a). For all scenarios (840/case), beam delivery with 21 starting phases were simulated. Motion mitigation effectiveness was quantified using D5-D95 (CTV) for all scenarios.

The interplay effect only marginally (2%) depends on initial 3D plan configuration, prescribed dose and layer switching time (Fig2). Impact of minimal MU however is pronounced for small spot-spacings, low prescription doses and high rescan number (Fig 1b and 2a). In addition, initial plan configuration (spot spacing) has a significant impact on 4D rescanning performance, especially for conventional fraction dose (1Gy/field), with larger spot spacings resulting in more effective and robust re-scanning. Rescanning also becomes more effective for hypofractionated dose schemes, independent of the layer switching time.

Our results demonstrate that increasing spot-spacing generally allows for more effective and robust motion mitigation when using layered rescanning.

P 120

Comparison study of radiation biology of carbon ion (12C6+) beam versus X-ray to NCI-H1975 lung adenocarcinoma cell line

R.F. Liu1, Z. Yang2, L. Shao3, H. Luo1, Q. Zhang1, X. Wang1

1Lanzhou Heavy ion Hospital, Radiation oncology center, Lanzhou/ Gansu province, China
2Lanzhou University, Basic medicine college, Lanzhou/ Gansu province, China
3Lanzhou University, The First Clinical Medical College, Lanzhou/ Gansu province, China

Objective: To investigate the radiosensitivity differences of carbon ion versus X-ray radiation H-1975 cell lines.

Method: EGFR mutation type lung adenocarcinoma cell line, NCI-H1975, was cultured . Respectively using 0, 1, 2, and 4 Gy carbon ion beam or X-ray to irradiate H-1975 cell line. Continue cultured to 24h and 48h, using cell clone forming test to detect cell clone rate, cell count assay (CCK 8) to detect cell inhibition rate and flow cytometry analysis to test cell cycle and apoptosis respectively. RT-PCR was used to measure mRNA expression of HIF1-a and VEGF.

Results: The results of cloning formation test and cell count test showed that Carbon ion has more inhibition effect than X-ray, and 4Gy group has the most significant inhibition effect. The inhibition effect was significantly increased with the increase of irradiation dose at 24h, 48h and 72h. Flow cytometry analysis results showed that the cell cycle was significantly blocked in the G2/M phase after carbon ion beam and X-ray irradiation, and blocking effect was most significant at 24h, Carbon ion beam has more significant blocking effect than X-ray in different dose. With the increase of radiation dose, apoptosis rate increased obviously. RT-PCR showed that with the increase of irradiation dose, the expression levels of HIF-1a and VEGF decreased accordingly, and carbon ion beam decreased more obviously than X-ray (p<0.05).

Conclusion: Carbon ion beam can more significantly inhibit the proliferation of H-1975 cell and induce cell to G2/M phase arrest than X-ray through significantly decreasing expression levels of HIF-1a and VEGF.

P 121

Laser-hybrid Accelerator for Radiobiological Applications (LhARA)

K. Long1, Hin Tung Lau1

1Imperial College London, Physics, London, United Kingdom

The ‘Laser-hybrid Accelerator for Radiobiological Applications',LhARA, is conceived as a novel, uniquely flexible facility dedicated to the study of radiobiology. The technologies that will be demonstrated in LhARA have the potential to allow particle-beam therapy to be delivered in a completely new regime, combining a variety of ion species in a single treatment fraction and exploiting ultra-high dose rates. LhARA will be a hybrid accelerator system in which laser interactions drive the creation of a large flux of protons or light ions that are captured using a plasma (Gabor) lens and formed into a beam. The laser-hybrid approach will allow radiobiological studies using a variety of ion species in completely new regimes.

LhARA will be developed in two stages . In the first stage, a programme of in vitro experiments (e.g. focusing on 2D cell survival and 3D spheroid/organoid growth in tumour-specific models, kinetics of DNA damage and repair) will be served with proton beams with energies between 10 MeV and 15 MeV. In stage two, the beam will be accelerated using a fixed-field accelerator (FFA). This will allow experiments to be carried out in vitro, and particularly in vivo (e.g. using tumour-specific xenograft mouse models), with proton beam energies of up to 125 MeV. In addition, ion beams, with energies up to ∼30 MeV per nucleon for carbon, will be available for in vitro and in vivo experiments. This paper presents the conceptual design for LhARA and the RandD programme by which the LhARA consortium seeks to establish the facility.

P 122

Carbon ion inhibiting esophageal squamous cell carcinoma proliferation and metastasis by regulated STAT3 pathway

H. Luo1, Z. Yang2, Q.N. Zhang3, X.H. Wang3, R.F. Liu3

1Gansu Provincial Cancer Hospita, Department of radiotherapy, Lanzhou, China
2Lanzhou university- Lanzhou, Basic medical college, Lanzhou, China
3Gansu Provincial Cancer Hospital, Department of radiotherapy, Lanzhou, China

Conventional radiation resistance is one of the main reasons for treatment failure in esophageal squamous cell carcinoma (ESCC). The superiority of heavy ion radiation in physics and biology has been increasingly highlighted in radiation therapy research. Signal transducer and activator of transcription 3 (STAT3) and its signal transduction pathway play an important role in the occurrence, development and metastasis of ESCC and are related to the development of ionizing radiation resistance in ESCC. Therefore, the aim of the present study was to investigate relationship between carbon ion inhibiting proliferation and metastasis of esophageal carcinoma cell and STAT3 signal pathway.The result demonstrated that carbon ion beams can significantly reduce cell viability and stimulate apoptosis in human ESCC cells in a dose-dependent manner. In addition, carbon ion beams induced G2/M phase cell cycle arrest in ESCC cells and dose-dependently inhibited tumor metastasis. Additionally, poorly differentiated Kyse-150 cells were more sensitive to the same carbon ion beam dose than moderately differentiated Eca-109 cells. Carbon ion beam exposure regulated the relative expression of metastasis-related molecules at the transcriptional and translational levels in ESCC cells. Carbon ion beams also regulated Ecadherin and MMP2 in the downstream STAT3 pathway and inhibited ESCC cell metastasis, which activated the STAT3 signaling pathway. This study confirmed the inhibition of cell proliferation and the metastasis effect of carbon ion beam therapy in ESCC cells. Additionally, we found that carbon ion beams inhibited human ESCC cell proliferation and metastasis in Eca-109 and Kyse-150 cells by regulating the STAT3 pathway.

P 123

A dedicated ripple filter for accurate measurements of cells survival irradiated with proton and carbon ion beams

F. Bourhaleb1, C. pardi1

1I-See Computing Ltd, I-See Computing Ltd, Turin, Italy

Background: The radiobiological cell survival measurements at the Bragg peak have the problem of uncertainties in the positioning of cell samples in the peak region. This due to the very sharp shape of the maximum of energy deposition for carbon ion beams(the Gaussian fit at the peak is less than 0.1mm sigma), and also proton beams at low energies.

Method: We designed for this specific purpose a special ripple filter in order to produce a small broadening of the peak region to allow a more reliable positioning of cells samples. Effects of the used ripple filter are accurately calculated. We used a Monte Carlo simulation on the top of GEANT4, well benchmarked for both proton a carbon ion beams.The beam line is a realistic one. It is principally composed of the exit window, the monitoring system, a first ripple filter and s second ripple filter. Actually instead of a detailed monitoring system (MS) we used an 'equivalent MS' in terms of water equivalent thickness.

Results: In this work we show results of carbon ion beam of 62 MeV (see Figure). Two ripple filters are designed in such away that the Bragg peak have a guassian shape of about 0.4 mm. We built two ripple filter of 0.75 mm thickness each, and with a triangular shape. The final ripple filter chosen to help radiation biology measurement is the one describe in the figure below, showing the minimum needed spread to get the cells samples irradiated still in the peak region.

P 124

FLASH proton irradiation of normal and breast cancer cells

D. Sanchez-Parcerisa1,2,3, Á. Gutierrez-Uzquiza2,4, P. Bragado2,4, S. España1,2, A. Espinosa1,2, C. Gutierrez-Neira5, P. Ibañez1,2, V. Sanchez-Tembleque1,2, J.M. Udias1,2, L.M. Fraile1,2

1Universidad Complutense de Madrid, Nuclear Physics Group, Madrid, Spain
2Instituto de Investigación Cínica San Carlos, IdISSC, Madrid, Spain
3Sedecal Molecular Imaging, Algete, Madrid, Spain
4Universidad Complutense de Madrid, Departamento de Bioquímica y Biología Molecular, Madrid, Spain
5Universidad Autónoma de Madrid, Centro de Microanálisis de Materiales, Madrid, Spain

Recent preclinical studies in mice and other animal models have demonstrated a protective effect of FLASH (very high dose-rate) radiotherapy in healthy tissues, while apparently not compromising its curative ability. Since clinical proton accelerators still require certain hardware adaptations to produce FLASH rates, biological experiments on FLASH proton therapy can be conducted in nuclear physics accelerators. The 5-MV tandem accelerator at CMAM (Madrid) offers an excellent setup for such experiments, with proton beams (up to 10 MeV) hundreds of times more intense than clinical beams and the capacity to irradiate samples with sub-millimeter precision.

Using specifically designed pinhole collimators, we can obtain a range of dose rates from 0.1 to 1000 Gy/s, covering both FLASH and standard-rate protontherapy. We developed a smart irradiation system with 3D-printed materials and robotic actuators, coupled with a specifically designed 2D treatment planning system for cell-culture samples (Fig1). Bean monitoring and dosimetry were performed using a Faraday cup combined with EBT3 films.

We used the irradiation system to deliver up to 8 Gy in healthy fibroblast cell lines (obtained from fresh mammary tissue from women undergoing mammoplasty), tumor-associated fibroblasts and MDA-MB-231 (triple negative) breast cancer cells. Initial results show a clear negative spatial correlation between delivered dose and cell survival (Fig2) for tumor cells irradiated at FLASH rates. Full analysis of the irradiated samples (ongoing) will include determination of DNA damage by H2AX foci formation, viability analysis using MTT, and study of additional markers, like HIF-1α.

P 125

Evaluation system for recurrent head and neck cancer

I. Kazuyo1, A. Sasaki2, Y. Ichikawa1, R. Ogawara3, M. Suda3, T. Hamano3

1Okayama university, Neutron Therapy Research Center, Okayama, Japan
2Okayama university, Graduate School of Medicine Dentistry and Pharmaceutical Sciences, Okayama, Japan
3National Institute for Quantum and Radiological Science and Technology, National Institute of Radiological Sciences, Chiba, Japan

Most of the patients with recurrent head and neck cancer (rHNC) are considered to have poor prognosis and consequently, palliative and supportive care in rHNC is usually chosen. The world first boron neutron capture therapy (BNCT) case of rHNC after surgery, X-ray therapy and chemotherapy, revealed the completely tumor regression and improvement in quality-of-life. In this study, the effectiveness of BNCT in resistant head and neck squamous cell carcinoma is evaluated.

Human oral squamous cell carcinoma (OSCC), 5-fluorouracil-resistant OSCC (5-FU-resistant OSCC) and X-ray resistant OSCC were used in vitro and vivo evaluation study. Boronophenylalanine (BPA) was used as the Boron-10. The boron accumulation was measured by Inductively Coupled Plasma Mass Spectrometry. In vitro tests were conducted at Neutron Exposure Accelerator System for Biological Effect Experiments (NASBEE). The boron uptake from BPA in each cell lines was more than 1μg/106 cells and the cell survival ratio of cells with BPA decreased exponentially with the increase of neutron fluence in all cell lines, in comparisons with the cell survival ratio of cell without BPA. The required neutron fluence to reduce survival to 10 % is 1∼2 × 1011/cm2 in each cells with BPA, on the other hand 4∼8 x1011 /cm2 in each cells without BPA.

This study showed the BNCT in rHNC is effective and these data is helpful to the evidence based BNCT. Moreover, cytology using this in vitro evaluation system would indicate the efficacy of BNCT in near future.

P 126

A cyclotron-based BNCT system and the current status of its regulatory process for the medical device in Japan

S. Masui1, T. Asano2

1Sumitomo Heavy Industries USA Inc., Sales and Marketing, Allentown, USA
2Stella Pharma Corporation, Management, Osaka, Japan

At last year's [2019] PTCOG in Manchester, U.K., we reported that the pre-submission review of the medical device application on a cyclotron-based BNCT system had started in Japan for head and neck cancer, and brain cancer. The current status of the application will be presented here.

A preview process called “Sakigake Designation System” by the Japanese Pharmaceuticals and Medical Devices Agency (PMDA) is a mechanism to expedite the approval process for breakthrough medical products. Sumitomo Heavy Industries's cyclotron-based BNCT system “NeuCure” and Stella Pharma's boron compound, boronophenylalanine or BPA (INN : borofalan(10B)), “SPM-011” were endorsed as breakthrough products in 2017, and the Sakigake preview process started in May 2019. Because a BNCT system in general is considered a combination product, only simultaneous use of NeuCure and SPM-011 will be allowed for public treatment within the clinical trials we conducted this time.

A Monte Carlo calculation code “NeuCure Dose Engine” for treatment planning has also been submitted to PMDA. It is critical to understand that a device, a drug, and treatment planning software need to be approved before public treatment.

As of the time of this abstract, after the Sakigake preview for head and neck cancer, the official application was submitted in October, 2019 and the review is now underway. We hope we could present further progress at the time of PTCOG 2020.

P 127

In vivo evaluation of a phenylboronic acid-installed, actively targeted novel BNCT agent against melanoma-bearing mouse model

Y. Matsumoto1, A. Kim2, N. Fukumitsu3, M. Suzuki4, Y. Nagasaki2, H. Sakurai1

1University of Tsukuba Hospital, Radiation Oncology Proton Medical Research Center, Tsukuba, Japan
2University of Tsukuba, Department of Materials Science- Graduate School of Pure and Applied Sciences, Tsukuba, Japan
3Kobe Proton Center, Department of Radiation Oncology, Kobe, Japan
4Kyoto University, Institute for Integrated Radiation and Nuclear Science, Kumatori, Japan

In BNCT, it is necessary to increase the accumulation ratio of B-10 compound in tumor and normal tissues. In this research, we developed novel boron-containing nanoparticle. The objective of this study was to introduce phenylboronic acid (PBA) on the nanoparticle surface and to target metastatic cancer cells with high sialic acid expression by the specific binding of PBA to sialic acid. We will report on the design of polymeric materials, the preparation of nanoparticles, and the evaluation in B16 melanoma tumor-bearing model mice. A block copolymer (PBA- poly(ethylene glycol)(PEG)-b-poly(lactic acid) (PLA)) was synthesized via anionic polymerization. After dissolution, nanoparticles were obtained by dialysis against dd water. The affinity of nanoparticles to sialic acid was quantified at the molecular level by surface plasmon resonance (SPR), and its cytotoxicity and interaction to cancer cell membranes were also examined. The antitumor effects of the nanoparticles on BNCT was evaluated using mice melanoma. In consideration of the difference in the bioavailability, nanoparticles and boronophenylalanine-fructose complex (BPA-f) were administered subcutaneously at 48 h and 2 h before irradiation. The nanoparticles showed a high affinity for the sialic acid immobilized surface of the cancers cell in vitro. In in vivo evaluation, the nanoparticles administration group exhibited the same efficacy as BPA at a boron dose corresponding to 1/100 of the BPA-f administration group. The nanoparticles are anticipated as a high-performance BNCT treatment.

P 128

BNCT (Boron Neutron Capture Therapy) mediated by BPA + Glutamine in the hamster cheek pouch oral cancer model

A. Monti Hughes1, I.S. Santa Cruz2, L.N. De Leo2, P.S. Ramos2, J.A. Goldinger1, S. Thorp3, P. Curotto4, E.C.C. Pozzi4, M.-C. Hsiao5, M.A. Palmieri6, M.A. Garabalino2, V.A. Trivillin1, A.E. Schwint1

1National Atomic Energy Commission CNEA- National Research Council CONICET, Radiobiology, Buenos Aires, Argentina
2National Atomic Energy Commission, Radiobiology, BsAs, Argentina
3National Atomic Energy Commission, Instrumentation and Control, BsAs, Argentina
4National Atomic Energy Commission, Research and Production Reactors, BsAs, Argentina
5Research and Development Center, Hi-Q Marine Biotech International Ltd.
6University of Buenos Aires, Biodiversity and Experimental Biology, BsAs, Argentina

Introduction: Some of the translational studies in our laboratory are devoted to assess different compounds as an adjunct to BNCT to improve therapeutic efficacy and/or reduce toxicity. Oligo-Fucoidan, a seaweed extract, significantly improved therapeutic efficacy of BNCT in an experimental oral cancer model but did not reduce mucositis, a dose-limiting radioinduced adverse effect (PTCOG 2019). Glutamine (GLN) depletion is associated to cancer development. In head and neck cancer patients, GLN deficit is accentuated by radiotherapy/radiochemotherapy side effects. However, no conclusive outcomes are reported for GLN in the treatment of oral mucositis, nowadays with no effective treatment available. Previous studies showed that GLN inhibited tumor development in cancerized hamster cheek pouches. The aim of the present study was to evaluate the potential capacity of GLN to reduce BNCT-induced mucositis and/or enhance therapeutic efficacy in the hamster cheek pouch oral cancer model.

Materials and Methods: Tumor bearing hamster cheek pouches were exposed to BPA-BNCT at 2.6 Gy to precancerous tissue +/- oral GLN administration (1 g/kg body weight per day during follow-up).

Results and Conclusion: Preliminary results showed that GLN did not reduce severe mucositis. However, GLN+BPA-BNCT improved the % of tumor complete responses vs BPA-BNCT at 7 and 14 days after BNCT (42% vs 16%, 68% vs 38%, respectively). Ongoing studies will assess new protocols to improve GLN radioprotective and therapeutic effects.

P 129

Albumin-based boron delivery to tumor by active targeting via integrin

H. Nakamura1, K. Kawai1, S. Sato1, M. Suzuki2

1Tokyo Institute of Technology, Institute of Innovative Research, Yokohama, Japan
2Kyoto University, Institute for Integrated Radiation and Nuclear Science, Osaka, Japan

Boron neutron capture therapy (BNCT) has been attracting growing interest as one of the minimally invasive cancer therapies. Mercaptoundecahydrododecaborate (Na2[B12H11SH]) and L-p-boronophenylalanine (L-BPA) have been used in BNCT for many years. L-BPA, in particular, has been widely used for the treatment of melanoma, brain tumor and head and neck cancer. However, development of new boron carriers is still strong requirements for cancers that are not able to be treated with L-BPA. Serum albumin accumulates in malignant and inflamed tissues due to enhanced permeability and retention (EPR) effect. Furthermore, it has been observed that tumor is the major site of serum albumin catabolism, thus serum albumin has been extensively investigated as a versatile drug carrier. We developed maleimide-functionalized closo-dodecaborate (MID) to conjugate it to bovine serum albumin (BSA). Surprisingly, MID was found to conjugate not only to free SH of cysteine residue but also to lysine residues in albumin under physiological conditions determined by MS/MS analysis. The highly boronated BSA showed high and selective accumulation in tumor. Significant tumor growth inhibition was observed in colon 26 tumor-bearing mice subjected to thermal neutron irradiation. Furthermore, we introduced cyclic RGD peptide (cRGD), which is known to strongly bind to αvβ3integrin overexpressing on many cancer cells and neovascularities, into free SH of Cys34 for active targeting to tumor. The integrin-dependent boron uptake was observed in U87 cells that overexpress αvβ3integrin. A higher BNCT effect was observed in U87 cells treated with cRGD/MID-BSA compared to MID-BSA.

P 130

Performance test of the new bolus in BNCT using cyclotron-based epithermal neutron source and 3D printed water phantom

A. Sasaki1, T. Takata2, Y. Kudo3, M. Suzuki4, Y. Sakurai2, T. Mitsumoto5, H. Tanaka2

1Kyoto University, Graduate School of Engineering, Osaka, Japan
2Kyoto University Institute for Integrated Radiation and Nuclear Science, Particle Radiation Medical Physics, Osaka, Japan
3Nissan Chemical Corporation, Material research laboratories, Tokyo, Japan
4Kyoto University Institute for Integrated Radiation and Nuclear Science, Particle Radiation Oncology, Osaka, Japan
5Sumitomo Heavy Industries- Ltd, Industrial Equipment Division, Tokyo, Japan

Introduction: In recent years, accelerator-based neutron sources for BNCT have been developing to treat deep-seated tumors using epithermal neutrons, which have higher energy than thermal neutrons. However, when treating superficially located tumors, the boron dose based on the reaction between 10B and thermal neutron is low because epithermal neutrons are not sufficiently thermalized. In this research, a new type of bolus has been developed to enhanced boron dose in the superficial region and an irradiation test using a 3D printed foot phantom was carried out.

Materials and Methods: Hydrogel material with the hydrogen component was used for bolus development. The foot phantom was made from 3D scanning data. The 150 mm * 150 mm bolus with a thickness of 20mm was affixed to the surface of the foot phantom. The phantom was irradiated using the C-BENS (Cyclotron-Based Epithermal Neutron Source) with a 12 cm diameter field size. The thermal neutron flux distribution at the phantom surface was measured using the gold activation method. The results of an irradiation test were compared with the simulation by the treatment planning system.

Results: The thermal neutron flux distribution at the phantom surface was measured. By comparing measured results with the results of treatment planning, the dose distribution on the affected skin surface area was evaluated.

Conclusion: A new type of bolus made from hydrogel for C-BENS has been developed and an irradiation test assuming the treatment for melanoma using a 3D printed foot phantom was carried out. The effect of the bolus was confirmed.

P 131

Boron compounds - pharmaceutical precursors in BNCT method

K. Wójciuk1, A. Bojanowska-Czajka2, E. Chajduk2, M. Dorosz3, G. Wójciuk4, J. Kocik1

1National Centre for Nuclear Research, Nuclear Facilities Operations Department, Otwock, Poland
2Laboratory of Nuclear Analytical Methods, Institute of Nuclear Chemistry and Technology, Warsaw, Poland
3National Centre for Nuclear Research, Radiological Metrology and Biomedical Physics, Nuclear Facilities Operations Department, Otwock, Poland
4Central Forensic Laboratory of the Police, Chemistry Department, Warsaw, Poland

The main goal of the project was to separate an optimal structure out of the group of polycrystalline boron compounds. The crucial feature of the chosen structure was the ability to penetrate into the cell or to be intercepted on its membrane. Fulfilling one of these conditions allows a selection of an appropriate carrier for the boron compound applicable in BNCT.

Within the project, distinct compounds have been studied. Undoubtedly, [C10H10] compounds and BPA have high water solubility, although utilizing of [C12H12] compounds (BSH and TMA/trimethylammonium 1-mercapto-1-carbadodecaborate) lead to an optimal effect in BNCT. BSH is easily water soluble, whereas TMA is not. Notwithstanding, TMA has various advantages over other compounds tested. It is highly stable in solutions of diverse pH and serum. Consequently, as a pharmaceutical it could be administered orally. Moreover, it is stable in all conditions tested and has the adequate lipophilicity. It does not exclude the ability to cross the blood brain barrier. The properties pointed out create an opportunity to enhance bioaccessibility and extending half-life of the compound.

Due to their structure, TMA and BSA are easily coupled with proteins. Additionally, by dint of showing poor EPR signal, TMA is detectable in biological structures. Both compounds crossed cell membrane and located in cytoplasm. Survivability of the cells correlated with cytotoxicity of the compounds tested. The apoptosis pathway is contingent on the concentration of the compound tested, not on incubation period.

We are planning to continue our research and make attempts to bind selected boron compounds with selective carriers.

P 132

Status of Protom synchrotrons for proton therapy

V. Balakin1,2, A. Bazhan1,2, V. Alexandrov2,3, A. Pryanichnikov2,3, A. Shemyakov2,3, A. Shestopalov2,3

1Protom Ltd., Administration, Protvino, Russian Federation
2Lebedev Physical Institute RAS, Physical-Technical Center, Protvino, Russian Federation
3Protom Ltd., Research and Development, Protvino, Russian Federation

Physical-Technical Center of P.N. Lebedev Physical Institute of RAS (PhTC LPI RAS) and Protom ltd. are engaged in development and implantation of synchrotrons for proton therapy complexes into clinical practice. There are two proton therapy complexes (PTC) “Prometheus” on the territory of Russia. That are fully developed and manufactured at PhTC LPI RAS and Protom. Every day patients with head and neck cancer get treatment at PTC “Prometheus”, which is located in the A. Tsyb MRRC, Obnisk, Russia. At the moment both complexes have accumulated more than 3 years of clinical experience. In addition, two facilities are based on the PhTC LPI RAS and Protom synchrotrons in the USA. One operates in the proton center in the McLaren hospital, Flint, and another one is as part of the “Radiance 330” facility in MGH, Boston. Both these centers are equipped with gantry-type setups. The first accelerator complex for proton therapy in Shilat, Israel was launched in August 2019. It is also based on Protom synchrotron. The important and distinctive characteristics of this proton synchrotron: low weight, compact size and low power consumption allow it to be placed in ordinary hospitals without the construction of any special buildings.

This report presents current data on developments of the PhTC LPI RAS and Protom ltd. That are related with proton synchrotrons. In addition, it provides data on the use of PTC “Prometheus” under the clinical conditions.

Distributed under Creative Commons CC-BY. Open Access