## Program Description and Objectives

The Particle Therapy Cooperative Group – North America (PTCOG-NA) was created to enhance collaboration between its members, create a platform for scientific exchange, and develop treatment guidelines, education, and training initiatives for particle therapy. PTCOG-NA will take steps towards establishing a particle therapy clinical trial collaborative group as well as instituting representation in industry technological innovations, involvement in health-care politics (as it pertains to particle beam therapy), and improvement in relationships with other professional societies in radiation oncology.

## Particle Therapy Cooperative Group – North America 2014 Committees

### Executive Committee

President: Eugen Hug

Vice President: Hesham E. Gayer

Secretary: Anita Mahajan

Treasurer: Carl Rossi

### Scientific Committee

William Hartsell

Robert Foote

Chris Hallemeier

Steven Schild

Susan McGovern

Stephanie Perkins

Robert Lustig

R. Chip Nichols

Nancy P. Mendenhall

Hassan Bentefour

Mike Herman

Martin Bues

Wei Liu

Ron Zhu

Heng Li

Jonathan Farr

## C1 CNS/Skull BaseO48: Minimal Acute Toxicities Following High-Dose Proton Therapy for Spinal Tumors

### CDH Proton Center

Background: Proton therapy for spinal malignancies can minimize acute toxicities by using posterior proton beams and positioning distal range anterior to target volume, which limits dose to anterior visceral structures, including esophagus, bowel and bladder. This study seeks to determine whether this dosimetric advantage of proton therapy impacts clinical outcomes in terms of acute toxicities.

Materials and Methods: All patients consecutively treated for a spinal malignancy on the prospective Proton Collaborative Group registry at a single proton center were reviewed. Cases receiving concurrent cranial irradiation or chemotherapy prior to or during proton therapy were excluded. All patients were treated either in the supine or prone position using left and right posterior oblique beam angles with distal edge placed anterior to the planning target volume to ensure at least 95% prescription coverage. Treatments were delivered at 1.8 CGyE/fraction. All patients were evaluated on a weekly basis and acute toxicities were prospectively graded in the electronic medical records using the Common Terminology Criteria for Adverse Events version 4.0.

Results: 19 patients were treated with proton therapy to the spine. Median age was 49 (range 12-74); 3 patients were less than 21 years old. 5 patients had chordoma/chondrosarcoma (dose range 70.4-75.8 CGyE); 1 patient had sarcoma (66.3 CGyE); 5 patients had a WHO grade II ependymoma (54.2-59.6 CGyE); 4 patients had a WHO grade I myxopapillary ependymoma (52.3-59.6 CGyE); 3 patients had benign or malignant peripheral nerve sheath tumor (57.9-59.8 CGyE); and, 1 patient had hemangioma (50.5 CGyE). 16 patients developed grade 1 dermatitis. Grade 2 desquamation was observed in 2 patients (11%), all of whom were treated to 72 CGyE, and grade 3 desquamation was observed in 1 patient (5%), treated to 75.8 CGyE. 2 patients (11%), all of whom were treated to ≥72 CGyE, developed grade 1 diarrhea. Two patients (10%), both of whom were treated to ≥72 CGyE, developed grade 1 urinary frequency. Two patients (10%) reported grade 1 fatigue. Aside from desquamation, no grade ≥2 acute toxicities were observed.

Conclusions: Proton therapy to >50 CGyE for spinal tumors is well tolerated, with no grade ≥2 non-dermatologic acute toxicities and grade 1 acute GI and GU toxicities only in patients treated to ≥72 CGyE. Dermatologic toxicities were observed in all patients, but severe only in patients treated to ≥72 CGyE.

## C2 EyeO49: 3-D treatment planning for proton therapy of ocular melanoma: evaluation of organs at risk

### CDH Proton Center, Rush University, Retina Consultants, CDH Proton Cancer Center

Purpose: Review surgical outcomes for patients undergoing pancreatectomy after 59.40 Cobalt Gray Equivalent (CGE) proton radiotherapy with concomitant capecitabine (1000mg PO BID) for unitially unresectable pancreatic adenocarcinoma.

Methods and Materials: From 4/20/10 to 9/30/13 15 patients with initially unresectable pancreatic cancer were treated with full dose proton therapy with concomitant capecitabine. All patients received 59.40CGE to gross disease, One patient also received 50.40CGE to the high risk nodal targets. There were no treatment interruptions and no chemotherapy dose reductions. 6 of these patients achieved a radiographic response sufficient to justify surgical exploration. 1 was identified as having intraperitoneal dissemination at the time of surgery and the planned pancreatectomy was aborted. 5 patients underwent resection. Procedures included: laparoscopic standard pancreaticoduodenectomy (3); open pyloris sparing pancreaticoduodenectomy (1); and open distal pancreatectomy with irreversible electroporation (IRE) of a pancreatic head mass (1).

Results: Median patient age is 60 years (range 51 to 67). Median duration of surgery was 419 minutes (range 290 to 484); Median estimated blood loss was 850cc (range 300 to 2000); Median ICU stay was 1 day (range 0 to 2); Median hospital stay was 10 days (range 5 to14); 3 patients were readmitted to hospital within 30 days after discharge for: wound infection (1); delayed gastric emptying (1); ischemic gastritis (1). 2 patients underwent R0 resections and demonstrated minimal residual disease in the final pathology specimen. 1 patient – after negative pancreatic head biopsies – underwent IRE followed by distal pancreatectomy with no tumor seen in the specimen. 2 patients underwent R2 resections. Only one patient demonstrated ultimate local progression at the primary site. Median survival for the 5 resected patients is 24 (range 10 to 30) months.

Conclusions: Pancreatic resection for patients with initially unresectable cancers is feasible after high dose proton radiotherapy with a high rate of local control and median survival of 24 months.

## C3 GenitourinaryO33: Depression Scores Pre and Post Treatment for 414 Men with Prostate Cancer using PCG Data

### CDH Proton Center, Proton Collaborative Group

Numerous studies have been done on the prevalence and effect of depression among patients undergoing radiation therapy. Some estimate prevalence as high as 50%. Depression among prostate cancer patients has been claimed as a cause of poor treatment outcomes and is believed to worsen if untreated. This review of self-reported symptoms aims to quantify this belief and either support, or discredit it. This can provide a guide for physicians as they evaluate and prioritize symptom management among this population. A prospective protocol was used to collect patient data related to self-reported levels of depression. Scores were collected pre-treatment at the patient's 3 month follow-up using the EPIC Quality of Life instrument. Pre and post treatment scores were evaluated for 414 men undergoing proton therapy. Pre-treatment depression scores averaged 9.67 (SD = 0.089) with a range of 2 to 11. Post-treatment depression scores averaged 9.88 (SD = 0.086) with a range of 2 to 11. A t-test was performed on the two data sets and results showed that the data was not significantly different (p = 0.07). Results indicate that there was no significant difference in the level of depression reported by patients before and after undergoing proton radiation therapy for prostate cancer. This large scale review shows that the prevalence of depression among prostate patients might not be as severe as thought, or that proton therapy might offer a sample of men unique to other forms of radiation therapy in regards to how frequently they present with depression. It also shows no significant difference in the level of depression reported by the patients pre-treatment and post-treatment, indicated that there was no significant detriment to not having a depression treatment or prevention intervention for these individuals.

## O17: Post-prostatectomy Proton Therapy versus IMRT

### University of Florida Proton Therapy Institute

Purpose: To compare the dose-volume distributions for proton radiotherapy (PT) and intensity-modulated radiotherapy (IMRT) in post-prostatectomy patients in order to determine the optimal treatment technique for normal tissue sparing (NTS).

Materials and Methods: Ten patients were accrued to an Institutional Review Board-approved study. Patients had CT and MRI simulation imaging with empty rectum (ER), rectal saline (RS) and saline-filled rectal balloon (SFRB). The prostate bed was contoured by one physician, according to RTOG guidelines for each patient, with ER, RS and SFRB set-ups using fused imaging data sets. PT treatment plans were then generated for ER, RS and SFRB data sets, and IMRT plans were done for ER and rectal balloon with air-density override (RBAO). Each treatment plan was optimized so that 95% of the PTV received 70 CGE/Gy, while minimizing rectal (RE) and bladder (BL) doses as much as possible. Dose Volume Histogram (DVH) data for five treatment approaches (PT-ER, PT-RS, PT-SFRB, IMRT-ER and IMRT-RBAO) in the ten patients was combined to yield median Dose Volume Histograms (mDVH). These data sets were then analyzed to determine the optimal treatment approach for RE sparing by comparing the relative volume dose to rectum at the following dose levels: 70, 50, and 30 CGE/Gy. Each treatment technique was then evaluated for BL sparing by comparing the relative volume dose to BL at the same dose levels. Finally, each technique was evaluated for integral dose.

Results: DVH values are indicated below. The median integral doses were 83, 81, 81, 140 and 144 Joules for PT-ER, PT-RS, PT-SFRB, IMRT-ER, and IMRT-RBAO, respectively (p <0.0001) (Table 1).

Conclusions: Proton Therapy has a significantly lower median integral dose, RE V50/30 (PT-SFRB) and BL V50/30 (PT-ER). Because of the closer correlation between radiation injury and rectal DVH data, rather than bladder DVH parameters, we prioritize the clinical impact of therectal DVH data and conclude that PT-SFRB appears to be the optimal treatment technique for post-prostatectomy irradiation. The further study of PT in postprostatectomy patients appears to be justified to establish baseline data for toxicity and patient reported outcomes, and to compare with similar data from patients treated with IMRT.

Table 1.

Median DVH Relative Volume Doses

## O18: Interim Analysis: Prospective Trial of Proton Standard-fractionation vs. Hypo-fractionation

### Proton Collaborative Group, CDH Proton Center, ProCure Proton Therapy Center – OKC

Purpose: Identify differences in terms of Quality of Life (QOL), International Prostate Symptom Score (IPSS), or adverse events (AE) among prostate cancer patients treated with either standard-fractionation or hypo-fractionation.

Materials and Methods: Seventy-six patients were randomized to 38GyE in 5 treatments (n=46) vs. 79.2GyE (n=32) over 44 treatments. All patients were defined to have low risk prostate cancer. All patients were treated with proton therapy using fiducial markers and image guidance.

Results: Median follow up for both groups was 12 months with 24 patients reaching follow-up of two years or more. Patient characteristics for both groups were similar with most patients being T1c (84%), all having a Gleason score of 6, and a PSA <10 (median 5.6). Pre-treatment mean IPSS was similar for both groups at 4.5. IPSS increased slightly in the 5-treatment arm to 8.2 at the two year follow-up time point while remaining the same for the 44-treatment arm. Expanded Prostate Index Composite (EPIC) GU symptoms were similar for each group before treatment at a mean of 26. Scores increased slightly in the 5-treatment arm at the two year follow-up , with a mean of 32, while remaining unchanged for the 44-treatment arm at a mean of 26 (p=0.046). EPIC reported GI symptoms and erectile dysfunction scores were similar between the two groups and remained unchanged over two years. No grade 3 or higher AEs were seen in either arm.

Conclusions: After prostate cancer treatment, most AE and QOL changes can be initially identified within the first 2-years. Patients treated with proton therapy in this randomized trial had excellent QOL, persistent low IPSS and no grade 3 or higher adverse events on either arm.

## O30: Study of Semen and Sperm Count following Treatment of Prostate Cancer with Proton Therapy

### University of Florida Proton Therapy Institute

Purpose: Younger men are seeking treatment with proton therapy (PT) as an alternative to prostatectomy in the management of localized prostate cancer. Consequently, there has been more concern regarding changing characteristics of the ejaculate and fertility following treatment with PT. Little data exist regarding ejaculate and sperm counts following any type of radiotherapy treatment among prostate cancer patients. We investigated the impact of treatment with double-scatter PT (dsPT) for prostate cancer on ejaculate and sperm counts.

Materials and Methods: From February 2010 through November 2013, 18 men with low- or intermediate-risk prostate cancer enrolled in an IRB-approved protocol and provided a semen sample prior to dsPT. Men were excluded from enrollment if they had high-risk prostate cancer, received androgen deprivation therapy, or were on alpha-blockers (due to retrograde ejaculation) prior to dsPT. After March 2010, all samples were evaluated at the same laboratory. Sperm counts of <15 million/ml were classified as oligospermia. Seventeen men provided a follow up specimen between 6-12 months following dsPT. One man was placed on alpha-blockers during follow up and was no longer eligible.

Results: At baseline, the median sperm count was 87 million/mL (0.9 to 320) and median volume was 1ml (0.5 to 4.5). Additionally, oligospermia was present in 1 patient and 10 men had low semen volumes <2 ml at baseline. During follow up, the median semen volume was 0.5 ml, including 3 men who were unable to provide sufficient volume of semen for analysis in follow up. Among the 14 men with semen analysis, the median sperm count was 56 million/mL (7 to 470) and only one had oligospermia. The median reduction in sperm count was 14.5 million/mL and in semen volume was 0.5 mL. Additional median changes that were found included median semen pH increasing by 0.5, 5% more immotile sperm, 13% fewer normal sperm morphology, and 10% fewer sperm with rapid progressive motility. Abnormal semen viscosity, liquefication, agglutination were found among some of the samples post treatment.

Conclusions: Men did not have oligospermia 6-12 months following dsPT indicating minimal scatter radiation to the testis during treatment. Changes in semen quantity and consistency may occur due to prostate irradiation, which could impact future fertility and/or sexual activity. These findings should be discussed with men considering dsPT.

## O41: What Men Want: National Survey Data on Prostate Cancer Treatment and Research Decision-Making

### University of Florida Jacksonville, University of Florida Proton Therapy Institute

Background: Little information is available on why men choose proton radiation therapy (RT) for prostate cancer over other treatment options. Additionally, it is unknown whether recruitment for randomized controlled trials (RCTs) comparing proton RT and conventional RT is feasible in prostate cancer. Therefore we conducted a survey on treatment and research decision making.

Materials and Methods: A 28-question survey was compiled using patient and provider input and covered the following domains: (1) Treatment decision-making; (2) Patient preferences; and (3) Research decision-making. The anonymous survey was distributed to men 50+ online using Survey Monkey, and using paper surveys to local clinics.

Results: A total of 1060 responses were received. Respondents were primarily white (91%), employed (54%), married (73%), and 89% had greater than a H.S. education. 78 of the respondents (7%) had a history of prostate cancer. For treatment decision-making, respondents ranked survival, likelihood of recurrence, and QOL as most important in making treatment decisions. Remaining active was ranked more important than sexual function, urinary function , bowel urgency, rectal bleeding, and cost of treatments. With a description of proton RT and conventional RT, those more likely to choose the description of proton RT over conventional RT were white, married, had private insurance, and higher education. Black men rated sexual function and cost as more important than white men in treatment decision-making. For research decision-making, 39% responded they would never participate in an RCT under any circumstance. However, 36% responded that they would agree to participate in an RCT of two different lengths of treatment. Those with higher education were more likely to respond that they would not participate in an RCT. Additionally, black men were more likely to say no to participation in an RCT.

Conclusions: This data suggested that remaining active was most important to men in considering treatment decisions and the impact of various symptoms and side effects, rather than the actual burden of the symptoms themselves. Our data also suggested that patients are hesitant to enroll in an RCT. These data have implications for research trial design, recruitment, participation, and clinical care and reimbursement. There were significant differences in responses for black men, certainly warranting further investigation.

## C4 GastrointestinalO11: Early results of a randomized clinical trial comparing protons to chemoembolization for HCC

### Loma Linda University

The application of proton beam radiotherapy in patients with hepatocellular carcinoma has demonstrated effectiveness in multiple retrospective reports and phase 2 clinical trials. Due to the availability of multiple treatment options for patients with limited stage HCC, there is a need for comparative trials to demonstrate which treatments produce the best outcomes in terms of disease control and toxicity. Following the completion of a phase 2 trial which showed promising results, a phase 3 randomized clinical trial was initiated comparing proton beam to TACE in patients who meet transplant criteria at LLUMC. This presentation will review the experience at LLUMC, describe the role of proton beam in HCC, review the treatment techniques currently in use, and present results of an interval analysis of the ongoing phase 3 clinical trial.

## O35: Proton therapy functional liver avoidance planning using 99mTc-sulfur colloid SPECT/CT

### University of Washington, SCCAPTC

Background: Radiotherapy (RT) for hepatocellular carcinoma (HCC) patients carries a risk for hepatotoxicity and is conventionally planned using anatomic-based information as defined on CT without consideration of potential spatial heterogeneity in liver function. We performed a proton radiotherapy (pRT) planning study to determine the feasibility and dosimetric gains of redistributing RT dose away from functional liver as interpreted by a measure of relative increased uptake of 99mTc-sulfur colloid (SC) on patient SPECT/CT images.

Materials and Methods: Ten HCC patients underwent SC SPECT/CT scans in the treatment position prior to pRT planning and registered to planning CT scans. Pencil beam scanning (PBS) pRT plans were optimized to deliver 37.5-60.0 Gy(RBE) over 5-15 fractions using single field uniform dose planning and robustness optimization. Functional avoidance plans were generated to differentially decrease dose away from functional liver volumes (FLV) defined on SC SPECT/CT images (Fig. 1A). FLVs were defined by a range of percentage threshold (43%-90%) of maximum SC uptake in the liver minus tumor. Relative differences in FLV sparing were compared between functional avoidance plans and conventional plans. Feasibility of functional avoidance plans was defined as a statistically significant negative Spearman's rank correlation (RS) between dose difference and SC uptake. Patient, tumor, and treatment planning characteristics were tested for statistical association to functional avoidance planning feasibility.

Results: Compared to conventional plans, functional avoidance plans reduced FLV mean dose by 23% (range 0 – 56%), and on average, resulted in a 3% FLV dose reduction for every 10% SC uptake increase (Fig. 1B and 2). Functional avoidance planning was feasible in the majority of patients with 60% of patients having RS < -0.5 (p < 0.01, range -1.0 – 0.2) and was particularly effective in 30% of patients (RS < -0.9). Fractionation regimen was the only variable found to be associated with functional avoidance planning feasibility: conventionally fractionated plans were more feasible than hypofractionated plans (p = 0.01).

Conclusions: Differential avoidance of functional liver regions defined by increased relative SC SPECT/CT uptake is feasible for pRT. Further investigation in a larger of cohort of patients may validate the clinical utility of functional avoidance planning for HCC pRT.

## O42: Early progress on a phase I study of Stereotactic Body Proton Therapy (SBPT) for Liver Metastases

### Loma Linda University

Background: Treatment of liver metastases with stereotactic body radiation therapy (SBRT) provides excellent local control but can be limited by liver dose constraints. Proton therapy offers potential for more liver sparing, but limited clinical data is available. A phase I trial was initiated at Loma Linda to assess feasibility identify safe and tolerable dose for stereotactic body proton therapy (SBPT).

Materials and Methods: Patients were enrolled in a standard 3+3 design with a starting dose of 36 GyE in 3 fractions over 1 week. Patients were limited to those having 1–3 non-lymphoma liver metastases that were <5cm and >2cm from GI tract. Early results from the first 3 patients treated at the first dose level are presented.

Results: 3 patients with a total of 4 lesions were enrolled and treated with 12GyE x 3 fractions for each lesion. No laboratory elevation of ALP compared to other transaminases or ascites development occurred within 2-3 months of treatment to indicate any signs or symptoms of radiation induced liver disease. There were 2 local failures with a local control rate of 50% at the first dose level. These occurred at 6 months and 19 months respectively. Both lesions were salvaged with a second course of stereotactic body proton therapy using 12GyE x 3 fractions. No acute toxicities were noted with salvage treatments and both lesions appear to be controlled at latest follow-up.

Conclusion: SBPT appears to be very safe and non-toxic at starting dose of phase I trial and warrants further testing with dose escalation. Current dose level of 16 GyE x 3 is under investigation.

## O51: Feasibility of pancreatectomy after high dose proton therapy for unresectable pancreatic cancer.

### University of Florida Proton Therapy Institute, Department of Surgery - University of Florida Cancer Center at Orlando Health, Department of Surgery - University of Florida Cancer Center - Gainesville, Department of Surgery - Mayo Clinic - Jacksonville, Department of Surgery - Baylor University, Department of Surgery - University of Louisville, Department of Hematology and Oncology - Mayo Clinic - Jacksonville, Division of Medical Oncology - University of Florida Health - Jacksonville

Purpose: Review surgical outcomes for patients undergoing pancreatectomy after 59.40 Cobalt Gray Equivalent (CGE) proton radiotherapy with concomitant capecitabine (1000mg PO BID) for unitially unresectable pancreatic adenocarcinoma.

Methods and Materials: From 4/20/10 to 9/30/13 15 patients with initially unresectable pancreatic cancer were treated with full dose proton therapy with concomitant capecitabine. All patients received 59.40CGE to gross disease, One patient also received 50.40CGE to the high risk nodal targets. There were no treatment interruptions and no chemotherapy dose reductions. 6 of these patients achieved a radiographic response sufficient to justify surgical exploration. 1 was identified as having intraperitoneal dissemination at the time of surgery and the planned pancreatectomy was aborted. 5 patients underwent resection. Procedures included: laparoscopic standard pancreaticoduodenectomy (3); open pyloris sparing pancreaticoduodenectomy (1); and open distal pancreatectomy with irreversible electroporation (IRE) of a pancreatic head mass (1).

Results: Median patient age is 60 years (range 51 to 67). Median duration of surgery was 419 minutes (range 290 to 484); Median estimated blood loss was 850cc (range 300 to 2000); Median ICU stay was 1 day (range 0 to 2); Median hospital stay was 10 days (range 5 to 14); 3 patients were readmitted to hospital within 30 days after discharge for: wound infection (1); delayed gastric emptying (1); ischemic gastritis (1). 2 patients underwent R0 resections and demonstrated minimal residual disease in the final pathology specimen. 1 patient – after negative pancreatic head biopsies – underwent IRE followed by distal pancreatectomy with no tumor seen in the specimen. 2 patients underwent R2 resections. Only one patient demonstrated ultimate local progression at the primary site. Median survival for the 5 resected patients is 24 (range 10 to 30) months.

Conclusions: Pancreatic resection for patients with initially unresectable cancers is feasible after high dose proton radiotherapy with a high rate of local control and median survival of 24 months.

## C5 Head & NeckO33: PBRT Results in Reduced Acute Toxicity Compared with IMRT When Treating the Ipsilateral Neck in HNC

### Memorial Sloan Kettering Cancer Center, ProCure Proton Therapy Center

Purpose/Objectives: The unique properties of PBRT may allow greater normal tissue sparing when compared with IMRT in head and neck cancer. In this study, we compare the dosimetry and acute treatment-related toxicities between 25 consecutive patients treated over a 2-year period with either PBRT or IMRT.

Materials/Methods: Between 01/2011 and 01/2013, 25 consecutive patients underwent unilateral irradiation for major salivary gland cancer or cutaneous squamous cell carcinoma metastases. The availability of PBRT during this period resulted in an immediate shift in practice from IMRT to PBRT, without any change in target delineation. The same tumor and normal tissue dose constraints were used for treatment planning and plan evaluation. Acute toxicities were assessed using the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0.

Results: Fourteen (56.0%) patients were treated with IMRT and 11 (44.0%) with PBRT. The groups were balanced in terms of baseline and treatment characteristics with no significant difference in median age (58.6 vs. 59.4, p=0.218), site of disease (90.9% vs. 92.9% parotid, p=0.859), unresectable disease (9.1% vs. 7.1%, p=0.859), perineural involvement (50.0% vs. 53.8%, p=0.855), neck node irradiation (45.5% vs. 71.4%, p=0.188), concurrent chemotherapy (18.2% vs. 35.7%, p=0.332), or median RT dose (66.0 CGE vs. 66.0 Gy, p=0.530) when comparing PBRT and IMRT patients.

There were no significant differences in the target volume size (parotid and ipsilateral neck 348.8 vs. 306.2 cc, p=0.994; parotid only 78.3 vs. 140.0 cc, p=0.592) or planning target volume max dose to 0.15 cm3 (109.5% vs. 109.7%, p=0.630) between PBRT and IMRT plans, respectively. PTV V95 > 95% was met for all cases. A lower median maximum brainstem (0.1 vs. 31.9 Gy, p<0.001) and spinal cord (1.4 vs. 36.9 Gy, p<0.001) dose and mean oral cavity (0.3 vs. 20.0 Gy, p<0.001), contralateral parotid (0.0 vs. 1.3 Gy, p=0.013), and contralateral submandibular (0.0 vs. 4.0 Gy, p=0.002) dose was noted for PBRT. PBRT resulted in a significantly lower mean larynx dose in patients undergoing ipsilateral neck irradiation, 8.8 vs. 26.2 Gy, p=0.012.

PBRT had significantly lower rates of grade 2 mucositis (9.1% vs. 57.1%, p=0.027), dysgeusia (9.1% vs. 78.6%, p=0.002), and nausea (9.1% vs. 64.3%, p=0.012). There was no difference in the rate of grade 3 dermatitis (27.3% vs. 35.7%, p=0.650) between PBRT and IMRT.

Conclusions: The unique properties of PBRT allow greater normal organ sparing without sacrificing target coverage when irradiating the unilateral neck. This dosimetric advantage translated into significantly lower rates of acute treatment-related toxicity including mucositis, nausea, and fatigue. Long-term follow-up is needed to determine whether this translates into fewer late effects.

## O39: Adaptive treatment for oropharyngeal cancer planned with proton pencil beam scanning

### University of Pennsylvania, the Christie NHS Foundation Trust, National Cancer Centre Singapore

Background: To investigate the requirement and optimal timing for adaptive re-planning in definitive treatment of locally advanced oropharyngeal cancer with proton pencil beam scanning (PBS).

Materials and Methods: Ten consecutive patients with locally advanced oropharyngeal cancer who had previously received simultaneous integrated boost IMRT (70Gy in 35 daily fractions over 7 weeks to planning target volume, PTV1) and synchronous chemotherapy with weekly kilovoltage cone beam CT (CBCT) verification imaging were included. The 10 CT scans were re-planned using RapidArc IMRT and proton PBS. Deformable image registration was used to deform the planning CT, target volumes and organs at risk contours onto each weekly CBCT. Target and OARs volumes were reviewed and modified by a radiation oncologist. The initial RapidArc IMRT and PBS treatment plans were forward calculated onto each corrected CBCT scan and dose-volume histograms were produced for target and OARs volumes.

Results: Proton PBS compared with RapidArc IMRT achieved lower mean doses to the contralateral parotid gland (14.8Gy versus 20.6Gy, p<0.05) and oral cavity (31.5Gy versus 43.0Gy, p<0.001). For proton PBS, mean delivered doses to several OARs significantly increased from weeks 3-4 of treatment; this was not observed in RapidArc plans. The respective overall increases in mean delivered doses (average weeks 1-7, compared with week 1) to the pharyngeal constrictor muscles, larynx and oral cavity were: 1.0Gy (SD ±1.3), 3.3Gy (SD ±3.3) and 1.7Gy (SD ±1.9).

Conclusions: Proton PBS was more sensitive to per-treatment anatomical changes, which resulted in modest increases in delivered doses to several OARs. A novel finding of this study was that in order to optimise OAR-sparing for locally advanced oropharyngeal cancer, routine adaptive proton re-planning prior to week 4 of treatment is advised. This information is especially useful given the current technical and practical challenges in implementation of adaptive proton therapy.

## O52: Proton beam re-irradiation for recurrent head and neck cancer: toxicity and preliminary outcomes

### Memorial Sloan Kettering Cancer Center, Mount Sinai Beth Israel, Montefiore Medical Center, ProCure Proton Therapy Center

Purpose: Locoregional control has been achieved with the use of re-irradiation (re-RT) using photon beams for recurrent head and neck cancer. However, the acute and late complications are of considerable concern. Since proton beams have negligible exit dose, proton beam radiation therapy (PBRT) should provide greater sparing of normal and previously irradiated tissue in comparison to photon beam, allowing for reduced acute and late toxicities. The purpose of this study was to report on our PBRT re-RT experience in patients with recurrent HNC.

Methods and Materials: A retrospective analysis of an on-going prospective data registry was performed; patients treated with PBRT for recurrent HNC that received at least one prior course of definitive head and neck RT were reviewed. Acute toxicity was assessed using the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0 and late toxicity (>90 days after RT completion) by the Radiation Therapy Oncology Group late radiation morbidity scoring system. Overall survival (OS) and local control (LC) rates were analyzed by Kaplan-Meier method.

Results: Between 2012 and 2014, 56 patients with recurrent HNC underwent re-RT with PBRT. The median length of time between prior RT and PBRT was 49.4 months (interquartile range 13.5 – 100.6 months). There were 14 patients (25%) who had two prior courses of RT, while four (7.1%) had three or more prior courses. Median follow-up length after PBRT was 4.0 months (interquartile range 2.3 – 5.8 months). The actuarial 6-month OS and LC rates were 98.2% and 80.3%, respectively.

Acute Grade 2 mucositis was experienced in 29%, dysphagia in 21%, and dysgeusia in 18%. Two patients experienced Grade 3 mucositis.

There was one patient who developed a sinocutaneous fistula during PBRT. A second patient experienced tumor bleeding during treatment requiring embolization; a subsequent pharyngocutaneous fistula developed one month after RT that resolved with conservative management. There were two deaths during or within three months of PBRT due to disease progression.

Data for late toxicity was available for 20 out of 33 evaluable patients. One patient developed a grade 4 chronic neck wound and required hyperbaric oxygen treatments, although during surgery prior to PBRT, a fresh skin flap was not placed over the re-RT fields. A second patient developed edema of the skull base that was managed with steroids.

Conclusions: Overall, re-RT with PBRT was well tolerated, with minimal severe acute and late toxicity. The preliminary treatment outcome result is also very encouraging. Further, due to the sparing of normal tissue resulting from the finite range of protons, head and neck cancer patients who had multiple prior courses of RT are also candidates. A randomized prospective trial is planned to better assess the potential benefit of re-RT using PBRT versus photon beam for recurrent HNC that has received prior radiotherapy.

## C6 PediatricO20: Referral Patterns and Utilization of Proton Therapy among Children in Florida

### University of Florida, Nemour's Children's Clinic

Purpose: With increasing evidence supporting the use of proton therapy in pediatric tumors, the number of children in the US treated with protons increased by 33% between 2010-2012. Currently the 3rd largest state in the US, Florida is served by a single proton therapy center and thus provides a unique setting to examine resource utilization and factors influencing referral.

Materials and Methods: 5,085 patients age 0-21 were registered in the Florida Association of Pediatric Tumor Programs (FAPTP) state cancer database between January 2007 and August 2013, which contains publicly available data on age, ethnicity, insurance status, as well as histology. 3,119 of these patients had diagnoses for which proton therapy would not be indicated and 28 had incomplete information, leaving 1938 patients with diagnoses for which proton therapy might be indicated. Estimates of the percentage of patients medically eligible for proton therapy were derived from the Central Brain Tumor Registry of the United States and published literature, then compared with the number of patients over the same time period actually referred to the University of Florida Proton Therapy Institute (UFPTI) for treatment.

Results: Over the past 5 years, only 181 of the 747 (24%) children in Florida registered in FAPTP and estimated to be medically eligible for proton therapy were referred to UFPTI (Table 1). All medically eligible patients were accepted for treatment. In the FAPTP estimate, median age was 10 years, 55% were male, and 70% were white, whereas in the proton therapy referrals, median age was 8 years, 50% were male, and 80% were white. In the FAPTP estimate, 39% of patients had commercial insurance, 34% were enrolled in Medicaid, and 27% were enrolled in CMS or uninsured compared with 63%, 43%, and 8%, respectively, in the proton therapy referrals.

Table 1.

Conclusions: This is the first state-wide study of its type and suggests that proton therapy is underutilized in Florida's pediatric patients. In particular, children who lack commercial insurance appear to be less likely to be referred for proton therapy, likely reflecting socioeconomic barriers to travel and relocation.

## O29: Pediatric Proton Therapy: Patterns of Care in 2013 across the United States

### University of Florida, Scripps Proton Therapy Center

Purpose/Objectives: Children are particularly prone to the late side effects of normal tissue irradiation. For this reason, pediatric solid tumors are a commonly cited indication for proton therapy worldwide. Since 2010, United States proton centers have contributed to a voluntary survey assessing patterns of care in the treatment of pediatric patients. This study reports the results from the 2013 survey.

Materials and Methods: A survey was developed and distributed to all clinical proton therapy facilities in the United States that were in operation throughout 2013. Anonymized patient information including age range, tumor site, diagnosis and other key data points were collected for each patient 18 years old or less treated between January 1, 2013 and December 31, 2013.

Results: There was a 100% response rate from the 11 US proton therapy centers operating in 2013. All facilities treated at least 14 pediatric patients (range, 14-157). A total of 722 pediatric patients were treated, up from 694, 613, and 465 patients in 2012, 2011, and 2010, respectively. The majority of patients (65%) were treated at four academic centers affiliated with large pediatric hospitals, up from 53% in 2012. Fifty-six percent of patients were under 10 years of age and 42% of patients required anesthesia, similar to past years. The six most common tumor diagnoses treated were medulloblastoma (n=102), ependymoma (n=88), low-grade glioma (n=85), rhabdomyosarcoma (n=85), craniopharyngioma (n = 57), and Ewing sarcoma (n= 41). Twenty-two percent of patients originated in countries outside the United States. Patient ages, tumor types, use of anesthesia, and percent of non-US patients were similar to 2011 and 2012 surveys. Across all centers, 186 patients (26%) were enrolled on multi-institutional registry studies. Overall, 138 patients participated in multi-institutional therapeutic trials; however, 109 (79%) of these patients were enrolled at 3 centers.

Conclusions: The total number of children treated at US proton centers continues to increase, rising 36% since 2010, consistent with the international perception that pediatric patients derive a relative benefit from this rare technology. The average patient continues to be a child <10 years old with a curable brain tumor or axial sarcoma. As the number of proton centers increases, broader engagement in collaborative research is necessary, particularly in the form of therapeutic studies.

## O55: Proton RT for Pediatric CNS Tumor Survivors: Correlating Dose to Structures with Neurocognition

### Massachusetts General Hospital

Background: Photon RT can adversely affect neurocognition in childhood brain tumor survivors. The supratentorial brain, cerebellum, left temporal lobe, and hippocampi are implicated as most sensitive to adverse neurocognitive late effects. Here, we investigate an association of several dosimetric indices of proton RT (PRT) dose to normal neuroanatomical structures with changes in long-term neurocognitive performance.

Materials and Methods: 114 PRT-treated pediatric brain tumor patients (median age = 7 yrs) completed both baseline and follow-up (>1 year) neuropsychological evaluations. DVHs of the temporal lobes (TL), hippocampi (HC), hypothalamus, brainstem, cerebellum, infratentorial brain, left (L) and right (R) supratentorial hemispheres (SH), and whole brain (WB) were used to calculate V20, V30, V40, V50, Dmax, Dmean, EUD-3, EUD-7, and EUD-15. For each patient, a temporal change in neurocognitive test performance was calculated using linear regression. The association of dosimetric indices of normal structures with the slope of changes in neurocognitive tests was evaluated using linear regression models.

Results: In preliminary analysis, structures that consistently showed a statistically significant association with neurocognitive performance include the L/R TLs, LHC, and the LSH. For example, Dmax, Dmean, and all EUDs for LTL, LHC, and LSH correlated with decline in FSIQ (see table). Additionally, Dmax and Dmean of the LTL (Pmax = 0.004, Pmean = 0.04) and LHC (Pmax = 0.01, Pmean = 0.03), as well as older age at Dx (P = 0.02), correlated significantly with changes in performance on the Peabody Picture Vocabulary Test (PPVT). Changes in the Children's Memory Scale (CMS) subtests also correlated with dosimetric indices to the L/R TLs and LHC: Dot Locations Long Delay with LTL (Dmax: P = 0.04, EUD-15: P = 0.04), LHC (Dmax: P = 0.03, EUD-15: P = 0.03); Sequences with RTL (V30: P = 0.03), LTL (V30, P = 0.03), LHC (V40, P = 0.04); Numbers with LTL (Dmax: P = 0.02, EUD-7: P = 0.04; EUD-15: P = 0.02). No significant correlation was observed between FSIQ and dose to WB. Correlations with VIQ, PIQ, processing speed, and motor function will be presented.

Conclusions: We have confirmed analogous photon studies showing the connection of the LTL and hippocampi to neurocognitive performance, specifically here in FSIQ, consolidation of memory, and receptive vocabulary. These findings have implications for optimizing treatment planning for minimal neurocognitive consequences.

## C8 ThoracicO2: Clinical implementation of intensity-modulated proton therapy in thoracic malignancies

### University of Texas M.D. Anderson Cancer Center

Background: Intensity-modulated proton therapy (IMPT) can improve dose conformality and better spare normal tissue over passive scattering techniques, but range uncertainties complicate its use, particularly for moving targets. We report our early experience with IMPT for thoracic malignancies in terms of motion analysis and management, plan optimization and robustness, and quality assurance.

Methods: 55 consecutive patients with primary/recurrent lung/mediastinal cancers received IMPT to a median 66 Gy(RBE). All patients were able to undergo definitive radiotherapy. IMPT was used when the treating physician judged that IMPT conferred a dosimetric advantage; all patients had minimal tumor motion (<5 mm) and individualized tumor-motion dose-uncertainty analysis and 4D CT-based treatment simulation and motion analysis. Plan robustness was optimized by using a worst-case scenario method. All patients had 4D CT re-simulation during treatment.

Results: IMPT produced lower mean lung dose (MLD), lung V5 and V20, heart V40, and esophageal V60 than IMRT (p<0.05) and lower MLD, lung V20, and esophageal V60 than passive scattering proton therapy (PSPT) (p<0.05). D5 to the gross tumor volume and clinical target volume was higher with IMPT than with intensity-modulated radiation therapy (IMRT) or PSPT (p<0.05). All cases were analyzed for beam-angle-specific motion, water equivalent thickness, and robustness. Beam angles were chosen to minimize the effect of respiratory motion and avoid previously treated regions, and the maximum deviation from the nominal dose-volume histogram values was kept at <5% for the target dose and met the normal tissue constraints under a worst-case scenario. Patient-specific quality assurance measurements showed that a median 99% (range 95%–100%) of the pixels met the 3% dose/3 mm distance criteria for the γ index. Adaptive re-planning was used for 26.5% patients. The most common site of failure was distant metastasis (21%), followed by local-regional recurrence (12%). 15% developed grade 2 or 3 skin reaction, 18% developed grade 2 or 3 esophagitis, and 15% developed grade 2 or 3 dyspnea. No patients manifested grade 4 or 5 toxicity of any type.

Conclusions: IMPT using 4D CT-based planning, motion management, and optimization was successfully implemented clinically and met our quality assurance parameters. IMPT is safe in thoracic cancers with tumor motion <5 mm based on physics analysis using our current approach.

Table 1.

Motion analysis, robustness analysis, and quality assurance results for all patients.

### University of Florida Proton Therapy Institute, Shands Hospital Division of Pulmonology, Shands Hospital Division of Medical Oncology, Shands Hospital Division of Interventional Radiology

Purpose: Review disease control, survival, and toxicity outcomes for patients undergoing thoracic reirradiation for recurrent NSCLC with proton therapy (PT).

Methods and Materials: From 5/25/10 to 5/8/14, 12 patients previously treated with definitive thoracic radiotherapy for NSCLC underwent reirradiation for recurrent disease [with (3) or without (9) concurrent chemotherapy]. Patients received PT as part of their initial radiotherapy (10), retreatment (8), or both (6). Patients not treated with PT for reirradiation were treated with SBXRT (5). Reirradiation targets defined relative to the initial treatment volume included 4 in field, 5 marginal and 3 out of field. Retreatment targeted parenchyma only lesions in 8 patients, nodal only in 3, and parenchyma and lymph nodes in 1. Initial radiotherapy was delivered with standard fractionation (6) or hypofractionation (6). Median dose for the initial treatment was 70Gy (range: 59.4-74Gy) in 1.8-2Gy/fraction or 54Gy (range: 30-60) in 6-12Gy/fx. Reirradiation was delivered using standard fractionation in 7 patients [median dose 60Gy (range: 50-70) in 2Gy/fraction] or hypofractionation in 5 patients [median dose 50Gy (range: 32-60) in 6-10Gy/fx]. One patient received reirradiation to two sites; 60CGE in 2Gy/fx (in field) and 50Gy in 10Gy/fx (out of field). Prospectively obtained toxicity was scored using CTCAE v4. Survival and disease control outcomes were estimated using Kaplan-Meier.

Results: Median follow-up for survivors is 13 (range: 6-44) months. Median age at reirradiation was 74 (range: 63-84) years. Median interval between radiation courses was 13 (range: 6-62) months. 2-year actuarial OS, PFS and MFS were 54%, 47%, and 47%, respectively. 2-year local control was 63%. Four patients (33%) experienced acute grade 2 side effects (esophagitis x2, dyspepsia x2, cough x1). There were no acute grade 3-4 side effects. One patient experienced a fatal hemorrhage within one month of reirradiation. This patient received PT to 60CGE for an infield hilar recurrence 5 years after receiving 70Gy. It is unclear if this patient's death was due to radiation toxicity or tumor necrosis/regression through a major vessel. There were no late ≥grade 2 side effects.

Conclusions: Preliminary data for proton therapy as part of reirradiation for recurrent NSCLC show encouraging

## O57: What Women Want: Results of National Survey on Breast Cancer Treatment and Research Decision Making

### University of Florida Proton Therapy Institute, University of Florida

Background: Data are sparse on patient preferences in treatment and research decision-making for breast cancer radiation therapy (RT). We surveyed American women to understand what influences treatment and research decision-making.

Materials and Methods: An anonymous survey was distributed online with the following domains: decision-making regarding treatment choices and research participation, patient preferences, and concerns about side effects. Parameters were set to capture women aged ≥45 years.

Results: 857 responses were received. Respondents were primarily white (90.7%), aged ≥60 (52.2%), and private insurance (62%). 7% were breast cancer survivors and 93% had no history of breast cancer. Quality of life, survival, and freedom from recurrence were the most important factors influencing treatment decisions, followed by avoiding heart and lung damage. Women ranked “remaining active” and “avoiding heart and lung damage” of greater importance than cost, convenience, or cosmesis. “Living with heart problems” was rated more bothersome than lymphedema, skin changes, and decreased activity. 21.8% responded they would not participate in a randomized control trial (RCT) under any circumstances. A written description of history and side effects of conventional RT (CRT) vs. proton RT (PT) was provided; 20.5% of women chose CRT and 71.9% PT. When a dosimetric picture was added to the description, only 8.2% of women chose CRT. 9.6% of women said they would agree to randomization in a study of 2 different types of RT. With a description and image of CRT vs. PT dose distributions, 7.8% of women responded yes to randomization and 43.8% responded no, while 40% were undecided. In comparison, 29.3% were willing to be randomized to 19 vs. 30 RT treatments and 18.7% responded no, with 40% undecided. Women with higher education were more likely to agree to a RCT. Married women were less likely to agree to RCT participation of CRT vs. PT.

Conclusion: In this national survey population, 21.8% would not participate in a RCT. They were less likely to consent to randomization between CRT and PT than to a 2-dose fractionation schema. Dosimetric images were influential in choosing CRT or PT, and therefore, should be included in the informed consent process for trials comparing CRT and PT. While RCTs represent the highest level of evidence, many patients are not inclined to participate in RCT; specifically, < 10% of American women appeared willing to participate in a RCT comparing PT and CRT.

## C9 Other (Breast)O21: Proton beam partial breast irradiation for early stage breast cancer: A not-so-costly alternative

### University of Texas M.D. Anderson Cancer Center

Background: Initial reports of proton based accelerated partial breast irradiation (P – APBI) show great promise as an option for the treatment of early stage breast cancer [1-4]. Proton therapy's reputation as an expensive treatment has led to the presumption that P – APBI is costly and could not represent a good value to the health care system. Our objective is to determine the cost of P – APBI relative to other established radiation treatments for early stage breast cancer.

Materials and Methods: We have compared Medicare allowable charges (MACs) across breast irradiation techniques and fractionations used nationally for the treatment of early stage breast cancer utilizing the 2014 standard Medicare payments for both professional and technical charges. Information on fractionation for each treatment is found in Table 1. MACs for each CPT code listed were obtained and totals for each treatment (to the nearest dollar) were calculated considering the frequency of the codes listed. Other related costs (e.g. prophylactic antibiotics for brachytherapy insertions) were not considered for analysis.

Results: Total, technical and professional MACs are shown on Figure 1. These include all CPT codes directly related to the delivery of radiation, treatment devices and special services. The exact dollar amounts are shown on Table 2. Total MACs range from $6,771 to$19,599. IMRT using standard fractionation to the breast has the highest allowable charges, exceeding the mean ($12,784) by over 50%. Total MACs for P-APBI ($13,833) are within 5.2% of WBI using standard fractionation followed by a sequential electron boost ($13,149). The treatment schedule with the lowest MACs is 3D-conformal APBI with photons/electrons ($6,771). Most of the variations were due to the technical components of treatment (range, $5,157 -$16,617), while professional MACs were within a narrower range (range, $1,614 -$3,245).

Conclusions: MACs for APBI with proton beam therapy are comparable to other APBI and WBI techniques. It has lower payer costs than WBI with IMRT and APBI with multi-lumen brachytherapy, and very similar ones to 6-weeks of field-in-field WBI with a sequential boost. Additionally, it is likely that the personal financial burden for patients is also lowered with a one-week regimen compared to the alternatives. As clinical experience is gained through clinical trials, payer costs for P – APBI should not be an impediment when considering treating patients with this emerging modality.

## O61: Outcomes and early toxicity in patients treated with proton therapy for breast cancer

### Memorial Sloan Kettering Cancer Center, Procure Proton Therapy Center, New Jersey

Purpose: Multiple dosimetric studies have shown that proton therapy (PT) can reduce exposure to the heart, lungs and contralateral breast for breast cancer patients undergoing radiotherapy, but clinical experience is limited. Here we report dosimetry and early toxicity data in a cohort of patients with breast cancer treated with postoperative proton radiation therapy.

Materials and Methods: From 3/2013 to 4/2014, 30 patients with non-metastatic breast cancer and no history of prior radiation were treated with proton therapy at a single proton center. There were 26 patients treated in the post-mastectomy setting and 4 patients in the post lumpectomy setting. Patient characteristics and dosimetry data were obtained through chart review. Proton therapy was delivered with uniform scanning beams using a four field technique. In general, 45 Gy (RBE) was delivered to the chest wall, regional nodes and supraclavicular fossa followed by a cone down of an additional 5.4 Gy (RBE) to the chest wall. Patients were seen weekly while on treatment and at 1 month after RT completion. Toxicity was assessed scored using CTCAE v4.0 with digital photographs taken at baseline and at each visit. Frequencies of toxicities were tabulated.

Results: Median dose delivered was 50.4 Gy (RBE) in 5 weeks. Target volumes included the breast/chest wall and regional lymph nodes including the IMNs (in 93%). No patients required a treatment break. There were no Grade 3 adverse events. Grade 2 dermatitis occurred in 20 patients (66.7%) with 8 patients (26.7%) experiencing moist desquamation. Grade 2 esophagitis occurred in 10 patients (33.3%). The median PTV V95 was 96.43% (range 79.39 – 99.60). The median heart V20, V5, and mean heart dose were 0.98% (0 – 6.0), 4.78% (0 – 14.40), and 0.88 Gy (RBE) (0.01 – 3.20), respectively. The median V20 and V5 of the ipsilateral lung were 16.50% (6.1 – 30.3) and 34.35% (22.5 – 53.8), respectively. The median contralateral lung V5 was 0.34% (0 – 5.30). The median maximal point dose to the esophagus was 45.65 Gy (RBE) (0 – 65.4). The median contralateral breast mean dose was 0.29 Gy (RBE) (0.03 – 3.50) and the median V5 was 1.46% (0 – 9.90).

Conclusions: Postoperative proton therapy after surgery is well tolerated with acceptable rates of skin toxicity. Proton therapy favorably spares normal tissue without compromising target coverage. Further follow up is necessary to assess for clinical outcomes and cardiopulmonary toxicities.

## C9 Other (General)O19: The Proton Collaborative Group's Prospective Chart Review Registry- Comments from the First 5 Years

### Proton Collaborative Group

The Proton Collaborative Group's (PCG) Evaluation Tracking Project: A Prospective Chart Review of Patients Treated with Proton Therapy (REG001-09) allows for the collection and analysis of data to evaluate the disease and treatment process for proton therapy patients. This collaborative project is currently being run at six centers and includes information on close to 3,500 patients representing a wide array of disease sites. The registry has been enrolling patients continuously since July, 2009 at a rate of 58 subjects per month. When looking at the past 6 months, this rate jumps to 117 subjects per month indicating heavy growth of participation from new and existing sites. Database quality assurance involves a process of automated queries in addition to monthly manual monitoring. Monthly data management includes source document verification for 100% of any grade 3 or higher adverse events, deaths, or treatment failures. Staff also runs checks for completeness of data and to ensure data falls within expected ranges. PCG members can request the use of the data, which is owned by the contributing physicians, for independent research. To date, 58 requests for use of this data have been approved. Resulting research has been presented at numerous conferences including ASTRO, AAPM, PTCOG, NASBS, ONS, ASCO, SIOP, ISPNO and AAMD. Data has also been used for publication in multiple journals including International Journal of Pharmacy and Technology, Journal of Applied Clinical Medical Physics, Practical Radiation Oncology, International Journal of Radiation Oncology*Biology*Physics, Radiotherapy and Oncology, International Journal of Cancer Therapy and Oncology, International Journal of Particle Therapy, and Red Journal. This project is a robust representation of collaborative research involving a devoted effort from both site and sponsor. Data is collected and managed in order to provide useful, quality information on the outcomes and practices of proton therapy.

## C9 Other (Lymphoma)O7: Dose to abdominal organs with proton therapy in diaphragmatic and infradiaphragmatic Hodgkin lymphoma

### University of Florida

Background: Several dosimetric studies have been conducted demonstrating clinically significant radiation dose reduction to thoracic organs at risk (OARs), such as the heart, lungs, and breasts, with the use of proton therapy compared with either 3D conformal radiation therapy (3DCRT) or IMRT among patients with Hodgkin lymphoma (HL). However, little data exists regarding dose reduction when using proton therapy to OARs within the abdomen. Recent reports have surfaced, demonstrating that HL survivors receiving abdominal radiotherapy as part of their treatment are at significantly higher risk of secondary gastric cancer and other late effects. This study investigates the dose reduction to the OARs in the abdomen with the use of proton therapy.

Methods: From June 2008 through December 2013, 12 patients with classical HL involving diaphragmatic or infradiaphragmatic regions were evaluated for treatment with radiotherapy and enrolled on an IRB-approved outcomes tracking protocol. Patients included 6 pediatric patients with stage III/IVS, 4 adults with stage II bulky mediastinal and diaphragmatic nodal involvement, and 2 patients with relapse involving the paraaortics (n=1) or spleen (n=1). All patients underwent treatment with proton therapy as consolidation. Comparative 3DCRT and IMRT plans were developed to evaluate differences in dose to the OARs. Pediatric patients were generally treated to 21 Gy (RBE), while adults were treated to 30-39.6 Gy (RBE). At a minimum, PTV D95% was greater than 95%. Although OAR dose limits were set for lung and heart, no OAR dose restrictions were enforced for abdominal OARs.

Results: Among the 12 patients, the median doses to the stomach with 3DCRT, IMRT, and proton therapy were 21, 14 and 6 Gy, respectively. The absolute dose reductions to the stomach when using proton therapy compared with 3DCRT and IMRT were 13 Gy (p = 0.0022) and 8 Gy (p = 0.0022). The median dose reductions with proton therapy compared with 3DCRT and IMRT were 4 Gy (p = 0.0022) and 6 Gy (p = 0.0022) for liver, 4 Gy (p = 0.0022) and 4 Gy (p = 0.0058) for pancreas, 5 Gy (p = 0.0051) and 4 Gy (p = 0.0051) for bowel, 4 (p = 0.0077) and 1 Gy (p = 0.0382) for left kidney, and 1 Gy (p = 0.0077) and 3 Gy (p = 0.0077) for right kidney.

Conclusions: Proton therapy reduces the dose to the stomach, liver, pancreas, and bowel compared with either 3DCRT or IMRT in patients with HL requiring abdominal radiotherapy. These dose reductions are expected to translate into lower risks of secondary cancers and may reduce the risk of hypertension (kidneys), diabetes (pancreas), and other late radiation toxicities in survivors compared with photon radiation.

## O8: Proton Therapy in the Management of Non-Hodgkin Lymphoma

### University of Florida

Background: Proton therapy is a highly conformal type of radiation therapy that can target tumor while sparing dose to surrounding normal tissues. Although several studies have evaluated the use of proton therapy in Hodgkin lymphoma, there is little data regarding outcomes of patients treated with proton therapy for non-Hodgkin lymphoma. This study reviews a single institution's experience managing patients with non-Hodgkin B-cell lymphoma treated with proton therapy.

Materials and Methods: Nine patients with non-Hodgkin B-cell lymphoma were treated with proton therapy from 2008 to 2013 on an IRB approved outcomes tracking protocol. Patients with indolent orbital lymphoma (n=4), either low grade follicular or MALT lymphoma, were treated definitively to doses of 24 to 30.6 CGE in 1.5 to 1.8 CGE fractions. Patients with primary mediastinal lymphoma were treated with consolidative doses from 30.6 to 41.4 in 1.5 to 1.8 CGE daily fractions following six cycles of R-CHOP chemotherapy (n=2) or after high-dose chemotherapy and autologous stem cell transplant (n=1). Patients with plasmablastic lymphoma included one with involvement of the paranasal sinuses and base of skull treated to 60 CGE in 1.2 CGE fractions twice daily following R-CHOP x6 cycles and one with plasmablastic lymphoma of the stomach who received 36 CGE in 2 CGE daily fractions following hyper-CVAD chemotherapy.

Results: The median follow up was 2 years. The four patients with orbital lymphoma all had a complete response to definitive proton therapy with no instances of local recurrence. Acute toxicities included grade 1-2 dermatitis, grade 1 fatigue, and grade 1-2 headache. Late toxicities were development of cataracts in two patients and grade 1 epiphora in one patient.

All three patients with mediastinal diffuse large B-cell lymphoma had an excellent response to consolidative proton therapy and no evidence of recurrent disease during follow up. No acute or late grade 2 toxicities occurred from proton therapy among these patients.

The patient with plasmablastic lymphoma of the stomach is without evidence of disease 18 months after treatment. The patient with plasmablastic lymphoma of the paranasal sinuses and base of skull had stable disease 8 months out, but then relapsed and died approximately 3 years later. This patient developed grade 2 dermatitis during treatment.

Conclusions: Proton therapy is a feasible and effective treatment for non-Hodgkin lymphoma. Early outcome results are favorable. Longer follow up and more patients are needed to confirm our findings.

## P3 Moving TargetsO4: Minimizing dose uncertainty for SSPT of moving tumors by optimizing the spot delivery sequence

### University of Texas M.D. Anderson Cancer Center

Background: Intensity-modulated proton therapy (IMPT) can result in a lower radiation dose to normal tissue than can intensity-modulated photon radiotherapy (IMRT), and it has been implemented for selected lung cancer patients. However, respiratory motion-induced dose uncertainty, which remains a major concern in the radiotherapy of lung and liver cancer patients, limits the utility of IMPT. Strategies such as re-scanning and tumor tracking have been proposed and studied; however, re-scanning can result in an unacceptably long delivery time, and tracking is not yet clinically available. We developed an analytical model of the spot-scanning beam delivery system and we proposed a novel delivery strategy for spot-scanning proton beam therapy to minimize the motion-induced dose uncertainty.

Materials and Methods: The effective delivery time for each spot position in the treatment plan was calculated on the basis of the parameters of the delivery system, including the time required for each spot, spot size, and energy. The dose uncertainty was then calculated as a function of the effective delivery time. The spot delivery sequence was optimized to maximize the effective delivery time and thus minimize the dose uncertainty. Two-dimensional measurements with a detector array on a 1D moving platform were obtained to validate the calculated results.

Results: We performed 346 2D measurements on a moving platform for different delivery sequences of a single-layer uniform pattern and patient treatment field. The measured dose uncertainty is a strong function of the delivery sequence: the worst delivery sequence results in a maximum absolute dose error of up to >90%, while the optimized delivery sequence results in a maximum absolute dose error of <5% for a single layer and <10% for a patient field with heterogeneity. The effective delivery time vs. the measured dose uncertainty follows the analytical formula.

Conclusions: With an optimized delivery sequence, dose uncertainty due to motion can be minimized in spot-scanning proton therapy, making IMPT a feasible treatment option for lung and liver cancer patients.

## P4 Quality Assurance (Device)O22: Characterization of a Quality Assurance device for hadron beam 2D profile and position measurements

### DE.TEC.TOR. Devices and Technologies Torino srl, CNAO Foundation

A novel QA monitor for hadron beam intensity, position, size and shape evaluation (miniQ-STRIP), was tested with both cyclotron and synchrotron proton beams. The two sets of data were acquired at the 230 MeV proton cyclotron IBA center (U.O. Protonterapia - Azienda Provinciale per i Servizi Sanitari - APSS) in Trento (Italy) and the 250 MeV proton synchrotron beamline at the Centro Nazionale Adronterapia Oncologica (CNAO), in Pavia (Italy). The device is conceived as a stack of 2 integral and 2 strips (in X and Y directions) ionization chambers with sensitive area 12.7x12.7 cm2 and an air gap of 2 mm. The applied voltage is -400 V. The repeatability of the integral chambers output was tested by keeping the beam at a fixed position and at different intensities and energies, showing standard deviations < 1 %. The spot barycenter was computed as weighted sum of the counts in the strips (1 mm wide with 0.1 mm of electrical insulation), in both X and Y directions. The maximum variation of the barycenter position with respect to its mean value was <0.1mm The uniformity of the integral ionization chambers was tested by moving the beam along a grid of equally spaced positions in both directions, showing a standard deviation <2.1% in the evaluation of the spot intensity. The dose rate independency was finally tested, scoring a standard deviation <0.1% when the current at the exit of the accelerator was varying of 20%. The device showed optimal performances with both cyclotron and synchrotron beams and can be used to speed routine dosimetry and Quality Assurance procedures up in clinical facilities.

## P4 Quality Assurance (Mevion)O13: Initial Technical and Clinical Experience with the world's first compact proton system

### Washington University

Background: On December 18, 2013 we treated our first patient with the Mevion S250 Proton Therapy System. This followed acceptance testing and commissioning, including; data acquisition, safety checks, establishing QA programs, mechanical and imaging checks, beam modeling validation, etc.

Materials: The Mevion S250 possesses a compact synchro-cyclotron attached to a gantry due to its small size (1.8m diameter) and light weight (22 tons) works in combination with a 6D robotic couch allowing any beam trajectory. The system currently customizes beams using passive scattering and provides 3D-2D imaging via fixed x-rays, with a portable CT being delivered to us after the writing of the abstract.

Our treatment planning system is Varian Eclipse, V11, for which beam modeling and validation comprised the core of commissioning time. The electronic medical record system is Elekta's MOSIAQ 2.5, which provides connectivity to the Mevion System, for setup, imaging and treatment.

Results: The treatment planning modeling and validation met a set gamma index of 1%, 1 millimeter. This was maintained above 90% for all field shaping system options, which are designed according to range, modulation, and field size. There is 1%/1mm criteria was not met for the first few cm of the buildup region.

The image quality is equal or better than current onboard imaging systems in our clinic. The imaging quality of Mevion's system surpassed expectations, for low and high contrast resolution.

The isocentricity of the gantry and couch was measured to be 0.4 and 0.6 mm, respectively. The QA program developed includes a 30 minute daily QA which includes checks for table isocentricity interlock, safety, imaging, and most importantly, dosimetric checks for three of the 24 options. Daily, we measure output, modulation flatness, range, field symmetry, and distal falloff, all with a single exposure per option.

The neutron dose was measured to be < 1.0 (mSvl)/Gyl for most clinical fields, thereby alleviating concerns for pediatric treatments.

Clinical start up over the first three months included a sampling of cases, namely; the central nervous system (including a pediatric anesthesia case for meduloblastoma), chrondrosarcomas of the pelvis, esophagus, mediastinum, liver, etc.

In terms of machine operation and reliability, the downtime has been 10% or less over the first three months, and < 3% for months 4-6.

Conclusions: We are ramping up to 20 patients per day, with the case load depending on the number of anesthesia cases. Overall, the Mevion S250 has added a vital delivery option to compliment other treatment machines within our National Cancer Institute Comprehensive Center.

## P4 Quality Assurance (Spot Scanning)O65: Experience with the stability of spot characteristics for spot-scanning proton beams

### ProCure Proton Therapy Center, CDH Proton Center, Shanghai Proton and Heavy Ion Center

Background: For spot-scanning proton beams, the stability of the spot characteristics such as spot size, tilt and position determines the accuracy of dose delivery. The size and tilt of elliptical spots in air are controlled mostly by the quadrupole magnet, while the positions of spots relative to a patient target depend on both the scanning magnet and the x-ray imaging system. In this study, the spot characteristics were measured and analyzed from July 2013 to July 2014, which included one major site configuration update, 2 scheduled cyclotron interventions, and 4 unscheduled cyclotron openings.

Materials and Methods: Fluence maps for a 9-spot asymmetric pattern distributed throughout a field of 16 cm x 18 cm were created for 18 proton energies (100.0-26.0 MeV). The spot fluence in air was measured using a CCD camera at three planes along the beam line axis (isocenter, and up/down stream 15 cm). The camera was first setup at the isocenter using kV x-ray imaging, and then shifted up/down stream using the robotic patient positioning system. The measured fluence maps were analyzed using in-house programs, which calculated the spot sizes, tilts, and offsets of the Gaussian shaped spot positions.

Results: Compared to the treatment planning system data, the year-averaged deviations of the spot sizes at the isocenter plane were 2.16% and 1.11% in the IEC gantry x and y directions respectively. The maximum deviation was 10.08% (IEC-x direction, 203.5 MeV) while the minimum deviation was 0.006% (IEC-y direction, 136.0 MeV), both at the isocenter plane. Consistent spot tilts of ~90 degrees were observed at the plane 15 cm upstream from the isocenter for most protons with energies < 200 MeV. Most spots were tilted less than 20 degrees at the plane 15 cm downstream from the isocenter. After the collinearity of the proton and x-ray imaging system isocenters were optimized, the offsets of the spot positions were < 1.5 mm at three planes. During the site configuration update, spot positions were found to be displaced by ~6 mm until the tuning parameters file was properly restored.

Conclusions: The stability of the spot characteristics was tested over a one-year period. Such a QA procedure is recommended to be performed at least monthly and any time after a database update or cyclotron intervention. For the spot size deviations, a tolerance of < 15% can be easily met with this system. The deviations of spot positions should be < 2 mm at any plane up/down stream 15 cm from the isocenter.

## P6 Treatment Planning (Breast)O67: Skin dose reduction with scanning beam proton accelerated partial breast irradiation

### Department of Radiation Oncology, Mayo Clinic, Rochester, MN

Purpose: Proton therapy is an attractive modality for accelerated partial breast irradiation (APBI) given the dose reduction to normal tissue when compared to 3DCRT and superior dose homogeneity when compared to brachytherapy. Recently, reported long term outcomes of patients treated with passive scatter proton therapy (PSPT) have revealed a higher than expected incidence of skin toxicity. Therefore, we evaluated the feasibility of scanning beam proton therapy (SBPT) for skin dose reduction in APBI.

Materials and Methods: Two-field SBPT plans were created for eight consecutive patients treated with 3DCRT APBI on the NSABP B-39/ RTOG 0413 trial using the Varian Eclipse v10 treatment planning system. The CTV was defined per protocol, and was restricted to within 5mm of the skin surface without skin constraints. Planned dose was 38.5 Gy in ten fractions and all SBPT plans were normalized for 100% of the CTV to receive at least 98% of the prescription dose. Robustness analysis was performed with range (+/- 3%) and position (+/- 5mm) uncertainty. Single (SFO) and multi-field optimization (MFO) was also performed for each plan.

Results: Mean maximum skin dose for all SBPT plans was 100.3% (98.6-100.6%) compared to 112.5% (102.2-110.7%) for photon plans. Mean skin D0.05cc was 38.2 Gy (37.2-40.2 Gy) for SBPT plans compared to 42.86Gy (38.7-56.5 Gy) for the corresponding photon plans, demonstrating a 10.9% reduction with SBPT. Even in the worst robustness scenario, skin dose was no worse than 98.6% of the max skin dose in photon plans. Mean ipsilateral breast V50% was 438.5 cc (202.7-773.6 cc) for SBPT plans and 906.34 cc (549.8-1411.8 cc) for 3DCRT plans. Ipsilateral lung V20Gy was also minimized with a mean of 0.3% (0-1.9%) for SBPT plans and 2.5% (0.6-4.6%) for associated photon plans. Two patients had left breast treatment. A decrease in heart dose was noted in SBPT plans with a mean V5Gy of 2.2% (0-6.2%) as compared to 9.3% (2-17.3%) for 3DCRT plans. SFO provided more robust CTV 95% coverage when compared to MFO plans.

Conclusions: SBPT is capable of delivering APBI with robust CTV coverage while reducing skin, lung and heart doses. Two-field SFO with small angle separation provides an acceptable compromise between skin sparing and robust CTV coverage. Studies are ongoing with skin constraints to determine whether further skin dose reduction is achievable without compromise of CTV coverage.

## O68: Treatment of Locally Advanced Breast Cancer: Case Studies of Proton and Photon Comparative Plans

### CDH Proton Center, Proton Collaborative Group

Background: Adjuvant treatment of locally advanced breast cancer with radiotherapy is known to reduce the risk of local disease recurrence. It is unknown which radiotherapy technique delivers the optimal therapeutic ratio, achieving appropriate target coverage while minimizing dose to normal tissues. This dosimetric analysis compares proton therapy (PT) with either uniform scanning (US) or pencil beam scanning (PBS), to a mixed photon/electron (MPE) technique.

Materials and Methods: Three patients with resected locally advanced breast cancer underwent treatment planning for adjuvant radiotherapy. All patients had the chest wall (CW)/reconstructed CW plus comprehensive regional lymphatics including internal mammary lymph nodes (IMN) targeted. Treatment plans were run using the PBS technique, US PT technique, and the MPE technique. CTV coverage specifications were D95=100%, and D99= 95%. With each of the three cases, dose volume histogram data was compared among the three planning types and the optimum results among the comparisons were noted (see attached table).

Results: The most notable differences between the planning types were seen in the coverage and hot spots of the PTV-Skin structure, as well as OAR doses to ipsilateral lung, heart, and skin.

Conclusions: In comparison to US PT and MPE, PBS PT appears to demonstrate an improved therapeutic ratio, decreasing dose to the ipsilateral lung, heart, and skin. Additionally, PBS is able to improve time utilization for both planning and potentially treatment as well.

## O69: Brachytherapy versus Uniform and Pencil Beam Scanning Protons for Partial Breast Irradiation

### CDH Proton Center

Background: Partial breast irradiation (PBI) is commonly used in the adjuvant treatment of early stage breast cancer. We compared the dosimetry of strut-based brachytherapy (BT) to uniform scanning (US) and pencil beam scanning (PBS) proton techniques.

Materials and Methods: 5 patients with left-sided breast cancer were treated with BT with a SAVI device; pts were treated to a PTV (catheter + 1 cm) to 34 Gy in 3.4 Gy BID fractions. Pts underwent CT scans for cavity evaluation prior to catheter placement; these CT's were used to create US and PBS plans. Non-coplanar 3-field plans were created for all pts with US (3D, forward planned) and PBS (single field uniform dose inverse planned) prescribed to 40 Gy(RBE) in 4 Gy(RBE) QD fractions per PCG BRE-007 protocol. Heterogeneity, coverage, and organ-at-risk doses were compared using t-tests.

Results: Mean PTV volume was larger for proton compared to BT plans (348.4 cc vs. 89.7 cc; p<0.001). For all proton and BT pts, PTV surface was ≤ 5 mm from the skin. BT plans were more heterogeneous than proton plans with mean V150 and V200 for BT at 41.68 cc and 19.14 cc, respectively, while V150 and V200 were 0 cc for both US and PBS plans (p<0.001). Maximum skin dose was lower with BT at 86.1% compared to 102.1% and 103.94% for PBS and US plans, respectively (p=0.004). Mean heart dose with protons was 5.8 cGy with PBS and 2 cGy with US compared to 191.34 cGy with BT (p=0.004). Mean lung dose was lower with PBS (22.2 cGy) and US (12.4 cGy) compared to BT (209.02 cGy) (p=0.002). There was no difference in heart V5 or V10 or lung V10 between BT and proton plans. There was no difference in coverage (D90, D100) between BT and proton plans. There were no differences in heterogeneity, coverage, or organ-at-risk doses between PBS and US plans except V5 to “elsewhere breast tissue” which was 49.9% for PBS versus 33.7% with US (p=0.027).

Conclusions: Proton PBI provided reduction in heart and lung dose compared to BT despite larger PTV volumes. While PBI with PBS could potentially reduce skin dose compared to US, our study did not demonstrate a difference. PTV volumes were ≤ 5 mm from the skin in all cases, preventing dose modulation at the surface. More studies are required to determine whether PBS may provide a benefit over US in patients with deeper cavities. BRE-007 is ongoing to evaluate the safety and efficacy of PBI with protons.

## O72: Potential Benefits Using Uniform Scanning & PBS Proton Therapy for Locally Advanced Breast Cancer

### CDH Proton Center, ROC

Background: Adjuvant treatment of locally advanced breast cancer (CA) with radiotherapy is known to reduce the risk of local disease recurrence. It is unknown which radiotherapy technique delivers the optimal therapeutic ratio. We compared uniform scanning (US) proton therapy (PT) to pencil beam scanning (PBS) PT to observe if these two techniques provided potential benefits for locally advanced breast CA patients.

Materials and Methods: 5 patients with resected locally advanced breast cancer underwent treatment planning for radiotherapy; 3 had reconstruction/implants post mastectomy and 2 did not have reconstruction prior to RT. All patients had the chest wall (CW) plus comprehensive regional lymphatics including internal mammary lymph nodes (IMN) targeted. Treatment plans were generated using the PBS technique and US conformal 3D PT technique. CTV coverage specifications were D95=100%, and D99= 95%. Dose volume histogram (DVH) analysis comparing the 2 techniques was performed on each patient, evaluating target coverage and doses to normal organs inlcuding heart, lung, esophagus, and skin. The US plans used matched upper and lower fields, with 2 match lines used for a total of 4 fields treated daily. The plans were optimized to cover the clinical target volume (CTV) distally with consideration of potential effects of range uncertainty and the 7 mm expanded planning target volume (PTV) laterally. The PBS approach utilized 2 beams, 10-15 degrees off enface in the medial and lateral directions to the PTV. The entire CTV was encompassed in a single field uniform dose (SFUD) beam, with both fields treated daily. PBS plans were initially optimized to cover the CTV in the distal direction and a 7mm margin lateral to the field direction, then the plans were recalculated to account for uncertainties and ensure robustness by that adding additional 3% of range to all layers plus 1 mm.

Results: The most notable differences between the planning types were seen in the coverage and homogeneity of the target volume, as well as OAR doses to ipsilateral lung, heart, esophagus, and skin. See attached table. The PBS plans took an average of 2 hours to plan, compared to an average of 10 hours for the US plans.

Conclusions: In comparison to US PT, PBS PT demonstrates a potential to improve the therapeutic ratio, decreasing dose to the ipsilateral lung, heart, and skin. Additionally, PBS appears to improve time utilization for both planning and potentially treatment as well.

## P6 Treatment Planning (Gastrointestinal)O75: 4D robustness of proton pencil beam scanning for esophageal cancer – a case study

### Mayo Clinic

Background: Our aim was to evaluate 4D robustness of various free-breathing delivery strategies employing pencil-beam scanning (PBS) for the treatment of esophageal cancer.

Materials and Methods: Proton planning was carried out for a patient with a distal esophagus tumor. To minimize uncertainties associated with GI variability and breathing motion, proton PBS plans were developed using either a single posterior field (1Fld) or two posterior oblique fields (2Fld) using single-field optimization. Low- (45 Gy in 25 fx) and high-dose (50 Gy) CTVs were contoured on extreme phases of a 4D-CT simulation to derive internal target volumes ITV45 and ITV50. Optimization target volumes, OTV45 and OTV50, were constructed via an isotropic 4mm expansion from ITV45 and ITV50, respectively. 3D plans were optimized on either the end-exhale 4D-CT phase (4D-50%) or the average (4D-Ave). The spot size (sig=4mm at Bragg peak) was varied by inserting a range shifter (RS; sig=8mm). Six plans were generated (1Fld-Ave; 1Fld-50%; 1Fld-Ave-RS; 2Fld-Ave; 2Fld-50%; 2Fld-Ave-RS) and normalized such that ≥98% of the ITV50 received ≥50 Gy. 4D plans were generated using in-house tools involving 1) binning of the layers/spots onto each of the ten 4D-CT phases; 2) 4D dose accumulation on 4D-50%. Step 1 included realistic synchrotron/patient timing/parameters. Step 2 incorporated deformation vector fields from deformable image registration (4D-N% to 4D-50%). A worst-case DVH evaluation was carried out simulating three separate fractional deliveries of each plan, with and without use of maximum-MU repainting (MaxMURP). This technique is a type of layer repainting achieved by restricting the maximum MU in the planning system to 0.005 MU (much less than the deliverable maximum MU).

Results: Fig. 3 shows an example of spot repainting distributions obtained with MaxMURP for 1Fld-Ave and 2Fld-Ave plans. Beam-on times with MaxMURP increased by factors of 1.9 and 1.3 for 1Fld-Ave and 2Fld-Ave plans, respectively. As demonstrated by Fig. 1, MaxMURP resulted in marked improvements in 4D target DVHs for 1Fld plans, with less significant improvements observed for 2Fld plans (inherently repainted). MaxMURP also improved 4D robustness of 1Fld plans to range overshoot into the heart (Fig. 2). Planning on 4D-50% did not confer 4D robustness advantages.

Conclusion: MaxMURP is likely an appropriate breathing motion mitigation strategy for esophageal PBS, allowing us to employ 1Fld plans and the smallest spot size available.

## P6 Treatment Planning (Spot Scanning)O73: Aggregated Plan Robustness Analyses for Spot-Scanned Pancreatic Stereotactic Body Proton Therapy

### Department of Radiation Oncology, Mayo Clinic, Rochester, MN, Department of Radiation Oncology, Mayo Clinic, Scottsdale, AZ

Purpose: Few effective treatment options exist for advanced pancreatic malignancies. We explored the planning feasibility and dosimetric advantage of spot-scanned stereotactic body proton therapy (SBPT). Proton plan robustness analyses under static and inter-breath-hold motion conditions were performed.

Materials and Methods: 10 patients were retrospectively selected; the GTV included only gross tumor volume. Isotropic CTVs were created assuming breath-hold scenarios. 2 beam lines were investigated, one with a nominal spot size (1σ) at relevant depths of 4 mm and the other with 6 mm; both single and multi-field optimizations (SFO v. MFO) were employed. Multiple optimization target volumes (OTV of 3, 5, and 7-mm expansion) were created. All plans were normalized to at least 98% of CTV covered by 40 Gy(E) in 5 fractions (Figure 1). 8 perturbed plans for translational isocenter shifts and stopping power uncertainties were generated, totaling 1,080 plans for robustness analyses. Custom DVH aggregation tools were developed for comparison. Inter-breath-hold motion was modeled under deformable dose accumulation.

Results: The duodenum was the most difficult organ-at-risk (OAR) for meeting dosimetric constraints, especially in the CTV+7 mm cases. Compared to IMRT, the integral and OAR doses were significantly better for proton plans (Table 1). The overall differences between various proton plans for a given OTV expansion were minor. Under static considerations, the combination of CTV+5mm, SFO, and a spot size of 6mm represented the best compromise among robustness, target dose coverage, and OAR sparing (Figure 2). Inter-breath-hold motion analyses showed that 4-mm spot-size and CTV+3mm resulted in significant dose inhomogeneity which was deleterious to plan integrity dynamically.

Conclusions: Benchmarked against IMRT planning, pancreatic SBPT offered comparable duodenal sparing with significant improvements in most OAR dosimetric metrics considered, which is clinically attractive. Under static considerations, MFO v. SFO differences were generally not statistically significant. Inter-breath-hold robustness analyses confirmed the preferred approach of utilizing a larger OTV expansion for up to 5 mm and larger spot size. The proton plan robustness evaluation has suggested that the ideal parameters for optimal spot size, dosimetric optimization technique, and optimization target volume expansion could be obtained for SBPT in patients with pancreatic cancer.

Table 1.

Dosimetric comparison of IMRT vs. SBPT (varying MFO vs. SFO and also spot size, σ) in pancreatic tumor treatment planning. Bold fonts represented the two best spot-scanned proton scenarios (all comparable to or better than their respective head-to-head photon plans) for each dosimetric index investigated. Clear dosimetric advantages were seen in SBPT plans.

## P6 Treatment PlanningO66: To what extent does a high Z implant affect the quality of a treatment plan in proton therapy (PT)?

### University Hospitals/Case Medical Center, Case Western Reserve University, University Hospitals

Background: To what extent does a high Z implant affect the quality of a treatment plan in PT? This question is particularly relevant in PT due to: (1) streaking artifacts in CT images, (2) CT calibration (HU-RSP) in the high Z region, (3) interface effect, and (4) selection of treatment technique. In this study, we focus on (2) and (4) in a treatment plan comparison study between proton and photon irradiations.

Materials and Methods: The CT scan of a spine patient treated with VMAT using 6 MV x rays is used retrospectively in this study. Proton planning uses the same contour set and dose-volume constraints as in VMAT. The metallic screws implanted between T4-T8 in the patient forms part of the PTV. A single PA field is employed due to the shape and location of the PTV. The is carried out with two HU-RSP curves: (1) default curve (highest RSP =1.7), and (2) extended curve (default curve + a metal point at HU=4000 and RSP=7.8 (steel). For each HU-RSP curve, three RSP values for the screws are used, 1.0 (simulating no screws), 3.5 and 7.8. Two treatment planning systems (TPS) are employed: Pinnacle (V.10) with a double scattering beam (DS) and pencil beam algorithm (PB), and Eclipse (V.13) with either a DS or a spot scanning beam and the superposition convolution algorithm (SC). For each combination of HU-RSP curve, RSP(screw) and TPS (total=18), a treatment plan is generated. DVH are calculated and compared for the different plans.

Results: A summary of the comparison of treatment plans is presented in the supporting document. Proton plans using on the default HU-RSP curve are superior to VMAT, even with RSP (Screw=3.5/7.8) since the highest RSP in the default curve is 1.9. When the extended HU-RSP curve is used, RSP(screw) has to be≤3.5 so that a plan can be generated. The VMAT plan in general is better than the proton plan with PB. However, SC produces DVH comparable to VMAT for the OARs but more uniform dose to the PTV.

Conclusion: Proton beam produces superior dose distribution compared to VMAT without the presence of high Z inhomogeneity. The presence of high Z implant in the spine requires the use of large compensator smearing and edge smoothing to cover the PTV, resulting in higher dose to heart and right lung compared to VMAT. If the implant material is known and can be covered by the proton beam from the accelerator, proton therapy is a viable and may be superior alternative. PB produces less favorable distribution compared to SC in this study.

## P7 OtherO76: Nanoscale radiation measurements in proton therapy beams

### Rice University, The University of Texas MD Anderson Cancer Center, Carleton University

Background: Fluorescent nuclear track detectors (FNTD) are radiation track detectors based on the fluorescence properties of Al2O3:C,Mg crystals. Ionizing radiation creates fluorescence centers in the FNTD (excitation and emission peaks at 635 and 750 nm, respectively). Confocal microscopy can be used to image the fluorescence and reconstruct 3D images of the particle tracks with nanometric resolution (~400 nm). If a cell layer is adhered to the FNTD's surface, tracks can be extrapolated back to the cell layer, making spatial co-localization of biological DNA repair response in cells with the particle path possible (Fig 1). In addition, track intensity may provide a metric for determining single particle LET, allowing further understanding the relationship between LET and relative biological effectiveness at the single cell level.

Materials and Methods: Two FNTDs were exposed to 0.1 MU in therapeutic proton beams under the following conditions: 1) 160 MeV at 12.5 cm depth; 2) 200 MeV at 18 cm depth; and 3) 160 MeV at 0.0 cm depth in water. The FNTDs were then imaged. Next, we re-exposed one of the FNTDs to 0.2 MU of protons at condition 1, and the other to 0.2 MU of protons at condition 2. Following each irradiation set, proton tracks in the FNTDs were imaged in an Olympus Fluoview FV1200 confocal microscope. Tracks were imaged using a 640 nm laser (~3 mW at sample position), 20 μs/pixel dwell time, photomultiplier at 700 V, 3.5× gain, 0% offset and 140 μm pinhole diameter. Images were processed to identify tracks and remove background. Tracks were fit using a 2D Gaussian function to obtain track intensities. The distribution of track intensities for each condition was determined by normalizing each intensity distribution to the lowest peak and subtracting the distributions. The obtained distribution of intensities was fit to a normal distribution. The peak of the distribution for track intensities were correlated to the dose averaged LET (D-LET) of the particles. LET values were determined using a Monte Carlo model of the nozzle.

Results: Fig 2a shows a 3D reconstruction of proton tracks in an FNTD. Fig 2b presents the maximum intensity projection of the image from Fig 2a. The peak track intensity shows an increase with increasing LET (Fig 3).

Conclusions: FNTDs can measure single proton tracks with nanometric resolution. The positive correlation between track intensity and LET suggests that track intensity may be a useful new metric for measuring single particle LET.

## P7 Other (MC)O78: Variations in proton LET distributions obtained using different Monte Carlo scoring techniques

### Carleton University, University of Texas MD Anderson Cancer Center

Background: Monte Carlo (MC) simulations are commonly used to study linear energy transfer (LET) distributions in therapeutic proton beams. A number of different methods have been implemented to score dose-averaged LET (D-LET) values of proton beams. The goal of this work is to compare the D-LET distributions obtained using different scoring techniques.

Materials and Methods: We used TOPAS, which is based on the GEANT4 simulation toolkit, for all of our simulations. We simulated a monoenergetic 150 MeV proton beam incident on a water phantom, and calculated D-LET as a function of depth in water using the following three scoring techniques: 1) score proton LET on a step-by-step basis using the ComputeElectronicDEDX() method in GEANT4's EmCalculator class, 2) score proton LET on a step-by-step basis by dividing the energy deposition during each step by the step length, and 3) score the proton energy spectrum, convert to an LET spectrum using a lookup table and calculate the average. We tracked and scored both primary and secondary protons in the simulations, with a secondary proton range production threshold of 0.05 mm in water. We also investigated the effects of varying the electron range production threshold from 0.01 mm to 10 mm.

Results: For a high electron range production threshold (10 mm), the three scoring techniques showed differences ranging from approximately 6-25% before the Bragg peak region. The two step-by-step scoring techniques (1 and 2) showed better agreement with each other than with the energy spectrum scoring technique (3). Techniques 1 and 3 were found to be insensitive to changes in the electron production threshold. However, the results were approximately 100% higher and had larger uncertainties using technique 2 at the lower electron production threshold.

Conclusion: Large differences were seen between D-LET values determined using each of the three scoring techniques. In addition, technique 2 was found to be sensitive to electron range production thresholds. Significant caution is required when comparing D-LET values obtained using different scoring techniques, and different MC transport parameters. Further research is warranted to determine the most accurate and reliable methods of scoring D-LET.

## P7 Other (Spot Scanning)O36: Spot Scanning Dose Delivery with Rapid Cycling Proton Beams from RCMS

### University Hospitals/Case Medical Center, UT Southwestern Medical Center, Indiana University

Purpose: A rapid cycling proton beam has several distinct characteristics that are superior to a slow extraction synchrotron: The beam energy and energy spread, beam intensity and spot size can be varied spot by spot. The feasibility of using a spot scanning beam from a rapid-cycling-medical-synchrotron (RCMS) at 10 Hz repetition frequency is investigated in this study for its application in proton therapy.

Materials and methods: The versatility of the beam is illustrated by two examples: (1) a cylindrical volume (diameter=5cm, length(SOBP)=5cm and depth of distal edge=11cm) embedded in a water phantom irradiated by a single field and (2) a spherical volume (diameter=4cm, center of sphere at depth=5cm) in a water phantom irradiated by two parallel opposed fields. A uniform dose distribution is to be delivered to the volumes. Geant4 Monte Carlo code is used to validate the dose distributions in each example.

Results: A transverse algorithm describing the distribution of beam spots to produce a uniform circular distribution of radius 10 arbitrary units at d=11cm in a water phantom is shown in Fig. 1. The parameters which describe the different spots are shown in Table 1. Longitudinally, different proton energies are used in successive transverse planes to produce the SOBP required to cover the volumes. In general, uniformity of dose distribution within 3% is obtained for the cylinder (Fig. 2) and 3.5% for the sphere (Fig. 3). The transverse algorithms requires only few hundred beam spots for each plane The algorithms may be applied to larger volumes by increasing the intensity spot by spot for the same delivery time of the same dose. The treatment time can be shorter than 1 minute for any field configuration and tumor shape.

Conclusions: The unique beam characteristics of a spot scanning beam from a RCMS at 10 Hz repetition frequency are used to design transverse and longitudinal algorithms to produce uniform distribution for any arbitrary shape and size of targets. The proposed spot scanning beam is more versatile than existing spot scanning beams in proton therapy with better beam control and lower neutron dose.

## C2 EyeP15: Dosimetry factors for a small field proton beam line for ocular treatment at MD Anderson CC

### The University of Texas M.D. Anderson Cancer Center

Aim: To determine dosimetric properties of the small field ocular beam line, including depth and lateral dose distributions, output dose/MU (d/MU) and other dosimetry parameters.

Materials and Methods: The proton beam enters the ocular beam line at 87 MeV. The maximum field size is 3.5 cm in diameter. The range (R) varies between 0.8 and 3.5 cm with 0.1 cm increments. Eleven range modulation wheels (RMW) provide modulation width (M) between 0.3 and 3.0 cm in steps of 0.1 cm for M ≤1.0 cm and of 1.0 cm for M ≥ 1.2 cm. The monitor chamber was calibrated for dose/MU based on the IAEA TRS-398 protocol with a PTW advanced Markus chamber. Measurements of relative depth dose and dose/MU for different options were performed with the advanced Markus chamber. Lateral dose distributions and field size were measured with an IBA SFD Hi-pSi diode and EBT-3 film for half-blocked fields.

Results: The measured range agreed with the nominal range within ± 0.03 cm, while the measured modulation width was within ±0.1cm. The lateral penumbra varied between 0.9 and 2.0 mm with air gap of 1.5 and 6 cm for the range of 2.5 cm and depth of 1.5 cm. The distal penumbra changed between 0.18 cm for RST of 0 cm and 0.22 cm for RST of 2.7 cm. The d/MU is expressed as d/MU = (d/MU)ref•RSF•SOBPF•FSF, where (d/MU)ref is d/MU for a reference RMW, RSF is the output factor change for RST change, SOBPF is the output factor change for M change, and FSF is the relative change in output factor with field size(1). Other factors appeared in the reference (1) were assumed to be unity because the measurements were done at the central axis of the field at the center of the SOBP. The (d/MU)ref was measured as 0.947 ± 0.009 cGy/MU with M of 3 cm and R of 3.5 cm. The RSF was expressed as RSF = 1.0 – 0.079RST – 0.044RST^2 ( Fig. 1). The measured SOBPF was fitted to the d/MU model by Kooy(2) and expressed as, SOBPF = 0.68(1 + 0.94r^0.44), where r = (R – M)/M (Fig. 2). FSF was 0.99 ± 0.01 for field sizes ≥ 0.8 cm in diameter.

Conclusion: We have measured dosimetric factors of the small field ocular beam line as part of our clinical commissioning process. Reference: (1)N. Sahoo et al., “A procedure for calculation of monitor units for passively scattered proton radiotherapy beams”, Med. Phys. 35(2008)5088 – 5097. (2) H. Kooy et al., “The prediction of output factors for spread-out proton Bragg peak fields in clinical practice”, Phys. Med. Biol. 50(2005) 5847-5856.

## P50: Proton Beam Radiation of Choroidal Melanomas and New Set-up Techniques to Improve Cosmetic Outcomes

### Loma Linda University Medical Center

Purpose: This retrospective review evaluated the local control and cosmesis of patients treated with proton beam for choroidal melanomas with respect to eyelid position in order to prevent eyelash alopecia.

Materials and Methods: Seventy-eight patients were treated to a dose of 70Gy in 5 fractions between December of 2010 and April 2000 with a setup of either eyes open or eyes closed in order to prevent eyelid alopecia. All patients had a pre-treatment evaluation and follow-up examinations by an ophthalmologist, and local control was assessed every six months by fundoscopic examination and a B Ultrasound. The toxicities and cosmesis were evaluated by physical exam and patient surveys.

Results: The local control was documented in 72 patients whereas late toxicities were documented in 61 patients. The local control was 86% for the entire cohort and an additional 4% developed a marginal recurrence or new primary. There was no difference in local control between the eyes open versus eyes closed set up (87.0% vs 85.4%, p = 0.86) and there was also no difference between COMS small vs med/large tumors (79.0% vs 86.8%, p = 0.415). The medium follow up time for local control was 3.2 years. Eyelid erythema and discoloration were more common when treated with eyes closed (p = 0.032) but dry eye complaints were equivalent between the two groups (p = 1.00).

Conclusion: Treatment through the eyelid caused more acute and late skin reactions, but the local control between the two techniques were equivalent. The closed eyelid position is a feasible treatment technique if the clinical situation warrants it.

## P79: Long-Term Outcomes after Proton Beam Radiation Therapy for Intraocular Vascular Tumors

### University of California, San Francisco, CA, The Tumori Foundation, San Francisco, CA

Background: The long-term clinical outcomes of proton beam radiation therapy (PBRT) for intraocular vascular tumors remain unclear. We aim to present long-term follow-up for patients with intraocular vascular tumors treated with PBRT.

Methods: We performed a review of our prospectively maintained database for all patients with intraocular vascular tumors treated with PBRT at our institution between 1994-2012. Patients with at least 6 months of follow-up data were included in the study.

Results: Our review identified 29 evaluable patients (29 eyes) with vascular tumors of the eye, with the most common diagnosis being choroidal hemangioma (n=22). Additionally, 3 diffuse choroidal hemangiomas associated with Sturge-Weber syndrome, 3 retinal angiomas, and 1 hemangioblastoma associated with von Hippel-Lindau syndrome were identified. The main analysis hereafter focuses on the choroidal hemangioma patients (n=22), with mean follow-up time of 78.5 months (range 12-177 months, median 69.2 months). Median age at diagnosis was 48.3 years (range 17-72). Initial tumor thickness, as measured by ultrasonography, ranged from 2.0 mm to 7.2 mm (mean 3.8 mm). Patients were treated using total doses between 18 and 24 Cobalt Gray Equivalent (CGE) in four fractions. Best corrected visual acuity improved or remained stable in 20 of 22 eyes (91%). Of the 16 eyes initially presenting with retinal detachment, 15 experienced reattachment (94%). Cystoid macular edema was initially documented in 16 eyes and completely resolved in 13 eyes (81%). Tumor thickness decreased in all eyes as of last examination. Notable complications occurred in five patients, with the most common being radiation retinopathy and atrophy. Nine patients received full dose to the disc and macula (range 18-24 CGE), none of whom had significant worsening of vision. The additional seven intraocular vascular tumors with variable histologies have also demonstrated tumor response after PBRT at mean follow-up of 29.5 months (median 24.4 months) with stable or improved visual acuity.

Conclusions: This series demonstrates consistent and long-term tumor response and visual preservation in patients with benign vascular tumors of the eye. The data presented represents the longest available follow-up in the literature to date for choroidal hemangiomas treated with PBRT.

## P81: Tumor & Dosimetric Predictors of Longterm Visual Outcome After Proton Beam Therapy of Uveal Melanoma

### University of California San Francisco, CA, Tumori Foundation, San Francisco, CA

Background/Objectives: To discern contributing tumor and critical structure dose-volume parameters affecting long-term visual outcomes after proton beam radiation therapy (PBRT) for uveal melanoma (UM) patients.

Materials/Methods: Using our prospectively maintained database, PBRT patients treated for UM by a single surgeon with visual acuity (VA) follow-up were reviewed (1995-2009). 512 patients with pre-treatment best corrected VA (BCVA) in the treated eye of 20/200 or better were evaluated for this study and patients were stratified into three groups as follows: 20/15-20/40 (n=363), 20/50-20/100 (n=114), and 20/200 (n=35). Patients received 56 GyE in 4 fractions. Median VA follow-up was 6.0 years (range 0.3 - 18). Loss in VA was defined as migration into a lower VA stratum including worse than 20/200 at most-recent post-treatment assessment compared with pre-treatment. Univariate and multivariate Cox's stepwise regression models adjusted for pretreatment VA were applied to identify independent predictors for loss in VA among patient, tumor, and dose-volume histogram (DVH) parameters.

Results: Univariate analysis indicated that loss in VA was significantly related to the following patient/tumor parameters: older age, diabetes, greater tumor height, larger tumor diameter, closer tumor distance to fovea and to optic disc. DVH analysis revealed that at ≥50% dose (≥28 GyE) to the lens, optic disc and macula, significant predictors of VA loss included >30% of the lens volume, 100% of the disc and >0% of the macula receiving dose. After adjusting for pretreatment VA, multivariate analysis identified the following independent predictors of loss in VA (likelihood ratio test; hazard ratio (HR), 95% confidence interval): increased tumor height (p<0.0001; HR 1.13 [1.08-1.18]), macula receiving ≥50% dose (p<0.0001; HR 1.87 [1.44-2.42] for 0 vs >0%), increased age at the time of PBRT (p<0.0001; HR 1.02 [1.01-1.03]), increased tumor diameter (p=0.02; HR 1.06 [1.02-1.10]) and 100% of the optic disc receiving ≥50% dose (p=0.02; HR 1.36 [1.05-1.77]).

Conclusions: For UM patients undergoing PBRT, the change in visual acuity from pretreatment BCVA may be predicted by patient age at treatment, tumor height, largest tumor diameter, and optic disc and macula dose-volume parameters. Further analysis is planned to evaluate VA change at various time points of follow-up as well as relationships for disc and macula sparing and dose effects.

Table 1.

Multivariate Predictors of VA Loss

## C3 GenitourinaryP1: Pathology Review for Prostate Cancer Patients Referred to the SCCA Proton Center

### University of Washington

Background: Treatment guidelines for patients with newly-diagnosed prostate cancer are largely determined by risk stratification (low, intermediate, high) which is based upon three parameters: T-stage, PSA value, and Gleason score. When treatments began in March, 2013, the center adopted the University of Washington policy of having the tumor pathology for all cancer patients reviewed centrally by the University of Washington pathology group. The purpose of this report is to evaluate the results of this review for the first 53 prostate cancer patients treated at the center and to determine the number of cases where changes in the Gleason score affected the recommended treatment.

Materials and Methods: All prostate cancer patients treated at the center were identified. Between March 13, 2013 and April 14, 2014, 53 prostate cancer patients with Gleason scores initially determined by outside pathologists were treated at the center. Tissue specimens were obtained for all patients and were reviewed by the University of Washington Department of Pathology. The results of this review were tabulated and analyzed.

Results: In 38 cases the original Gleason score was confirmed. In 7 cases the original Gleason score was upgraded and in 8 cases the original Gleason score was downgraded. The greatest number of changes occurred for patients who were originally categorized as low risk with a Gleason score of 6 (3 + 3) and upon reclassification had their Gleason score changed to 7 (3 + 4). In the case of patients who were originally classified as intermediate risk via a Gleason score of 7, two were downgraded from a (4 + 3) to a (3 + 4) while two became low risk having a Gleason score of 7 (3 + 4) changed to a Gleason score of 6 (3 + 3). Four patients originally classified as having high risk disease became intermediate risk upon review. Of three patients with initial scores of 8 (4 + 4), two were shifted to scores of (4 + 3) while one was shifted to a score of (3 + 4). The greatest change occurred in a patient referred with a Gleason score of 9 (4 + 5) who was shifted to a Gleason score of 7 (4 + 3).

Conclusions: Twenty eight percent of reviewed patients had significant changes in their Gleason scores which changed their risk category. This impacted recommended treatment in terms of the use of androgen deprivation therapy and/or the radiation fields that were treatment.

## C4 GastrointestinalP5: Proton Radiotherapy Offers Reduced Normal Tissue Exposure in Pancreatic Cancer

### Loma Linda University

Background: Pancreatic cancer is a highly aggressive malignancy. Chemoradiotherapy (CRT) is utilized in many cases to improve locoregional control; however, toxicities associated with radiation can be significant given the location of the pancreas. The purpose of this study is to quantify the dosimetric changes seen when using photons or protons in patients receiving CRT for cancer of the pancreas.

Materials and Methods: Ten patients with pancreatic head adenocarcinoma treated between 2010 and 2013 were evaluated in this study. All patients were simulated with contrast-enhanced CT imaging. Separate treatment plans using proton radiotherapy, intensity-modulated radiation therapy (IMRT), and 3D photon radiotherapy modalities were created for each patient. Dose-volume histograms were calculated and analyzed in order to compare plans between the three modalities. The organs at risk (OAR) being evaluated in this study are the kidneys, liver, small bowel, and spinal cord. To determine statistical significance, ANOVA and two-tailed paired t-tests were performed for all ata parameters.

Results: The proton radiotherapy plans resulted in significantly lower doses delivered to the kidneys, liver, small bowel, and spinal cord. This finding was seen consistently across the parameters analyzed in this study.

Conclusions: For patients receiving CRT, the proton plans are technically feasible and dosimetrically appealing with superior OAR sparing. Proton radiotherapy may improve the therapeutic ratio for patients receiving CRT for pancreatic cancer.

## P6: Proton therapy reduces perioperative cardiopulmonary complications in esophageal cancer treatment

### Loma Linda University

Background: Radiation-induced perioperative complications are frequently seen in patients receiving multimodality therapy for esophageal cancer. The purpose of this study is to investigate methods of minimizing radiation-induced perioperative cardiopulmonary complications.

Materials and Methods: Ten patients with esophageal cancer treated with radiation therapy between 2010 and 2013 were included in this study. Three separate plans were generated for each patient: 3D proton plan, 3D photon plan, and IMRT plan. A dose of 50.4Gy given in 28 fractions was delivered to the PTV allowing 90% isodose coverage of at least 95% of the PTV. Dose distributions of proton, 3D conformal photon, and intensity modulated radiation therapy (IMRT) treatment plans in patients with esophageal carcinoma were examined, focusing specifically on dose reduction to cardiac and pulmonary structures. Dose-volume histograms were calculated and analyzed in order to compare plans between the three modalities.

Results: The 3D proton plans showed decreased dose to partial volumes of the entire heart, arteries, and left ventricle in comparison to both the IMRT and 3D photon plans. The IMRT plans showed decreased dose delivered to the LAD artery, pericardium, and left atrium in comparison to the 3D photon plans (See Table 1).

Conclusions: There was no difference between IMRT and 3D-CRT with regard to dose delivered to the heart and lungs. However, proton radiotherapy did demonstrate lower overall doses delivered to various cardiopulmonary structures. Proton radiotherapy may help reduce normal tissue toxicity and perioperative morbidity for patients receiving radiotherapy as part of their esophageal cancer treatment.

## P43: Early progress on a phase II study of proton chemotherapy for resectable esophageal cancer

### Loma Linda University

Background: Preoperative chemoradiotherapy appears to increase respectability, produce tumor downstaging, and improve local control, disease-free survival, and overall survival compared to surgery alone. Proton therapy can deliver equivalent radiation dose as photon techniques while reducing the dose of radiation to adjacent normal tissue has the potential to decrease radiation-related toxicity and still provide adequate coverage to the tumor and adjacent lymph nodes. A phase II trial was initiated at our institution to assess toxicity and efficacy of a regimen incorporating a proven systemic regimen, carboplatin /paclitaxel, with the conformal proton modality, followed by definitive surgery.

Materials and Methods: Patients with T1-3N0-1M0 squamous cell carcinoma or adenocarcinoma of the esophagus or esophagogastric junction that are deemed surgical candidates received 40 GyE in 20 fractions to the tumor and regional nodes with proton radiotherapy and concurrent carboplatin/paclitaxel chemotherapy. Early results from the first 5 patients enrolled are presented.

Results: Five patients were enrolled and completed proton chemotherapy without significant acute toxicities aside from mild esophagitis, fatigue and weight loss. Three patients underwent surgical resection 5-12 weeks after proton chemotherapy. One patient had evidence of distant metastases prior to surgery. One patient experienced significant improvement in clinical symptoms after proton chemotherapy that he refused definitive surgical treatment. All 3 surgical patients had R0 resections and all had partial pathological response. With a median follow up of 11 months for those that completed the entire study treatment, 1 had distant metastatic failure and all 3 remain alive.

Conclusion: Preoperative proton chemotherapy for locally advanced resection esophageal or esophagogastric junction tumors appears promising with the potential to reduce toxicity, particularly perioperative late toxicity to lungs and heart. Continued accrual on the clinical trial is expected to provide data on toxicity, efficacy and quality of life measurements.

## P45: Preoperative proton therapy for soft tissue sarcoma reduces risk of acute wound complications

### Loma Linda University

Background: Benefits of preoperative radiotherapy for extremity soft tissue sarcomas are often weighed against the risks of acute wound complications. Proton therapy with improved and more homogenous dose distribution may potentially reduce risks of acute wound complications with appropriate planning.

Materials and Methods: A phase II study of preoperative proton therapy for soft tissue sarcomas has been implemented at our institution delivering 50 GyE in 25 fractions in 5 weeks followed by definitive limb-sparing surgery within 8 weeks of completion of radiotherapy. A 61 year old male with a 5 x7 cm high grade pleomorphic sarcoma of the left anterior thigh was enrolled onto the trial. In order to reduce risk of acute wound complications, surgical assessment prior to preoperative radiotherapy provided planned surgical approach and potential reconstruction options to aid in radiotherapy treatment planning. Volumes of skin based on surgery input were used as dose avoidance structures. Steep dose gradients from the sharp penumbra created from conformal apertures of passively scattered proton fields were mainly used for skin sparing techniques.

Results: Two passively scattered proton fields (Figure 1) were employed with aperture editing to limit dose to surgical approach/reconstruction (light green) while delivering prescription doses to the target. The majority of the surgical approach volume is outside of the 5 GyE (purple) isodose line. The V20GyE and V5GyE for the surgical approach volume was less than 10% and 25% respectively. Patient did not experience any wound complications and had minimal fibrosis or skin changes. Photos from immediate post-op (Figure 2A), 25 days post-op (Figure 2B) and 11 weeks post-op (Figure 2C) show excellent healing and cosmesis.

Conclusions: Proton therapy has excellent dose distribution properties that may reduce the risk of acute wound complications with appropriate planning to minimize radiation dose to volumes of skin/soft tissue involved with the surgical approach and/or reconstruction.

## C5 Head & NeckP38: Proton Radiation Therapy for Sinonasal Mucosal Melanoma

### University of Florida Health

Purpose: To report results and toxicity of patients treated with high-dose proton radiotherapy (RT) for mucosal melanoma with definitive intent.

Materials and Methods: We conducted an IRB approved retrospective review of 7 adult patients treated for primary mucosal head and neck melanoma at the UF Proton Therapy Institute to determine their treatment outcomes. Seven patients with a median age of 60 years (range, 37-80) were treated with external-beam proton RT. Primary tumor locations were nasal cavity (5), ethmoid sinus (1), and maxillary sinus (1). Six patients underwent gross total resection prior to proton therapy and 1 patient had a biopsy due to unresectable disease from intracranial extension. In the postoperative setting, proton therapy was initiated with a median of 53 days after surgery (range, 33-85). The primary target volume was defined as surgical bed plus a margin for microscopic disease and setup uncertainties. Median proton therapy dose to the primary PTV HR was 74.4 CGE (range, 69.6-74.4). Six patients were treated with elective nodal radiation to the neck; median dose was 50 Gy (range, 48-50). No patient had concurrent or neoadjuvant chemotherapy. Complications were graded according to the CTCAE, v4.0.

Results: Median followup was 1.4 years from the end of RT (0.5-6.6 years). The cause specific survival and overall survival were equivalent: 83% at 1 year and 67% at 1.5 years. There were 3 patients who died of metastatic disease, and all died within 25 months after RT. There was local recurrence in a patient who was treated with 69.6 CGE at 2 months after completing RT, who also died of distant disease. The local-regional control rate was at 1 and 1.5 years was 86% while the freedom from distant metastases was 71% at 1 and 1.5 years. One patient experienced grade 2 epiphora and grade 4 visual toxicity, leading to loss of useful vision in the treated eye. One patient had brain necrosis found with brain metastases following RT that required steroids. Two patients report grade 2 sinusitis, 1 patient has a grade 2 oral cavity fistula, and 1 patient reports grade 2 sensorineural hearing loss. No contralateral visual complications occurred.

Conclusions: Proton therapy can be utilized to escalate dose in sinonasal mucosal melanoma while sparing uninvolved optic structures. Metastatic progression remains the dominant form of treatment failure. Our data supports that postoperative RT given above 70 CGE may help increase local-regional control in this disease.

## C6 PediatricP53: Prospective Study on Pediatric Patients with Atypical Teratoid Rhabdoid Tumors (ATRT)

### CDH Proton Center, Oklahoma City Procure Proton Therapy Center, Ann and Robert H Lurie Children's Hospital of Chicago

Purpose: ATRT is a rare and aggressive Central Nervous System (CNS) tumor usually presenting in very young children. Aggressive treatments have improved outcomes. Such strategies have included radiation therapy. However, at such a young age, short and long term radiation toxicities are prevalent. We prospectively enrolled pediatric CNS ATRT patients onto the Proton Collaborative Group registry protocol to evaluate the efficacy and toxicities of proton radiation therapy in this population.

Materials and Methods: 13 consecutive pediatric ATRT patients were treated with at the Central DuPage Hospital Proton Center and the Oklahoma City Procure Proton Therapy Center between March 2010 – December 2013 utilizing 3D Conformal Proton Therapy.

Results: 13 patients were evaluated. They were all 3 years of age or younger (4.4 – 37.7 months). Eight patients had gross total resections, while 4 had subtotal resections along with another 1 not documented. Nine patients received multiagent intensive chemotherapy per the Dana Farber Cancer Institute regimen while 4 had treatment either on or per ACNS 0333 protocol with intensive multiagent chemotherapy along with stem cell transplants. Radiation was to local fields for 10 patients, while 3 had craniospinal irradiation. The mean follow up was 14.9 months (range of 1-43 months) and median follow up 14.2 months. At last follow up, 11 patients were alive without evidence of disease. Only 4 children had grade 3 toxicities (all acute nausea, vomiting and anorexia during radiation therapy that responded to steroids). Proton therapy was able to reduce the dose to the cochlea, optic chiasm, hippocampus, temporal lobes and integral whole brain.

Conclusion: The initial results on the largest prospective series of CNS ATRT patients treated with proton therapy seem to be favorable. The aggressive treatment regimens utilizing proton beam therapy yield proven efficacy and improved toxicity profiles, which is critically important in this young patient population with such an aggressive disease.

## P82: Outcomes and acute toxicities of proton therapy for pediatric atypical teratoid/rhabdoid tumor of the CNS

### MD Anderson Cancer Center, Houston, TX; Department of Pediatrics, Baylor College of Medicine, Houston, TX

Purpose: Atypical teratoid/rhabdoid tumor (AT/RT) of the central nervous system is a rare cancer primarily affecting children younger than age five. Because patients are young and receive intensive chemotherapy, there is concern regarding late radiation toxicity, particularly as survival rates improve. Therefore, there is interest in using proton therapy to treat these tumors. This study was undertaken to investigate outcomes and acute toxicities associated with proton therapy for AT/RT.

Materials and Methods: The records of 31 patients with AT/RT treated with proton radiation from October 2008 to August 2013 were reviewed. Demographics, treatment characteristics and outcomes were recorded and analyzed.

Results: Median age at diagnosis was 19 months (range, 4 – 55 months), with median age at radiation start of 24 months (range, 6 – 62 months). Seventeen received local radiation with median dose of 50.4 GyRBE (range, 9 – 54). Fourteen received craniospinal radiation; half received 24 GyRBE or less and half received 30.6 GyRBE or higher. For patients receiving craniospinal radiation, the median tumor dose was 54 GyRBE (range, 43.2 – 55.8). Twenty-seven (87%) completed the planned radiation. With median follow-up of 24 months for all patients (range, 3 – 53 months), median progression-free survival was 20.8 months and median overall survival was 34.3 months. Five patients (16%) developed clinical findings and imaging changes in the brainstem one to four months after radiation consistent with radiation reaction; all resolved with steroids or bevacizumab.

Conclusions: This is the largest report of children with AT/RT treated with proton therapy. Preliminary survival outcomes in this young pediatric population are encouraging compared to historic results, but further study is warranted.

## P84: Neurocognitive Outcomes in Relationship to Hippocampal and Brain Doses after Partial Brain Proton Radiotherapy in Children

### MD Anderson Cancer Center, Houston, TX; Department of Pediatrics, Baylor College of Medicine, Houston, TX

Objectives: We evaluated the relationship between IQ and hippocampal RT doses in pediatric patients with brain tumors treated with proton radiotherapy (PRT).

Methods: IQ scores were abstracted for patients treated with partial brain PRT who underwent serial neurocognitive testing. Hippocampal delineation was performed according to RTOG guidelines. The PRT dosimetric data were compiled. The relationships between IQ and 1) median hippocampal doses and 2) percentage of hippocampus receiving 50, 30 and 10 GyRBE were evaluated using general linear mixed models while controlling for mean brain-GTV dose, CTV volume, tumor location, and age-at-RT.

Results: Data were collected for 25 patients (13 males, 12 females) receiving PRT from 2007-2012. The mean interval between first-last IQ testing was 2.1 y (range 0.9-6.0). The median age-at-PRT was 9.4 y (range 1.7-15.4 y). Tumors were supratentorial in 17 patients (9 glial, 5 craniopharyngioma, 2 germinoma, 1 meningeal tumors) and infratentorial in 7 patients (3 anaplastic ependymoma, 2 medulloblastoma, 2 glial tumor). 17, 2 and 6 tumors were midline, right and left, respectively. The median dose was 50.4 GyRBE (range 45.0-60.0 GyRBE). The median CTV volume was 43.3 cc (range 12.3-234.4 cc). The median mean brain-GTV dose was 9.4 GyRBE (range 3.9-19.7 GyRBE). The median mean right and left hippocampal doses were 17.0 and 16.3 GyRBE (range 0.2-54.8 and 0.0-46.2 GyRBE), respectively. After controlling for the factors mentioned above, patients who received any right hippocampal dose >50 GyRBE exhibited significant IQ decline (6.4 points/y, p=0.01). Patients who received 30 GyRBE to ≥10% of the right hippocampal volume also experienced significant IQ decline (4.0 points/y, p<0.01), while patients with less right hippocampal dose/volume maintained stable IQ over time (p=0.67). Left hippocampal dose/volume was not associated with IQ decline in this cohort.

Conclusions: Preliminary findings suggest right hippocampal dose is associated with steeper IQ decline post-PRT in this cohort of patients.

## C7 SarcomaP44: Early report of a phase II trial for preoperative proton therapy in soft tissue sarcomas

### Loma Linda University

Background: Preoperative radiotherapy for soft tissue sarcoma is a standard option with benefits of lower radiation doses and smaller volumes that may lead to reduced long term chronic side effects and extremity function when compared to postoperative radiotherapy options. Proton therapy with superior and more homogenous dose distribution has the potential to further reduce risk of chronic side effects and possibly reduce acute wound complications.

Materials and Methods: Patients with histologically proven primary soft tissue sarcoma of the extremity or body trunk (excluding retroperitoneum) that were deemed resectable by an oncologic surgeon and no evidence of regional or distant metastases were enrolled on a phase II study for preoperative proton therapy. Patients were treated with 50GyE in 25 fractions in 5 weeks followed by definitive limb-sparing surgery within 8 weeks of completion of radiotherapy. Radiotherapy target volumes included gross tumor with a 2cm longitudinal and 1cm radial margin for clinical target volumes along with appropriate expansions for setup uncertainty and proton beam specific range and penumbra uncertainty. Discussion regarding surgical approach and reconstruction techniques allows potentially limiting radiation dose to these areas to reduce the risk of acute wound complications.

Results: Three patients with a median follow up of 7 months have been enrolled on the trial to date. Acute toxicities have been mild with 2 patients with grade 2 skin erythema, 1 patient with grade 1 skin erythema, 1 patient with grade 1 edema and 1 patient with grade 1 fatigue. One patient had a positive margin and underwent a second surgery to obtain clear margins. This patient did have a seroma at the surgical site after 2nd surgery and had wound complications. At last follow up, all 3 patients had no evidence of local or regional disease recurrence.

Conclusions: Proton therapy for preoperative sarcoma is safe with minimal toxicity with good local control with early follow up. Surgical mapping and dose limitation may reduce risk of wound complications. Late chronic side effects manifest years after therapy necessitating longer follow up. Further investigation is warranted.

## C9 OtherP9: Clinical Applications for the Use of an Innovative Proton Treatment Chair

### CDH Proton Center

Background: The aim of this presentation is to introduce the audience to the treatment chair. This will include specifications, indications for use, and the advantages and disadvantages of the chair. It will also discuss the processes and workflow the CDH Proton Center has implemented for using the treatment chair; including simulation, planning, and treatment. Furthermore, the presentation will describe the treatment of a patient using the treatment chair versus a patient treated in the lying position; comparing patient comfort, positioning, planning considerations, and net treatment time. Finally, the presentation will discuss the future direction of the treatment chair's use and development.

Materials and Methods: The CDH Proton Center uses a treatment chair in conjunction with a Fixed Beam or Incline Beam room to treat intra-cranial and ocular targets in the seated position as an alternative to the standard supine or prone treatment positions.

Results: Using the treatment chair has given the CDH Proton Center the ability to treat intra-cranial targets in the seated position for patients who cannot lie flat for treatment. It also gives the option of utilizing the inclined beam delivery system rather than the gantry room, which is commonly reserved for more complex cases.

Conclusions: The CDH Proton Center uses a proton treatment chair as a viable way to treat intra-cranial tumors. The CDH Proton Center has also created processes and a workflow for the use of the chair to better refine and improve the overall patient experience, and allow more patients access to proton therapy.

## P26: Evaluation of Standard Beam Delivery Devices in Proton Intracranial Radiosurgery

### Loma Linda University

Background: To evaluate the use of standard apertures and range shifters for the treatment of brain metastasis in proton stereotactic radiosurgery.

Materials and Methods: Five localized brain metastasis patients previously treated using our intracranial proton stereotactic radiosurgery procedure (i.e. with a custom aperture and bolus), were randomly selected from our patient cohort. The custom aperture and bolus treatment plans were used as the standard of care in this case and comparative treatment plans using the standard aperture and range shifter concept were generated. Gantry/table angle and the number of treatment beams were optimized as part of this study to evaluate the ability of the standard aperture/range shifter system to deliver a comparable treatment to the patient. Conformity index, homogeneity index, isodose volumes and integral dose were all evaluated to determine the degree of conformity of the plans created and for comparison to the custom aperture/bolus treatment modality.

Results: The generated treatment plans demonstrated that the standard aperture and range shifter combination could be used to produce comparable conformity index and isodose volumes to the custom aperture/bolus case in four out of the five patients studied. In two of the patients a comparative conformity index was achieved by optimizing the angles of the 3 treatment beams, while in two of the cases 1 or 2 additional beams were required. Additionally, this system exhibited efficiency gains over the custom aperture bolus system in reducing the time necessary for treatment planning, device manufacture and QA.

Conclusion: This work demonstrated that largely spherical shape of brain metastasis makes this target well suited to an application of standard apertures, while additionally providing efficiency gains in device manufacture and QA for treatment.

## P58: Complications of Endoscopic Interventions for Prostate Cancer Patients Following Radiation Treatment

### CDH Proton Center

Background: A future standard in proton radiation therapy for prostate cancer patients will be to verify that a patient has had a screening colonoscopy at least one year prior to radiation. Post treatment prostate cancer patients have between a 5% and 20% chance of experiencing symptoms related to Chronic Radiation Proctitis. A baseline colonoscopy allows for a particular level of discrimination when evaluating if a patient is experiencing symptoms of mild proctitis or a more problematic gastrointestinal disease process post treatment. This particular case study evaluates three different patients that experienced mild to moderate symptoms of radiation proctitis and rectal hemorrhage, proceeded to have a colonoscopy with unnecessary biopsies or cauterization done, and as a result experienced more intense symptoms within the following months.

Materials and Methods: Three prostate cancer patients were studied. The first patient was a 64 year old male whom experienced symptoms 9 month after cessation of treatment. He had a flexible sigmoidoscopy with cauterization which led to increased tenesmus and BRBPR. The second patient was a 70 year old male whom experienced symptoms 8 months post treatment, which prompted a colonoscopy with biopsies of an anorectal ulcer. At 12 months the patient experienced greater pain and bleeding with concerns that a fissure was developing at the site of the endoscopically biopsied area. The final patient was a 57 year old man whom experienced symptoms at 16 months and proceeded to have a colonoscopy with biopsies of an anterior rectal wall ulceration. Since the colonoscopy, he continued to experience increased rectal hemorrhage.

Results: It has become increasingly apparent that those who had areas of radiation proctitis or rectal ulcers biopsied or cauterized during colonoscopies experienced more severe symptoms after having the procedure completed.

Conclusion: Open communication between the gastroenterologist and radiation oncologist is an important component in preventing radiated areas from being unnecessarily biopsied or cauterized. When this dialogue is initiated early in the process and involves the patient, radiation oncologist, gastroenterologist and registered nurse, we have a greater chance of preventing future complications. Confirming that the patient has a good understanding of symptoms to watch for, when to consult a gastroenterologist, and the appropriate conversation that needs to take place is crucial.

## P1 Beam CalibrationP24: Evaluation of the dosimetric properties of a diode detector to proton radiosurgery

### Loma Linda University, San Diego State University

Background: Small fields typically encountered in proton radiosurgery require high spatial resolution dosimetric measurements, especially below 1-2 cm diameters where the limits of standard metrology devices are exceeded. Radiochromic film provides high resolution but exhibits a LET dependent response and requires post processing and special handling. Promising alternatives are diode detectors with small sensitive volumes (SV) that are capable of high resolution and real time dose acquisition. In this study we evaluated the PTW PR60020 proton dosimetry diode in radiation fields relevant to radiosurgery applications.

Methods and Materials: Beams with energies of 127 and 157 MeV and initial diameters of 8 and 20 mm were delivered using single-stage scattering and three modulations (0, 15, and 30 mm) to a water tank in our Gantry 1 treatment room. Precursory experiments evaluated diode response as a function of dose, dose rate, and diode orientation (ie, diode positioned with beam parallel to (axial mode), or perpendicular to detector axis (edge-on mode)). Depth dose profiles were measured using the diode in axial mode or a PTW Markus N23343 ion chamber. Transverse dose profiles were measured using the diode in edge-on mode or EBT2 film. Finally, analogous Monte Carlo simulations were performed using Geant4.

Results: Diode response was linear with respect to dose, uniform with dose rate, and essentially independent of orientation. Close agreement of ion chamber (‘gold standard') and diode depth dose profiles suggests diode LET dependence was negligible. Similarly, plots of the diode to Markus dose ratio vs dose-weighted lineal energy suggest that diode response is essentially LET independent downstream and over the SOBP of the Markus depth dose curve. Also, while not possible with the ion chamber, accurate diode dose measurements of 8 mm diameter beams were obtained. Finally, in edge-on mode data was acquired with spacing as low as 0.25 mm in the penumbra of transverse profiles.

Conclusion: The PR60020 diode response was linear with dose and dose rate, and not dependent on orientation. Comparison with ‘gold standard' ion chamber data suggests a negligible LET dependence over particle ranges relevant for clinical radiosurgery. The small SV allows measurements of small fields without partial volume effects. Edge-on, the diode is capable of sub-millimeter resolution (on par with film) that is essential for small fields and high dose gradients (eg, penumbra, distal edge).

## P32: Application of a Pixelated Silicon Detector to Proton Radiosurgery

### Loma Linda University, San Diego State University, University of Malaya, University of Wollongong

Background: To evaluate the applicability of a pixelated silicon detector known as the dose magnifying glass (DMG) to small field proton dosimetry and compare its performance against commercially available metrology techniques and Monte Carlo.

Materials and Methods: Proton radiosurgery requires metrology apparatus that have a high spatial resolution and a stable (or well characterized) response to LET. To meet this need the DMG was developed at the University of Wollongong. This device is a pixelated silicon strip detector comprising an array of 128 phosphor implanted n+ strips on a p-type silicon wafer. A 100 micron pitch device was tested at Loma Linda University Medical Center with proton radiation energies associated with radiosurgery applications. For these tests the DMG was mounted in a water tank with depth dose and lateral profiles collected simultaneously using the linear sensitive volume array for comparison with commercial ion chamber and diode detectors as well as Geant4 simulation results. The DMG response to varying dose rate was also evaluated and reported.

Results: The DMG performed well in these tests, providing real-time depth dose and lateral profile information. When compared with data from commercial measurement systems and Monte Carlo, agreement as a function of depth was very good. These results indicated that the device is largely LET independent in its response to proton radiation indicating potential for deployment in this radiation modality. Using the central channel of the DMG allowed for accurate point dose measurements without errors associated with partial volume sampling, while the full array could be used to provide real time profile information of fields below 1.5 cm diameter without the need for detector scanning. The DMG also exhibited a uniform response to dose rate up to 12 Gy/min, with under response noted for dose rates above this value.

Conclusion: The DMG is a useful device for proton therapy, in particular providing real-time data for small fields associated with proton radiosurgery applications. The device exhibited minimal variations in response as a function of LET and presents an interesting direction for development of real time dosimetry of both depth and lateral dose profiles. Future developments of the detector to provide larger linear detector arrays could widen the application of such a device to large field proton dosimetry and also beam characterization for pencil beam scanning applications.

## P2 Image GuidanceP34: A Simple Method For Evaluating Setup Tolerances On A 6D Couch

### Procure New Jersey

Background: The use of 6D robotic couch can make it difficult to determine if the patient is within the PTV margin due to the complicated relationship between rotations and the translations. Our goal was to provide a simple method help therapists know if they are within the PTV margin.

Methods: Therapists are able to obtain from the IGRT software the magnitude of the displacement vector V = where x, y and z are the displacements between the current setup and the target setup in the orthogonal coordinate system. A rotation is approximated by a displacement of rPTVθ where r is the maximum radius of the PTV and θ is the angle of rotation in radians. However, a more conservative and simplifying assumption is to take rs as the snout radius. Also, since there are uncertainties in the actual alignment between the x-ray, proton and robot isocenters, we approximate this value as σiso=1 mm.

Results: To be within the margin, PTV>. By keeping rs as a constant for each snout, we can generate charts that V as a function of .The resulting straight line shows when the therapists need to shift and when the patient is within the PTV margin. To simplify the procedure, therapists take the largest rotation in pitch, yaw, and roll to be operationally.

Conclusions: We have generated a simple graphical relationship that allows therapists to determine quickly if a patient is within the setup tolerance. This procedure can be generalized to conventional radiation therapy applications as well.

## P63: Pro and Con about the Surface Image Application in Proton Therapy

### Procure Proton Therapy Center

Background: The surface imaging system mainly refer to the Align RT from Vision RT Ltd. United Kingdom. It has been used around dozen photon and proton centers in North America. Regarding the workflow, different center has different way to implement the application into its own clinic. After reviewing about fifteen institute application worldwide, four methods are proposed in our center per our proton center installation and clinical application. The reliability of each method was discussed.

Materials and Method: The four methods are IGRT (CT image as reference guided therapy) in which we imported the DICOM CT simulation image of one rigid phantom into the Align RT system. Three fields were planned in the XIO treatment planning system. The six dimension couch shift data were collected corresponding to three different beam positions. The comparison with orthog-X rays predicted couch shift data has been performed. SGRT (Setup position image as reference image guided therapy) is the method where the reference image was acquired while the phantom was at initial setup position. Then Vision RT predicted couch shift data was analyzed at three different beam positions. TGRT (Treatment position image as reference image guided therapy) is the one in which three reference images were acquired respectively at three different couch position on the day one treatment. MGRT (Monitoring image guided radiation therapy) is the method in which the reference image was acquired at each daily beam treatment position. During treatment the real time tracking infrared light was turned on. The Align RT system will automatically tracking the offset on time.

Results: All the results are based on the ROI which are created per Align RT recommendation. The offset data depends on the size of ROI. The part of our results are listed in table 1-3.

Conclusion: Align RT at current status is reliable only on the MGRT. It is not reliable to be implemented into our clinical application to perform IGRT, SGRT, TGRT procedure.

Table 1.

Six dimension data shift for nine fractions for beam 1.

Table 2.

Six dimension data shift for nine fractions for beam 2.

Table 3.

Six dimension data shift for nine fractions for beam 3.

## P4 Quality AssuranceP64: Implementation and Quality Assurance of Cranial Spinal Irradiation with a Compact Proton Machine

### Washington University in St. Louis, Barns and Jewish Hospital

Introduction: Craniospinal irradiation (CSI) is technically challenging due to the complex target volume. Proton therapy offers significant clinical benefits. We have recently implemented proton therapy for CSI with a single room compact proton machine at our institution. This is the first implementation of CSI using a single-room proton system with a 190 degree rotating gantry and 6-degree robotic couch. Implementation of proton for CSI entails numerous QA issues including dosimetric measurements, range, and dosimetric gap verification. This study reports our experience with the implementation and QA of CSI.

Materials and Methods: Four consecutive pediatric patients requiring CSI were treated with proton therapy. Patients were immobilized supine with a thermoplastic mask and treated using a Qfix BoS table. The planning CT was performed with 1.5 mm slices. Treatment beams include two opposed lateral oblique cranial fields and two or three PA spinal fields, depending on the patient length. Every five fractions, a junction shift was applied and junction shifts were typically ~1 cm apart. A perfect matching of brain fields with an upper spine field can be achieved by matching the penumbra by manipulating air gap and/or compensator thickness. Daily setup verification was with orthogonal KV images at the level of the brain, upper, middle and lower spine on a daily basis. To confirm proper field matching, radiochromic films were placed at the junction locations. The field matching was measured in both coronal and sagittal planes. Dose at junction area was measured using a 2D planar ion chamber array. The distal-fall off with compensators for each beam was verified using a separate radiochromic film within plastic water.

Results: Agreement of dose distributions in plan and delivery was observed with gamma passing rate larger than 95%. For junctions in the overlap region, the difference in width between plan and measured was less than 1 mm, with maximum difference up to 8% of prescription dose. This indicated up to 2% of uncertainty of dose in junction region when washed out with three additional junction shifts. The range difference between measured and shown in the treatment planning system was within 2 mm. One daily treatment with 2 lateral brain fields and 3 spine fields was completed within one hour.

Conclusion: A technique for delivering craniospinal irradiation to supine patients has been developed on a compact proton machine.

## P80: Integrated High-Resolution 3D QA System for IMPT Using Laser CT of Reusable Polymer Gel Dosimeters

### M.J. Maryanski, L. Lin, A. Kassaee, S. Avery

Background: Tight margins, steep dose gradients, and LET effects, typical in IMPT, challenge most dosimeters. Even small spatial errors may result in significant departures from prescribed doses in PTV and/or OAR regions, potentially leading to complications and/or recurrences. Hence, there is an urgent need for high-resolution 3D pre-treatment QA systems, capable of detecting errors anywhere in the treatment volume, preferably without decreasing patient throughput or increasing the costs. One candidate QA system (CrystalBall™ from MGS) is based on computerized laser CT of spherical soft-tissue-equivalent polymer gel dosimeters whose optical density increases with local dose, with LET - independent dose-response. The most recent development is a re-usable polymer gel dosimeter, in which the dose image spontaneously fades away within 12 hrs, allowing for the dosimeter to be re-used many times over, thereby greatly reducing the costs of patient QA.

Method: A user-friendly new model of the laser CT scanner was placed at the Roberts Proton Therapy Center, within one minute walking distance from the treatment rooms. A PBS nozzle developed by IBA was used for proton irradiation, with nominal energies between 103 and 140 MeV. All dosimeters are air-tight sealed inside molded Pyrex glass spheres, transparent red-colored for light-protection, with 2.5 mm thin wall. A special acrylic/polyurethane phantom, with 3D image-registration and image-guidance markers, holds the spheres in one unique, reproducible position. A 3D QA analysis software running on Windows 7 features DICOM import, 3D image registration, dose calibration, QA analysis, and a summary QA report spreadsheet automatically generated in Excel. A standard PBS plan was designed for the system test and delivered to several permanent-image and re-usable dosimeters as well as to the DigiPhant™ phantom (from IBA), for comparison.

Results: The overall per-patient QA time was found to be less than 1 hr including the combined setup/beam time under 10 minutes, the laser CT scan, data analysis and report generation. The QA measurement covers a spherical phantom volume up to 200 mm diameter. Median dosimetric accuracy was 99%. Typical isotropic spatial resolution is 1 mm.

Conclusion: An economical, integrated system for high-resolution 3D QA for IMPT has been developed and tested. Mailed QA service provided by the manufacturer is a practical option when on-site laser CT scanning resources are unavailable.

## P83: Quality assurance of scanned proton beam delivery systems with a 3D liquid scintillator detector

### MD Anderson Cancer Center

Background: Scanned proton beams present challenges for commissioning and quality assurance (QA) measurements. Ideally, the lateral profile, depth dose distribution, and scanning magnet accuracy at each beam energy should be measured initially as a part of commissioning and then regularly as a part of the QA program. However, in practice only a subset of these measurements can be performed due to the large number of beam energies, busy clinical schedules, and the limitations of current detectors. Liquid scintillator-based volumetric detectors have the potential to overcome these limitations. They exhibit high resolution, fast response, and immediate readout over a large detection volume.

Methods: The detector used in this study consists of a 20×20×20 cm3 cubic tank of scintillator, two CCD cameras, a light-tight housing, and a mirror to redirect the light from one face of the scintillator tank to one of the cameras (Fig. 1). Proton beams with 60 different energies ranging from 72.5 MeV to 168.8 MeV were measured in a single delivery sequence, requiring less than 3 minutes of beam time. A 3×3 grid of proton pencil beam spots (PPBS) with 5 cm spacing at 5 different energies was also measured.

Results: The mean of the differences between the measured and nominal beam ranges was 0.10 mm, with a standard deviation of 0.11 mm. Figure 2 shows the measured full width at half-maximum (FWHM) of 60 proton beams, compared with data measured with a pinpoint ionization chamber in a water tank. The FWHM values followed the same trend as the ionization chamber measurements, but were 1 mm narrower, possibly due to decreased detector size effects. The differences between the nominal lateral beam positions and the measured positions ranged from 0.22 mm to 2.64 mm, with the errors increasing with distance from the center of the detector. The standard deviation of the position measurement at each beam location was less than 0.37 mm.

Conclusions: The novel detector described in this study can provide fast, high-resolution measurements of the beam range, lateral profile, and lateral position for therapeutic scanned PPBS. The ability of the detector to rapidly measure the beam range and lateral profile of many PPBS makes it particularly promising as a tool for scanned proton pencil beam QA. This measurement efficiency could save valuable time in busy proton therapy clinics and facilitate more comprehensive commissioning and QA measurements than current techniques can provide.

## P5 ShieldingP25: Dose delivered to gantry-mounted electronics in proton therapy

### Loma Linda University

Background: To evaluate the scattered and secondary radiation fields present in and around a passive proton treatment nozzle to assess any electronic complications that may arise.

Materials and Methods: Landauer Luxel dosimeters were used to evaluate the radiation field around one of the proton passive scattering nozzles at Loma Linda University proton therapy center. These detectors use optically stimulated luminescence technology in conjunction with CR-39 to measure doses from X, gamma, proton, beta, fast and thermal neutron radiation. The dosimeters were stationed at various positions around the gantry pit and on racks on the gantry itself to evaluate the dose to electronics. Wax shielding was also employed on some detectors to evaluate the usefulness of this material as a dose moderator. To create the scattered and secondary radiation field in the gantry enclosure, a polystyrene phantom placed at isocenter and irradiated with 250 MeV protons to a dose of 1.3 kGy over 16 hours.

Results: The measured dose equivalent ranged from 100-6000 mrem, with proton/photon, thermal neutron, fast neutron and overall dose equivalent evaluated. The position of the detector/electronics relative to both isocenter and also neutron producing devices such as the collimators, first and second scatters definitely had a bearing on the dose received. Interestingly the addition of 1 inch thick wax shielding decreased the fast neutron component by almost 50%, yet this had a corresponding increase in thermal neutron dose of 100% and a 50% increase in photon/proton dose as there was no B-10 component to capture thermal neutrons.

Conclusions: The data obtained in this study will benefit future upgrades and facility designs by identifying mounting positions for electronics that minimize radiation dose.

## P6 Treatment PlanningP70: Comparison of two different planning techniques for IMPT of Head-and-Neck cancers

### M. D. Anderson Cancer Center, Chang Gang Memorial Hospital

Purpose: To evaluate plan quality two beam planning techniques for intensity modulated proton therapy (IMPT) for oropharynx and nasopharynx cancer patients.

Methods and Materials: Three oropharynx and three nasopharynx cases were included in this preliminary study. The method A included three fields with right-anterior-superior-oblique(RASO), left-anterior-superior-oblique(LASO), and posterior(PA). The method B used three fields with right-posterior-oblique(RPO), left-posterior-oblique(LPO), and anterior(AP). In method A, each of the three fields covers the entire target volume. For method B, the AP field covers only supra-clavicle nodes and the two posterior oblique fields cover the upper target from middle neck. The identical optimization constraints were used for both methods. With the same target coverage, the conformity index(CI), homogeneity index(HI) and doses to critical organs were compared.

Results: For oropharynx cases, the target coverage was more conform and the dose was more homogenous in the method A plans in comparison to Method B (average CI: 0.435 vs. 0.347 and average HI: 0.076 vs 0.099). However, the mean doses to some critical organs were higher in the method A plans than Method B (the right parotid gland: 27.4 vs 25.6 Gy (RBE), left parotid gland: 26.7 vs. 21.8 Gy (RBE), oral cavity 25.1 vs. 21.0 Gy (RBE), and larynx 30.0 vs. 29.6 Gy (RBE)). On the other hand, the maximum dose to the spinal cord and brain stem were lower in the method A plans compared to method B (34.6 vs 38.4 Gy (RBE) and 34.0 vs. 38.7 Gy (RBE)). For nasopharynx cases, the target coverage was slightly more conform and the dose was significantly more homogenous in the method A plans than Method B (average CI: 0.355 vs. 0.340 and average HI: 0.077 vs 0.140). The mean doses to some critical organs were slightly lower in the method A plans than Method B (the right parotid gland: 30.3 vs. 31.6 Gy (RBE), left parotid gland: 25.1 vs. 25.7 Gy (RBE), oral cavity 15.1 vs. 18.3 Gy (RBE), and larynx 20.8 vs. 21.8 Gy (RBE)). The maximum dose to the spinal cord and brain stem were significantly lower in the method A plans compared to method B (spinal cord: 33.0 vs. 42.0 Gy (RBE) and brain stem: 53.0 vs. 57.4 Gy(RBE)).

Conclusion: For nasopharynx, the method A displayed some advantages in normal tissue sparing in comparison to the method B. For oropharynx patients, the method B had lower doses to critical organs than method A.

## P71: Prostate treatment plan comparisons for Protons, TomoTherapy and Cyberknife

### University Hospitals Case Medical Center, Rhode Island Hospital, St. Elizabeth's Hospital

Background: A Total of 28 cases of previously treated patients with prostate CA using helical Tomotherapy (HT) (22 patients) or Cyberknife (CK) (6 patients) treatment machines were selected chronologically between the years of 2009 and 2011 and replanned using parallel-opposed beams for Proton Therapy (PT).

Materials and Methods: Proton data used an IBA double scattering beam model that was made available from the Philips Corporation as a pre-510K approved Pinnacle treatment planning system. Comparison for the coverage of the PTV (expanded CTV) and doses to bladder, rectum and femoral heads were used to evaluate the plans. RTOG 0815 (prostate IMRT) and 0938 (prostate SBRT) were used as an independent “benchmark” for plan robustness. No retrospective re-optimization was attempted for the HT and CK treated cases to meet these criteria.

Results: In a head-to-head comparison, PT planning results showed greater uniformity in coverage of the expanded CTV since CK planning criteria customarily required the prescribed dose to the 75 – 85 % isodose line (IDL) and some HT cases were prescribed to a mean dose to the PTV at 95% coverage where as PT has been prescribed to the 98% IDL as specified in the RTOG protocols. It was found that in the group of 18 eligible patients, 84 % of plans were able to meet RTOG protocol normal tissue dose constraints for all 3 modalities. Among the reasons for failure for the remaining 10 plans to meet the RTOG prescription dose criteria was due to: 1) original dose prescription was not eligible for RTOG inclusion, 2) ranging out of the protons in larger patients since the planning data set provided did not have high enough energy protons (Bragg peak max – 28.0 g/cm 2) or 3) large overlap between expanded CTV and rectum and bladder from a lateral perspective.

Conclusions: Summary findings indicated that for the patients selected, RTOG benchmark for inclusion was met for a substantial majority for all 3 modalities. Conformality indices for the prescription dose/target volume for HT and CK were marginally but consistently superior compared to parallel-opposed passive scatter PT plans. It is hypothesized that PT superiority in targeting is resultant from the beams rapid distal fall-off whereas increased conformality for HT and CK treatment plans is attributed to the 360 degree helical delivery and out-of-plane convergent beam configuration respectively. Additionally, the use of a rectal balloon may increase conformality for some plans.

Support data – Wessels et. al – Prostate treatment plan comparison for Proton, TomoTherapy and Cyberknife Radiation Therapy

Tables 1 a and b – Comparison of HT and Proton Therapy plans according to RTOG 0815 guidelines

## P74: Dosimetric impact of dose calculation grid size in proton planning based on head phantom study

### ProCure Proton Therapy Center

Purpose: To investigate the dosimetric impact of different dose calculation grid size in a brain proton plan based on IROC head proton phantom and compare with the TLD measurements.

Materials and Methods: In this study, we have utilized the anthropomorphic head proton phantom provided by IROC. The imaging, treatment planning, and beam delivery were carried out following the guidelines provided by IROC. First, the computed tomography (CT) image of the phantom with imaging insert was obtained. The imaging insert was used to delineate the target structure (simulating a brain tumor). Proton plan was generated in XiO treatment planning system using three equally weighted fields (vertex, right lateral, and left superior oblique) with an objective of achieving: (1) total dose of 6 Gy (RBE) to at least 95% of the gross target volume (GTV) (4.05 cc), (2) minimum dose of 5.1 Gy (RBE) to the GTV, and (3) maximum dose of 6.6 Gy (CGE) to 0.03 cc of the GTV. Proton dose calculations were performed using dose calculation grid size ranged from 1 to 3 mm, with an increment of 0.5 mm. For the proton beam delivery, the imaging insert was replaced by the dosimetric insert, which contained TLD capsules at 2 locations (right and left) near the center of the target. The phantom with the dosimetric insert was irradiated to approximately 6 Gy (RBE) using uniform scanning proton beams. The irradiated TLDs were read by the IROC. The calculated doses from the XiO were compared to the measured TLD doses provided by the IROC.

Results: For the right TLD, on average, the calculated dose was higher than the measured dose by 1.49% (range, 1.42% to 1.56%). For the left, TLD, the calculated dose was higher than the measured dose by 2.63% (range, 2.55% to 2.68%). For both the right and left TLDs, there was no distinct trend exhibiting the dependence of grid size on dose calculation accuracy when calculated dose of different grid sizes (1, 1.5, 2, 2.5, and 3 mm) were compared to the measured TLD doses (±3% uncertainty)

Conclusion: The results based on our phantom study showed that XiO might overestimate the dose, and the difference between the calculated dose of various grid sizes and measured dose from TLDs was within 3%. The difference in the calculated dose among the smallest grid size (1 mm) and the largest grid size (3 mm) was typically less than 0.1%.

## P7 OtherP12: 3D Patient Setup Assistant Tool for Proton Therapy Treatment Planning

### University of Texas M.D. Anderson Cancer Center

Background: Proton therapy system has many bulky movable components of gantry, snout and couch. Due to large variety of the in-room components (the snout extension, couch and compensator) and patient sizes, it is necessary to check clearance and confirm geometrical availability of a particular planning setup. Unfortunately, most treatment planning systems do not have the information of all objects in the treatment area. To avoid interference, in-room checks are performed after daily treatment is finished. However, this is labor intensive and prolongs the planning process.

Materials and Methods: We developed an interactive 3D patient setup assistant tool for treatment planning. The code is based on C#. The software renders all objects (3D models for compensator and energy absorber are also prepared as well as common components such as tables, couches and nozzles). The hardware objects are defined in a popular 3D data exchange format that is used by many CAD tools. The patient surface is obtained from the 3D CT data that are imported together with the treatment plan, which contains the information of accessories and plan isocenter. The software automatically aligns the plan isocenter to the machine isocenter. All parameters (e.g. couch position, gantry angle, patient position) can be manually adjusted. The program calculates the smallest clearance between patient/table and snout, and is able to project an outline of the each field. In addition, the program warns the user about machine limits.

Results: The viewer shows the location and amount of the smallest clearance for each field in the plan. User can virtually simulate the patient setup in the 3D viewer and can therefore immediately see which gantry or couch angle require a certain snout position. This gives the dosimetrist or physicist a quick feedback for optimal available beam angles and improves the overall treatment plan. The equipment parameter and treatment isocenter can be changed individually to see which configuration allows a safe treatment. Using this tool reduces unnecessary feedback time for checking physical setup in some cases.

Conclusions: Interactive 3D patient setup assistant tool helps to check the feasibility of patient and equipment setup without specific knowledge about the equipment dimensions and motion. It greatly reduces the feedback time for checking the planning setup and can improve treatment plans. The software framework can be easily adapted to alternative accelerator manufactures.

## P27: wbeam: A User Friendly Web-based Interface for Monte Carlo Proton Beam Computer Simulations

### Loma Linda University

Background: Monte Carlo computer simulation (MCCS) of the passage of protons through patient treatment nozzles and tissue allows an accurate prediction of the radiation dose deposited in normal and target tissues. This is useful in research, but because such simulations can provide a rigorous verification of planned dose distributions produced by treatment planning software, there are potentially important clinical benefits. However, even in the research environment technical barriers tend to limit the routine use of MCCS except for the case of skilled programmers. The purpose of this project is to provide an intuitive user interface to our existing MCCS system that will allow non-programmers to perform simulations of protons delivered in the treatment rooms at our institution.

Materials and Methods: A modern web based interface was designed using the Django and Bootstrap3 web frameworks. User interaction with the interface initiates calls to our existing Geant4-based in house Monte Carlo simulation program that allows simulations of particles through experimental and clinical beam lines. The simulations are run on multi-core server hardware and various analyses are performed on the simulation output files using Python code called from the interface.

Results: Multiple pages allow for selection of relevant and familiar proton beam parameters for a simulation at varying levels of customization and flexibility: The ‘General Page' contains ‘canned' parameter sets chosen from eg discrete drop down menus. An ‘Advanced Page' allows greater customization for experienced users (eg, users can create a custom secondary scattering foil from a file description). A ‘Phase Space Page' allows for setup of simulations generated from phase space files (ie, pre-recorded particle momentum, energy and position at preset locations along the treatment nozzle) resulting in reduced simulation times. An ‘Analysis Page' allows users to analyze simulation results and create various types of plotted output and data. Finally, the system outputs a file that can be used to rerun the simulation.

Conclusion: wbeam is an easy to use and intuitive interface based on modern web technologies that allows non-programmers to create computer simulations and thus expands the community of MCCS users at our institution. The ability to easily produce accurate simulations of radiation dose delivered to planning target volumes promises to benefit both researchers and clinicians.

## C1 CNS/Skull BaseOP47: Outcomes following large-volume re-irradiation using proton therapy for recurrent malignant glioma

### CDH Proton Center, Northwestern University Feinberg School of Medicine

Background: The therapeutic benefit of targeting T2/FLAIR in addition to contrast-enhancing (CE) tumor during re-irradiation for recurrent malignant glioma is attenuated by augmented toxicity. Given its steep dose fall-off and narrow penumbrae, proton therapy (PT) minimizes the volume of brain parenchyma outside target volume, potentially permitting safer delivery of large-volume re-irradiation. This study seeks to explore this hypothesis through a retrospective review of overall survival (OS) and toxicity outcomes of patients treated with large-volume PT re-irradiation.

Materials and Methods: From 2/2011 to 11/2013, 19 consecutive adult patients with recurrent glioma treated with PT re-irradiation at a single institution were retrospectively analyzed. Planning target volume (PTV) included T2/FLAIR and CE abnormalities. Covariates assessed for OS prediction were age, gender, KPS at time of PT, number of pre-PT recurrences, grade at initial diagnosis, interval between PT and prior radiotherapy, PT dose, PT PTV, bevacizumab failure, concurrent use of temozolomide and/or bevacizumab, and post-PT radiation necrosis. OS time from PT start was estimated with Kaplan-Meier analysis; comparisons used log-rank statistic. Multivariate analysis used the Cox proportional hazards model.

Results: Median age was 42 and median KPS was 90. Median number of pre-PT recurrences was 2 (range 1-9). Median interval between PT and prior radiotherapy was 36.1 months (mos) (range 6.9-162.9). 12 patients (63%) were bevacizumab-refractory. Median PT dose was 50.4 CGE and median PTV was 224.2 cc. 5 patients (26%) remain alive. Median OS was 9.4 mos overall, 6.6 mos amongst bevacizumab-refractory patients, and 12.3 mos amongst bevacizumab-naive patients. Prior bevacizumab failure (hazard ratio (HR) 3.79; P=0.047), shorter interval since prior radiotherapy (HR 1.04; P=0.02), and Grade 4 disease (HR 4.17; P=0.03) were prognostic of inferior OS. One patient had grade 3 radiation necrosis in the setting of PT re-irradiation for progressive brainstem glioma. One patient had grade 2 radiation necrosis, and another had grade 2 stroke. No other grade ≥3 toxicities were observed.

Conclusion: Large-volume PT re-irradiation for recurrent glioma is safe and associated with promising OS outcomes, particularly in the setting of bevacizumab-refractory tumors.

## C3 GenitourinaryOP40: Patient Interviews on Prostate and Breast Cancer Treatment and Research Decision Making

### University of Florida Jacksonville, University of Florida PTI

Background: Although treatment choices, side effects, and outcomes are considered in treatment decision making, we know little about specific side effects, specific outcomes, and other factors that help patients in both treatment and research decision making.

Materials and Methods: To better understand specific factors important to patients, semi-structured interview was conducted with 6 breast cancer survivors and 6 prostate cancer survivors, covering the following topics: (1) Treatment decision- making; (2) Research decision-making; and (3) Impact of side effects. All interviews lasted approximately 90 minutes, and were conducted in a private office by a Behavioral Researcher not involved in the patient's clinical care.

Results: For treatment decision making, breast cancer survivors considered survival first, and then cited other factors including secondary tissue damage, QOL, and recurrence in making treatment choices and utilized their trust in their medical team in the process. Prostate cancer survivors responded that activity level and avoiding long term side effects were most important, and utilized family and friends in their decision making process. For research decision making, only 1 of 6 breast cancer survivors would agree to be randomized to different types of radiation therapy (protons vs X rays), but 3 of 6 agreed to be randomized to different fractionated treatment lengths. For the prostate cancer survivors, none of the males would be willing to be randomized to different types of radiation therapy (e.g., protons vs X rays) for their treatment , however, all 6 responded that they were willing to be randomized to different fractionated treatment lengths. All breast and prostate cancer survivors stated that they would want all information (including dosimetry pictures) on all side effects and potential outcomes to make their decisions. For symptoms and side effects, all of the breast cancer survivors stated that they were willing to undergo temporary side effects for improved long term outcomes. All of the prostate cancer survivors responded to specific side effects with concerns about how they impacted daily activity levels, rather than the symptom itself.

Conclusions: Our individual interviews, yielded data on the different specific factors involved in treatment and research decision-making. These data may impact informed consent discussions and the structure of comparative effectiveness research.

## C4 GastrointestinalOP10: Clinical Outcomes of Proton Beam Re-irradiation for Esophageal Cancer

### University of Pennsylvania

Background: Re-irradiation to the thorax and esophagus carries the risk of radiation (RT)-induced complications. Proton therapy (PRT) may offer an advantage in the re-irradiation (re-RT) setting due to the lack of exit dose and potential sparing of normal tissues.

Materials and Methods: Between 6/2010 and 2/2014, 13 patients (pts) with a prior history of thoracic RT and newly diagnosed or locally recurrent esophageal cancer in or near prior treatment fields began proton beam re-RT on a prospective registry protocol. Toxicity was graded according to the Common Toxicity Criteria version 4.0. Acute toxicity was defined as occurring within 90 days from start of PRT.

Results: The median follow-up was 11.3 months (1-23) from the start of re-RT. Mean age was 68 yrs (53-91). ECOG PS was 0 (n=6), 1 (n=4), and 2 (n=3). 9 pts received re-RT for esophageal cancer recurrence and 4 pts developed esophageal cancer as a new primary after prior RT for a different primary cancer. 9 pts had adenocarcinoma and 4 had squamous cell carcinoma. Median CTV size was 155 cc (33-390 cc). 10 pts received concurrent chemotherapy. The median interval between RT courses was 31 months (10-307). The median re-RT dose was 54.0 Gy (RBE)(50.4-61.2) and the median cumulative dose was 108.9 Gy (76-129.4 Gy) with a median cumulative cord maximum dose of 48 Gy (43-74 Gy). Of the 9 pts who presented with symptomatic disease, 4 pts had complete resolution of symptoms (sx), 3 had diminished sx, 1 had stable sx and 1 had progressive sx. The median time to sx recurrence was 11 months. Maximum acute toxicity was grade 2 (31%), 3 (31%), 4 (31%), and 5 (8%). The grade 5 toxicity was an esophagopleural fistula and respiratory failure unlikely related to RT. The grade 4 toxicities were 3 neutropenic toxicities and 1 COPD exacerbation. The grade 3 non-hematologic acute toxicities included dysphagia, dehydration and dyspnea related to pneumonia. There were 4 late grade 3 toxicities: heart failure, esophageal stenosis requiring dilation and PEG tube dependence. The median time to distant failure was 18 months after starting re-RT. The median time to local failure was 14 months. 6 pts were controlled locally within the retreated fields. The median overall survival for the pt population was 14 months.

Conclusion: Results using PRT for re-irradiation of esophageal cancer shows that symptomatic and local control can be achieved in some pts where re-RT with conventional photon beam therapy may not be an option.

## C8 ThoracicOP37: Early Experience with Intensity-Modulated Proton Therapy for Lung-Intact Mesothelioma: A Case Series

### UT MD Anderson Cancer Center

Purpose: To describe our experience implementing intensity-modulated proton therapy (IMPT) for treating lung-intact malignant pleural mesothelioma (MPM), including patient selection, treatment planning, dose verification, and process optimization.

Methods and Materials: Seven patients with epithelioid MPM were reviewed. Most had undergone pleurectomy/decortication, but one had had only biopsy. Four patients received IMPT and three received intensity-modulated radiation therapy (IMRT) to 45 Gy in 25 fractions, with a boost to 60 Gy for gross disease. Treatment plans for the other modality were created for dosimetric comparisons. Quality assurance processes included dose verification and robustness analysis. Image-guided setup was done with the first isocenter, and couch shifts were applied to reposition to the second isocenter. All patients had lung perfusion scans before treatment simulation. Doses to organs at risk were compared between techniques with Wilcoxon's signed rank test.

Results: Treatment was well tolerated and completed without breaks. IMPT plans were designed with 2 isocenters, 4 beams, and ≤64 energy layers per beam. Dose verification processes were completed in 3 hours. Total daily treatment time was approximately 45 minutes (20 minutes for setup and 25 minutes for delivery). The mean percent perfusion to the contralateral lung before treatment was 79%. IMPT plans had lower mean doses to the contralateral lung (5.0 vs 1.7 Gy, P=0.01), heart (24.9 vs 14.5 Gy, P=0.01), esophagus (28.4 vs 23.4 Gy, P=0.02), liver (24.8 vs 12.4 Gy, P=0.02), and ipsilateral kidney (15.6 vs 7.7 Gy, P=0.01).

Conclusions: Our initial experience showed that IMPT was feasible for routine care of patients with MPM and intact lungs. Further study is necessary to determine whether the differences identified by the planning study comparison translate into improved clinical outcomes.

## C9 OtherOP60: Breast Cancer Treatment with Uniform Scanning Proton Therapy on an Incline Beam Line

### CDH Proton Center

Background: This retrospective series evaluates the acute toxicity outcomes for locally advanced breast cancer (LABC) patients treated with uniform scanning proton therapy (USPT) on an incline beam line (IBL).

Materials and Methods: Between February and May 2014, 13 patients with LABC received USPT targeting either chest wall (N=10) or intact breast (N=3) plus comprehensive regional lymphatics including the internal mammary lymph nodes (IMN). Patients were treated with a new technique on the IBL using superior and inferior anterior superior oblique (ASO) fields alternated every other day with a superior and inferior en face anterior oblique (AO) field. Patients were simulated in a supine position with the ipsilateral arm above the head, immobilized in an alpha cradle with the thorax rotated 30 degrees to approximate an en face 30 degree angle achieved with the IBL AO fields. Toxicity was prospectively recorded using CTCAE v.4.0 weekly during treatment and at 2 and 4 weeks following treatment completion. Dosimetric parameters were retrospectively recorded to assess target coverage, hot spots, and dose to the esophagus and skin.

Outcomes: Median follow-up was 2.4 months. 12 women and 1 man were treated with USPT for stage IIB (N=3) or stage III (N=10) LABC. 6 were left-sided; 7 were right-sided. Of the 10 post-mastectomy patients, 3 were reconstructed and treated after the permanent implant was placed. Initial fields were planned to 50.4 Gy (N=10) or 45 Gy (N=3) with 5 patients prescribed a boost of 10 Gy. Median total dose received was 50.51 Gy, (range; 45.11-60.53 Gy). All patients experienced radiation dermatitis: grade 1 (N=3); grade 2 (N=9); and grade 3 (N=1). 2 patients did not complete the full course of prescribed PT due to skin toxicity. Of the 10 patients with grade 2-3 dermatitis, 5 had a V110>3.4% (range, 4.7-29.4%); and CTV dmax to 0.5cc >57.69 CGE (range, 58.3-59.46). 6 patients experienced grade 1 chest wall pain; 5 of these required narcotics. 4 patients experienced grade 1 esophagitis; 1 required narcotics. All had esophageal dmax dose greater than the cohort median (42.56 CGE). 2 patients experienced skin infection requiring antibiotics 2 weeks following treatment completion.

Conclusions: These early results demonstrate the feasibility of LABC treatment with a 2-field daily IBL technique. Acute toxicity results appear acceptable. Longer follow-up is needed.

## OP62: Developing a Cost Effectiveness Model for Comparing Radiation Modalities for Stage III NSCLC

### University of Washington

Background: To facilitate a cost-effectiveness analysis of proton beam vs. photon beam radiotherapy, we created a multi-parametric model including patient, disease, treatment, and post-treatment toxicity data using stage III NSCLC as initial proof of principle.

Materials and Methods: An influence diagram was used to model radiation delivery and its acute toxicity at 6 months. The ID consists of an action node (protons vs. photons), a Bayesian network to calculate the joint probabilities of the parameters, and cost nodes to compute the costs of therapy and toxicities.

Results: Input parameters were grouped into three categories: patient-, tumor-, and treatment-specific. Patient factors included age, sex, smoking status, and pre-treatment lung function. Tumor-specific factors included histology, tumor/nodal stage, and location. Treatment-specific factors included chemotherapy type, radiation modality, and established dose-volume parameters for toxicity. Conditional probabilities for photon therapy toxicities were obtained from review of large prospective trials. Proton therapy toxicity probabilities were determined from the largest series of proton therapy in Stage III NSCLC and from published dose-volume models. Toxicity distributions for each input factor were estimated from toxicity meta-analyses for pneumonitis and esophagitis. The overall probability of toxicities was determined for a range of these input parameters. These probabilities are used to calculate the overall cost of treatment (costs of radiation therapy plus costs of intervention for each toxicity grade) across a range of toxicity rates and input parameters. The higher initial costs of proton therapy may be offset by toxicity reduction and its associated costs but depends on the initial assumptions and input parameters of the model.

Conclusions: We created a novel cost-effectiveness model that can examine the overall costs associated with photon and proton radiation for Stage III NSCLC. This model incorporates pre-treatment and radiation planning parameters, which impacts toxicity and costs. Through this model, we can obtain population- and patient-specific costs associated with therapy, which forms the basis for investigating the utility or benefits of different modalities in clinical trials or recommending competing treatments for individual patients.

## P4 Quality AssuranceOP14: In-Vivo Proton Range Verification Using A Collimated Gamma Detector Array

### Arizona State University, Mayo Clinic

Proton range uncertainty remains one of the most severe limiting factors in proton beam therapy. In this work we aim to reduce proton range uncertainty by measuring proton range during patient treatment. Many studies pertaining to the feasibility of in-vivo proton range verification via detection of prompt gamma radiation have been performed, but detectors designed for clinical use continue to represent a difficult challenge. To build a detector for clinical use (pencil beam spot scanning), a collimated gamma detector array is being constructed. An MCNPX simulation was first conducted which suggested the feasibility of such a design to measure the Bragg Peak with an uncertainty of 2 to 4 mm per spot. To better understand the scintillation optics, Geant4 was utilized as a model to mirror and guide the construction progress. As a way of calibrating the Geant4 simulation and configuring the data acquisition electronics, a two-crystal (CsI Tl-doped) detector array was built. Promising features planned in the data acquisition are pulse shape discrimination and time of flight, both for eliminating background neutrons. With a calibrated Geant4 model and sophisticated data acquisition techniques, the final detector design will be optimized and tested. A progress update will be presented along with the simulated and experimental results obtained.

## OP16: Three Dimensional Gamma Criterion for PBS Patient Specific Quality Assurances

### ProCure Proton Therapy Center, New Jersey

Background: The existence of distal penumbra of a proton beam necessitates 3D Gamma evaluation for modulated scanning proton treatment plans. A software program is presented which evaluates the dosimetric agreement between measured and calculated patient-specific proton QA plans.

Materials and Methods: Gamma criterion is the standard in determining the agreement between measured and calculated doses for photon IMRT plans. It is especially useful in evaluating lateral penumbra where rapid dose fall-offs, coupled with minute lateral misalignment, can cause artificially large percentage error. By simultaneously considering dose and distance, the Gamma criterion renders dosimetrically relevant evaluation on actual dose distribution. In contrast to x-rays, spot scanning proton beams can be modulated in the beam direction, generating a true 3D dose distribution. Evaluation of this 3D dose distribution would naturally require an extension of the Gamma criterion along the beam direction. For each patient-specific modulated scanning proton plan, we compare the calculated 3D dose distribution to measurements made by a 2D ion chamber array at varying water-equivalent depths. No interpolation is needed since depth along the beam direction is directly incorporated into the 3D Gamma comparison. Both absolute and relative (i.e. Van Dyk) comparison are available, for user specified gamma criteria. Additional functionality such as auto-alignment, lateral shifts, and inclusion threshold are also provided.

Results: Ten patient plans with a total of 32 beams were measured and compared using this 3D Gamma method. All evaluations are done with 3%/3mm criteria with a 10% threshold and Van Dyk normalization. Point of normalization is set at the middle of the field for Single-Field Uniform Dose (SFUD) plans. For Intensity Modulated Proton Therapy (IMPT) plans, the normalization point is selected from regions of low dose gradient. Effect of temperature and pressure variation is also taken into account by scaling the measured dose with a standard output measurement. Results obtained from this 3D Gamma software were compared with those obtained using the 2D Gamma program from the commercially available ion chamber array. Higher passing rates are observed at the distal penumbra for IMPT plans as expected, while good agreements were found for both SFUD and IMPT plans at locations proximal of the distal fall-off. Passing rates using the commercially available 2D Gamma program can be improved by interpolating the 2D dose distribution at a slightly different depth, corroborating the use of 3D Gamma criterion on the distal penumbra.

Conclusions: 3D Gamma criterion was implemented and demonstrated for modulated scanning proton plans. The rapid dose fall-offs, in both lateral and distal penumbra, were successfully taken into account by our approach. Clinical implementation of this program can be used to evaluate the effects of range uncertainty along the beam direction, as well as lateral setup errors.

## P31: A study comparing rectal immobilization devices on intra-fraction motion for prostate proton therapy

### CDH Proton Therapy Center

Background: Intra-fraction motion during prostate radiation therapy treatments can affect the relative position of the treatment volume and dose given to normal tissue structures. Maintaining a reproducible setup and internal stabilization can be achieved with variations in organ filling, especially in the rectum. To better evaluate the effect of rectum stabilization on prostate mobility, a study was completed comparing our current rectal immobilization with commercial and in-house made rectal immobilization.

Materials and Methods: Data was collected from 95 prostate cancer patients treated with proton therapy. Following fiducial placement, each patient was simulated with either a commercial water filled rectal balloon (90cc, 110cc, or 120cc), a 120cc in-house design water filled rectal balloon, or our current immobilization, free water (100cc), inserted in the rectum approximately 10cm. Two orthogonal portal images were taken daily to match the fiducial markers within a 1mm margin. After treatment delivery, a second set of orthogonal portal images were taken. If the fiducial markers were displaced from the original position, a positional magnitude was calculated to determine intra-fractional fiducial motion, using offsets from three axes: lateral, anterior-posterior, and craniocaudal. An average vector magnitude was calculated for each patient and each rectal immobilization device to compare intra-fractional movements of the target volume.

Results: The mean magnitude (in mm) for the 90cc rectal balloon was 1.10 (SD=1.33). The mean magnitude for the 110cc rectal balloon was 0.78 (SD=1.08). The mean magnitude for the 120cc rectal balloon was 0.96 (SD=1.23). The mean magnitude for the 120cc in-house rectal balloon was 2.29 (SD=1.46). The mean magnitude for the 100cc free rectal water was 1.43 (SD=1.43). An ANOVA was performed to test the differences in the mean magnitude among the five groups. This yielded statistically significant results (p < .05). A second ANOVA was performed to compare the differences in mean magnitude among the three commercially available rectal balloons and free water, which also yielded statistically significant results (p < .05).

Conclusion: Based on our study, commercial rectal balloon devices are more effective internal immobilization devices compared to free water inserted into the rectum. All three volumes studied reduced intra-fraction motion, improved target localization, and did not result in any significant acute toxicity to the patient.

## P7 OtherOP28: Single-plane Magnetically Focused Narrow Elongated Small Field Proton Beams

### Loma Linda University

Background: The original clinical goal of the project was to deliver narrow elongated proton beams to the spinal cord for potential neuropathic pain treatments. Previous Monte Carlo simulations suggested that using a single quadrapole magnet could produce beams of elliptical cross section with superior dose distribution properties and greater efficiencies than collimated beams (McAuley et al Phys. Med. Biol. 58 (2013)). The parameters of the simulated magnets were chosen to mimic k=3 Halbach cylinders that are available commercially as assemblies of rare earth permanent magnetic materials. In the present study, results from experiments with actual prototype magnets are presented.

Materials and Methods: Magnets consisting of 24 segments of radiation hard samarium-cobalt (Sm2Co17) adhered into hollow cylinders were manufactured. A single focusing magnet was placed on a positioning track on our Gantry 1 treatment table and 15, 12 and 8 mm diameter proton beams with energies and modulation relevant to clinical radiosurgery (127 to 186 MeV, and 0 to 30 mm modulation) were delivered to a water tank. Dose distributions were measured using a PTW diode detector and Gafchromic EBT2 film. Longitudinal and transverse dose profiles were analyzed and compared to data from Monte Carlo simulations analogous to experiments.

Results: The narrow elongated focused beam spots showed high elliptical symmetry indicating high magnet quality. Monte Carlo simulations were in good agreement with diode and film data. When compared to unfocused beams (ie, the focusing magnet removed), peak-to-entrance depth dose ratios were 11 to 14 % larger (depending on presence or extent of modulation). However, when using a more relevant comparison with beams collimated with an elliptically shaped collimator, the peak-to-entrance ratios were 26 to 38 % larger and per particle efficiency was ~ 2 fold greater than the later collimated beams.

Conclusion: Our results suggest that the use of rare earth magnet focusing assemblies is feasible and could improve dose-sparing of normal tissue and organs at risk while delivering enhanced dose to small elongated targets. Such magnets are small, inexpensive, do not require power or cryogenic cooling, have low external magnetic fields and could be incorporated in place of the aperture in a standard proton treatment nozzle.