Hodgkin lymphoma (HL) survivors are susceptible to long-term side effects from definitive treatment with combination chemotherapy and radiation therapy (RT). Because late effects from RT correlate to radiation treatment volume, stage III/IV pediatric HL requiring large radiation fields are at greater risk for late complications. This study investigates the dosimetric impact of proton therapy (PT) to spare the organs at risk (OARs) in stage III and IV HL.
From October 2011 through January 2014, 7 pediatric patients with stage III or IV HL involving both supradiaphragmatic and infradiaphragmatic underwent treatment planning with 3-dimensional conformal radiation therapy (3DCRT), intensity-modulated radiation therapy (IMRT), and PT to a dose of 21 Gy after ABVE-PC chemotherapy. Plans were normalized to meet treatment parameters including target coverage and dose to OARs.
All 7 patients were treated with PT. PT had the lowest integral body dose (101 J; range, 55–139) compared with 3DCRT (180 J; range, 110–301) and IMRT (179 J; range, 96–298). PT led to a clinically significant absolute dose reduction compared with 3DCRT and IMRT in the heart (median, 8.8 Gy and 4.5 Gy), breasts (6.1 Gy and 5.2 Gy), lungs (3 Gy and 3.1 Gy), stomach (11.5 Gy and 7.6 Gy), liver (3.9 Gy and 6.4 Gy), and bowel (5 Gy and 4.7 Gy, respectively). A smaller reduction was seen in the thyroid (0.7 Gy and 0.2 Gy), pancreas (2.5 Gy and 1.2 Gy), left kidney (3.6 Gy and 1.2 Gy), and right kidney (0.9 Gy and 3.2 Gy, respectively). No benefit was seen for the esophagus (−0.2 Gy and −0.1 Gy).
Integral dose and dose to OARs were reduced with PT compared with 3DCRT and IMRT. Many of these absolute dose reductions are clinically meaningful and should lower the risk of late effects in this high-risk population.
In the United States, there are currently approximately 175,000 Hodgkin lymphoma (HL) survivors. Although HL is rare in the general population, it is one of the most prevalent childhood cancers with over 1,000 cases diagnosed each year, of which approximately 50% are stage I or II and 50% are stage III or IV. Fortunately, even with advanced disease, 80% of patients survive at least 5 years and the 10-year survival rate approaches 76% [1–3].
Standard-of-care treatment for pediatric patients with stage III or IV HL in the United States is chemotherapy, with involved-field or involved-site radiation therapy reserved for patients receiving less-intensive chemotherapy regimens or those with incomplete or slow responses to chemotherapy [4–6]. Unfortunately, HL survivors are at a very high risk of long-term side effects from their definitive treatment [7–21]. In fact, Oeffinger et al  demonstrated that among more than 10,000 childhood cancer survivors, HL survivors endured some of the highest cumulative incidences of grade 3 to 5 severe, life-threatening, and fatal complications at a 28-year interval, including secondary malignant neoplasia (SMN) and cardiac toxicity. Because these sequelae are a function of exposure to both chemotherapy and radiation therapy, investigators have been seeking ways to reduce chemotherapy and radiation damage to normal tissue from HL treatment regimens [4–6, 18, 22–25].
Studies have already shown the ability to reduce radiation exposure of nontargeted tissue by reducing the prescribed radiation dose and the target volume without sacrificing event-free survival rates; however, another way to achieve this goal is through the use of more conformal radiation techniques, including intensity-modulated radiation therapy (IMRT) and proton therapy (PT) [4, 6, 12, 26, 27]. Although researchers have demonstrated the dosimetric advantages of using IMRT and PT in patients with stage I or II HL [12, 26, 27], to our knowledge, no study has evaluated their use in patients with advanced-stage HL. Due to the challenges of delivering RT to large treatment fields in pediatric patients with stage III HL and stage IV HL with both supradiaphragmatic and infradiaphragmatic involvement (IVsi), conventional 3-dimensional conformal radiation therapy (3DCRT) using opposing anterior-posterior and posterior-anterior fields remain the standard treatment approach. The present study investigates dose reduction to the organs at risk (OAR) with IMRT and PT in 7 consecutive pediatric patients with stage III or IV HL treated at the University of Florida Proton Therapy Institute (Jacksonville, FL).
Patients and Methods
Under institutional review board approval, the pertinent medical records of 7 consecutive pediatric patients with stage III or IVsi HL previously treated with PT were reviewed. Eligible patients were younger than 18 years old, had an initial prechemotherapy staging computed tomography (CT) scan and either 18F-Flourodeoxyglucose positron emission tomography (PET)/CT or Gallium-67 citrate scans, and either complete or partial remission within 8 weeks of starting radiation therapy. Table 1 and Table 2 outline patient characteristics and prechemotherapy initial nodal sites as determined on staging work up. All patients had a complete response by PET/CT or Gallium scan by the end of treatment.
The methods for this investigation were based on a protocol established in a previous University of Florida study and with the goal of updating a prior case report . Treatment set-up aimed to reproduce the positioning at the initial staging scans to facilitate the fusion of initial, response, and treatment planning CT images. Patients were positioned supine on the treatment couch and both a Vac-Lok bag (CIVCO Medical Solutions, Kalona, IA) and facemask were used for immobilization.
A 4-dimensional (4D) CT simulation was performed followed by a 3-dimensional (3D) CT scan with intravenous (IV) contrast (Philips Brilliance, Philips Medical Systems, Madison, WI). Patients were instructed to breathe normally and no attempt to restrict respiratory motion was made because no significant target movement due to respiratory motion was detected. All of the respiratory-phase scans and reconstructed average images were transferred with the IV 3D CT and the initial staging PET-CT scan (if available) to a MIMVista workstation (MIMVista Corp, Cleveland, OH). Patients were treated with either involved-site or involved-node radiation therapy to all the sites of initial involvement . According to guidelines [28, 29], gross tumor volume (GTV) was contoured on the average 4D CT, and the clinical target volume (CTV) was defined as the prechemotherapy involved disease. The CTV was then uniformly expanded by 5 to 10 millimeters to create the planning target volume (PTV).
Using the averaged 4D CT for optimization and dose calculation, dosimetric plans for 3DCRT, IMRT, and PT treatment for each patient were prepared, totaling 21 plans. For PT, target and OAR volumes were imported into an Eclipse proton treatment planning system (Varian Medical Systems, Palo Alto, CA) and a 3D conformal proton therapy plan was created. Six of the patients with mediastinal disease for whom breathing motion was a concern had double-scatter plans developed and 1 patient with neck, bilateral axilla, and pelvic disease without breathing motion issues had a uniform-scanning plan developed. Although setup errors along the beam direction have no effect on dose distribution, a smearing factor was included in the proton range compensator design to account for intrafraction and interfraction motion perpendicular to the beam direction. Distal and proximal margins were calculated for each proton beam (proximal/distal-CTV-depth [mm] x 1.025 + 1.5 mm) to account for CT-number-to-proton-stopping-power conversion uncertainties and daily beam-delivery variations. Field apertures were calculated with a block margin to account for the proton beam penumbra at the target depth, and range compensators were calculated using an 8- to 10-mm smearing margin and up to 10-mm border smoothing values. The 3DCRT and IMRT treatment plans were calculated by exporting the scans and contours of the target and OARs used in the PT plans from MIMVista to a Pinnacle treatment planning system (Philips Healthcare, Andover, MA). For each plan, optimization objectives were selected to ensure that the required target coverage received high relative weights and were adjusted during an interactive modeling process to maximize OAR sparing. For 3DCRT, we utilized a combination of 6-, 10-, and 15-MV photon beams in opposing anterior-posterior and posterior-anterior beams. For IMRT, plans followed direct machine parameter optimization with approximately 10 segments per beam and a minimum segment area of 6 cm2. Several of the patients required matching fields with a match-line change performed halfway through treatment due to the length of the treatment area for the 3DCPT, 3DCRT, and IMRT plans.
All patients received a total radiation dose of 21 Gy (relative biological effectiveness [RBE]) in 1.5-Gy (RBE) daily fractionations for a total course of 14 days. Treatment plans were normalized and target coverage aimed to meet strict requirements for prescription dose (CTV D99% = 100% and PTV D95% ≥ 95%) and for OAR protection, including lung V20 <30%, mean lung dose <14 Gy (RBE), and mean heart dose <20 Gy (RBE). OARs were prioritized as follows: women, heart > breast > lung > thyroid > esophagus > body; men, heart > lung > thyroid > esophagus > body. Nontargeted tissues were defined as (body – PTV), and integral body dose was reported in joules and derived from multiplying the mean dose to the nontargeted body and multiplying the product by the volume of the non-targeted body.
No testing was done for statistical significance owing to the limited number of patients involved in the study.
The relative reduction and absolute reduction (AR) in the integral body dose with PT was 46% and 83 J compared with 3DCRT, while the relative reduction and absolute reduction in the integral body dose with PT was 44% and 78 J compared with IMRT. Figure 1 depicts a 3DCRT, IMRT, and PT color wash dose distribution for a patient within the cohort. The OAR dose-volume histogram (DVH) is depicted in Figure 2A and B. The DVH demonstrates substantial reductions in the volume of the OARs receiving low doses of radiation when PT is used, while the 3DCRT and IMRT plans deliver low-dose radiation to a much larger volume. On the other hand, PT irradiates more of the body than IMRT in the region receiving 20 Gy and higher for most OARs.
Table 3 and Table 4 represent the mean OAR doses for 3DCRT, IMRT, and PT for the entire cohort along with the absolute dose reduction to the OARs when comparing PT to 3DCRT, PT to IMRT, or IMRT to 3DCRT. For all 7 patients, PT delivered the lowest dose to all of the OARs except the esophagus compared with 3DCRT. This included a large AR for the heart (median, 8.8 Gy), breast (6.1 Gy), lungs (3.0 Gy), stomach (11.5 Gy), bowel (5.0 Gy), liver (3.9 Gy), and left kidney (3.6 Gy), and a small AR for the thyroid (0.7 Gy), pancreas (2.5 Gy), and right kidney (0.9 Gy). When compared with IMRT, all 7 patients had the lowest dose to the heart, breast, lungs, stomach, liver, and right kidney with PT. In some cases IMRT offered lower doses to the esophagus, thyroid, pancreas, bowel, and left kidney compared with the PT plan. A large AR with PT versus IMRT was seen to the heart (median, 4.5 Gy), breast (5.2 Gy), lungs (3.1 Gy), stomach (7.6 Gy), liver (6.4 Gy), bowel (4.7 Gy), and right kidney (3.2 Gy), while smaller reductions were seen to the thyroid (0.2 Gy), pancreas (1.2 Gy), and left kidney (1.2 Gy).
This study demonstrates important reductions to OAR doses and integral body doses with PT, while maintaining appropriate coverage of the target volume in patients requiring both supradiaphragmatic and infradiaphragmatic radiation. By reducing the dose to nontargeted tissue, PT appears to be the most appropriate technique to deliver consolidative radiation to pediatric and adolescent patients with stage III and stage IVsi HL because it is expected to cause fewer late effects among this high-risk patient cohort.
Pediatric and adolescent patients with stage III and IVsi receiving consolidative radiation therapy require some of the largest radiation fields in oncology, putting them at the highest risk of long-term side effects of radiation that may not develop for 20 to 40 years after completing radiation therapy. In a cohort of over 4,000 patients, Tukenova et al  found that death due to SMN was associated with increased integral doses of radiation. More specifically, patients exposed to greater than 150 joules had a statistically significant adjusted relative risk of 5.2 for developing carcinomas and a relative risk of 12.6 for developing sarcomas. In our study, PT reduced the integral dose by close to 50% compared with 3DCRT and IMRT and, furthermore, none of the PT plans delivered an integral dose greater than 150 joules. Consequently, we can deduce that PT should lower these patients' risk of secondary sarcomas and carcinomas based on the reduction in integral body dose.
Female survivors of HL are at a significantly higher risk of secondary cancers primarily due to an increased risk for developing secondary breast cancers. Bhatia et al  found that childhood survivors of HL have a seventy-five times greater risk of breast cancer and most of those breast cancers develop within the field of radiation. Although using smaller radiation therapy fields that omit the uninvolved axilla in female HL patients, such as involved-field treatment, has already shown a reduction in breast cancer development, it remains a problem in patients with axillary involvement . In our investigation, PT reduced the volume of breast tissue treated to 5 Gy by 34% compared to 3DCRT and by 54% compared to IMRT. These volume reductions are important considering epidemiological studies have observed an increased risk of secondary breast cancers among Hodgkin lymphoma survivors with radiation doses as low as 4 Gy  to the breast. Consequently, the dose reduction to the breasts demonstrated with PT is expected to reduce the risk of secondary breast cancer among HL survivors. Although prior studies evaluating PT in patients with predominantly localized stage I/II HL have demonstrated less substantial reductions in breast dose, this is likely because fewer patients in that study had axillary involvement and, thus, the overall dose to the breasts was low for all three modalities . Since the patients in this study included just stage III and IVsi patients, it is unsurprising that all 4 of the female patients would have a component of axillary/subpectoral disease at presentation, which would lead to substantially higher breast doses with either 3DCRT or IMRT that could be significantly decreased with PT. This is similar to the dose reduction seen by Andolino et al  in pediatric female patients.
HL often presents with mediastinal involvement, so lung irradiation poses an almost unavoidable dilemma. Lung radiation has been associated with an increased risk of pneumonitis, pulmonary fibrosis, pulmonary disease, and, most concerning, lung cancer after HL treatment [10, 11, 16, 18]. Salloum et al  showed that the relationship between radiation and SMNs in HL patients was highest for the development of lung cancer, for which the relative risk was 10.7. Within a population-based cohort of over 19,000 individuals treated for HL, another study found that lung SMN risks were elevated in all radiation categories starting at doses of 5 Gy or more, and the risks increased with increasing doses to the lung . In the present study, proton therapy reduced the volume of lung receiving 5 Gy by 11% when compared with 3DCRT and 35% when compared with IMRT. Consequently, it is expected that the radiation dose reduction to the lungs is likely to reduce the risk of a secondary lung cancer among HL survivors. These findings are similar to prior studies demonstrating a reduction in radiation dose to the lungs in stage I/II HL patients with PT .
In addition to SMN, cardiotoxicity leads to a significant reduction in the survival of HL patients [10, 14, 25]. In a study that included over 14,000 5-year survivors of cancer diagnosed under the age of 21 years, Mulrooney et al  showed that the HL subgroup had the single highest hazard ratio for congestive heart failure, myocardial infarction, pericardial disease, and valvular abnormalities. Moreover, average radiation doses above 15 Gy were associated with over a 2-fold increase in the hazard ratio of each of these cardiac sequelae. In this study, only PT sufficiently mitigated the dose to meet this target for OAR protection as the median for 3DCRT, IMRT, and PT were 20.5 Gy, 16.3 Gy, and 11.3 Gy. Furthermore, PT reduced the volume of heart receiving 5 Gy or more by 39% when compared with 3DCRT and 40% when compared with IMRT. PT reduced the volume of heart receiving 15Gy by 34% when compared with 3DCRT and 4% when compared with IMRT. Consistent with previous investigations in patients with stage I/II mediastinal HL , these results reaffirm PT's ability to attenuate radiation exposure to cardiac structures more effectively than either 3DCRT or IMRT for stage III and IVsi HL and, therefore, would be expected to reduce long-term complications [12, 27].
Despite prior studies demonstrating lower doses to the esophagus in patients with more limited involvement of HL , the same could not be said in patients with stage III/IVsi involvement. In these patients, the need to match PT fields within the mediastinum may raise the esophagus dose higher than that seen in the stage I/II patients where matching fields weren't required as often. Nevertheless, the higher dose seen in some of the PT plans to the esophagus was no greater than 1.2 Gy in any patient and on average was only 0.2 Gy higher than the 3DCRT esophagus dose and 0.1 Gy higher than the IMRT esophagus dose. Consequently, the dose reductions to the other OARs compensates for this concern.
Because patients with advanced-stage disease can present with involvement on both sides of the diaphragm, nontargeted abdominal structures risk irradiation. In a cohort of over 5,500 HL patients, there was an increased risk of gastrointestinal cancers, including stomach (relative risk, 28.9) and colon (relative risk, 13.8) cancers . Of note, we observed that in our investigation patients with splenic involvement and/or paraaortic involvement appreciated a substantial margin of OAR dose reduction to abdominal structures. As illustrated in a cohort of over 5,100 patients that included survivors of testicular cancer or HL, van den Belt-Dusebout et al  found a linear dose-response relationship for stomach cancer with radiation doses up to 40 Gy. Additionally, they observed that mean stomach doses greater than 20 Gy were associated with a relative risk of 9.9 for SMN compared with doses below 11 Gy. In the present study, the mean stomach dose was 21.5 Gy (median) with 3DCRT, demonstrating a considerable increased risk of a stomach cancer with this treatment approach. Although all of the IMRT and PT plans had mean stomach doses below 20 Gy, the IMRT plan had a much higher stomach dose (14.6 Gy, median) compared with the PT dose (7.0 Gy). Thus, while both may decrease the incidence of stomach cancer, PT could lead to even further reductions. Additionally, it is likely that the lower stomach dose would decrease the risk of acute side effects, including nausea, vomiting, and fatigue, seen with radiation to the stomach.
Unintentional radiation to other abdominal OARS, including the liver, kidneys, pancreas, and bowel, is also likely to contribute to late effects among long-term survivors, such as kidney disease , liver disease , bowel obstruction , and diabetes . A direct dose response is still unclear and data is still emerging. Yet despite the limited data, PT still resulted in overall lower mean doses to the kidneys, pancreas, liver, and bowel compared with 3DCRT and IMRT.
While PT appears to reduce OAR doses and integral doses in pediatric patients with stage III HL, as with any new burgeoning technology, diligent scrutiny must accompany this treatment option. As Hall et al  notes, to increase radiation fields to clinically practical sizes, some institutions currently use passive double-scattering delivery techniques, which entails passing the initial proton beam through a scattering foil, potentially introducing uncharged particles called neutrons into the field [36, 37]. Despite the initial apprehension caused by introducing neutrons, studies have shown only small differences in the amount of secondary neutrons generated with PT techniques as compared to photon treatments, suggesting that double-scattering techniques contribute a marginal influence if any on the dose distribution [38, 39]. Furthermore, no clinical study has supported these concerns . As proton centers transition to active scanning techniques that do not require foil scattering, the production of secondary neutrons will become less of a concern.
Because of its small sample size, the present analysis may not have considered all of the potential disease sites in patients with stage III HL. Nevertheless, we included patients with involved head and neck, thoracic, abdominal, pelvic, and inguinal nodes, thereby representing a wide range of classic locations typically seen at presentation.
As this investigation supports, PT reduces radiation to nontargeted tissues in stage III HL pediatric patients without compromising target coverage. This improvement in OAR sparing and reduction in integral dose suggests an increased margin of safety against the long-term side effects of radiation therapy, and establishes the foundation for the clinical use of PT in this patient population.
ADDITIONAL INFORMATION AND DECLARATIONS
Conflict of Interest Statement: The authors have no conflicts of interest to disclose.
Acknowledgments: We would like to thank Jessica Kirwan and the editorial staff of the University of Florida Department of Radiation Oncology for editing and preparing the manuscript for publication, and Natalie Getman and Debbie Louis for dosimetric assistance on the study.