Irradiating Residual Disease to 30 Gy with Proton Therapy in Pediatric Mediastinal Hodgkin Lymphoma

In the 1970s, the first protocols of chemotherapy with consolidative low-dose involved-field radiotherapy in pediatric Hodgkin lymphoma (PHL) were developed to reduce the potential long-term complications of full-dose, definitive radiotherapy, especially as it affected bone growth [1]. During the past half century, PHL protocols have encouraged enrollment of older patients; use of newer systemic therapy combinations—including targeted agents and immunotherapy—and use of smaller radiotherapy fields, such as involved-site radiotherapy (ISRT); or no radiotherapy. When radiotherapy has been used, the standard dose has remained between 21 and 25.5 Gy [2, 3].

Response-adaptive trials in PHL have identified rapidly responding patients for whom radiotherapy may not provide a meaningful benefit when weighed against its potential late effects [2, 3]. However, those trials also identify patients who do not respond well to chemotherapy and may benefit from more-aggressive systemic therapies and higher radiotherapy doses. Studying the pattern of relapse can be helpful in determining whether relapse rates can be improved with additional radiation (which manages local relapses) or systemic therapy (which manages distant relapses). In the largest intermediate-risk PHL study conducted, Children's Oncology Group study AHOD0031), the 4-year, event-free survival was 85% and, among the relapses, 88% occurred in field after 21 Gy [4]. Importantly, despite similar rates of overall relapse, the in-field relapse rate was twice as high as that seen in adult patients with intermediate-risk Hodgkin lymphoma (HL) who received 30 to 36 Gy and relapsed out of field [5].

Because of concerns about the late effects from radiotherapy, oncologists hesitate to increase the radiotherapy dose in any protocol. However, if a higher radiation dose could be delivered without an increase to the dose to the organs at risk (OARs), oncologists might consider allowing higher doses. Recent studies have demonstrated excellent outcomes with smaller involved-node radiotherapy and ISRT fields [6]. Therefore, the next logical step is to evaluate even smaller radiation therapy (RT) treatment fields, such as the postchemotherapy anatomic disease volume [7].

The present study evaluated the potential implication to the OARs of escalating the postchemotherapy radiation dose in PHL patients to 30 Gy compared with 21 Gy, either by increasing the prescription dose or through a boost field to the gross residual disease.

After institutional review board approval and with informed consent, 10 PHL patients with mediastinal involvement enrolled in an institutional prospective study and treated with consolidative ISRT were selected for this planning study.

All patients had ISRT field volumes developed following International Lymphoma Radiation Oncology Group guidelines and were planned to receive 21 Gy at 1.5 Gy/fraction [6]. The ISRT target volumes included residual anatomic disease seen at the time of computed tomography (CT) simulation of >1 cm gross tumor volume (GTV_residual) in dimension. The ISRT clinical target volume (CTV_ISRT) included the GTV_residual and locations of previously involved lymph nodes seen during staging positron emission tomography (PET)/CT scan with a margin to account for differences in positioning between the fusion of the PET/CT staging scan and the CT simulation. An ISRT internal target volume (ITV_ISRT) was developed from the 4-dimensional CT simulation and a planning target volume (PTV)_ISRT margin of 7 mm added to the ITV. For the purposes of the study, the GTV_residual was expanded per the internal GTV_residual. The ITV_residual was the internal GTV_residual without further expansion; the PTV_residual included a 5-mm margin for the residual ISRT (rISRT) (Figure 1A and 1B).

Three additional RT treatment plans were retrospectively developed for each patient. Plan 1 treated the ISRT to 21 Gy (ISRT_21Gy). Plan 2 treated the ISRT to 30 Gy (ISRT_30Gy). Plan 3 treated the ISRT to 21 Gy with a 9-Gy boost delivered to the rISRT (rISRT_boost). Finally, Plan 4 treated the rISRT only to 30 Gy (rISRT_30Gy).

Radiation treatment plans were developed with passive-scatter proton therapy techniques designed using Eclipse software (Eclipse Foundation, Ottawa, ON Canada), as previously described [8]. In short, the 3-dimensional conformal proton plans were designed to treat ISRT_21Gy and rISRT_boost with a robustness of 2.5% range uncertainties and setup errors of 5 mm. For the cases resulting in a range margin smaller than the PTV margin, no effort was made to cover the PTV along individual beams. Laterally, custom-made collimators shaped the individual proton beams to the PTV with a margin for the penumbra. Range compensators were smeared to allow target coverage under tissue and density variations on the proton beam paths. All plans were designed and calculated on the average 4-dimensional CT image. Depending on the size and complexity of the targets, 1 to 3 proton fields were used per plan. The proton beams with the shortest path in healthy tissue were favored. Plan priority was set to the target coverage. The OAR importance was set to the heart, breast (female), and lung. Plan 2 was a scaled-up version of plan 1 from 21 to 30 Gy radiobiological equivalents (GyRBE). Plan 3 included plan 1 as the initial setup and an additional boost plan. Plan 4 was the boost plan scaled up from 9 to 30 GyRBE. All plans were normalized to cover 99% of the ISRT volumes with the prescribed dose and maintain the coverage on inhale and end of exhale during the 4-dimensional phases. Moreover, 95% of the PTV was covered by ≥95% of the prescribed dose.

Excel software (Microsoft Corporation, Richmond, WA) was used for basic statistical calculations. Doses to each OAR were recorded, and relative dose differences were calculated between each plan.

We found that treating patients with ISRT to 30 Gy, rather than 21 Gy, increased the relative dose to the target and the thoracic OARs by 42% for ISRT_30Gy. The mean heart dose was, on average, 8.5 Gy with ISRT_21Gy and increased by 3.6 Gy with ISRT_30Gy. However, the mean heart dose only increased by 15% when the rISRT_boost was used, and dropped by 46% for the rISRT_30Gy plan. The corresponding doses to the various thoracic OARs among the 4 different treatment plans are shown in the Table.

Consolidative radiotherapy has been shown to improve overall survival over chemotherapy alone for patients with pediatric Hodgkin lymphoma [18]. Although higher doses of RT have been avoided in PHL to limit late effects, the greatest concern is OAR dose. Mean heart and lung doses of 10 and 12 Gy [9, 10], respectively, are considered reasonably tolerable, whereas doses of 18 to 20 Gy to the bone can stunt growth [11]. Our results demonstrate that higher RT doses to the target can be delivered while reducing the dose to the OARs with a smaller target volume, as with the rISRT_30Gy plan.

Identifying the ideal radiation dose is important to achieving the best chance of local control and relapse-free survival. In the adult population, randomized controlled studies have demonstrated that, although 20 Gy is sufficient for favorable early stage adult HL after 2 cycles of Adriamycin (doxorubicin), bleomycin, vinblastine, and dacarbazine (ABVD), patients with unfavorable early stage adult HL required doses of 30 Gy with 4 cycles of ABVD [12, 13]. Conversely, radiotherapy dose has not, to our knowledge, been studied in the PHL population, leaving us to ponder the results of the Children's Oncology Group study AHOD 0031, wherein after IFRT to 21 Gy, the primary site of relapse (>90%) occurred within the radiotherapy site [4]. These findings left the authors to suggest that 21 Gy may be insufficient for controlling disease and that higher doses (ie, 30 Gy in adults), may be necessary to control poorly responding disease. The present study describes ways of safely escalating this radiation dose.

We expect patients to relapse within the postchemotherapy volume. Limited data, however, are available to support that theory. Postchemotherapy residual disease volumes have been treated in other protocols (NCT01868451, NCT01920932), including the GHSG H15 study [14], which reported excellent outcomes in adults. Field reductions or boost fields have been used in numerous protocols during the past 30 years, but only a randomized phase 3 study can answer whether these target volumes provide similar levels of control to more standard ISRT volumes. Unfortunately, such a study is not feasible in this rare disease for which the focus is on exploring targeted therapies and immunotherapies. For RT to remain relevant in future PHL studies, a smaller RT volume must be adopted and evaluated just as ISRT/involved-node radiotherapy was adopted 10 years ago alongside new RT technologies, such as intensity-modulated radiation therapy, proton therapy, and the breath-hold technique [15–18]. Understanding relapse patterns within these new field designs will help us understand whether target volumes have become too small and whether the doses we use are appropriate. We may find that treating even smaller fields is feasible [19].

Although the present study cannot define which patients benefit most from RT, the appropriate RT field size, or the appropriate total dose in the treatment of adult PHL, it does report 2 important findings: first, an additional 9-Gy boost to residual postchemotherapy disease can be added to a standard 21-Gy ISRT plan with minimal increase in the radiation dose to the OARs; second, by treating only the postchemotherapy volume to 30 Gy, we can reduce the dose to the OARs in a clinically meaningful way compared with 21 Gy of ISRT.

Conflicts of Interest: Bradford S. Hoppe, MD, MPH, and Nancy P. Mendenhall, MD, are Associate Editor and Editor-in-Chief of the International Journal of Particle Therapy, respectively. In addition, Dr Hoppe is a scientific consultant for Merck & Co, Inc, and Bristol-Myers Squibb. Raymond B. Mailhot Vega, MD, MPH, has received a travel grant from IBA.

Acknowledgments: The authors wish to thank Robin Cacchio, RN, for research administration assistance; Jeffrey Glidden, Nataly Getman, and Debbie Louis for dosimetric assistance; and Christopher Stich for editorial assistance. Bradford S. Hoppe, MD, MPH, is grateful for the support of the James E. Lockwood, Jr, Professorship.

Funding: The authors have no funding to disclose.

Ethical Approval: All patient data were collected under internal review board–approved protocol.