An excess mortality risk analysis of proton beam versus optimal photon radiotherapy for mediastinal Hodgkin lymphoma: who may benefit most?

Abstract

Purpose/Objective(s): Previous studies have shown that proton beam therapy (PBT) can reduce radiation exposure of normal tissues in patients with mediastinal lymphoma compared to photon radiotherapy (RT). However, most studies were small, did not use intensity modulated radiotherapy (IMRT) which avoids low dose bath and focused mainly on dosimetric endpoints. Our purpose was to compare PBT to best available photon RT in a large cohort to identify which patients benefited most from PBT both dosimetrically and in terms of predicted excess mortality risk (EMR) from cardiovascular disease (CVD) and second cancers (SC). Materials/Methods: Between 2015 and 2019, 80 supradiaphragmatic Hodgkin lymphoma patients with mediastinal disease were treated with pencil beam scanning PBT in deep inspiration breath hold (DIBH). We replanned this cohort using Butterfly-VMAT (BVMAT) in DIBH to provide a comparison between clinical PBT and “best-plan” photon RT. We predicted EMR from CVD and SC (breast, lung and esophagus) using established dose-response relationships and mean organ doses. Doses to cardiac substructures (left ventricle, valves) and to carotids were also included in the prediction of EMR from CVD, as was the chemotherapyrelated EMR. The two RT methods were compared using within-patient multiple regression. Results: Compared to BVMAT, PBT substantially reduced cardiac doses and EMR from CVD but only for those patients whose clinical target volume (CTV) overlapped longitudinally with the heart by >40% (n Z 23). For those with 40% overlap (n Z 57), there was no substantial difference between PBT and BVMAT. Mean lung dose and EMR from lung SC was reduced substantially for all patients with PBT, but an additional benefit was observed for those with axillary disease (n Z 25). The EMR from breast cancer was halved for female patients with axillary disease, but this was a small effect in absolute terms. (Table 1) The integral dose was doubled with BVMAT compared to PBT (4.7 Gy from 2.3 Gy). Conclusion: PBT reduced radiation exposure to normal tissues to varying degrees in different patients. Only patients with >40% longitudinal overlap of CTV with the heart received substantial reduction of EMR from CVD. EMR from lung SC was reduced for all patients, and those with axillary involvement received the greatest benefit in terms of EMR from lung and breast SC. The EMR from breast SC, however, was low for both methods, attributable to the excellent sparing of the female breasts with both RT techniques.