Dose-Toxicity Relationship Algorithm for Reirradiation: A Novel Tool for ‘How to Treat in a ‘No-Treatment’ Zone’

Abstract

Reirradiation is an increasing problem. Due to longer-life expectancies with better anti-cancer therapies, larger numbers of patients are candidates for reirradiation because of delayed recurrence. Patients are also more commonly undergoing multiple treatment courses within the same body site as part of subsequent metastatic local consolidation (LCT). However, little is known about the safety of reirradiation throughout the body. Given the wide-range of dose/fractionation schedules, and lack of a common conversion format for dose-toxicity relationship assessment, a void has existed for physicians wishing to pursue reirradiation. This is most commonly observed at the time of planning directive placement and knowledge of toxicity thresholds for organs at risk (OARs) in the reirradiation setting. To address this, we created a novel algorithm to convert nominal dose (Gy) from prior treatment plans to corresponding BED and EQD2 values, at voxel size of 2.5mm. Using this algorithm, and an alpha/beta of 3 for normal tissues, a proof of principle analysis was performed on 4 reirradiation NSCLC patients receiving 1) 50Gy in 4 fraction therapy followed by 70Gy in 10, 2) 50Gy in 4 followed by 50Gy in 4, 3) 70Gy in 10 followed by 70Gy in 10, and 4) 50Gy in 4 followed by 60Gy in 30. All cases were in the lung and same lobe. Following the conversion of each plan to its corresponding BED and EQD2 map, deformable registration was used to create a summed composite reirradiation BED/EQD2 plan for each patient. This allowed for composite – and anatomically accurate -- dosimetric EQD2 and BED assessment for each OAR. Evaluation of the algorithm-generated maps revealed accurate conversion of Gy to BED/EQD2 at each voxel. This enabled evaluation of dose summation received by organs at risk in the reirradiation setting (e.g. mean heart BED dose, max BED dose to major vessels, mean esophageal and lung BED). These dose characteristics were then compared to the toxicities experienced by patients following reirradiation to understand what dose thresholds might exist for OARs, regardless of fractionation scheme(s) chosen by physicians. To this end, the novel BED/EQD2 algorithm was successfully used to assess toxicity relationships in OARs for 15 patients receiving re-SABR after initial SABR, and 26 patients receiving curative chemoradiation (CRT) after initial SABR. This relatively simple, but highly clinically relevant, algorithm allows for assessment of dose-toxicity relationships using a common format in the reirradiation setting. Using this algorithm, multiple cohorts of reirradiation patients are being analyzed at our center to characterize dose-thresholds for reirradiation. Using BED/EQD2 reirradiation thresholds, physicians everywhere can calculate dose received to OARs during the first radiation course and, after selecting fractionation for the second, be able to generate evidence-based planning directives for organs at risk in the reirradiation setting.