Population-Based Real-World Outcomes of Post-Operative Adjuvant Brain Cavity Radiotherapy Versus Observation

To evaluate the factors influencing the outcomes of patients who underwent surgical resection of brain metastasis followed by either surveillance or post-operative stereotactic radiosurgery/fractionated radiotherapy (SRS/SFRT), a retrospective multi-center chart review was performed on all patients who underwent brain metastases resection in British Columbia between 2018 and 2020. Patients with prior whole-brain radiotherapy were excluded from the study. The primary study endpoints included local recurrence, distant intracranial control, radionecrosis (RN), leptomeningeal disease (LMD), and overall survival (OS). The Kaplan-Meier method was used to analyze survival. The Cox proportional hazards model was used to perform univariable (UVA) and multivariable (MVA) analyses to identify predictors of local control. A total of 113 patients met the inclusion criteria. A total of 31 patients received adjuvant SRS/SFRT to the surgical cavity, while 82 went on observation. The 12-month local control was 69% (50-88%) for the SRS/SFRT cohort and 31% (18-45%) for the observation cohort (p < 0.001). The 12-month distant intracranial control was 44% (26-63%) for the SRS/SFRT cohort and 46% (30-62%) for the observation cohort (p = 0.9). Sensitivity analysis did not show a difference in overall survival (p = 0.6). En bloc resection (p < 0.05), resection without residual disease (p < 0.05), and SRS/SFRT (p < 0.001) were predictive of local control on MVA. Three SRS/SFRT patients (10%) and two observation patients (2%) developed LMD. Four SRS/SFRT patients experienced RN (13%), with no grade 3 or higher toxicities observed. Post-operative SRT outcomes based on real-world population data are consistent with the data from clinical trials and support the established guidelines. For patients requiring surgical resection of brain metastasis, en bloc gross total resection should be encouraged when feasible to reduce local recurrence.