Outcomes of Brain Metastases with Suspicious Imaging Undergoing Resection to Evaluate for Radionecrosis vs. Tumor Progression

Abstract

In patients treated with stereotactic radiosurgery (SRS) for brain metastases, radiographic changes on surveillance imaging may result from treatment effect/radionecrosis (RN) or tumor progression. Distinguishing between these processes is critical to appropriate management. We report long-term outcomes for a cohort of patients who demonstrated radiographic progression on serial imaging after initial radiation and ultimately underwent resection to inform further management. A retrospective chart review identified 76 patients with an associated 82 brain lesions between 2009 and 2022 that were initially treated with SRS, then demonstrated suspicious imaging changes developing through at least two scan time points that led to pathologic confirmation of either tumor or RN. We report clinical outcomes and details of further treatments. Of the 82 lesions, 55 (67.1%) were found to be pathologically-confirmed viable tumor and were treated with repeat radiation and 27 (32.9%) were found to be strictly RN and conservatively managed. Over half of the lesions (14/27) ultimately found to be radionecrotic required use of steroids pre-operatively due to neurologic symptoms. Among the 27 that were found to be RN, the most common histology was melanoma (33.3%, n = 9). The most common dose fractionation regimen was 20 Gy in 1 fx (n = 11, 40.7%; range: 16-20 Gy x 1Fx), and the median BED (10) was 50.4 Gy (IQR 41.6 - 50 Gy). None of these lesions required further intervention with median post-surgery follow up of 24.4 months (range 1-104 months). There were 55 instances (in 51 patients) of pathologically-confirmed recurrent/progressive tumor who were consequently treated with repeat radiation with either Cs-131 brachytherapy (12 (21.8%)) or SRS (43 (78.2%)). The most common histology was NSCLC (37.2%, n = 19). The most common fractionation for repeat irradiation with SRS was 8 Gy x 3 fx (n = 15, 27.3%), followed by 5 Gy x 5 fx (n = 10, 18.2%), and 4 Gy x 5 fx (n = 8, 14.6%). Four individuals each had two lesions that were re-irradiated for local recurrence. Among patients treated with re-irradiation, the median follow-up to local failure was 15.2 months (95% CI 7.3-26.6 months). Radionecrosis was confirmed on pathology in 4/55 (7.2%) of lesions. The median follow-up from date of SRS2 to local failure was 14.1 months (95% CI 7.6-24.3 months). The 2-yr local control rate was 74.8% (95% CI 61.7-90.7%). We recommend cautious monitoring of possible progression after radiosurgery, with consideration of resection for continuous progression, as a significant proportion of radiographic progression are ultimately pure RN. Management determined by pathology (observation for RN; additional radiation for confirmed tumor) leads to excellent control.