RDTA-19 IDENTIFYING PREDICTIVE FACTORS FOR LOCAL RECURRENCE VS. RADIATION NECROSIS IN BIOPSY-PROVEN PROGRESSIVE LESIONS AFTER STEREOTACTIC RADIOSURGERY FOR BRAIN METASTASES

Abstract BACKGROUND Stereotactic radiosurgery (SRS) remains the standard of care for the initial management of brain metastases (BM), offering excellent local control with fewer neurocognitive side effects reported. However, when radiographic progression occurs post-SRS, distinguishing radiation necrosis (RN) from local recurrence (LR) remains challenging. Identifying clinical and treatment-related factors associated with RN or LR may aid in more informed decision-making. In a smaller cohort, increased time to progression was associated with RN, and we aimed to reassess this relationship in a larger patient population. METHODS We retrospectively analyzed patients treated with SRS for BM between 2011–2022 who later underwent biopsy, with or without LITT, for imaging-suspected LR vs. RN. We collected patient demographics, clinicopathologic characteristics, SRS-treatment parameters, and biopsy results. RESULTS Clinical data was available for 145 patients with 151 progressive lesions; a smaller subset (n = 108) had detailed radiosurgery parameters available. Post-SRS biopsy confirmed RN in 93 lesions (61.6%) and LR in 58 (38.4%). Median age at SRS was 62 years (range: 27–92), with a median interval to biopsy of 13 months (range: 3–81) and a median prescribed dose of 20.0 Gy (range: 14.0–30.0). By multivariate analysis, patients aged ≥ 65 years at SRS with lesions biopsied ≥ 9 months post-SRS were significantly associated with RN (38/52 lesions, 73.1%). Additionally, in the radiosurgery subset, a V12 ≥ 4.64 cc favored RN, while a maximum dose ≥ 22.32 Gy was associated with higher odds of LR. CONCLUSION The natural history of LR and RN post-SRS differ such that in patients ≥ 65.0 years at the time of SRS with lesion progression ≥ 9 months post-SRS and higher reported V12, RN is more likely, whereas patients < 65 years with earlier progression are more likely to have LR. These findings highlight the value of incorporating predictive factors into clinical decision-making to optimize patient management.