Utilization patterns and outcomes of radiotherapy for patients with germ cell tumor relapsing with brain metastases, with a focus on stereotactic radiosurgery.

5032 Background: Brain metastases (BM) from germ cell tumors (GCT) are a poor-prognostic feature. Nonetheless, many patients (pts) relapsing with BM achieve long-term disease control. Radiotherapy (RT) is a key component of GCT BM management, alongside salvage chemotherapy (CT), but minimal data are available to guide its use. Outcomes of stereotactic radiosurgery (SRS) alone, without whole brain RT (WBRT), are seldom reported. We describe treatment patterns and outcomes of pts with GCT who received RT for BM relapse. Methods: Male pts with extracranial GCT who received RT for BM relapse between 2005 and 2023 were included. Four subgroups were defined based on incorporation of RT into the broader BM treatment strategy: RT without concurrent salvage CT (Group 1), RT with salvage conventional-dose CT (CDCT) (Group 2), RT with high-dose CT (HDCT) (Group 3), and salvage RT for progressing BM initially treated with CT alone (Group 4). RT approach included WBRT and SRS. Primary outcomes were overall survival (OS) and intracranial progression (IP) from RT completion. Local progression (LP) after SRS was also assessed. Associations with outcomes were modeled with Cox regression and competing risk regression, accounting for death. Results: Sixty pts were included. Median age at BM diagnosis was 28 years. Median follow-up among those alive was 87 months. At 3 years after RT, OS was 39%, and IP was 44%. Groups 1, 2, 3, and 4 included 20, 17, 13, and 10 pts, respectively. Groups were not significantly associated with other baseline features, including RT approach. Outcomes by Group are reported (Table). OS was not significantly different in any Group. There was a trend for worse IP in Group 4, compared with all other groups (HR 2.2, 95% CI 0.99-5.1, p = 0.05). WBRT and SRS alone were used in 32 and 27 pts, respectively. RT approach was not significantly associated with other baseline features. Outcomes by RT approach are reported (Table). OS and IP were not significantly different. Among 53 BM treated with SRS, LP was 8% at 3 years. Among 27 pts treated with SRS, ≥ 2 BM (13 pts) predicted increased risk of death (HR 4.9, 95% CI 1.5-16, p = 0.01) and IP (HR 3.6, 95% CI 1.2-11, p = 0.02). This was not observed after WBRT. Conclusions: We found heterogeneous patterns of RT use for BM relapse: RT alone, RT with CDCT, RT with HDCT, and salvage RT after initial CT alone. Long-term survival and intracranial control are achievable with each strategy and with both WBRT and SRS. This analysis includes the largest reported series of GCT BM treated with SRS; local control after SRS is excellent, but caution is advised for pts with multiple BM, given elevated risk of IP and death. Table: see text