CNSC-91. UNCOVERING DIAGNOSTIC UNCERTAINTY IN PRESUMED CNS METASTASES: A REAL-WORLD ANALYSIS OF SUBSEQUENT MALIGNANCY DIAGNOSIS PATTERNS

Abstract BACKGROUND In patients with systemic cancer, new brain lesions are often presumed metastatic and treated empirically with radiation, risking misdiagnosis. We evaluated how often alternative malignancies are later found and compared outcomes between patients receiving radiation alone versus surgery plus radiation. METHODS Using the TriNetX global research network, we identified adults treated for brain metastasis or leptomeningeal disease between 2005 and 2024. Patients were stratified into radiation-only (n = 72,701) and surgery plus radiation (n = 15,131) cohorts. We assessed rates and timing of subsequent diagnoses of primary brain tumors or second systemic malignancies, along with survival outcomes, using cumulative incidence, Kaplan–Meier curves, and log-rank testing. RESULTS A subsequent malignancy was identified in 16.3% of the surgery plus radiation cohort (n = 2,471) and 8.4% of the radiation-only group (n = 6,102) (p < 0.0001). Primary brain tumors occurred in 9.5% vs. 5.1%, and second systemic malignancies in 3.3% vs. 2.4%, respectively. Median time to diagnosis was shorter in the surgery cohort across all conditions (primary brain tumor: 60 vs. 95 days; second malignancy: 30 vs. 37 days; any malignancy: 55 vs. 70 days; all p < 0.001). In a matched cohort, survival from the time of CNS metastasis diagnosis was longer in the surgical group (682 vs. 465 days, p < 0.0001), but no difference was observed when indexed to the time of second malignancy diagnosis (679 vs. 851 days, p = 0.97). CONCLUSIONS Some patients treated empirically for presumed CNS metastases are later found to have other malignancies. Earlier detection in surgical cases highlights the importance of tissue diagnosis, while CSF-based liquid biopsy offers a non-invasive alternative when surgery is not an option.