Abstract / Summary
Abstract Background The role of molecular alterations in predicting local and distant progression following brain metastasis resection is poorly understood. In this study, we investigated clinical and molecular predictors of intracranial progression following brain metastasis resection. Methods A retrospective analysis of 228 patients who underwent brain metastasis resection at University of Pittsburgh Medical Center was performed. Tumor specimens underwent next-generation sequencing using the UPMC Oncomine Comprehensive Assay. Bootstrap forest analysis was used to identify top genetic predictors by variable importance for local and distant progression. Cox proportional hazards models were used to identify independent predictors. Results Local progression occurred in 110 patients (48.2%) and distant progression in 98 patients (43%). The most frequent genetic alterations were TP53 mutations (61.6%), KRAS mutations (31.2%), and CDKN2A/B copy number alterations (24.6%, 23.2%). For local progression, bootstrap forest analysis identified PTEN, STK11, KRAS, FGF3, and ATM alterations as top predictors. On multivariable analysis adjusting for primary malignancy type, independent predictors of local progression included RB1 mutations (HR 3.03, p = 0.003), PTEN copy number alterations (HR 0.25, p = 0.02), subtotal resection (HR 2.1, p = 0.04), and postoperative systemic therapy (HR 0.4, p = 0.002). For distant progression, PTEN, TERT, STK11, CDKN2AB, and TP53 alterations were top predictors. Postoperative systemic therapy (HR 0.45, p = 0.01) was the only independent predictor of distant progression in multivariable analysis. Conclusion These findings suggest that integrating molecular profiling with clinical factors may contribute to intracranial progression risk stratification after brain metastasis resection, though expansion of the dataset with histology-stratified analyses is warranted to confirm these observations.