Abstract
INTRODUCTION: Central nervous system (CNS) metastasis in patients with urothelial carcinoma (UC) is uncommon and poorly understood. We aimed to explore the clinical behavior and outcomes of this unique patient population. MATERIALS AND METHODS: We performed a retrospective analysis of patients with UC and CNS metastasis, treated in our institution (2006-2018), along with an exploratory patient-point meta-analysis of a similar patient population derived from a comprehensive literature review. Data regarding diagnosis, management, and outcomes were extracted. Overall survival, time to CNS metastasis (TTCM), and residual survival (RS) from CNS involvement to death were calculated (Kaplan-Meier method). Cox regression was used for testing key clinicopathologic associations. RESULTS: We identified 20 "institutional" and 154 "literature" patients with adequate data granularity for analysis. Median TTCM was 17.7 (institutional cohort) and 10 (literature cohort) months. Most patients who developed CNS metastases had previous non-CNS metastasis (15/20 [75%] and 103/154 [67%], respectively). CNS lesions without previous history of metastasis were identified in 5/20 (25%) and 33/154 (21%) cases and those patients had a shorter TTCM. CNS lesions in the absence of known UC history were also documented in 18/154 (12%) literature cases. Multifocal CNS disease was associated with shorter RS in both cohorts in univariate, but not multivariate, analysis. CONCLUSION: We observed a variability in disease presentation and course, with a subset of patients showing an early predilection for CNS insult, potentially reflecting a diverse underlying biology. Genomic profiling studies, elucidating the molecular landscape, and driving future treatments should be considered in this setting.