Cerebrospinal fluid circulating tumor DNA profiling for risk stratification and matched treatment of central nervous system metastases.

Genomic profiling of central nervous system (CNS) metastases has the potential to guide treatments. In the present study, we included 584 patients with non-small-cell lung cancer and CNS metastases and performed a comprehensive analysis of cerebrospinal fluid (CSF) circulating tumor DNA (ctDNA) with clinicopathological annotation. CSF ctDNA-positive detection was independently associated with shorter survival than negative detection (hazard ratio (HR) = 1.9, 95% confidence interval (CI) = 1.56-2.39; P < 0.0001). Matched tumor-CSF analysis characterized the CSF private molecular features causing poor survival (HR = 1.64, 95% CI = 1.15-2.32, P = 0.006). A multimetric CSF ctDNA prognostic model integrating CSF ctDNA features and clinical factors was developed for risk-stratifying CNS metastases and validated in an independent cohort. Among patients with treatment histories available, those positive for a driver alteration by CSF ctDNA showed a survival benefit from CSF-matched therapy (HR = 0.78, 95% CI = 0.65-0.92, P = 0.003). Longitudinal monitoring by CSF identified CNS-specific resistant mechanisms and a second matched targeted therapy indicating improved survival (HR = 0.56, 95% CI = 0.35-0.91, P = 0.018). These findings support the clinical value of CSF ctDNA for risk-stratifying CNS metastases and guiding therapy.