Abstract
Brain metastases have unique genetic mechanisms which enable them to engraft and grow either within the brain (solid) or on the surface of the brain and nerves (leptomeningeal disease; LMD) A solid brain metastasis occurs in up to 45% of all cancer patients compared to an approximate 5% incidence of LMD. Unlike solid brain metastases, LMD cannot be resected and has a worse survival, often times weeks to months after diagnosis. Why metastatic cells favor one brain site over the other has yet to be determined. To date, there has been no systematic study of the genetic differences between LMD and solid brain metastases. Through whole exome sequencing we compared the mutational landscape of lung-to-brain metastases in eight cytology-positive LMD samples and 38 solid brain tumor metastases. Cerebral spinal fluid and normal control samples (blood or saliva) of patients with LMD were collected at Stanford Hospital. Samples underwent DNA extraction, indexed library preparation, exome enrichment, and sequencing with pair end-reads on an Illumina NextSeq system. LMD mutations were compared to exome sequencing data of solid brain metastases in the Cancer Genome Atlas. We identified a subset of recurrently mutated genes (e.g. EGFR, TAS2R31, CACNA1I) that occured more frequently in our LMD samples. Another subset of genes (e.g. KRAS, LRP1B, CSMD3) were more frequently mutated in the solid brain metastasis samples. TP53, MUC16, MUC17 were among the genes shared across LMD and solid lung-to-brain metastases. No recurrent mutations were shared between all LMD samples, highlighting LMD heterogeneity. Solid brain metastases and LMD appear to harbor distinct genetic mutations, suggesting unique mechanisms for growth in these unique brain niches. A better understanding of the genomic events facilitating lung-to-brain metastases will improve targeted treatment options, and eventually patient outcomes.