The next generation of cerebrospinal fluid (CSF)-based molecular diagnostics: Improving sensitivity and actionability in breast and lung cancer patients with CNS involvement.

Abstract

e14502 Background: Liquid biopsy has emerged as a minimally invasive and cost-effective strategy to assess cancer biomarkers without the risk of complications associated with surgical biopsies. Once a tumor has metastasized to the brain, circulating tumor cells (CTCs) and circulating tumor DNA (ctDNA) can be found in the cerebrospinal fluid (CSF). We analyzed CSF samples from patients(pts) with primary lung or breast cancer with either brain (BM) or leptomeningeal metastases (LM). Here we report the analytical and clinical validation of Target Selector CSF assays. Validation testing included pre-analytical and analytical steps. Methods: CSF was collected prospectively from pts with a prior solid tumor diagnosis and suspected BM or LM. CTCs were captured utilizing a primary ten antibody cocktail followed by biotinylated secondary antibodies that bind selectively to CTCs followed by staining with cytokeratin (CK), CD45, streptavidin and DAPI. CTCs were captured in a microfluidic channel,classified as either CK+ or CK-. Cell-free total nucleic acids (cfTNA) was extracted from CSF supernatant and underwent both Target Selector™ single gene and next-generation sequencing (NGS) lung and breast multi-gene testing to assess for molecular alterations. For NGS, data analysis was performed using Torrent Suite with annotation and curation by Ion Reporter and Oncomine Knowledgebase Reporter software. Results: The Target Selector CTC platform assays performed on clinical samples (n = 89) resulted in clinical accuracy = 85.4%, clinical precision (intra-assay, inter-assay, inter-operator, and inter-instrument) = 100% for each measure, clinical sensitivity = 80.0%, clinical specificity = 96.6%, positive predictive value (PPV) = 98%, negative predictive value (NPV) = 70.0%, and analytical specificity = 96.0% (acceptance criteria was 95%) at a limit of detection of 2 CTCs. For molecular analyses, Target Selector™ platform assays resulted in clinical accuracy = 87.4%, clinical sensitivity = 85.2%, clinical specificity = 88.3%, PPV = 76.7%, and NPV = 93.0%. Conclusions: Target Selector is a viable, sensitive, reproducible platform for CTC detection and molecular analysis of CSF samples from patients with breast or lung cancer with CNS metastases especially as the sensitivity of CSF cytology is low and MRI findings can be equivocal. Identifying CTCs and molecular alterations can help characterize both tumor genomic evolution as well as guide treatment following cancer metastasis to the CNS.