CSF ctDNA analysis guides molecular reclassification of diffuse glioma patients

BACKGROUND: There is a critical need for better strategies to diagnose and monitor patients with gliomas. Liquid biopsies are less invasive than tissue biopsies and are amenable to serial collection at multiple timepoints. However, the analysis of circulating tumor DNA (ctDNA) in cerebrospinal fluid (CSF) from patients with gliomas remains underutilized in clinical practice due to technical challenges. METHODS: We collected CSF samples from 43 patients with various types of gliomas and evaluated ctDNA derived from CSF with a next generation sequencing (NGS) panel interrogating 600 cancer-related genes. RESULTS: Total cell-free DNA (cfDNA) concentration obtained from CSF (0.4-5 mL) ranged from 0.09 to 60 ng/µL. Mutations were detected in 26/43 (60%) samples including 19/43 (44%) with single nucleotide variants (SNV), 11/43 (25%) with copy number variants (CNV), 8/43 (18.6%) with frameshifts and 1/43 with fusion (2.3%). Mutations were not identified in 17 CSF samples, and most of these samples (15/17; 88%) had suboptimal DNA input (< 30 ng). We identified diagnostic alterations in CSF including mutations in IDH1 (3/6 IDH-mutant gliomas), TERTp (7/29 glioblastomas), EGFR amplification (5/29 glioblastomas), and H3-3 A (4/4 pediatric high-grade gliomas). Some of these mutations, in combination with other clinical parameters, allowed us to identify specific CNS tumor types based on the 2021 World Health Organization (WHO) classification of CNS tumors, resulting in the reclassification of 3 patients. Furthermore, we observed a significant association between the CSF-ctDNA detection and imaging features. CONCLUSION: This study highlights the potential clinical utility of CSF-ctDNA analysis for evaluating and reclassification of patients with gliomas.