CSF proteomics for differentiation of brain malignancies.

Abstract

2021 Background: Accurate diagnosis and prognostication of intra-axial brain tumors hinges on invasive brain sampling, which carries risk of morbidity. Minimally invasive sampling of proximal fluids, also known as liquid biopsy, can mitigate this risk. Although the cerebrospinal fluid (CSF) is the ideal liquid biopsy source, the traditionally high volumes required for impactful analyses have deterred progress. The objective of this study was to identify diagnostic and prognostic CSF proteomic signatures in glioblastoma (GBM), brain metastases (BM), and central nervous system lymphoma (CNSL). Methods: CSF samples were retrospectively retrieved from the Penn State Neuroscience Biorepository and profiled using shotgun proteomics with ultra-low sample volumes. Proteomic signatures were identified using machine learning classifiers and survival analyses. Results: Using 30µL CSF volumes, we recovered 800 unique peptides across 73 samples [20 normal pressure hydrocephalus (NPH, non-tumor control), 22 GBM, 17 BM, and 14 CNSL]. Externally-validated proteomic-based classifiers identified malignancy with AUROC of 0.94 and distinguished individual tumor entities from others with AUROC ≥0.96. More clinically relevant triplex classifiers, comprised of just 3 peptides, distinguished individual tumor entities with AUROC ≥0.90. Novel biomarkers were identified among the top classifiers, including TFF3 and CACNA2D2, and characterized using single-cell RNA sequencing data. Survival analyses validated previously implicated prognostic signatures, including blood brain barrier disruption. Conclusions: Reliable classification of intra-axial malignancies using ultra-low CSF volumes is feasible, which has ramifications for longitudinal tumor surveillance. Novel biomarkers identified here necessitate future validation. Based on emerging evidence, upfront implantation of CSF reservoirs in brain tumor patients warrants consideration.