209 Detection and Characterization of Glioma and Therapy-associated Biomarkers in CSF. (1st April 2022)

• Record Type: Journal Article
• Authors: Carlstrom, Lucas P.; Riviere-Cazaux, Cecile; Casabella, Amanda; Rahman, Masum; Burns, Terry C.
• Is Part Of: Neurosurgery. Volume 68(2022)Supplement 1
• Page Start: 64
• Page End: 64
• Publication Date: 2022-04-01
• Subjects: Nervous system -- Surgery -- Periodicals; 617.48005
• DOI: 10.1227/NEU.0000000000001880_209
• Languages: English
• ISSNs: 0148-396X

Abstract

Abstract : INTRODUCTION: Molecular characterization of human gliomas currently requires access to tissue samples. While tumor-associated cell-free DNA has been detected in CSF and cell-free DNA, utility is limited to those harboring specific detectable mutations. Gliomas with IDH-mutations produce high levels of D-2-Hydroxyglutarate (D2-HG), a stable and quantifiable oncometabolite that is correlated with tumor burden as assessed within the tumor mass. However, the reliability of D2-HG in CSF as an in vivo pharmacodynamic biomarker is unknown. METHODS: A brain tumor CSF biorepository has been launched, including prospective samples obtained intra-operatively and via ventriculoperitoneal shunts, Ommaya reservoirs, or lumbar punctures at clinically-relevant time points. Analyses included targeted and untargeted metabolomic analyses, and PCR for patient-specific chromosomal junctions utilizing primers specific to junctions predicted to be stable over time. RESULTS: Twenty-one CSF samples were utilized for one or more analyses. Patients with IDH-mutant gliomas had over twenty-fold higher levels of D2HG in CSF (median 4.1 uM, range 1.6-13.2 uM, n = 7) compared to IDH-wild type gliomas (median 0.19 uM; range 0.89-0.35, n = 14) (p < 0.001). A total of 13 CSF samples from 8 patients were evaluated via targeted and untargeted untargeted metabolomics. 282 metabolites were identified consistently in all 13-samples. In addition to IDH-mutant status, signatures were identified conveying patient identity Correlation and Principle component analyses identified distinct clustering of CSF samples by patient identify, prior radiation status, and clinical/collection variables including plasma contamination, and NPO status. Samples obtained after prior radiation were enriched for curated metabolite sets including TCA cycle and Warburg effect. Relative CNS levels of several pharmacologic agents (acetaminophen/levetiracetam/mannitol/caffeine) could also be quantified from the untargeted metabolomic profiles. Multiple unique patient tumor-specific chromosomal junctions were robustly quantifiable in all evaluated CSF samples, but no plasma samples, correlating with tumor burden across serial samples. CONCLUSION: Mutation-specific oncometabolites, cell-free DNA mutations and signatures of prior therapy were sensitively and dynamically detected in intra-operative and ventricular CSF samples. CSF may provide an underutilized resource to monitor glioma burden and phenotype.