BIOM-03. PLASMA BIOMARKERS FOR NON-INVASIVE GLIOBLASTOMA DIAGNOSIS: A MULTIMARKER STRATEGY WITH GFAP, UCH-L1, NFL, TAU, MPO, AND PMN ELASTASE

Résumé fourni par la source

Abstract Glioblastoma (GBM) is the most aggressive primary brain tumor in adults, with poor prognosis and limited non-invasive diagnostic options. Current standards such as imaging and tissue biopsy present challenges in sensitivity, accessibility, and patient burden. In the context of the NucleoGLIO initiative, we evaluated the diagnostic performance of six plasma biomarkers—GFAP, UCH-L1, NfL, Tau, MPO, and PMN Elastase—in a retrospective cohort of 61 GBM patients and 102 healthy controls (CHU-UCL Namur Ethical Agreement 229/2023). Biomarker levels were measured using high-sensitivity digital immunoassays (Simoa, Quanterix®) and ELISA platforms, followed by statistical analyses including ROC curve generation and multivariate logistic regression. All biomarkers showed significantly higher levels in GBM patients. GFAP demonstrated the strongest individual diagnostic power, with an AUC of 0.99 and nearly complete group separation. NfL and UCH-L1 also performed robustly, with AUCs of 0.95 and 0.92, respectively. Tau and MPO showed moderate performance, while PMN Elastase yielded limited discriminatory value. Multimarker models integrating UCH-L1, NfL, and Tau improved diagnostic accuracy (AUC = 0.96), and the inclusion of GFAP led to perfect group separation, underscoring its potential as a standalone marker. Predictive modeling further enabled the identification of clinically actionable thresholds, achieving negative predictive values (NPV) approaching or reaching 100%, especially for GFAP (cut-off: 119.9pg/mL) and combinations of UCH-L1, NfL, and Tau (cut-off, respectively: 23.0pg/mL, 11.6pg/mL and 0.8pg/mL). Notably, adding demographic variables such as age and sex had minimal impact on model performance, suggesting the robustness of the biomarker approach. These findings support the development of a minimally invasive, blood-based diagnostic tool for GBM, with the potential to reliably exclude disease in at-risk individuals and guide further clinical decision-making. Overall, this study demonstrates that a multi-analyte plasma biomarker panel, particularly one incorporating GFAP, offers a highly promising strategy for improving early detection and monitoring in glioblastoma care.