Biomarker-Guided Dynamic Liquid Biopsy System for Enhanced Detection of Glioblastoma Recurrence via Blood–Brain Barrier Permeability Optimisation

Abstract Background: We propose a biomarker-guided dynamic liquid biopsy system to enhance glioblastoma (GBM) recurrence detection by optimising blood–brain barrier (BBB) permeability timing for circulating tumour deoxyribonucleic acid (ctDNA) analysis. Current approaches often suffer from low sensitivity due to unpredictable ctDNA shedding and static sampling. Methods: The proposed method integrates real-time measurements of BBB permeability biomarkers including albumin index, glial fibrillary acidic protein (GFAP) and S100B into a composite score that dynamically triggers ctDNA analysis when permeability exceeds patient-specific thresholds. This replaces fixed sampling with a decision engine predicting optimal windows aligned with peak BBB disruption. The hardware implementation combines high-sensitivity immunoassays with droplet digital polymerase chain reaction and next-generation sequencing, supported by a Kubernetes-managed workflow orchestrator for seamless integration. Key innovations include the use of GFAP and S100B kinetics as proxies for BBB dynamics, a closed-loop feedback mechanism between permeability scores and ctDNA analysis parameters and a microfluidic platform for parallel biomarker and ctDNA processing. Results: Clinical validation demonstrates higher sensitivity and specificity, supporting earlier and more accurate recurrence detection. The system adapts to inter-patient variability in BBB physiology, representing a meaningful advance in precision oncology and offering potential applications in other central nervous system malignancies with intermittent ctDNA release. Conclusion: The biomarker-guided dynamic liquid biopsy improves GBM recurrence detection by aligning ctDNA analysis with real-time BBB permeability, enhancing sensitivity and specificity and enabling earlier monitoring.