Raman spectroscopy for glioma diagnosis toward precision neuro-oncology: A systematic review

INTRODUCTION: Gliomas are the most common primary malignant tumors of the central nervous system and remain difficult to manage because of diffuse infiltration and molecular heterogeneity. As glioma classification increasingly relies on molecular biomarkers, interest has grown in label-free technologies capable of providing real-time biochemical information during surgery. Although Raman spectroscopy has emerged as a promising optical approach, its diagnostic performance and translational readiness remain incompletely defined. METHODS: This systematic review was registered in PROSPERO and conducted according to PRISMA 2020 guidelines. PubMed, Embase, Scopus, and Web of Science were searched for studies published between 2015 and 2025 evaluating Raman-based technologies for glioma diagnosis. Two reviewers independently performed study selection, data extraction, and risk-of-bias assessment using QUADAS-2. Owing to methodological heterogeneity, findings were synthesized using a SWiM-guided narrative approach. RESULTS: Forty-six studies met the eligibility criteria. Across conventional Raman spectroscopy, surface-enhanced Raman spectroscopy, stimulated Raman histology, stimulated Raman scattering, and multimodal platforms, Raman-based technologies consistently demonstrated strong performance for tissue discrimination, tumor margin assessment, molecular characterization, and intraoperative tissue evaluation. Recent studies increasingly integrated artificial intelligence for automated tissue classification and molecular prediction. However, substantial heterogeneity in acquisition protocols, analytical workflows, and validation strategies, together with predominantly retrospective single-center studies and limited external validation, reduced confidence in clinical generalizability. QUADAS-2 identified the greatest risk of bias in the patient-selection and index-test domains. CONCLUSIONS: Raman spectroscopy is evolving from an optical imaging technique into a molecular decision-support platform for precision neuro-oncology. Future clinical translation will depend on methodological standardization, prospective multicenter validation, explainable artificial intelligence, and integration into routine neurosurgical workflows.