Illuminating glioma surgery: A meta-analysis of Raman spectroscopy for intraoperative decision-making

Abstract

Precise tumor characterization is key in glioma surgery. Raman spectroscopy offers real-time, molecular-level tissue analysis for intraoperative guidance. This meta-analysis synthesizes current applications of Raman spectroscopy in glioma surgery. Systematic literature search in PubMed/MEDLINE and the Cochrane Library through March 15, 2025, using the algorithm: “Raman AND (brain tumor OR glioma OR glioblastoma) AND (surgery OR intraoperative)” retrieved 206 studies. Studies evaluating intraoperative Raman spectroscopy for glioma diagnosis, using histopathology/molecular biology as the reference standard, were included. Meta-analysis, following PRISMA-DTA/STARD guidelines, compared Raman spectroscopy to histopathology/molecular biology. Pooled analyses assessed Raman spectroscopy’s ability to: (1) delineate tumor margins; (2) grade high-grade (HGG) versus low-grade gliomas (LGG); (3) distinguish astrocytomas from oligodendrogliomas; (4) classify IDH-wildtype (IDHwt) versus IDH-mutant (IDHmut) gliomas; and (5) discriminate gliomas from other brain tumors. From 206 studies, 19 were used in pooled analyses. For tumor vs. normal tissue, pooled sensitivity was 95.66%, and specificity was 86.13%, with heterogeneity due to varying definitions of “normal” tissue. Differentiation of HGG vs. LGG had Limited data; one study showed 91% sensitivity and 87% specificity. Astrocytoma vs. oligodendroglioma differentiation showed a pooled sensitivity of 89.9% and specificity of 86.4%. IDHwt vs. IDHmut glioma classification had a pooled sensitivity of 91.4% and specificity of 90.4%. For glioma vs. other brain tumors, pooled sensitivity was 91.2% and specificity was 91.6%. Raman spectroscopy shows potential for intraoperative glioma characterization. Standardization is needed for better comparison across studies.