Predicting molecular stratification in gliomas using intravoxel incoherent motion MRI: a systematic review and meta-analysis focusing on IDH1 mutation.

Abstract

Modern glioma management increasingly relies on molecular classification, but tissue-based approaches have inherent limitations including invasiveness, sampling error, and surgical risks. Intravoxel incoherent motion (IVIM) MRI offers a promising radiogenomic alternative by noninvasively quantifying diffusion and microvascular perfusion, though current evidence remains fragmented across studies. To systematically evaluate the discriminatory ability of IVIM MRI parameters in predicting genetic and histopathological biomarkers in gliomas through a hybrid approach combining quantitative meta-analysis and narrative synthesis. Following PRISMA guidelines, four major databases were searched through May 2026. Twelve eligible studies assessing IVIM parameters against molecular or histopathological markers were included. Quality was assessed using QUADAS-2. A formal quantitative meta-analysis was conducted for IDH1 mutation status, while other markers (e.g., ATRX, 1p/19q, and Ki-67) were narratively synthesized due to pronounced methodological heterogeneity. Across the included studies, true diffusion coefficient (D) emerged as the most consistent discriminator among the IVIM-derived parameters. Our quantitative synthesis demonstrated that D values-particularly minimum D evaluated within tumor hotspots-were significantly higher in isocitrate dehydrogenase 1 (IDH1)-mutant gliomas, suggesting a modest diagnostic advantage over conventional apparent diffusion coefficient (ADC) based on the limited available intra-cohort data. Furthermore, narrative synthesis confirmed D as a promising predictor of ATRX loss and inversely correlated with proliferation indices (Ki-67/MIB-1), generally exhibiting slightly stronger correlation coefficients compared to standard ADC. In contrast, the pseudo-diffusion coefficient (D*) and perfusion fraction (f) exhibited high between-study heterogeneity; although biologically linked to microvascular perfusion, their diagnostic capabilities varied significantly across studies and often failed to reach statistical significance. IVIM-derived D serves as a promising and biologically grounded noninvasive biomarker for glioma molecular stratification. While microvascular metrics (D* and f) demonstrate context-dependent value influenced by methodological variability, future studies must establish whether full bi-exponential IVIM modeling provides true incremental clinical value over simpler, perfusion-corrected mono-exponential diffusion techniques. Ultimately, advancing this field toward precision neuro-oncology will be contingent upon strict protocol standardization and multicenter validation.

Abstract as supplied by the indexing source. © the authors / publisher.