Diagnostic Performance of Data-Driven MRI Habitat Analysis for Preoperative Prediction of IDH Status in Adult Gliomas: A Systematic Review and Diagnostic Accuracy Meta-Analysis

Background/Objectives: Isocitrate dehydrogenase (IDH) status is important for the classification and prognosis of adult gliomas. Whole-tumor MRI summary measures may obscure intratumoral spatial heterogeneity, whereas data-driven habitat analysis identifies imaging-defined tumor subregions and may better capture this heterogeneity. This study aimed to evaluate the diagnostic performance of MRI habitat analysis for preoperative prediction of IDH status. Methods: We searched PubMed, Embase, Web of Science Core Collection, China National Knowledge Infrastructure (CNKI), and Wanfang Data from database inception through 22 August 2026. Eligible studies used data-driven MRI habitat analysis to predict IDH status in adult gliomas. Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool. Sensitivity and specificity were pooled using a bivariate random-effects model. We also performed sensitivity analyses, univariable meta-regression, subgroup analyses, and Radiomics Quality Score 2.0 (RQS 2.0) assessment. The protocol was registered in PROSPERO (CRD420261489645). Results: Eighteen reports were included in the systematic review, and 13 independent studies involving 1236 patients contributed to the meta-analysis. Pooled sensitivity and specificity were 0.78 (95% CI, 0.70–0.84) and 0.85 (95% CI, 0.79–0.89), respectively, with an SROC AUC of 0.89 (95% CI, 0.86–0.91). I2 was 82.1% for sensitivity and 71.3% for specificity. Excluding two studies with approximately reconstructed 2 × 2 tables did not materially change the pooled estimates. Radiomics studies and studies using ≥4 habitats showed higher specificity and lower sensitivity, but their overlap limited separate interpretation of feature type and habitat number. Conclusions: MRI habitat analysis showed promising diagnostic performance for preoperative prediction of IDH status in adult gliomas. However, limited external validation and incomplete assessment of reproducibility and clinical utility remain barriers to clinical translation. Future prospective multicenter studies should further evaluate habitat stability, model calibration, and clinical net benefit.