Beyond the visual T2-FLAIR mismatch: quantitative imaging markers of IDH mutation in Gliomas-A retrospective diagnostic accuracy study.

BACKGROUND: Accurate noninvasive prediction of isocitrate dehydrogenase (IDH) mutation status is essential for the classification, prognostication, and management of diffuse gliomas. PURPOSE: To evaluate quantitative and semi-quantitative T2-FLAIR-based MRI parameters for differentiating IDH-mutant from IDH-wildtype gliomas. METHODS: This retrospective study included 29 patients with histopathologically confirmed diffuse gliomas who underwent preoperative 1.5-T brain MRI. Quantitative parameters (rT2, rFLAIR, and mismatch [rT2-rFLAIR]) were calculated using region-of-interest measurements relative to contralateral normal parenchyma. A semi-quantitative T2-FLAIR mismatch score (1-4) was assessed on subtraction images. Groups were compared using the Mann-Whitney U test. Diagnostic performance was evaluated using ROC analysis, including comparison of correlated AUCs. Univariable and exploratory adjusted logistic regression analyses were performed, and interobserver agreement for the mismatch score was assessed. RESULTS: Nineteen tumors were IDH-wildtype and ten were IDH-mutant. IDH-mutant gliomas showed significantly higher rFLAIR and rT2 values (p = 0.016 and p < 0.001, respectively). The mismatch score was significantly higher in IDH-mutant tumors (median [IQR]: 3 [3-4] vs. 1 [1-2], p < 0.001). ROC analysis demonstrated excellent performance for the mismatch score (AUC = 0.95) and rT2 (AUC = 0.94). A mismatch score ≥ 3 yielded 90% sensitivity and 89.5% specificity. Both the mismatch score and rT2 showed the strongest associations with IDH mutation status. Interobserver agreement was substantial (weighted κ = 0.67; 95% CI: 0.43-0.83). The AUCs of the mismatch score and rT2 did not differ significantly (p = 0.861). Quantitative mismatch was higher in IDH-mutant tumors but did not reach statistical significance (p = 0.070). In exploratory models adjusted for age and lesion size, rT2 and mismatch score remained significantly associated with IDH-mutant status. CONCLUSION: Quantitative and semi-quantitative T2-FLAIR-based parameters showed promising performance for differentiating IDH-mutant from IDH-wildtype gliomas in this cohort. However, because 1p/19q co-deletion status was unavailable, these findings require validation in cohorts with complete molecular classification.