Peritumoral Edema Outperforms Solid Tumor for Predicting TERTp Mutation in IDH Wild-Type Gliomas: A Multiparametric Diffusion MRI Histogram Study with Prognostic Analysis.

RATIONALE AND

OBJECTIVES

To investigate the preoperative predictive value of histogram features of diffusion-weighted imaging (DWI), diffusion kurtosis imaging (DKI), and intravoxel incoherent motion (IVIM) parameters for isocitrate dehydrogenase (IDH) and telomerase reverse transcriptase promoter (TERTp) mutation status in adult diffuse gliomas, and to identify independent prognostic factors for overall survival (OS).

MATERIALS AND

METHODS

This retrospective study included 404 patients who underwent DWI, DKI, and IVIM imaging. Region-specific logistic regression models were developed using stratified 70:30 training and hold-out test splits. Logistic regression and receiver operating characteristic curves were used to assess predictive performance. Decision-curve, synthetic minority over-sampling technique (SMOTE) sensitivity, and bootstrap stability analyses were also performed. Prognostic factors were evaluated using Cox regression in 343 patients with follow-up.

RESULTS

For IDH prediction, areas under the curve (AUCs) ranged from 0.874 to 0.920 in solid tumor and 0.902 to 0.936 in peritumoral edema, without significant between-region differences. For TERTp prediction in IDH wild-type gliomas, peritumoral edema models achieved higher AUCs than their solid tumor counterparts: DWI, 0.893 versus 0.677 (P = 0.019); DKI, 0.916 versus 0.754 (P = 0.036); IVIM, 0.934 versus 0.819 (P = 0.022); and combined, 0.948 versus 0.819 (P = 0.020). Decision-curve analysis favored the peritumoral edema combined model, and the regional discrimination pattern persisted after SMOTE balancing. In the overall cohort, peritumoral edema Kskewness, D*10th, and fentropy were independently associated with OS, alongside age, IDH status, O6-methylguanine-DNA methyltransferase promoter methylation, and Ki-67 LI. Within the IDH wild-type subgroup, age, TERTp status, and Ki-67 LI were independent prognostic factors.

CONCLUSION

Diffusion histogram models provided high discrimination for IDH status in both regions. Peritumoral edema models offered superior TERTp prediction in IDH wild-type gliomas, while selected peritumoral features provided prognostic information in the overall cohort.