Multiparametric MRI with DWI, SWI, and DSC-PWI for preoperative prediction of Ki-67, IDH, and ATRX in adult-type diffuse gliomas

Imaging-based assessment of proliferative and molecular features in adult-type diffuse gliomas may support preoperative tumor characterization. This study evaluated conventional MRI combined with diffusion-weighted imaging (DWI), susceptibility-weighted imaging (SWI), and dynamic susceptibility contrast perfusion-weighted imaging (DSC-PWI) for their associations with Ki-67 index, IDH status, and ATRX expression status. This retrospective study included 126 patients with adult-type diffuse gliomas who underwent conventional MRI, DWI, and SWI; rCBV from DSC-PWI was available in 59 patients. Quantitative ADC and rCBV measurements were independently performed by two readers. Histopathological and molecular data were obtained from pathology reports. Ki-67 expression was dichotomized as low (< 10%) and high (≥ 10%). ROC analysis assessed the diagnostic performance of MRI parameters for high Ki-67 expression and WHO grade ≥ 3 tumors; sensitivity analyses were performed using alternative Ki-67 thresholds of ≥ 15%, ≥ 20%, ≥ 25%, and ≥ 30%. Potential selection bias related to rCBV availability was assessed, and multivariable logistic regression was performed. Among 126 patients, 84.1% had WHO grade 4 tumors. ADCmin, rADCmin, ADCmean, and rADCmean were significantly lower, whereas rCBVmax and ITSS grade were significantly higher, in tumors with high Ki-67 expression (all p < 0.001). High Ki-67 expression was also associated with necrosis, peritumoral edema, and heterogeneous enhancement. ADC-derived parameters differed significantly between IDH-mutant and IDH-wildtype tumors. rCBVmax was higher in tumors with retained ATRX expression ( p = 0.016). rCBVmax showed the highest diagnostic performance for high Ki-67 expression (AUC = 0.967; 95% CI: 0.920–1.000). In multivariable analysis, lower ADCmin and rADCmin remained independently associated with high Ki-67 expression, whereas ITSS grade did not. Interobserver agreement was excellent for quantitative ADC and rCBV measurements. Multiparametric MRI features, particularly diffusion-derived metrics and rCBVmax, were associated with proliferative activity in adult-type diffuse gliomas and may support preoperative noninvasive tumor characterization, but validation in larger and more balanced cohorts is required.