The value of morphological magnetic resonance imaging features combined with diffusion kurtosis imaging for predicting Ki-67 expression levels in high-grade gliomas.

Background: Accurate preoperative prediction of Ki-67 expression levels is especially crucial for developing appropriate individualized treatment plans and evaluating prognoses in patients with high-grade gliomas (HGGs). Although previous studies have shown that magnetic resonance imaging (MRI) can preoperatively predict Ki-67 expression levels in glioma, the optimal parameters and predictive performance remain controversial. In particular, there is insufficient research on the value of morphological MRI (mMRI) features and diffusion kurtosis imaging (DKI) for the preoperative assessment of Ki-67 expression levels in HGG. This study aimed to investigate the value of combining mMRI features with DKI for preoperative prediction of Ki-67 expression levels in HGG. Methods: A total of 52 patients who were diagnosed with HGG by surgical pathology and who underwent conventional MRI and DKI scans were included in the study. The clinical and pathological characteristics, mMRI features, relative mean diffusivity (rMD), relative mean kurtosis (rMK), relative radial kurtosis (rKr), relative axial kurtosis (rKa), and relative fractional anisotropy (rFA) were compared between the Ki-67 high- and low-expression groups in HGG. Receiver operating characteristic (ROC) curves were plotted, and the areas under the curve (AUCs), as well as the sensitivities and specificities, were calculated. A nomogram for the prediction of Ki-67 expression levels in HGG was developed on the basis of the pivotal parameters from the mMRI and DKI. Calibration and decision curve analysis were used to evaluate the nomogram. Results: The differences in tumor grade, subventricular involvement (SVI), boundary, diffusion restriction, enhancement, rMD, rMK, rKr, and rKa between the Ki-67 high- and low-expression groups in HGG were statistically significant (P<0.05). When the mMRI and DKI parameters were employed for individual diagnostics, the rMK exhibited the highest diagnostic efficiency, with an AUC value of 0.777 and a sensitivity and specificity of 80.0% and 67.6%, respectively. The diagnostic performance of the DKI model was superior to that of the mMRI model, with an AUC value of 0.834 and a sensitivity and specificity of 86.7% and 67.6%, respectively. The combination of the mMRI features and DKI yielded the optimal diagnostic performance, with an AUC value of 0.892 and a sensitivity and specificity of 100% and 67.6%, respectively. The C index for the nomogram was 0.874. The calibration and decision curve analysis confirmed that there was good consistency between the probability predicted by the nomogram and the actual probability and good clinical utility. Conclusions: The combination of the mMRI and DKI is useful for noninvasive preoperative prediction of HGG Ki-67 expression levels, and the nomogram may help in clinical decision-making for HGG patients.