Detecting the muscle invasiveness of bladder cancer: An application of diffusion kurtosis imaging and tumor contact length

Purpose: To evaluate the diagnostic efficacy of diffusion kurtosis imaging (DKI) parameters and tumor contact length (TCL) among clinical and radiological factors for preoperative prediction of muscle-invasive bladder cancer (MIBC). Method: A total of ninety-seven patients underwent 3.0 T MRI scan with propeller fast spin-echo T2WI, echo planar imaging diffusion-weighted imaging (DWI), and dynamic contrast-enhanced imaging (DCE). Two radiologists independently viewed multiparametric MRI (mpMRI) of each patient, graded the VI-RADS, drew the region of interest (ROI) and measured TCL. Interclass correlation coefficients (ICCs), Kappa statistics, Kolmogorov-Smirnov test, Mann-Whitney U tests, chi-square tests, logistic regression analyses, HosmerLemeshow tests, receiver operating characteristic curve (ROC) analysis, and area under the curve (AUC) were applied. Results: The mean K-app of NMIBC group (0.62 +/- 0.01) was significantly lower than that of MIBC group (0.79 +/- 0.08). The mean TCL of MIBC group (4.66 +/- 1.89) was significantly larger than TCL of NMIBC group (1.88 & PLUSMN; 1.50) (all p < 0.01). At the corresponding cut-off, AUC of TCL, K-app, VI-RADS and the combination of K-app and TCL were 0.87, 0.92, 0.90, and 0.95, respectively. TCL and K-app were risk factors of BC muscle invasion at both univariate and multivariate analysis. Conclusions: K-app performed better than conventional DWI in predicting MIBC. K-app and TCL were independent risk factors of MIBC and could complement VI-RADS for predicting muscle invasion. The combination of K-app and TCL had the largest AUC and highest accuracy among all parameters.