Diffusion Kurtosis Imaging Helps to Predict Upgrading in Biopsy-Proven Prostate Cancer With a Gleason Score of 6.

Abstract

The purpose of this study was to investigate whether diffusion kurtosis imaging (DKI) is useful for predicting upgrades in Gleason score (GS) in biopsy-proven prostate cancer with a GS of 6. A total of 46 patients with biopsy-proven GS 6 prostate cancer, 3-T DWI results, and surgical pathologic results were retrospectively included in the study. DWI data were postprocessed with monoexponential and DK models to quantify the apparent diffusion coefficient (ADC), apparent diffusion for gaussian distribution (D_app), and apparent kurtosis coefficient (K_app). The volume of the lesions, prostate-specific antigen (PSA) level, and diffusion variables (ADC_min, D_appmin, K_appmax, ADC_mean, D_appmean, and K_appmean) were evaluated. PSA and DKI were combined as a parameter in a logistic regression model. The utility of these parameters in predicting an upgrade in GS was analyzed with ROC regression. The rate of GS upgrade was 50.0% (23/46). The GS upgrade group had significantly lower ADC_min (p = 0.007), ADC mean (p = 0.003), D appmin (p < 0.001), and D_appmean (p = 0.001) values and significantly higher K_appmax (p = 0.003), K_appmean (p = 0.005), and PSA (p = 0.004) values than the group that did not have an upgrade. Among single parameters, K_appmax had the highest ROC AUC value (0.819, p < 0.05), and among all the parameters and models, PSA-K_appmax had the highest AUC (0.868, p < 0.05) and Youden index (0.6522). The results showed that DKI may help in prediction of GS upgrade in biopsy-proven GS 6 prostate cancer. The comprehensive consideration of DKI and PSA may be a promising approach to predicting GS upgrade.