Comparison of PI-RADS 2, ADC histogram-derived parameters, and their combination for the diagnosis of peripheral zone prostate cancer

The purpose of this study was to compare the PI-RADS V2 scores, ADC histogram-derived parameters, and their combination for the diagnosis of clinically significant peripheral zone prostate cancer (PCa). The IRB approved this retrospective study of 47 men who underwent 1.5 Tesla endorectal prostate magnetic resonance imaging (MRI). Informed consent was waived. Two readers identified and scored MRI lesions using PI-RADS V2. Their mean, median, 10th, 25th, 75th percentile ADC values, and normalized ratio were also calculated. Multilevel logistic regression and receiver-operating characteristic (ROC) curve analyses assessed their diagnostic performance. Clinically significant PCa was defined as tumor volume over 0.5 cc and Gleason grade of 4 or 5 on prostatectomy. The area under the ROC curve (A (z)) of the overall and diffusion-weighted imaging (DWI) PI-RADS V2 scores were 0.69 and 0.84 (reader-1), and 0.68 and 0.73 (reader-2). The A (z) of ADC parameters ranged from 0.68 to 0.75 for both readers. Compared to other predictors, DWI PI-RADS V2 yielded the highest A (z) for identification of significant cancer; but, except for reader-1 75th percentile ADC, the differences were not statistically significant (P > 0.05). Adding ADC parameters to PI-RADS V2 scores did not improve their diagnostic ability. DWI PI-RADS V2 score may a better predictor of clinically significant PCa than the overall PI-RADS V2 score, but its diagnostic performance was not significantly improved by the addition of objective ADC value measurements.