Quantitation of hypoechoic lesions for the prediction and Gleason grading of prostate cancer: a prospective study

Purpose Hypoechoic lesions are not included as indicators for prostate biopsy. To discriminate the features of hypoechoic lesions, we investigated the ultrasonographic characteristics of hypoechoic lesions using numerical analysis in image. In addition, we evaluated previously suggested subjective parameters on hypoechoic lesion. Methods We performed one-core targeted biopsy (TBx) for each hypoechoic lesion in up to two lesions in each patient before the 12-core systemic biopsy was obtained between July 2015 and May 2016. Image analysis data were analyzed using grayscale values and Hounsfield units (HU) to measure heterogeneity. Subjective evaluation of hypoechoic lesions including hypoechoicity, irregularity, vascularity, and microcalcification was also validated. Results Of 157 patients (median age = 67.1 years, median prostate-specific antigen = 6.21 ng/mL) included in the study, 77 (49.0%) were diagnosed with prostate cancer (PCa), and 39 (17.0%) diagnoses were confirmed by the results of targeted cores. The existence of hypoechoic lesions was not a final predictor for PCa detection. In multivariate analysis using a combination of clinical and quantitative image analyses, the grayscale value was identified as a significant predictive factor for the presence of PCa and high-grade disease (Gleason score >= 7) on target lesions. The combination of clinical and image variables had the highest area under the curve (0.890) for detecting PCa in TBx. Conclusions The proposed method for the quantitation of hypoechoic lesions using grayscale images and HU is simple. Combined with the current clinical approaches, quantitative scoring of lesions can be useful for detecting PCa and making more precise diagnoses.