Three-dimensional grayscale ultrasound: Evaluation of prostate cancer compared with benign prostatic hyperplasia

Objectives, To compare the accuracy of the detection, localization, and staging of prostate cancer using transrectal three-dimensional (3D) grayscale ultrasonography (3D-US) with conventional transrectal two-dimensional grayscale ultrasonography (2D-US). Methods. Fifty patients with clinical localized prostate cancer scheduled to undergo radical retropubic prostatectomy and 50 patients with clinical benign prostatic hyperplasia underwent transrectal ultrasound investigations (2D and 3D). The prostate images were retrospectively analyzed by two ultrasound experts unaware of the clinical findings. The images of the prostate cancer group were correlated with the wholemount histologic specimens of the prostate. Results, All percentages are given for experts 1 and 2. The sensitivity, specificity, and accuracy for the detection of prostate cancer without considering the definitive localization of the tumor for 2D-US was 72% and 76%, 50% and 54%, and 63% and 64%, respectively; for 3D-US, the rates were 82% and 88%, 40% and 42%, and 61% and 65%. The sensitivity, specificity, and accuracy of the combination of 2D-US with 3D-US was 88% and 90%, 36% and 38%, and 62% and 64%, respectively. The sensitivity, specificity, and accuracy for the exact localization of the prostate tumor for 2D-US was 44% and 46%, 50% and 54%, and 47% and 50%, respectively; for 3D-US, they were 52% and 62%, 40% and 42%, and 46% and 52%. The staging of prostate cancer using SD-US was correct in 49% (expert 1) and in 57% (expert 2) of patients. No difference was observed between 2D-US and 3D-US for accurate staging. Both experts judged the interpretation of 3D-US images as superior to that of 2D-US images. Conclusions. Although 3D-US had statistically significant increased sensitivity in the detection of lesions and decreased specificity compared with 2D-US, 3D-US did not result in significant clinical improvement in the detection and staging of prostate cancer. UROLOGY 57: 914-920, 2001. (C) 2001, Elsevier Science Inc.