Predictive accuracy of transrectal ultrasound-guided prostate biopsy: Correlations to matched prostatectomy specimens

Objectives. To characterize observed differences in Gleason score between prostate biopsy and corresponding radical retropubic prostatectomy (RRP) specimens. Methods. One hundred consecutive clinically localized prostate cancers diagnosed by transrectal ultrasound-guided biopsy (TRUS-Bx) and treated with RRP were reviewed. All specimens were evaluated in blinded review by a single expert uropathologist and contrasted with the initial histologic analysis, performed by multiple pathologists. Results. Mean Gleason score of TRUS-Bx specimens for blinded review and at initial evaluation were 6.6 +/- 0.1 and 6.0 +/- 0.1 (P < 0.001). Corresponding RRP values were 6.8 +/- 0.1 and 6.5 +/- 0.1 (P < 0.03). Differences in Gleason score between TRUS-Bx and RRP at initial evaluation were significant (P < 0.02), but not in blinded review (P = NS). In blinded review, TRUS-Bx correctly predicted RRP histology for 88% of men with lesions scored as Gleason 5 to 7 and 41% of men with well- (Gleason score of 2 to 4) or poorly differentiated (Gleason score of 8 to 10) lesions (P < 0.01). Conclusions. TRUS-Bx does not accurately reflect RRP histology when predicting well- or poorly differentiated lesions. Prostate cancer treatment algorithms should not be predicated upon biopsy histology alone. Histologic interpretation is more accurate and precise when performed by a single experienced uropathologist. (C) 1997, Elsevier Science Inc.