Transperineal sonography guided biopsy of the prostate: critical review of 1107 cases.

Purpose: To assess the predictive value for a positive biopsy of different indicators (rectal examination, transrectal ultrasonography, total PSA, PSA density).
Material and Methods: Positive predictive value was assessed on 1107 consecutive US-guided transperineal biopsies performed from 1991 to 2001 (cancer=344, dysplasia (PIN)=64, atypical hyperplasia=4, benign hyperplasia=686, inadequate=9) with univariate and multivariate analysis.
Results: Increasing age (chi square for trend 52.2, p <0.001), positive rectal examination (chi square 233, df=1, p<0.001) or ultrasonography (chi square 191, df=1, p<0.001), total PSA (chi square for trend 68.9, p<0.001) and PSA density (cutoff 0.15, chi square 104, df=1, p<0.001; cutoff 0.20, chi square 104, df=1, p<0,001) were all significantly associated to the likelihood of a positive biopsy outcome. Multivariate analysis stresses the independent role of Psa density over total PSA. If the parameters studied had determined the biopsy, spared benign biopsies [positive rectal examination=505 (66%), positive ultrasonography=467 (61%), PSA>4=124 (16%), PSA>10=159 (74%), PSA density >0,15=426 (62%), PSA density >0.20=517 (75%)] would not have justified the amount of delayed cancer biopsies [positive rectal examination=103 (29%), positive ultrasonography=55 (15%), PSA>4=42 (12%), PSA>10=569 (46%), PSA density >0,15=73 (25%), PSA density >0,20=107 (37%)].
Conclusions: The parameters currently available prior to biopsy, if used alone, allow no reliable prediction of biopsy outcome. Positive predictive value, particularly for PSA density, allows a better evaluation of biopsy indication, particularly for random sextant biopsies in the 4-10 ng/ml PSA range, which are frequently negative.