Can the number of cores with high-grade prostate intraepithelial neoplasia predict cancer in men who undergo repeat biopsy?

OBJECTIVES: To evaluate whether the presence of, or the number of cores containing, high-grade prostatic intraepithelial neoplasia (PIN) found in men who underwent initial extended multisite biopsy could predict which men would have prostate cancer on subsequent repeat biopsies. METHODS: Between June 1997 and January 2003, 1086 men underwent initial prostate biopsy for early detection of prostate cancer using an extended multisite biopsy scheme. Of these, 175 men without cancer underwent at least one repeat biopsy (range one to three; median interval between biopsies, 3 months). Among these 175 patients, 47 had high-grade PIN on initial biopsy. RESULTS: The initial extended biopsy identified cancer in 33.8% (367 of 1086) and high-grade PIN in 20.8% (226 of 1086). The incidence of high-grade PIN only in patients found to have cancer on initial biopsy was 29.7% (109 of 367). The presence of high-grade PIN was associated with concurrent prostate cancer at the initial biopsy (P <0.0001). Overall, repeat biopsy identified cancer in 18.3% of the 175 men. Of the 47 men with high-grade PIN, 5 (10.6%) were found to have cancer on repeat biopsy. The number of biopsy specimens positive for high-grade PIN on initial biopsy was not associated with the likelihood of prostate cancer on repeat biopsy. Multivariate logistic regression analysis showed that neither the presence of high-grade PIN nor the number of cores containing high-grade PIN on initial biopsy were predictors for prostate cancer on repeat biopsy. CONCLUSIONS: The number of cores positive for high-grade PIN was not predictive for cancer on repeat biopsy.