Relationship between systematic biopsies and histological features of 222 radical prostatectomy specimens: lack of prediction of tumor significance for men with nonpalpable prostate cancer.

PURPOSE Because of the recent increase in nonpalpable prostate cancer (clinical stage T1c) in men, preoperative needle biopsy findings have had an important role for treatment decisions. We examine the correlation among histopathological features of 6 systematic biopsies and radical prostatectomy specimens in which 1 investigator reviewed all histological sections. MATERIALS AND METHODS We studied a total of 450 men with clinical stage T1c prostate cancer from whom needle biopsies were matched with radical prostatectomy specimens, and selected 222 patient biopsies that were obtained from 6 or more separate regions of the prostate. The pretreatment parameters of serum prostate specific antigen (PSA), PSA density, number of positive needle biopsies, distribution of positive cores, linear cancer length, and percent Gleason grade 4/5 on the biopsy were determined and compared with histopathological features of prostate cancer in the radical prostatectomy specimens. All biopsies and radical prostatectomies were evaluated morphologically at the department of urology. RESULTS Of the 222 men the largest cancer was clinically insignificant in 23 (10%), as measured by a cancer volume of less than 0.5 cc. Cancer volume in the prostatectomy specimen was significantly related to all parameters in the biopsy, with the surprising exception of cancer distribution in the positive biopsies. However, all of these correlations with cancer volume were weak, with Pearson's correlation squared (R(2)) multiplied by 100 less than 10%. Unfortunately, tumor grade on the biopsy agreed with the prostatectomy specimen in only 81 of 222 (36%) cases. Grade assessment with needle biopsy underestimated the tumor grade in 102 (46%) cases and overestimated it in 39 (18%). No single parameter in the biopsy was a predictor of tumor significance, as measured by a cancer volume of greater than 0.5 cc. However, the best model to predict a tumor less than 0.5 cc in volume was the combination of a single positive core with cancer length less than 3 mm. that contained no Gleason grade 4/5. The use of PSA or PSA density in combination with needle biopsy findings did not enhance prediction of tumor significance. CONCLUSIONS These results indicate a weak and disappointing correlation among all pathological features of 6 systematic biopsies and radical prostatectomy specimens. The combination of 1 positive core with cancer length less than 3 mm. that contains no Gleason grade 4/5 is probably the best predictor of prostate cancer less than 0.5 cc in men with nonpalpable tumors, a cancer volume that occurred in only 10% of the 222 (23) men.