Substratification of Stage T1c prostate cancer based on the probability of biochemical recurrence

Objectives. To evaluate the influence of preoperative prostate-specific antigen (PSA), biopsy Gleason sum, and prostate biopsy quantitative histologic findings on the probability of biochemical failure in an attempt to identify criteria to substratify Stage T1c prostate cancer more accurately. Methods. We reviewed the records of 1149 patients who underwent prostatectomy for T1c disease between 1988 and 2000. Biochemical recurrence (PSA 0.2 ng/mL or greater) defined the endpoint in this study. Recursive partitioning analysis was used to establish cutpoints for preoperative PSA level, biopsy Gleason sum, number of positive biopsy cores, and maximal percentage of any single biopsy core involved with cancer. These cutoff values were then evaluated using Kaplan-Meier estimations to determine the probability of remaining biochemically recurrence free. Results. Using a PSA cutpoint of 10 ng/mL or a biopsy Gleason sum of 7, two groups of patients were identified (T1 cl and T1 cll). The rate of freedom from PSA recurrence at 3, 5, and 10 years after surgery for T1 cl was 98%, 96%, and 96%, respectively, and for T1 cll was 86%, 83%, and 73%, respectively (P <0.001). For T1 cll patients, the greatest percentage of cancer in a single biopsy core was found to be a predictor of biochemical failure on multivariate analysis and, using a cutoff value of 50%, further stratified the PSA recurrence-free rates for the men in group T1 cll (90% and 85% versus 75% and 56% at 5 and 10 years after surgery, respectively, P = 0.03). Conclusions. The results of this study demonstrate that within Stage T1c there are two populations of patients with significantly different recurrence probabilities: T1 cl (Gleason sum less than 7 and PSA 10 ng/mL or less) and T1 cl I (Gleason sum 7 or greater or PSA greater than 10 ng/mL). Furthermore, using a cutpoint of 50% of cancer in a single core of biopsy tissue, additional risk stratification is afforded to men with higher risk "T1cll" cancer. (C) 2002, Elsevier Science Inc.