Comparison of percentage of total prostate needle biopsy tissue with cancer to percentage of cores with cancer for predicting PSA recurrence after radical prostatectomy: results from the SEARCH database.

OBJECTIVES: Tumor volume in the prostate needle biopsy is an important prognosticator for patients with prostate cancer. However, the best method to measure tumor volume in the prostate needle biopsy is unknown. We compared the total percentage of biopsy tissue with cancer to the percentage of cores positive for their ability to predict adverse pathologic findings and biochemical failure after radical prostatectomy (RP).
METHODS: A retrospective survey of 355 patients from the Shared Equal Access Regional Cancer Hospital database treated with RP between 1990 and 2002 was undertaken. Multivariate analysis was used to compare the percentage of cores and percentage of tissue with cancer to the standard clinical variables of age, prostate-specific antigen (PSA) level, biopsy Gleason score, and clinical stage for their ability to predict positive surgical margins, non-organ-confined disease, seminal vesicle invasion, and time to PSA recurrence after RP.
RESULTS: On multivariate analysis, the percentage of tissue with cancer significantly predicted non-organ-confined disease and seminal vesicle invasion, but the percentage of cores did not significantly predict any of the pathologic features examined. In separate multivariate analysis, only the percentage of tissue with cancer, but not the percentage of cores with cancer, significantly predicted PSA failure. Moreover, when compared in the same multivariate analysis, only the percentage of tissue with cancer (hazard ratio 8.25, 95% confidence interval 3.06 to 22.22, P <0.001) was a significant predictor. The area under the receiver operating curves for predicting PSA failure was significantly greater for the percentage of tissue with cancer (0.697) than for the percentage of cores (0.644, P = 0.022). Cutpoints for the percentage of tissue with cancer (less than 20%, 20% to 40%, and greater than 40%) and the percentage of cores (less than 34%, 34% to 50%, greater than 50%) both provided significant preoperative risk stratification for biochemical failure, although the percentage of tissue with cancer cutpoints provided better risk stratification (higher hazard ratios and lower P value). Cutpoints for the percentage of tissue with cancer but not the percentage of cores positive further stratified patients who were at low (P = 0.041), intermediate (P = 0.002), and high (P = 0.023) risk on the basis of the PSA level and biopsy Gleason score.
CONCLUSIONS: The percentage of tissue with cancer was better than the percentage of cores at predicting advanced pathologic features and PSA recurrence after RP. Unlike the percentage of cores, the percentage of tissue with cancer cutpoints further stratified low, intermediate, and high-risk patients on the basis of PSA level and biopsy Gleason score. Although the percentage of tissue with cancer is a slightly more cumbersome measurement than the percentage of positive cores, it provided statistically and clinically superior preoperative risk stratification for biochemical failure after RP.