Positive resection margins may not reflect the true margin in patients undergoing radical prostatectomy

The aim of the present study was to evaluate the hypothesis that a positive resection margin (RM1) of an excised specimen may not reflect the true margin in patients that have undergone radical prostatectomy (RP). Between September 2003 and March 2011, 370 Japanese patients underwent an antegrade RP at the National Kyushu Cancer Center (Fukuoka, Japan), however, 95 of these patients were excluded from the study due to a history of receiving hormonal therapy or insufficient preoperative clinical data. The incidence of biochemical failure (BCF) was evaluated using multivariate analysis, which revealed that the preoperative prostate-specific antigen (PSA) level, pathological tumor stage, RP Gleason score and a PSA nadir <0.008 ng/ml were significant predictors (P=0.0065, 0.0006, 0.0002 and <0.0001, respectively). By contrast, an RM1 was not found to be a significant predictor of BCF, while the parameter with the highest hazard ratio (HR) was a PSA nadir <0.008 ng/ml (HR, 10.055; 95% confidence interval, 5.005-20.200). From the 56 cases that were RM1, 41 cases (73.2%) exhibited a PSA nadir <0.008 ng/ml. There were 42 cases (75.0%) in which only one site was identified to be RM1; among these cases, no significant difference was observed between a PSA level <0.008 ng/ml and a PSA level ≥0.008 ng/ml at the RM1 site (apex, P=0.1460; base, P=0.1384; anterior, P=0.3870; and posterolateral, P=0.5040). There were 14 cases (25.0%) in which multiple sites were RM1; these cases were classified by the number of sites that were RM1 (one vs. multiple) and no significant difference was observed between a PSA level <0.008 ng/ml and a PSA level ≥0.008 ng/ml (P=0.6090). Based on these results, an RM1 of an excised specimen may not reflect the true margin in patients that are treated with RP, specifically in cases where the PSA level is identified to decrease to below the postoperative measurement threshold value (PSA nadir <0.008 ng/ml).