Conventional assessment of needle biopsy specimens is more useful than digital image analysis of proliferation and DNA ploidy in prediction of positive surgical margins at radical prostatectomy.

OBJECTIVES

The preoperative prediction of the likelihood of positive surgical margins (+SMs) at radical retropubic prostatectomy (RRP) may be useful for counseling and determining the surgical approach. The aim of this study was to assess the additional value of digital image analysis (DIA) of ploidy and proliferation on needle biopsies, in addition to the known preoperative predictors of +SMs at RRP.

METHODS

We identified 454 patients treated by RRP at our institution from 1995 to 1998 for prostate cancer verified by transrectal ultrasound-guided biopsy, with a specimen adequate for DIA. Patients receiving preoperative hormonal therapy were excluded. The clinical features, transrectal ultrasound-guided biopsy findings, and DIA evaluation of MIB-I immunostaining and DNA ploidy were assessed in a multivariate logistic regression model to predict for +SMs at RRP.

RESULTS

The mean +/- SD age at treatment was 64.5 +/- 6.5 years, the percentage of positive cores was 40.4% +/- 24.3%, the median prostate-specific antigen level was 6.3 ng/mL (range 0.6 to 112.0), median biopsy Gleason score was 6 (range 4 to 9), and median percentage of diploid nuclei was 67% (range 0% to 100%). Of the 454 patients, 185 (40.7%) had +SMs; this finding was time dependent (1995 to 1996, 45% and 1997 to 1998, 31%; P = 0.004). Univariately, preoperative prostate-specific antigen, biopsy Gleason score, extent of cancer on biopsy, MIB-1 expression, percentage of diploid or nondiploid nuclei, and year of surgery were predictive for +SMs. On multivariate analysis, the preoperative prostate-specific antigen level, biopsy Gleason score, percentage of positive cores, and year of surgery remained significant.

CONCLUSIONS

The results of our study have shown that the likelihood of +SMs at RRP is best predicted on the basis of conventional prognostic factors. The DIA features of needle biopsies did not provide additional predictive power.