Predictability of PSA failure in prostate cancer by computerized cytometric assessment of tumoral cell proliferation

Objectives. To evaluate the relationship of DNA ploidy and cell proliferati on (CP) with Gleason score (GS) and clinical outcome in prostate cancer. Methods. Sixteen patients with benign prostatic hyperplasia (BPH) and 65 pa tients with prostate cancer classified by GS (four groups: 2 to 4, 5 to 6, 7, and 8 to 10) were studied. All patients with carcinoma underwent prostat ectomy and were separated into prostate-specific antigen (PSA) failure and nonfailure groups (failure if PSA 0.1 ng/mL or more three times after surge ry). Tumoral CP (Ki-67 inmunostaining and SG(2)M phase) and DNA ploidy were evaluated by computerized cytometry. Results. BPH were diploid with low CP (8% SG(2)M cells or less). Carcinomas were either diploid with high CP (greater than 8% SG(2)M cells) or aneuplo id. CP was significantly higher (P <0.001) in tumors with CS 7 or greater t han in tumors with GS less than 7 (mean percent Ki-67 cells 18.3% versus 7. 8%, respectively). PSA failure increased with GS (7.1% in GS 2 to 4, 21% in GS 5 to 6, 28.6% in GS 7, and 50% in GS 8 to 10), as well as with aneuploi dy (18.5% in diploid tumors versus 72.7% in aneuploid tumors). Those experi encing PSA failure had significantly higher (P <0.001) CP than those not fa iling (mean percent Ki-67 cells 24% and mean percent SG(2)M 30.4% versus 8. 7% and 13.5%, respectively). Cox regression analysis showed GS, DNA ploidy, Ki-67, and SG(2)M to each be univariately prognostic for time to PSA failu re; however, Ki-67 and SG(2)M were more highly significant (P <0.0001 for b oth) than CS (P = 0.007) or DNA ploidy (P = 0.002). After adjusting for eit her SG(2)M or Ki-67 measures of CP, neither ploidy nor GS contained additio nal prognostic value. Conclusions. Tumor CP and DNA ploidy can be reliably determined in prostate cancer by computerized cytometry. On the basis of our preliminary results, CP correlates well with GS and predicts PSA failure better than DNA ploidy or GS. UROLOGY 53: 931-958, 1999, (C) 1999, Elsevier Science Inc. All righ ts reserved.