CALCULATED PROSTATE CARCINOMA VOLUME - THE OPTIMAL PREDICTOR OF 3-YEAR PROSTATE-SPECIFIC ANTIGEN (PSA) FAILURE FREE SURVIVAL AFTER SURGERY OR RADIATION-THERAPY OF PATIENTS WITH PRETREATMENT PSA LEVELS OF 4-20 NANOGRAMS PER MILLILITER

BACKGROUND. In this study, the authors evaluated whether a clinically relevant stratification of prostate specific antigen (PSA) failure fre e survival (bNED) after definitive local therapy could be made for pat ients with prostate carcinoma clinically classified as T1 or T2 and pr etreatment PSA levels of 4-20 ng/mL. METHODS. Multivariate Cox regress ion analysis and Kaplan-Meier analysis were performed for clinically l ocalized prostate carcinoma patients who presented with PSA levels of 4-20 ng/mL. Three hundred forty-eight of the pa tients were managed de finitively with conventional external beam radiation therapy (median d ose, 67 gray), whereas 547 of the patients were managed definitively w ith a radical retropubic prostatectomy. The outcome tested was time to posttreatment PSA failure. The clinical predictors evaluated included the standard paradigm (PSA, biopsy Gleason score, and clinical stage) ; type of local therapy; and a newly defined factor, the calculated pr ostate cancer volume (cV(Ca)). RESULTS. Time to posttreatment PSA fail ure was equivalent (P = 0.52) independent of the type of local therapy . The cV(Ca) (P < 0.0001), pretreatment PSA (P = 0.003), and clinical classification of T2c (P = 0.04) remained significant predictors of ti me to posttreatment PSA failure in multivariate analysis. CONCLUSIONS. The staging system described herein, which is based on cV(Ca) and PSA , may optimize patient selection for definitive local therapy and entr y onto randomized clinical trials examining the use of adjuvant hormon al or chemotherapy in patients with clinically localized disease who p resent with PSA levels of 4-20 ng/mL. Validation of this staging syste m by other investigators is currently underway. (C) 1998 American Canc er Society.