PROGNOSTIC FACTORS FOR CLINICALLY LOCALIZED PROSTATE CARCINOMA - ANALYSIS OF 938 PATIENTS IRRADIATED IN THE PROSTATE-SPECIFIC ANTIGEN ERA

BACKGROUND. Pretreatment serum prostate specific antigen (PSA) level i s a powerful prognostic factor for clinically localized prostate carci noma. The traditional prognostic factors, T classification and Gleason score, appear to have been relegated to a minor position. This study was conducted to evaluate the relative prognostic roles of PSA, T clas sification, and Gleason score in a large cohort of men irradiated in t he PSA era. METHODS. The authors analyzed the outcome for a group of 9 38 men with T1-T4, NO or NX, MO prostate carcinoma who received defini tive external beam radiation therapy as their only initial treatment d uring tile period 1987-1995. The T classifications were as follows: T1 , 283 (30%); T2, 360 (38%); T3/T4, 295 (31%). Gleason scores were as f ollows: Gleason 2-6, 580 (62%); Gleason 7, 224 (24%); Gleason 8-10, 12 2 (13%). Pretreatment PSA levels (ng/mL) were as follows: PSA less tha n or equal to 4, 167 (18%); PSA 4 to less than or equal to 10, 363 (39 %); PSA 10 to less than or equal to 20, 259 (28%); PSA > 20, 149 (16%) . At a mean follow-up of 43 months (range, 6-106 months), disease outc ome specified as relapse/rising PSA, local recurrence, or metastasis w as analyzed using univariate and multivariate techniques. RESULTS. The 6-year actuarial incidences of relapse/rising PSA, local recurrence, and metastases were 48%, 27%, and, 6%, respectively, In multivariate r egression, pretreatment PSA level, T classification, and Gleason score were each independently highly significantly (P < 0.001) correlated w ith every endpoint. Pretreatment PSA level was the most significant va riable for rising PSA and local recurrence, and T classification was t he most significant variable for metastatic relapse. Using relapse/ris ing PSA as the endpoint, the authors formulated a highly significant 6 -tier prognostic grouping based on PSA, T classification, and Gleason score, as follows: Category I: T1/T2, PSA less than or equal to 4, and Gleason 2-6 (relapse rate, 6%); Category II: T1/T2, PSA less than or equal to 4 and Gleason 7-10, or PSA 4 to less than or equal to 10 and Gleason 2-7 (relapse rate, 30%); Category III: T1/T2, PSA 4 to less th an or equal to 10 and Gleason 8-10, or PSA 10 to less than or equal to 20 and Gleason < 8 (relapse rate, 40%); Category IV: T3/T4, PSA < 10 (relapse rate, 46%); Category V: T3/T4, PSA 10 to less than or equal t o 20, and Gleason < 8 (relapse rate, 57%); Category unfavorable: any T , PSA > 20 and any Gleason, or PSA 10 to less than or equal to 20 and Gleason 8-10 (relapse rate, 88%). CONCLUSIONS. The establishment of T classification and Gleason score as independent prognostic factors bri dges an apparent gap between an older era and tile current PSA era. PS A has supplemented, rather than supplanted, the utility of the traditi onally established prognostic factors for clinically localized prostat e carcinoma. (C) 1997 American Cancer Society.