The 1989 patterns of care study for prostate cancer: Five-year outcomes

Purpose: Five-year results from the 1989 patterns of care study (PCS) for p rostate cancer are now ready for analysis. The PCS was initiated to determi ne national averages for treatments and examine outcomes prospectively; the 1989 prostate study is the first to have collected pre- and post-treatment serum PSA data. Methods and Materials: Six hundred patients treated with radiotherapy with curative intent for prostate cancer at 71 separate Institutions in the year 1989 made up the study population. Three hundred ninety-one cases were ful ly analyzable. Pretreatment patient and tumor characteristics were as follo ws: of the 391 analyzable, 255 had pretreatment PSA values obtained, and 24 5 had a Gleason's sum (GS) reported. Three hundred fifty-eight were Caucasi an, 24 African-American, and 3 Hispanic (also 6 unknown). One hundred three patients had PSA <10, 60 had PSA 10-19, and 92 presented with PSA >20. Nin ety-seven patients were from Radiation Therapy Oncology Group (RTOG), Commu nity Cancer Centers (CCC), or teaching institutions; 141 patients were from other hospital-based, nonteaching institutions; and 153 were from freestan ding radiation oncology facilities. Seventy-one patients were T1, 203 T2, a nd 100 T3/4. Twenty-four out of 391 patients also received neoadjuvant horm one therapy. Survival curves were constructed using Kaplan-Meier methods, a nd differences between groups were tested for significance using the log-ra nk test. For cumulative incidence curves, Gray's test was used to investiga te failure distributions between groups. The variables entering Cox model f or multivariate analysis included age, race, T stage, pretreatment PSA, and GS. A patient was considered a PSA failure if the treating radiation oncol ogist reported it as such. Results: With a median follow-up of 5.7 years, the 5-year biochemical no ev idence of disease (bNED) and overall survival were 56% and 79% respectively for Stage T1, 52% and 81% for T2, and 36% and 63% for Stages T3 and T4 com bined. As expected, higher pretreatment PSA, GS, and T stage were all progn ostic of poorer outcome. On univariate analysis, bNED survival was adversel y impacted by T stage (p = 0.009), pretreatment PSA (p = 0.0035), and by th e GS (p = 0.0038). Cause-specific failure was significantly lower for highe r T stage (p = 0.014), GS (p = 0.001), and also pretreatment PSA (p = 0.000 4). Overall survival was significantly lower in patients with higher T stag e (p = 0.017) or GS (p = 0.0191), but not pretreatment PSA (p = 0.284). On multivariate analysis, pretreatment PSA was found to be statistically signi ficant in association with bNED survival, and GS was associated with overal l survival, cause-specific survival, and distant metastasis. Few late compl ications were reported: 13/391 and 13/391 Grade 2-3 gastrointestinal (GI) a nd genitourinary (GU) complications respectively, with two patients having required surgery with or without a permanent colostomy. Conclusion: For a representative cross-section of institutions in the Unite d States, radiotherapy achieved high rates of bNED and CSS in selected grou ps of prostate cancer patients. When studied retrospectively, increased pre treatment PSA was a strong predictor of both biochemical failure and death due to prostate cancer. New strategies for patients with high-stage, high-g rade tumors and/or pretreatment PSA >20 deserve testing. (C) 2001 Elsevier Science Inc.