Assessing the variability of outcome for patients treated with localized prostate irradiation using different definitions of biochemical control

Purpose: Biochemical control using serial posttreatment serum prostate specific antigen (PSA) levels is being increasingly used to assess treatment efficacy for localized prostate cancer. However, no standardized definition of biochemical control has been established. We reviewed our experience treating patients with localized prostate cancer and applied three different commonly used definitions of biochemical control to determine if differences in therapeutic outcome would be observed. Methods and Materials: Between January 1987 and December 1991, 480 patients with clinically localized prostate cancer received external beam irradiation (RT) using localized prostate fields at William Beaumont Hospital. The median dose to the prostate was 66.6 Gy (range 58-70.4) using a four-field or are technique. Pretreatment and posttreatment serum PSA levels were recorded. Over 86% (414 of 480) of patients had a pretreatment PSA level available. Three different definitions of biochemical control were used: (a) PSA nadir < 1 ng/ml within 1 year of treatment completion. After achieving nadir, if two consecutive increases of PSA were noted, the patient was scored a failure at the time of the first increase; (b) PSA nadir < 1.5 ng/ml within 1 year of treatment completion. After achieving nadir, if two consecutive increases of PSA were noted, the patient was scored a failure at the time of the first increase; (c) Posttreatment PSA nadir < 4 ng/ml without a time limit. Once the nadir was achieved, if it did not rise above normal the patient was considered to be biochemically controlled. Clinical local control was defined as no palpable prostate nodularity beyond 18 months, no new prostate nodularity, or a negative prostate biopsy. Results: Median follow-up was 48 months (range 3-112). Pretreatment PSA values were correlated with treatment outcome using the three definitions of biochemical control as well as clinical local control. Pretreatment PSA values were stratified into five groups (Group 1: PSA < 4; Group 2: PSA 4-10; Group 3: PSA 10-15; Group 4: PSA 15-20; and Group 5: PSA > 20), and 5-year actuarial rates of biochemical control were calculated using the three biochemical control and one clinical local control definitions. For Group 1, 5-year actuarial rates of biochemical control were 84%, 90%, 91%, and 96% for Definitions 1-3 and clinical local control, respectively. For Group 2, 5-year actuarial control rates were 45%, 54%, 74%, and 92% for the four definitions, respectively. For Group 3, 5-year actuarial control rates were 26%, 31%, 63%, and 100% for the four definitions, respectively. For Group 4, 5-year actuarial control rates were 24%, 24%, 50%, and 100% for the four definitions, respectively. Finally, for Group 5, 5-year actuarial control rates were 5%, 14%, 15%, and 89% for the four definitions, respectively. Depending on the definition used, statistically significant differences overall in outcome rates were observed. Differences between all four definitions for all pairwise comparisons ranged from 5 to 53% (p < 0.001). Conclusion: When different definitions of biochemical control are used in assessing treatment outcome, significantly different rates of success are noted. Until a standardized definition of biochemical control is adopted, differences in treatment outcome cannot be meaningfully compared. Copyright (C) 1996 Elsevier Science Inc.