The impact of irregularly rising prostate-specific antigen and "impending failure" on the apparent outcome of localized prostate cancer following radiotherapy

Purpose: To examine the impact of irregularly rising prostate-specific anti gen (PSA) and "impending" biochemical failure on the apparent rate of bioch emical relapse following radiotherapy for localized prostate cancer. Methods and Materials: We analyzed the outcome of 572 patients with T1/T2 p rostate cancer treated with radiotherapy alone at the Princess Margaret Hos pital (median follow-up, 4.21 years). Biochemical outcomes were analyzed us ing 2 different definitions of failure: (I) the American Society for Therap eutic Radiology and Oncology (ASTRO) definition, and (2) a modified definit ion that included 2 consecutive rises in PSA, with a minimum rise of 1.5 ng /mL above the nadir, or a nadir value of greater than 4 ng/mL, Patients wer e defined as having "impending failure" when the last 2 PSA measurements ta ken demonstrated 2 consecutive rises. Results: Two-hundred and thirty patients (40%) met the ASTRO definition of failure; 258 patients (48%) failed by the modified definition (p = 0.001), Live-year biochemical relapse-free rate (bNED) rate was 55% using the ASTRO definition, and 49% using the modified definition. This difference in 5-ye ar bNED was greatest for patients with high-risk disease (ASTRO definition 30% vs. modified definition 15%). Twenty-four of the 38 additional cases id entified as biochemical failures by the modified definition had irregularly rising PSA levels; 14 were "impending failures." These additional 38 patie nts had a median PSA elevation 5.4 ng/mL above the nadir, and a high risk o f subsequent clinical failure (4-year clinical failure-free rate of 63%). T he ASTRO definition had a sensitivity of 87% and specificity of 74% for pre dicting clinical relapse. The modified definition had a sensitivity of 95% and a specificity of 70%, Conclusion: A definition of biochemical failure that includes an absolute a llowable rise in PSA above the nadir can identify patients with rising PSA who are at substantial risk of clinical relapse, but who are not defined as biochemical failures by the ASTRO definition. This is particularly true fo r patients with high-risk disease. The use of a uniform definition of bioch emical failure is crucial to ensure that differences in apparent outcome ar e not due to differences in the definition of relapse. Currently, the ASTRO definition should remain the standard. Large cohort studies with long foll ow-up can be utilized to optimize the definition of biochemical failure fol lowing radiotherapy for prostate cancer, (C) 2001 Elsevier Science Inc.