Purpose: We studied prostate-specific antigen (PSA) changes after radiation with or without neoadjuvant androgen deprivation to determine posttreatment PSA scenarios in which false-positive biochemical failures (FPBF) are most likely to occur. Methods and Materials: In the Trans-Tasman Radiation Oncology 96.01 Group trial, patients with T2b, 2c, 3, 4 N0 prostate cancer were randomized to 3 or 6 months goserelin and flutamide (STAD) before and during 66 Gy to the prostate and seminal vesicles (XRT) or to XRT alone. Piecewise longitudinal changes in PSA before relapse were characterized and quantified to determine which might cause FPBF calls. Results: Between 1996 and 2000, 802 eligible patients were randomized. Of these, 492 met the criteria for American Society for Therapeutic Radiology and Oncology (ASTRO) failure and 467 for Phoenix failure. Seventy-seven ASTRO fails and 39 Phoenix fails were deemed false positives (FPs). The majority of FPBFs were associated with the “plateauing” in PSA values that follow posttreatment nadir. FPBFs were particularly common in men treated with STAD, in whom small, consecutive PSA rises before or during this phenomenon triggered 56 FP ASTRO fail calls. In these men, the Phoenix fail criteria triggered only 15 FPBF calls. However, the Phoenix criteria were more vulnerable than ASTRO to short-term isolated PSA rises during plateau, which resulted in 15 Phoenix fail calls but only 3 FP ASTRO fails. Conclusions: The Phoenix definition avoided 50% of FPBF calls that occurred with the ASTRO definition. Failures should be confirmed by further PSA rises before investigation and treatment is considered.
International Journal of Radiation Oncology Biology Physics Vol. 74, Issue 2, p. 404-411