Background: We sought to determine whether inter-patient variations in pattern of PSA changes after radiation exist and, if so, are they prognostically significant. Methods: In the Trans-Tasman Radiation Oncology Group (TROG) 96.01 randomized controlled trial, patients with T2b,c,3,4 N0 prostate cancer (PC) were randomised to 0, 3 or 6 months maximal androgen deprivation prior to 66 Gy to the prostate and seminal vesicles (XRT). Patterns of anatomical site of failure were one of the trial endpoints. Serial serum PSA’s were mandated at all follow-up visits. Pattern recognition software was developed to characterize PSA response “signatures” (PRS) after therapy in individual patients. Results: By 2000, 270 eligible patients were randomised to radiation alone. Individual patient PSA values were observed to descend after radiation according to one of two characteristic “signatures”: single exponential (PRS Type 1), non-exponential (PRS Type 2). Compared to PRS Type 1, men with PRS Type 2 (50% of the group) had lower PSA nadir (nPSA) levels (p < .0001), longer doubling times on relapse (p = .006) and significantly lower rates of local (hazard ratio [HR]: 0.47, 95% confidence interval [0.30–0.75], p = .0014) and distant failure (HR: 0.25[0.13–0.46], p < .0001), death due to PC (HR: 0.20[0.10–0.42], p < .0001) and death due to any cause (HR: 0.37 [0.23–0.60], p < .0001). PRS retained its powerful prognostic significance in Cox models that incorporated all key pre-treatment covariates and nPSA. Conclusions: PRS reflect the presence of tumor phenotypes that vary substantially in their clinical behavior and response to XRT. Molecular characterization is now necessary.
Radiotherapy and Oncology Vol. 90, Issue 3, p. 382-388