Dose response in prostate cancer with 8-12 years' follow-up.

PURPOSE This communication reports the long-term results of the original group of prostate cancer patients who participated in the first prospective Fox Chase Cancer Center radiation dose escalation study for which 8-12 years of follow-up is now available. METHODS AND MATERIALS Between March 1, 1989 and October 31, 1992, 232 patients with clinically localized prostate cancer received three-dimensional conformal radiotherapy only at Fox Chase Cancer Center in a prospective dose-escalation study. Of these patients, 229 were assessable. The 8-, 10-, and 12-year actuarial rates of biochemical control (biochemically no evidence of disease [bNED]), freedom from distant metastasis (FDM), and morbidity were calculated. The Cox proportional hazards model was used to assess multivariately the predictors of bNED control and FDM, including pretreatment prostate-specific antigen (PSA) level (continuous), tumor stage (T1/T2a vs. T2b/T3), Gleason score (2-6 vs. 7-10), and radiation dose (continuous). The median total dose for all patients was 74 Gy (range 67-81). The median follow-up for living patients was 110 months (range 89-147). bNED control was defined using the American Society for Therapeutic Radiology and Oncology consensus definition. RESULTS The actuarial bNED control for all patients included in this series was 55% at 5 years, 48% at 10 years, and 48% at 12 years. Patients with pretreatment PSA levels of 10-20 ng/mL had statistically significant differences (19% vs. 31% vs. 84%, p = 0.0003) in bNED control when stratified by dose (<71.5, 71.5-75.6, and > 75.6 Gy, respectively) on univariate analysis. For the 229 patients with follow-up, 124 (54%) were clinically and biochemically without evidence of disease. Sixty-nine patients were alive at the time of last follow-up, and 55 patients were dead of intercurrent disease. On multivariate analysis, radiation dose was a statistically significant predictor of bNED control for all patients and for unfavorable patients with a pretreatment PSA <10 ng/mL. For the patients with a pretreatment PSA level of 10-20 ng/mL, the radiation dose was a statistically significant predictor across all groups. No radiation dose response was seen for those patients with a pretreatment PSA level >20 ng/mL, although large numbers of patients are required to demonstrate a difference. The radiation dose, Gleason score, and palpation T stage were significant predictors for the entire patient set, as well as for those with pretreatment PSA levels between 10 and 20 ng/mL. The FDM rate for all patients included in this series was 89%, 83%, and 83% at 5, 10, and 12 years, respectively. For patients with pretreatment PSA levels <10 ng/mL, all four covariates (radiation dose, Gleason score, pretreatment PSA, and palpation T stage) were significant predictors of distance metastasis. Using the Radiation Therapy Oncology Group morbidity scale, no difference was noted in the frequency of Grade 2 and 3 genitourinary and Grade 3 gastrointestinal morbidity when patients in this data set were stratified by radiation dose. However, a significant increase occurred in Grade 2 gastrointestinal complications as the radiation dose increased. CONCLUSION The long-term results of the original Fox Chase radiation dose escalation study with >9 years of median follow-up confirm the existence of a dose response for both bNED control and FDM. The dose response in prostate cancer is real, and the absence of biochemical recurrence after 8 years demonstrates the lack of late failure and suggests cure.