Postradiotherapy PSA nadirs fail to support dose escalation study in patients with pretreatment PSA values < 10 ng/ml.

Abstract

With three-dimensional conformal therapy, doses > 75 Gy have been delivered to the prostate with acceptable levels of morbidity; however, higher doses do appear to increase late gastrointestinal (GI) and genitourinary (GU) morbidity. Because patients with pretreatment prostate-specific antigen (PSA) values < 10 ng/ml can achieve 3-year actuarial bNED control rates of 90% after treatment with external beam radiotherapy to doses < 71 Gy, one might question the need for further dose escalation in this population. In this report, we examined the relationship between dose and PSA nadir for 90 patients with pretreatment PSA values < 10 ng/ml entered into a dose escalation study from March 1987 to October 1992. We wanted to see if nadir response data would predict a different outcome from our 3-year bNED control reports. All patients were treated with external beam radiotherapy to ICRU reporting point doses of 6,598 cGy to 7,895 cGy (median of 7,068 cGy). Minimum follow-up was 36 months (median, 47 months). Seven hundred thirty-nine posttreatment PSA nadir values were analyzed, yielding an average of 8.2 values per patient. Estimates of rates of bNED control and time to reach a posttreatment PSA of 1.0 ng/ml were calculated using the Kaplan-Meier product limit method. The log-rank test was used to evaluate differences in rates according to dose levels. Linear regression and Cox proportional hazard modeling were used to relate dose to bNED control on a continuum. Escalating doses from 66 to 79 Gy failed to increase the percentage of patients achieving nadir values < 1 ng/ml and similarly failed to increase the 3-year actuarial bNED control. Linear regression (P = .81) and the chi-square test of association (P = .23) supported the lack of a dose effect on nadir continuously and categorically, respectively, and the Cox regression model supported the conclusion that dose on a continuum has no effect on bNED control (P = .34). Furthermore, time to reach a posttreatment PSA level of 1.0 ng/ml was not statistically dependent on dose level (P = .13). Based on this study and prior reports demonstrating a dose response for late GI/GU morbidity, we question whether further dose escalation in this subgroup of patients is justified.