Limitation of prostate-specific antigen doubling time as a predictor of outcome in hormone-naive prostate cancer
4566 Background: Prostate-specific antigen doubling time (PSADT) is used to monitor progression after primary therapy for prostate cancer. Increasingly in this population, PSADT is an efficacy measure in therapeutic trials, many of which are uncontrolled. Examination of PSADT response on placebo in a randomized controlled study may evaluate the utility of PSADT as an outcome measure in patients with hormone-naive prostate cancer. Methods: Patients with biochemical failure post radical prostatectomy were enrolled in a randomized, double-blind, placebo-controlled trial of atrasentan 10 mg. Eligibility criteria included baseline PSA between 0.4 ng/mL and 5 ng/mL with a PSADT ≤12 months based on two PSA values at least 2 weeks apart. On-treatment PSADT was compared between treatment groups in patients with at least 3 on-treatment PSA values collected every 12 weeks while on study drug. Categorical analysis comparing PSADT between groups was performed using Cochran-Mantel-Haenszel methodology. Mean change from baseline in PSA velocity (log PSA vs time) was analyzed using a one-way analysis of variance with treatment as the factor. Results: Of 222 patients enrolled, 209 had 3 PSA values on study: N=107 for placebo; N = 102 for atrasentan. Baseline PSA was lower for patients in the placebo arm than for those in the atrasentan arm (mean of 1.4 ng/mL [SE 0.1] vs 1.8 ng/mL [SE 0.1] P =.012). On-treatment PSADT was similar between treatment arms (P =.363); the proportion of patients whose PSADT was >1 year on study was 55% with placebo and 51% with atrasentan. More notably, PSADT lengthened on study for most patients in both arms (76/107, 71% for placebo and 68/102, 64% for atrasentan). PSA velocity was also protracted for patients in both treatment arms (placebo: slope = 1.47 at baseline to 0.96, P < .001; atrasentan: slope = 1.56 at baseline to 1.26, P =.017). Conclusions: The fact that over 70% of patients receiving placebo experienced lengthened PSADT suggests that, in the absence of a control arm, changes in PSADT from baseline are not a reliable measure of treatment effect in trials in early prostate cancer. PSADT data from trials in this population should be interpreted with caution. [Table: see text]