Pre-treatment PSA doubling time (PSADT) predicts biochemical response to chemotherapy in patients with androgen sensitive PSA progression after local therapy

Abstract

14528 Background: As phase III studies are developed in patients with androgen sensitive (AS) PSA progression after local therapy, efforts to stratify patients based on likelihood of benefit is critical. In an effort to define a pretreatment PSADT cutoff that predicts chemotherapy response in AS patients for use in hypothesis testing in phase III studies, we assessed patients with pretreatment PSA data in two of our prior published phase II studies of docetaxel (D) or mitoxantrone (M) chemotherapy in patients with AS PSA progression without metastasis. Methods: A retrospective analysis of PSADT was performed on 36 pts with AS PSA progression following local therapy without metastasis with available pretreatment PSA values (23 pts with D and 13 pts with M). PSADT was calculated as [LN(2)]/b, where b is the slope of the regression of LN(PSA) vs time. A t-test was used to compare the PSADT of responders vs non-responders; ANOVA was used for the combined data. A logistic regression analysis with ROC curves was used to evaluate the value of PSADT in discriminating response. Results: Mean PSADT of patients with ≥50% decrease in PSA with therapy was 204 vs. 99 days (P = 0.02). For D only, mean PSADT of pts with any PSA decrease (mean 52%) vs. increase was 211 vs. 115 days (P = 0.04). The mean PSADT of pts with ≥50% PSA decrease only on D was 203 vs. 174 days (P = 0.31). Overall, logistic regression analysis showed that PSADT is a good test to discriminate which pts will have some decrease in PSA while on treatment (P = 0.03). ROC curves (AUC = 0.807) for pts with a decline in PSA on treatment showed that the best balance of sensitivity and specificity was for a PSA DT cutoff of 70 days (sens = 70.0%, spec = 71.4%). Sensitivity was highest at a DT of 50 days (85%), and specificity at 180 days (93%). Conclusions: This study suggests that pretreatment PSADT may be a useful predictor of response to therapy in patients with AS PSA progression after local therapy, but will require larger confirmatory studies. A PSADT cutoff of 70 days can be considered to stratify patients in future trials in this patient population. No significant financial relationships to disclose.