Association of prostate-specific antigen (PSA) trajectories with risk for metastasis and mortality in non-metastatic castration-resistant prostate cancer (nmCRPC).

Abstract

27 Background: Currently there is insufficient guidance for the management of nmCRPC. This study assessed patient risk of developing metastases and death based on their PSA levels over time. Methods: This was a retrospective study using the Optum electronic health record database (1/1/2007 – 4/30/2016) in men ≥18 years. nmCRPC was defined as a PC diagnosis, 2 rising PSA levels ≥1 week apart, testosterone < 50 ng/dL (post-PC diagnosis), and no ICD-9/10 code or therapy indicating metastasis. Patients were required to have ≥1 PSA record per 3-month period for 9 months following nmCRPC diagnosis. Group Based Trajectory Modeling (GBTM) was used to group patients based on similar PSA trends over 9 months. The association of these PSA groups with metastasis/mortality risk was measured using multivariate Cox proportional hazard regression models. An overall trend for metastasis and mortality across the groups was also tested. Results: From a total of 729 patients included, 4 distinct groups were identified: Group 1 (49% of patients), 2 (32%), 3 (14%) and 4 (5%). Group 1 had the lowest PSA (7 ng/mL) at nmCRPC diagnosis and steady PSA during the 9-month follow-up. In contrast, Groups 2, 3 and 4 had higher PSA at nmCRPC diagnosis (17, 61, 513 ng/mL respectively) and rising PSA during follow-up. There was a trend of increasing metastasis and mortality risk (p < 0.001 for both trends) with the higher PSA groups. For metastasis, Hazard Ratios (HRs) and 95% confidence intervals (CIs) were 1.7 (1.3-2.2), 3.5 (2.5-5.0), 1.8 (0.7-4.7) in Groups 2, 3 and 4, respectively, vs. Group 1. For mortality, HRs (95% CIs) were 1.9 (1.4-2.5), 2.6 (1.8-3.7), 4.5 (2.4-8.4) in Groups 2, 3 and 4, respectively, vs. Group 1. Metastasis-free survival (MFS) independently predicted mortality risk. Patients developing metastasis within 1 year had 4.4-fold greater risk for mortality (95% CI = 2.2-8.8) than those who remained MFS at year 3. Conclusions: A large proportion of nmCRPC patients with PSA increases during the follow-up period had significantly increased risk for metastasis and mortality, with MFS predicting mortality risk. Periodic measurement of PSA may better inform management of nmCRPC.