Castration resistance and high-risk disease among nonmetastatic (M0) prostate cancer (PC) patients on androgen deprivation therapy (ADT).

1581 Background: Prostate-specific antigen (PSA) is a well-known PC biomarker. In the M0 setting, rising PSAs despite ADT is an indication of the development of castration-resistant prostate cancer (CRPC). In this disease state, men are at risk for developing bone metastasis (BM), which is associated with significant morbidity and may negatively affect survival. Using real-world data, we explored PSA-based criteria to identify patients who develop CRPC while on ADT and the subsets that may be at increased risk of BM. Methods: We used the Oncology Services Comprehensive Electronic Records (OSCER) database, which includes electronic medical record (EMR) data on cancer patients from 328 urology and oncology clinics in the US. Eligible patients were adult men with M0 PC with ≥1 PSA recorded between 3/1/2010 and 2/28/2011 and currently receiving ADT (gonadotropin-releasing hormone agonists or bilateral orchiectomy) for ≥6 months (mos). We defined CRPC as two sequential PSA rises while on ADT and high risk for BM as any PSA ≥8 ng/mL or PSA doubling time (DT) ≤10 mos, as described by Smith MR et al, Lancet 2012. We explored subsets of CRPC patients who may be at even higher risk of BM using PSA thresholds (≥8 ng/mL and ≥20 ng/mL) and DT (≤4, 6, 8, and 10 mos). Results: Of 1,818 men with M0 PC receiving ADT ≥6 mos, 36% (N=646) met the CRPC definition, of whom 80% (N=517) had PSA ≥8 ng/mL and/or PSA DT ≤10 mos (high risk). PSA DT alone explained 63% (44% / 70%) to 93% (65% / 70%) of subgroup eligibility (Table), and emerged as a main driver in defining increased risk of BM for CRPC subsets. Conclusions: In this analysis of EMR data, over one-third of men with M0 PC on ADT met criteria for CRPC, and most CRPC patients (80%) may be considered at high risk for BM. Requiring ≥3 PSAs to define CRPC may be a limitation; however, because PSAs are closely monitored in patients on ADT, these definitions of CRPC and high risk may be useful in practice. These data suggest that PSA DT may be a more clinically meaningful measure of defining CRPC subsets than absolute PSA thresholds. [Table: see text] .