(025) Phenotype Development and Characterization of Patients with Prostate Cancer in Varying Disease Phases Using Real-World Data

Background: The treatment landscape for prostate cancer (PC) is evolving with introduction of next-generation anti-androgens. Contemporary evidence on the clinical characteristics and treatment patterns of patients with PC in the real-world setting are limited, partly due to real-world database limitations to identify patients in different disease phase.

Objective: To develop phenotypes for patients with PC in varying disease phases using US commercial claims databases and to assess clinical characteristics and use of anti-androgens in these patients.

Methods: We developed algorithms to identify patients in four disease phases of PC: 1) non-metastatic castration-sensitive PC (nmCSPC), 2) non-metastatic castration-resistant PC (nmCRPC), 3) metastatic castration-sensitive PC (mCSPC), and 4) metastatic castration-resistant PC (mCRPC). Metastasis was defined based on condition or measurement codes, castration-resistance was defined using proxies including rising prostate specific antigen (PSA) following androgen deprivation therapy (ADT, i.e. surgical or medical castration), clinical observation, or use of medication indicated for CRPC. These phenotypes were assessed in the Optum's de-identified Clinformatics® Data Mart Database (Clinformatics®),
Merative MarketScan® Commercial Claims and Encounters Database, Multi-State Medicaid, and Medicare Supplemental Databases. Using Clinformatics®, we also examined patient characteristics within 365 days prior to index (i.e. the first date when the cohort entry criteria were met) and treatment utilization following index during the time from January 1, 2018 to the last database observation.

Results: Cohorts of patients in different PC disease phase were identified (nmCSPC [n=165,124], nmCRPC [n=17,243], mCSPC [n=10,952], mCRPC [n=9,901]), with median age ranging 72-75 years. Across all disease phases, commonly observed comorbidities included essential hypertension (72-78%), hyperlipidemia (51-54%), type 2 diabetes (31-37%), anemia (13-27%), and urinary tract infection (12-19%). Commonly observed treatment classes included localized treatments (radical prostatectomy or radiotherapy) (28-40%), ADT (21-67%), and anti-androgens (e.g., abiraterone, bicalutamide, apalutamide, enzalutamide, darolutamide) (10-44%). Anti-androgens were more commonly used for treatment of metastatic PC (36-44%), with Abiraterone (14-21%) and Bicalutamide (14-22%) most often observed, as compared with non-metastatic PC (10-21%).

Conclusion: Algorithm-based phenotypes in claims databases can identify patients with PC by clinically relevant phases. Development of these phenotypes will improve future research on PC using real-world data.