Background: Second-generation androgen receptor inhibitors (ARIs) are the preferred treatment for patients (pts) with non-metastatic castration-resistant prostate cancer (nmCRPC). In this setting, an important therapeutic goal is to minimize adverse events (AEs) and prolong time on ARI treatment. This is the first study to assess real-world (RW) utilization and outcomes of all currently available ARIs, darolutamide (daro), enzalutamide (enza), and apalutamide (apa) in pts with nmCRPC. Methods: DEAR (NCT05362149) is a retrospective observational chart review using electronic medical records from the Precision Point Specialty network of US urology practices. Eligibility criteria included nmCRPC diagnosis, no previous treatment with a novel hormonal agent, and ARI treatment initiation for the first time from August 2019 to March 2022. Pts were classified into 3 cohorts based on the first prescribed ARI in the nmCRPC stage. We present an interim report including the primary composite endpoint (DIS/MET): proportion of pts who discontinued initial ARI treatment (DIS) or progressed to metastasis (MET); and secondary endpoints: time to DIS/MET and underlying reasons. Results: In total, 666 pts were included (daro/enza/apa, n=276/280/110). The proportion of pts starting treatment with daro increased during the study period relative to enza or apa, which declined or remained stable, respectively. Median age (80/79/80 years), White race (68%/69%/74%) and median PSA doubling time (6.7/6.0/7.8 months) at baseline were similar in the daro/enza/apa cohorts, respectively. Median length of follow-up was similar. Fewer pts had a DIS/MET event during the study in the daro cohort (35.9%), compared with enza or apa (52.1%/50.9%); this trend was seen within 6 to 18 months after ARI start. Median time to DIS/MET was 33.4/20.8/18.5 months for daro/enza/apa, respectively. Most common reasons for DIS/MET in the daro/enza/apa cohorts were AEs (8.3%/15.0%/12.7%), death (3.3%/5.4%/2.7%), and MET (14.9%/22.9%/18.2%). Conclusions: This RW study in the US shows that, despite similar baseline characteristics, fewer nmCRPC pts treated with daro had DIS/MET (especially due to AEs/MET) versus enza and apa, and the time to this event was longer in the daro cohort. These findings may be attributed to daro being a structurally distinct ARI with low potential for blood–brain barrier penetration.

CI, confidence interval; DIS, discontinuation; MET, metastasis; NR, not reached.