Real-world prevalence and outcomes of androgen receptor ligand-binding domain mutations in prostate cancer.

93 Background: Anti-hormonal therapies targeting the androgen receptor (AR) pathway are the standard of care treatment for patients with prostate cancer (PC). However, patients can develop resistance to androgen receptor pathway inhibitors (ARPI), and one mechanism is due to AR ligand-binding domain mutations (AR LBDm). While evidence suggests AR LBDm may play a key role in disease progression, limited data on AR LBDm is available across PC clinical states. This study used large real-world data to assess mutations within 12 AR LBD genes reported in the literature supporting a potential resistance mechanism to ARPI and their association with patient survival. Methods: The Guardant INFORM de-identified real-world clinical-genomic database was utilized, which links comprehensive circulating tumor DNA (ctDNA) results (using the next-generation sequencing Guardant360 assay [G360]) to claims data. The study included U.S. PC patients who had G360 testing between June 2014 and June 2023. A 6-month baseline period before the first G360 test (index) was used to obtain clinical data. Prevalence of AR LBDm was calculated for the presence of 12 mutations (L702H, T878A, H875Y, F877L, W742C, T878S, D891H, W742L, V716M, D891Y, S889G, and V716L) in the AR LBD region and 95% confidence intervals (CIs) were estimated based on Poisson distribution. Association between AR LBDm and real-world overall survival (rwOS) was assessed using log rank tests and Cox proportional hazards models adjusted for age, race/ethnicity, year of test, comorbidity indices, metastasis status, and PC clinical state. Patients with repeated G360 testing were evaluated for AR LBDm status over time. Results: In 16,757 PC patients with a valid ctDNA test result, overall prevalence of AR LBDm was 14.8% (95% CI: 14.2-15.4%). The prevalence of AR LBDm was higher in patients with metastatic castration-resistant PC, older than 75 years, with higher comorbidity index, and received multiple lines of PC treatment. PC patients with AR LBDm had shorter median rwOS (16.2 months) after index compared to patients without those 12 AR LBDm (27.4 months) and patients with undetectable ctDNA (53.1 months) (Table). In 2,030 (12%) of patients with >=2 ctDNA tests (median interval=7.5 months), AR LBDm emerged in 11% of patients and was lost in 4% of patients. Conclusions: This large real-world study shows that AR LBDm varied by patient clinical and treatment characteristics and occurred in early PC states. Consistent with previous research, AR LBDm was associated with worse OS in PC patients. As AR LBDm status can change over time, continuous monitoring of AR LBDm, especially for patients with poor treatment responses, may improve patient outcomes. rwOS by AR LBDm status. AR LBDm Status N No. of Events Median rwOS in Months (95% CI) P-value Undetectable ctDNA 2,211 429 53.1 (48.6, 65.9) <0.0001 Without any of the 12 AR LBDm 12,064 3,704 27.4 (26.5, 28.4) With any of the 12 AR LBDm 2,482 1,050 16.2 (15.1, 17.2)