Abstract PS4-02-20: Clinical and genomic biomarkers of capivasertib response in patients with hormone receptor-positive, HER2-negative metastatic breast cancer (HR+/HER2- MBC)

Abstract Introduction: HR+/HER2- MBC remains incurable, with therapeutic resistance limiting treatment durability. Capivasertib, the first FDA-approved AKT inhibitor for MBC with PIK3CA/AKT1/PTEN alterations, is now widely utilized. However, predictors of response remain poorly understood. This study examined baseline genomic alterations and prior therapy exposures for associations with progression-free survival (PFS) in patients treated with capivasertib. Methods: This retrospective cohort included patients from three U.S. centers treated between November 2023 (capivasertib approval) and June 2025 (data cutoff). Follow up continued until progression, toxicity-related discontinuation, death, or data cutoff. Eligible patients had HR+/HER2- MBC treated with capivasertib plus fulvestrant and baseline next-generation sequencing, as defined by testing within 6 months of therapy start. Two cohorts were analyzed: 1) patients with baseline circulating tumor DNA (ctDNA) sequencing only, and 2) patients with baseline ctDNA or tissue sequencing. Patients were identified via pharmacy claims; clinical and sequencing data were extracted and de-identified. The primary variable examined was the presence of baseline oncogenic genomic alterations; variants of uncertain significance and genes altered in 3 patients were excluded. Other exposures included prior lines of metastatic therapy and prior PI3K inhibitor (PI3Ki) use. The primary outcome was PFS, analyzed using the Kaplan-Meier method and Cox regression. Results: In cohort 1, we identified 45 patients treated with capivasertib and fulvestrant with baseline ctDNA-only sequencing. For this group, median age was 67, 96% were postmenopausal, and two-thirds had visceral and/or brain metastases. Median number of prior lines of therapy was 2; 96% had received a CDK4/6 inhibitor, 60% fulvestrant, 51% chemotherapy, and 22% prior PI3Ki. Median PFS in cohort 1 was 253 days (95% CI 147-392); 21 patients remained on therapy at data cutoff (approximately half had 6 months follow up). Baseline PIK3CA mutations (n=30) trended toward longer PFS (305 vs 194 days; HR 0.58, 95% CI 0.24-1.47); PTEN alterations (n=6) showed no PFS difference (HR 1.09, 95% CI 0.12-4.53), and AKT1 mutations (n=2) were too few to analyze. Pathogenic CCND1 alterations (n=4; 3 amplifications, 1 truncating mutation) were associated with shorter PFS (101 vs 305 days; HR 4.91, 95% CI 1.22-15.40), as were truncating ARID1A mutations (n=4; HR 6.35, 95% CI 1.19-23.45). ESR1 mutations (n=12), all co-occurring with a PI3K pathway alteration (PIK3CA, PTEN, or AKT1), trended toward worse PFS (147 vs 356 days; HR 1.86, 95% CI 0.66-4.82). PFS was numerically longer in patients with ≤1 prior line of therapy vs 2-3 prior lines (305 vs 202 days; HR 1.17, 95% CI 0.36-3.53) or ≥4 prior lines (305 vs 151 days; HR 1.44, 95% CI 0.51-4.06). Prior PI3Ki use trended toward shorter PFS (147 vs 305 days; HR 1.67, 95% CI 0.64-4.07). In cohort 2, 53 patients had baseline ctDNA or tissue sequencing; demographic and clinical features were consistent with cohort 1. FGFR1 alterations, too infrequent to analyze in the ctDNA-only group, trended toward shorter PFS (n=3; 88 vs 253 days; HR 2.33, 95% CI 0.58-6.92). Other clinical-genomic associations remained similar. Conclusions: In this study of patients with HR+/HER2- MBC treated with capivasertib and fulvestrant, alterations in CCND1 and ARID1A in ctDNA were associated with poorer PFS, while ESR1 and PI3K pathway co-mutations trended toward worse outcomes. Patients treated in the second-line or earlier, and those without prior PI3Ki exposure, may derive greater benefit. These findings are exploratory and warrant validation but suggest specific biomarkers that may correlate with risk for early progression on an AKT inhibitor. Citation Format: M. R. Lloyd, E. L. Podany, Z. Bagheri, E. Rizk, A. Dedeoglu, A. Putur, C. Ping, J. Hesse, A. Golden, S. Addison, J. J. Tao, N. Vidula, R. O. Abelman, K. K. Clifton, I. Sanidas, A. Varkaris, D. Juric, N. Vasan, C. X. Ma, G. M. Wulf, L. W. Ellisen, S. A. Wander. Clinical and genomic biomarkers of capivasertib response in patients with hormone receptor-positive, HER2-negative metastatic breast cancer (HR+/HER2- MBC) abstract. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS4-02-20.