Abstract RF7-05: Capivasertib plus fulvestrant in hormone receptor-positive (HR+) advanced breast cancer (ABC): exploratory ctDNA analyses from the Phase 3 CAPItello-291 trial

Abstract Background: Based on the results of the global Phase 3 CAPItello-291 trial, the pan-AKT inhibitor capivasertib (C) is recommended in combination with fulvestrant (F) for patients (pts) with PIK3CA/AKT1/PTEN-altered HR+ ABC progressing after ≥1 line of endocrine therapy-based treatment. The primary analysis showed clinically meaningful improvement in progression-free survival (PFS) in both overall and PIK3CA/AKT1/PTEN-altered populations (identified based on tumor sequencing) treated with C + F vs placebo (pbo) + F. Here, we report exploratory ctDNA analyses. Methods: Eligible pts with HR+/HER2- ABC had progression ≤12 months of (neo)adjuvant aromatase inhibitor (AI) treatment, or while on prior AI-based treatment for ABC. Pts were randomized 1:1 to C + F or pbo + F. PIK3CA/AKT1/PTEN alteration status was determined in tumor tissue collected at study entry assessed by FoundationOne®CDx NGS, and in baseline blood samples for ctDNA analysis using Guardant InfinityTM which also reports methylation-based ctDNA tumor fraction (TF). Results: Of 708 pts in the overall population, 658 (92.9%) had a ctDNA sample for analysis. Overall concordance between ctDNA and tissue was 83.3% (Table). No/low shedding (TF0.1%) was found in 49% of ctDNA samples from pts with alterations detected in tissue but not ctDNA (n=63), compared with 1% of ctDNA samples from pts with alterations detected in both tissue and ctDNA (n=207). Alterations were detected by ctDNA alone in 72/700 pts (10.3%). Of these, 42 had no tissue result; the remaining 30 alterations detected in ctDNA but not tissue were enriched for PTEN: 18 pt samples had ≥1 PTEN alteration, most large genomic rearrangements. In the 279/658 (42%) pts with detected ctDNA alterations, baseline characteristics were generally well-balanced between treatment arms: median (range) age was 58 (29-86) years, 80.3% were postmenopausal, 77.4% had prior CDK4/6i and 19.4% prior chemotherapy for ABC, 53.4% had co-occurring ESR1m. ECOG PS was worse in the C + F (PS 1: 38.0%) vs pbo + F (PS 1: 26.3%) arm. In the ctDNA-altered subgroup, PFS hazard ratio with C + F vs pbo + F was 0.43 (95% CI 0.33-0.56); co-occurring ESR1m detected 0.52 (0.36-0.76); ESR1m not detected 0.36 (0.23-0.54). In pts with an alteration detected in either tissue or ctDNA (n=359), PFS hazard ratio was 0.49 (0.38-0.62), in pts non-altered by ≥1 method (n=335) was 0.74 (0.58-0.95). Conclusions: ctDNA analysis offers a minimally invasive option to identify PIK3CA/AKT1/PTEN alterations in HR+ ABC, particularly for pts without suitable tumor tissue analysis. However, ctDNA testing alone may miss alterations, such as in cases of low ctDNA shedding (TF 0.1%), warranting a confirmatory tissue test. The PFS benefit of C + F vs pbo + F in the ctDNA-altered group was consistent with the primary tissue-based analysis. Clinical benefit with C + F was observed irrespective of ESR1m. Citation Format: N. C. Turner, M. Oliveira, S. J. Howell, F. Dalenc, W.-P. Chung, J. Cortés, H. L. Gomez Moreno, X. Hu, K. Jhaveri, P. Krivorotko, G. Lerzo, S. Loibl, S. Morales Murillo, Z. Nowecki, M. Okera, Y. Park, J. Sohn, M. Toi, L. Zhukova, E. DeBruin, A. Nardone, A. Spencer, M. Fulford, G. Schiavon, H. S. Rugo. Capivasertib plus fulvestrant in hormone receptor-positive (HR+) advanced breast cancer (ABC): exploratory ctDNA analyses from the Phase 3 CAPItello-291 trial 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 RF7-05.