Proteomics-based phosphorylated AKT S473 and PI3K/AKT inhibitors in patients with hormone receptor–positive (HR+) metastatic breast cancer (MBC).

Abstract

e13053 Background: Multiple PIK3CA/AKT/mTOR (PAM) pathway inhibitors have been FDA-approved for HR+ MBC with PIK3CA mutation or PTEN loss. There is a clinical unmet need for better characterization of biomarkers for responses and guidelines. We hypothesize that pretreatment phosphorylation sites of the PAM pathway may associate with clinical activity of PAM inhibitors. This retrospective analysis is designed to evaluate the clinical utility of a proteomics-based reverse phase protein microarrays (RPPA) assay in patients with HR+ PIK3CA mut+ MBC receiving PAM inhibitors. Methods: Through an IRB-approved protocol, a retrospective cohort of 107 patients with HR+ PIK3CA mut+ MBC receiving PIK3CA/AKT inhibitors was identified between January, 2019 and December 2025 at Cedars Sinai Medical Center. Of these, 20 had pretreatment from metastatic sites(n = 19) or primary breast tissue (n = 1) with FFPE tissue block available. The blocks underwent laser capture microdissection, lysis, and protein quantification via RPPA. Samples were stained for six AKT/mTOR pathway phosphoproteins and values converted to percentages based on an internal breast tumor reference population. Within this cohort, analyte values above the cohort median were defined as positive. The predictive value of AKT/mTOR protein phosphorylation for progression-free survival (PFS) or time to treatment discontinuation (TTD) following treatment with PI3K and AKT inhibitors (capivasertib, alpelisib, and inavolisib) was assessed in pre-treatment case samples using Wilcoxon rank-sum and Spearman’s Rho analysis. Results: Phosphoproteomic profiling identified significant clinical stratification, with AKT S473-positive samples (n = 9) demonstrating significantly longer mean PFS/TTD (10.3 months) compared to AKT S473-negative samples (n = 11, 5.25 months; Wilcoxon p < 0.05). Spearman’s Rho analysis confirmed a moderate correlation between AKT S473 levels and PFS/TTD (ρ = 0.38). Other AKT/mTOR pathway analytes did not show significant associations with clinical outcomes. Genomic analysis revealed PI3K/AKT/PTEN mutations in 95% (19/20) of samples; 18 with PIK3CA mut, and 1 with AKT mut. There is no clear association between genomic mutation vs PFS; no association between AKT/mTOR pathway analytes and the genomic alterations, likely due to limited sample size. Conclusions: In this pilot study, AKT S473 positivity emerged as a potential predictive biomarker in PI3K/AKT inhibitor-treated breast cancer, where PI3K/AKT/PTEN genomic mutations were ubiquitous. This suggests that the quantification of phosphorylated AKT may be a better predictor of response to treatment with PI3K or AKT kinase inhibitors than genomic alteration status. Further validation in a larger cohort is warranted to confirm the clinical utility of this approach.