Abstract PS2-07-07: Relationship of early ctDNA dynamics with response to inavolisib alone or in combination with endocrine therapy +/- CDK4/6 inhibitor in PIK3CA-mutated HR+, HER2- metastatic breast cancer from the first-in-human phI/Ib trial

Abstract Background: Rapid determination of patient response to cancer therapy is crucial for timely treatment decisions, and may enable faster clinical trial decision-making to accelerate drug development. Traditional response assessments require imaging over several weeks. Across multiple indications and therapies, circulating tumor (ct)DNA dynamics have demonstrated the potential to serve as earlier response endpoints. Here, we evaluate the relationship between patient response and changes in ctDNA after 15 days of treatment in the first-in-human PhI/Ib study (NCT03006172) of the recently FDA-approved p110α-inhibitor, inavolisib. Methods: ctDNA was isolated from plasma samples collected from patients immediately prior to treatment (baseline) and at Cycle 1 Day 15 (C1D15) of treatment. Sequencing was performed with the Foundation Assay for Circulating Tumor DNA (FoundationACTTM) or FoundationOne® Liquid CDx. Responders were defined as those patients with confirmed complete or partial response per RECIST v1.1; non-responders had progressive disease. Logistic regression modeling was used to explore associations between ctDNA metrics and response. Maximally selected rank statistics were used for determining the optimal progression-free survival (PFS) cutpoints. Key ctDNA metrics including total number of alterations (totalAlts) and maximum (max) or sum of PIK3CA variant allele frequency (VAF) were computed for the subset of responder/nonresponder samples at baseline (n = 128) and C1D15 (n = 98). For the 38 responder/nonresponder patients with paired baseline: C1D15 samples, log2 and % change were also calculated. Finally, PFS stratification was evaluated for the 100 patients who had both paired baseline: C1D15 samples and PFS data. Results: Univariate logistic regression to test the association between response/non-response and ctDNA metrics identified % change in PIK3CA VAF (p = 0.006 for PIK3CA sumVAFs; p = 0.007 for PIK3CA maxVAF) and C1D15 totalAlts (p = 0.034) as associated with response. The median % change between baseline and C1D15 PIK3CA sumVAF was -58.41% for responders, and 0% for non-responders. Percent change in PIK3CA maxVAF and sumVAFs were highly correlated with one another (rho = 0.996 , p 2.2e-16), so the % change in PIK3CA sumVAFs was prioritized moving forward. Further analyses showed C1D15 totalAlts did not significantly correlate with PIK3CA sumVAFs (rho = 0.186, p = 0.063), and C1D15 totalAlts did not remain significantly correlated with response in a multivariate model (p = 0.058). We tested the ability of % change in PIK3CA sumVAFs to stratify patients by PFS and identified 0% change to be the optimal cutpoint that effectively stratified patients (HR=10.4, 95% CI, 4.6-23.6), with 10 patients showing 0% change (median 43.74%) vs. 90 patients with ≤ 0% change (median -61.18%). Stratification by quartile or median changes did not effectively stratify patients by PFS. Conclusions: Together, our results suggest that early changes in ctDNA, specifically the % change in PIK3CA VAF between baseline and C1D15, are associated with objective response to inavolisib-containing regimens and may enable early stratification of patient response that is ultimately predictive of PFS. Additional research with larger cohorts and encompassing the evaluation of tumor fraction (not available for this dataset) are warranted to confirm these findings and to understand the applicability to other targeted therapies for HR+, HER2- breast cancer. Citation Format: S. Hilz, M. K. Accordino, P. L. Bedard, A. Cervantes, V. Gambardella, E. P. Hamilton, A. Italiano, K. L. Jhaveri, D. Juric, K. Kalinsky, I. E. Krop, M. Oliveira, C. Saura, P. Schmid, N. C. Turner, E. S. Sokol, M. Childress, R. S. Huang, Z. J. Assaf, J. Aimi, S. Royer-Joo, J. L. Schutzman, K. E. Hutchinson. Relationship of early ctDNA dynamics with response to inavolisib alone or in combination with endocrine therapy +/- CDK4/6 inhibitor in PIK3CA-mutated HR+, HER2- metastatic breast cancer from the first-in-human phI/Ib 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 PS2-07-07.