Abstract PS4-02-02: Ultrasensitive ctDNA-based MRD monitoring predicts relapse in postoperative HR+ inflammatory breast cancer

Abstract Background: Inflammatory breast cancer (IBC) is a rare, aggressive subtype characterized by rapid progression and a high risk of relapse despite curative-intent therapy. Outcomes are particularly poor for patients with residual disease following neoadjuvant treatment. Circulating tumor DNA (ctDNA)-based assays offer a noninvasive means of detecting minimal residual disease (MRD) and may enable real-time, risk-adapted decisions regarding adjuvant therapy. However, feasibility and performance data for personalized ctDNA monitoring in IBC remain limited. Methods: We prospectively enrolled adults with hormone receptor (HR)-positive (≥10% ER or PR) stage III IBC who did not achieve a pathologic complete response (pCR) to neoadjuvant chemotherapy to a single-arm phase II trial (NCT02971748). Eligible patients had no prior immunotherapy or active autoimmune disease and received pembrolizumab 200 mg IV every three weeks plus physician-chosen endocrine therapy for up to 24 months, starting during or after radiation. The primary endpoint was 2-year event-free survival (EFS); secondary endpoints included safety, toxicity, and overall survival (OS). As an exploratory analysis, the Myriad Precise ultrasensitive MRD assay was used to perform whole-genome sequencing of paired tumor/normal DNA and to subsequently design 1,000-variant capture panels. Plasma was collected postoperatively, serially during therapy, and at end-of-treatment (EOT) for ctDNA analysis. Quantitative tumor fraction (TF) and quality control (QC) metrics were tracked longitudinally. Results: Of 23 enrolled patients, 19 (83%) provided sufficient tumor or normal tissue for personalized panel design. A total of 153 out of 154 plasma samples (99%) passed all quality control (QC) metrics. ctDNA was detected in 42 of 154 samples (27%), with 23 (96%) ctDNA-positive calls below 0.01% tumor fraction (TF), underscoring the need for ultrasensitive detection. One patient who cleared ctDNA detection early remains recurrence-free (early molecular clearance). Two patients who were negative at baseline became MRD positive 10 and 15 months prior to clinical progression. For all patients with disease progression 24 months, we observed persistent and rising ctDNA levels ahead of radiographic relapse with lead times up to 18 months. In contrast, all patients who were MRD negative at the post-operative baseline and throughout treatment remained progression-free at 24 months. Correlative analyses of baseline MRD status and end-of-treatment (EOT) conversion patterns with EFS and overall survival (OS) are ongoing and will be presented. Conclusions: Personalized ctDNA MRD testing is feasible and highly sensitive in the postoperative HR-positive IBC setting, with 99% QC pass rate and detection capability below 0.001% TF. Preliminary data show that ctDNA kinetics correlate with clinical outcomes and ctDNA detection often precedes radiographic relapse by months. These findings support incorporating MRD monitoring into adjuvant trials for real-time therapeutic adaptation in high-risk breast cancer. Citation Format: R. Upadhyay, A. Alexander, Q. Ye, K. Chen, C. Yam, G. J. Hogan, S. Zhang, M. L. LaBella, S. Ratzel, The MDACC Inflammatory Breast Cancer Team, A. Lucci, W. A. Woodward, B. Lim. Ultrasensitive ctDNA-based MRD monitoring predicts relapse in postoperative HR+ inflammatory breast cancer 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 PS4-02-02. Acknowledgments: The participants included in the current project were recruited through a Merck supported investigator-initiated trial (NCT02971748). The WGS and ctDNA analyses were performed by the Myriad and MDACC teams through a Myriad and MDACC strategic alliance. The Morgan Welch IBC Research and Clinic Program is funded by Texas State.