Abstract PS1-10-16: Longitudinal genomic profiling of circulating tumor DNA in metastatic invasive lobular (ILC) and no special type (NST) breast cancer

Abstract Background Liquid biopsy enables minimally invasive cancer monitoring through blood circulating tumor DNA (ctDNA) sequencing. We analyzed ctDNA sequencing data of 127 blood samples from 27 patients with hormone receptor positive (HR+) metastatic invasive lobular breast carcinoma (ILC) and 10 patients with HR+ no special type/invasive ductal carcinoma (NST/IDC) from the UPMC Hillman Cancer Center APOLLO biobank collection. In APOLLO collection, patients’ blood is collected at each disease progression, enabling longitudinal analysis of ctDNA dynamics. In this study, we focused on longitudinal analysis of ctDNA dynamics in patients with ILC. Methods We analyzed 31 longitudinal NST samples (2-6 per patient, median 3) and 96 ILC samples (2-7 per patient, median 3). DNA was extracted from plasma and sequenced using shallow whole genome sequencing, with a mean coverage of 0.21x. We used ichorCNA to determine the tumor fraction (TFx) of samples, and used CNVkit to annotate gene level copy number variation (CNV) and filtered with a list of genes that have most frequent copy number aberration in breast cancer. GISTIC2.0 was applied to compare chromosomal segment aberration of ILC and NST cases. Gene level CNVs were collapsed to each patient, and continuity index (1 - number of transition/maximum possible transitions) was applied to measure the longitudinal consistency of gene CNVs. Results Patients with ILC had a median overall survival of 43 months (range 6-124 months) with a median of 5 treatment lines (range 2-11). IchorCNA analysis showed 85 samples (91.4%) had TFx 3% (mean TFx 12.5%). Tumor fraction fluctuates across time, and shows correlation with the patient treatment scheme. Patient-level analysis demonstrated predominant copy number gains (copy number ≥3) versus losses (copy number 2). MDM4, ELF3, IKBKE and IL10 were the four most gained genes across 88% of patients for each gene. ESR1 was amplified in 40% and ERBB2 in 32% of patients. The top copy number loss gene was ZFHX3 (68% of patients), followed by TP53, MAP2K4 and NCOR1 (totaling 64% of patients). We compared ILC cases to NST/IDC cases in genomic CNV. We discovered a chromosome 11q13.3 amplicon enriched in ILC cases. This region showed preferential gain in genes CCND1, FGF3, FGF4, and FGF19, suggesting a potential ILC oncogenesis driver locus. Additional ILC and NST/IDC samples are being prepared for sequencing to validate these findings. Conclusions This study demonstrates the dynamics of ctDNA in metastatic ILC, which can be used to infer tumor evolution in concordance with treatment scheme. ILC-enriched chromosomal signatures may suggest potential unique oncogenic drivers. We are also developing a transformer-based model to predict gene expression pattern via genomic copy number changes using public databases, and fine-tune it with in-house metastatic breast cancer sequencing dataset. We will validate the model with liquid biopsy sequencing data and paired tumor sequencing data to identify robust predictors that can stratify patients’ prognosis. Our study highlights the potential clinical utility for ctDNA-based cancer monitoring to advance precision oncology in metastatic breast cancer. Citation Format: D. Liu, R. Kumar, J. Hooda, A. C. Chang, L. Miller, M. Rosenzweig, J. M. Atkinson, A. Stalnaker, V. Gao, S. Oesterreich, A. V. Lee, M. Balic, J. Foldi. Longitudinal genomic profiling of circulating tumor DNA in metastatic invasive lobular (ILC) and no special type (NST) 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 PS1-10-16.