Deciphering ctDNA release influencing factors in luminal breast cancer via single-cell spatial mapping.

Abstract

e12596 Background: Circulating tumor DNA (ctDNA) is a promising non-invasive biomarker for detecting and monitoring HR+/HER2- breast cancer. However, its clinical utility is limited by tumor heterogeneity and dynamic release mechanisms. Understanding factors influencing ctDNA release is critical for improving its diagnostic and prognostic applications. Methods: We prospectively collected tumor tissues and matched blood samples from ten treatment-naive, II-III stage HR+/HER2- breast cancer patients, integrating single-cell full-length transcriptomic, mutational, and spatial transcriptomic data with plasma ctDNA mutations. Using a deep learning-based multi-omics framework, we computationally mapped ctDNA release at single-cell resolution by tracking clonal mutations co-existing in tumor subpopulations and plasma ctDNA. We systematically analyzed contributions of gene expression programs, functional driver mutations, spatial architectures, and tumor microenvironment (TME) cell-cell interactions to ctDNA release dynamics. Results: We identified eight epithelial states with distinct functional and spatial characteristics. ctDNA release was influenced by gene expression, functional mutations, spatial subclones, and TME composition (all p < 0.05). Mesenchymal luminal progenitors with stemness features demonstrated high ctDNA release potential and exhibited significantly enriched B-cell interactions. PIK3CA mutations enhanced ctDNA release through TME remodeling and increased tumor-associated macrophage phagocytosis. Perivascular distance analysis revealed tumor regions distal from vasculature showed elevated ctDNA release, linked to hypoxia and cellular stress pathway activation (all p < 0.01). Unsupervised clustering based on TME interactions stratified patients into five prognostic subgroups, with immunomodulatory TME profiles associated with poor overall survival and higher ctDNA release in METABRIC dataset (n = 1257, HR = 1.8, 95% CI: 1.2-2.5, p < 0.001). Conclusions: Our study pioneers high-resolution spatial mapping of ctDNA release at single-cell level for the first time, provides insights into ctDNA release heterogeneity from cellular subpopulations, genomic alterations, and immune interactions. We identified luminal progenitors as high ctDNA release potential and elucidated how mutations and TME interactions influence ctDNA release. We stratified patients according to TME characteristics and analyzed the impact on tumor function, prognosis, and ctDNA release. These findings highlight the importance of tumor heterogeneity in ctDNA studies and its potential for precision medicine.