Decoding the pre-malignant niche in cholangiocarcinoma: A single-cell multi-omic atlas from inflammation to invasion.

e22555 Background: The dismal prognosis of extrahepatic cholangiocarcinoma (eCCA) is primarily attributable to late-stage diagnosis. Although chronic biliary inflammation is a well-established predisposing condition, the specific cellular and epigenetic mechanisms that orchestrate the progression from benign inflammation to invasive malignancy remain elusive, hindering the development of early interception strategies. Methods: We performed integrated single-cell RNA sequencing (scRNA-seq) and single-cell ATAC sequencing (scATAC-seq) on freshly procured human tissue samples representing the full pathological spectrum: normal bile duct, primary sclerosing cholangitis, pre-neoplastic lesions with dysplasia, and eCCA (perihilar and distal). Bioinformatic analyses included cross-sample integration, cell-type annotation, trajectory inference, and regulatory network construction. Results: Our analysis uncovered a co-evolutionary trajectory marked by escalating genomic instability and transcriptional reprogramming. Notably, somatic copy number alterations (SCNAs) were detectable within the epithelial compartment as early as the stage of chronic cholangitis, with their prevalence and genomic complexity increasing substantially towards carcinoma. Pre-neoplastic lesions displayed extreme multi-layered heterogeneity, indicative of a branching evolutionary model with divergent clinical fates. Through integrative analysis of monotonic expression patterns and stage-specific chromatin dynamics, we delineated a core transcriptional circuitry driving this transformation and derived a focused gene signature exhibiting strong discriminatory potential for early detection. Conclusions: This study provides the first high-resolution, single-cell multi-omic map of eCCA pathogenesis, defining the cellular and epigenetic continuum from inflammation to carcinoma. Our work elucidates fundamental mechanisms of inflammatory-driven oncogenesis and delivers a translatable framework for risk stratification and early interception.