Transcriptomic-based molecular classification of ampullary adenocarcinoma

Ampullary adenocarcinoma (AMPAC) is a rare and heterogeneous malignancy with markedly variable clinical outcomes, underscoring the urgent need for molecularly informed classification and personalized therapeutic strategies. Current AMPAC classification relies primarily on morphological and immunohistochemical criteria, which lack prognostic accuracy and fail to capture the underlying molecular and tumor microenvironment heterogeneity. In this study, we integrated bulk RNA sequencing, single-cell RNA sequencing (scRNA-seq), and clinicopathological data to elucidate the molecular architecture of AMPAC and introduced a transcriptomic-based molecular classification (AMS). The AMS categorized patients into two molecular subtypes: mesenchymal AMPAC (AMS-M) and classical AMPAC (AMS-C). The AMS-M demonstrated increased epithelial-mesenchymal transition (EMT) and stromal activation, correlating with basal-like subtypes associated with unfavorable prognosis. In contrast, AMS-C tumors exhibited metabolic and differentiation programs with relatively preserved immune infiltration. Importantly, we identified MUC16 as a pivotal biomarker specifically overexpressed in AMS-M tumors. MUC16 knockout markedly reversed EMT, attenuated invasive and migratory capacities, and restored chemosensitivity in both cell lines and their xenograft models. Collectively, our findings establish AMS as a prognostically and therapeutically informative molecular classification framework for AMPAC, unveil the critical role of the tumor microenvironment in AMPAC heterogeneity, and provide a translational foundation for precision oncology in this rare tumor.