TITLE:
Research Progress on the Correlation and Predictive Value of Multimodal Ultrasound Imaging Features with Immunohistochemical Subtypes of Intrahepatic Cholangiocarcinoma
AUTHORS:
Meilin Qin, Hui Wang, Yan Zhang, Xuan Gao
KEYWORDS:
Intrahepatic Cholangiocarcinoma, Multimodal Ultrasound, Contrast-Enhanced Ultrasound, Immunohistochemistry
JOURNAL NAME:
Journal of Biosciences and Medicines,
Vol.14 No.4,
April
17,
2026
ABSTRACT: Intrahepatic cholangiocarcinoma (ICC) is the second most common primary liver malignancy and is characterized by substantial biological heterogeneity and poor prognosis. Immunohistochemical (IHC) biomarkers, including cytokeratin‑19 (CK‑19), Ki‑67, and S100P, provide important information regarding tumor differentiation, proliferative activity, and aggressive behavior. However, definitive evaluation of these biomarkers typically relies on postoperative pathological examination or invasive biopsy procedures. Therefore, identifying reliable non‑invasive imaging biomarkers capable of predicting IHC phenotypes before treatment is of considerable clinical importance. Multimodal ultrasound (US), integrating conventional B‑mode imaging, color Doppler flow imaging (CDFI), contrast‑enhanced ultrasound (CEUS), and elastography, enables comprehensive characterization of tumor morphology, vascular perfusion dynamics, and mechanical tissue properties. Increasing evidence suggests that these imaging features reflect underlying tumor microenvironment and molecular characteristics. In particular, CEUS enhancement patterns—such as arterial‑phase peripheral rim enhancement, heterogeneous perfusion, and early washout—together with increased stiffness measured by elastography, have been associated with aggressive biological behavior and higher proliferative activity in ICC. Recent advances in quantitative imaging analysis, radiomics, and artificial intelligence (AI) have further expanded the potential of ultrasound‑based biomarkers for predicting tumor biology. By extracting high‑dimensional imaging features and integrating clinical variables, machine‑learning models have demonstrated promising performance in predicting Ki‑67 expression, microvascular invasion, and tumor subtype. This review summarizes current evidence regarding the correlation between multimodal ultrasound features and IHC markers in ICC, discusses emerging quantitative imaging techniques, and highlights future directions for developing imaging‑based predictive models that may facilitate individualized treatment planning and precision oncology.