Artificial intelligence in hepatocellular carcinoma: imaging-based subtyping and prediction of treatment response and prognosis.
Authors
Affiliations (6)
Affiliations (6)
- Memorial Sloan Kettering Cancer Center, New York, USA. [email protected].
- Southern Radiology Consultants, Baton Rouge, USA.
- Banner - University Medical Center Tucson, Tucson, USA.
- Department of Radiology, University of Colorado Anschutz Medical Campus, University of Colorado, Boulder, USA.
- University of Chicago, Chicago, USA.
- University of North Carolina School of Medicine, Chapel Hill, USA.
Abstract
Hepatocellular carcinoma (HCC) is the most common primary liver malignancy, accounting for nearly 90% of primary liver cancers worldwide. It is a biologically heterogeneous malignancy with variable aggressiveness and treatment response. Imaging is central to the diagnosis and staging of HCC but offers limited insights into tumor biology. Advances in artificial intelligence (AI) and radiomics enable the extraction of quantitative imaging features from CT and MRI which can be used to assess tumor heterogeneity, predict response to therapy, and aid risk stratification. In addition, emerging evidence suggests that certain imaging features can aid in distinguishing HCC subtypes, raising the possibility that imaging could extend beyond diagnosis and staging to subtype classification. For example, rim arterial phase hyperenhancement, corona enhancement, intratumoral arteries, peritumoral hypointensity on hepatobiliary phase imaging, tumor in vein, and necrosis are associated with macrotrabecular-massive HCC. This review summarizes the current landscape of AI and radiomics in advancing imaging-based HCC assessment, specifically for HCC subtype classification and treatment response prediction and prognostication. Current progress, limitations, and future directions for integrating AI and radiomics into HCC management are also discussed.