Abstract
The aggressive subtype of hepatocellular carcinoma (HCC) is associated with a poor prognosis, and histopathological biopsy is the current method used for its diagnosis and tumour microenvironment analysis. Hence, we constructed a radiomics-based artificial intelligence model with robust predictive performance and explored the underlying biological characteristics by analysing mRNA data. The predictive performance was validated in two external centres, yielding areas under the curve ranging from 0.79 to 0.84, and their ability to predict progression-free survival (PFS) was evaluated. Radiogenomics analysis revealed that the high-risk group exhibited increased cell proliferation and tumour immune suppression. KIT inhibitors may serve as potential therapeutic drugs, whereas ADAM9 and PTK2B are key genes influencing patient prognosis. The artificial intelligence model developed from MRI has emerged as a dependable method for predicting aggressive HCC, with further biological exploration offering the potential to augment its clinical utility.