Can imaging data predict cancer severity? This study explores the use of CT-based radiomics to predict the pathological grade of hepatocellular carcinoma (HCC), a common type of liver cancer. The research aims to construct and validate radiomics models using contrast-enhanced CT (CECT) scans to distinguish between low- and high-grade HCC. Data from 242 patients with pathologically confirmed HCC were used to construct radiomic models, using univariate analysis and LASSO regression. Combined models incorporating clinical factors and radiomics scores were also developed. The arterial phase and portal venous phase (AP+VP) radiomics model demonstrated the best performance in predicting HCC pathological grade, achieving AUC values of 0.981 in the training dataset and 0.842 in the validation dataset. The study concludes that low- and high-grade HCC can be distinguished with good diagnostic performance using a CECT-based radiomics model, providing a non-invasive tool to improve diagnosis and inform treatment strategies. This research contributes to the field of **clinical oncology**, highlighting the potential of **radiology** and **diagnostic imaging** in improving the management of **liver cancer**.
Published in Frontiers in Oncology, this study aligns with the journal's focus on neoplasms, tumors, and oncology. The research explores the application of radiomics for predicting the pathological grade of hepatocellular carcinoma, fitting within the scope of the journal.