MRI Parameters and Clinicopathologic Features in Predicting Hepatocellular Carcinomas With Aggressive-Related Marker Expression, Defined by Epithelial Cell Adhesion Molecule and Glypican-3 Status

YanZhuo Li,1,* Lei Huo,2,* LianLian Liu,1 NingYang Jia,2 Bin Song1 1Department of Radiology, Minhang Hospital, Fudan University, Shanghai City, 201100, People’s Republic of China; 2Department of Radiology, The Eastern Hepatobiliary Surgery Hospital, Naval Medical University, Shanghai City, 200438, People’s Republic of China*These authors contributed equally to this workCorrespondence: Bin Song, Department of Radiology, Minhang Hospital, Fudan University, No. 170, Xinsong Road, Shanghai City, 201100, People’s Republic of China, Email [email protected] NingYang Jia, Department of Radiology, The Eastern Hepatobiliary Surgery Hospital, Naval Medical University, No. 225, Changhai Road, Shanghai City, 200438, People’s Republic of China, Email [email protected]: Epithelial cell adhesion molecule (EpCAM) and glypican-3 (GPC3) are associated with a poor survival rate in hepatocellular carcinoma (HCC). Predicting aggressive HCC non-invasively should be investigated using preoperative gadoxetic acid-enhanced MRI and clinical characteristics.Methods: This retrospective study examined 134 patients with confirmed cases of HCC who had undergone gadolinic acid-enhanced MRI before surgery. The data were collected between November 2017 and December 2021. Significant parameters were subjected to quantitative analysis and univariate testing. HCC characterized by poor differentiation and progenitor traits was identified by the positive expression of EpCAM and/or GPC3. Independent influencing variables for different subtypes of invasive HCC were identified by multivariate analysis. The diagnostic performance of variables was measured using ROC analysis.Results: Cases with EpCAM and GPC3 expression were classified into four subtypes based on IHC staining. Among them, EpCAM+/GPC3+ HCC had the highest risk. Significant predictors of HCC expressing aggressive markers included tumor size (p = 0.01), AFP level (p = 0.008), MVI (p = 0.003), satellite nodule (p = 0.002), proliferative pattern (p = 0.04), and arterial peripheral enhancement (p = 0.03). By integrating all of these relevant variables, the resulting diagnostic accuracy was enhanced to 0.914 (95% CI 0.830– 0.960) with a sensitivity of 0.964 (95% CI 0.922– 1.000) and a specificity of 0.833 (95% CI 0.365– 0.911).Conclusion: The research demonstrated that combining biochemical and radiological data allows for accurate prognosis and patient management, along with non-invasive detection of aggressive HCC with high accuracy.Keywords: glypican-3, epithelial cell adhesion molecule, hepatocellular carcinoma, magnetic resonance imaging