Abstract: Objective The aim of this study was to establish an imaging diagnosis model for atypical hepatocellular carcinoma (aHCC) and mass intrahepatic cholangiocarcinoma (mICC) differentiation. Methods 62 patients with aHCC and 31 patients with mICC were encountered in 904th Hospital, Joint Logistics Support Force between May 2019 and May 2024, and all patients underwent contrast-enhanced CT scan to record imaging parameters, and Gd-EOB-DTPA enhanced magnetic resonance (MR) scan to record main and concomitant features of tumors. Hinting parameters for diagnosis were analyzed by multivariate Logistic regression, and diagnostic efficacy was analyzed by receiver operating characteristic (ROC) curve. Results Imaging skewness, energy, entropy, mean value, TP enhancement characteristics, T2WI central hyperintensity, multifocal hyperintensity, tumor (false) capsule, fibrous septum, DWI target sign and "EOB cloud" sign in mICC lesions were significantly different as compared to in aHCC lesions (P<0.05); multivariate Logistic regression analysis showed that serum CA19-9 level (OR=3.473, 95%CI: 1.298～9.290), energy (OR=0.166, 95%CI: 0.075～0.366), entropy (OR=5.319, 95%CI: 2.447～11.558), mean value (OR=2.587, 95%CI: 1.491～4.487) and "EOB cloud" sign (OR=4.527, 95%CI: 1.960～10.453) were independent factors for predicting mICC (P<0.05); C-index of the diagnostic model based on the Logistic results was 0.836, and the diagnostic correction was very close to the ideal curve (P>0.05); ROC analysis showed that the area under the curve (AUC) of the model based on imaging feature was 0.865 (P<0.05), with the specificity of 80.3% and the sensitivity of 85.6%. Conclusion Imaging energy, entropy, mean value and "EOB cloud" sign are independent parameters for the differential diagnosis of aHCC and mICC, which warrants further clinical investigation.

Key words: Hepatocellular carcinoma, Intrahepatic cholangiocarcinoma, Computed tomography, Magnetic resonance imaging, Differential diagnosis