A nomogram model integrating LI-RADS features and radiomics based on contrast-enhanced magnetic resonance imaging for predicting microvascular invasion in hepatocellular carcinoma falling the Milan criteria

• DOI: 10.1016/j.tranon.2022.101597
• Publication Date: 2022-12-08T10:14:36Z
• Authors: Hai-Feng Liu; Yan-Zhen-Zi Zhang; Qing Wang; Zu-Hui Zhu; Wei Xing
• Keywords: Nomogram

To establish and validate a nomogram model incorporating both liver imaging reporting data system (LI-RADS) features contrast enhanced magnetic resonance (CEMRI)-based radiomics for predicting microvascular invasion (MVI) in hepatocellular carcinoma (HCC) falling the Milan criteria.In total, 161 patients with 165 HCCs diagnosed MVI (n = 99) or without 66) were assigned to training test group. MRI LI-RADS characteristics selected by LASSO algorithm used Rad-score models, respectively, independent integrated develop model. The predictive ability of was evaluated receiver operating characteristic (ROC) curves.The risk factors associated (P<0.05) related larger tumor size, nonsmooth margin, mosaic architecture, corona enhancement higher Rad-score. areas under ROC curve (AUCs) feature 0.85 (95% CI: 0.78-0.92) 0.74-0.95), those 0.82 0.73-0.90) 0.80 0.67-0.93) groups, respectively. presented improved AUC values 0.87 0.81-0.94) group 0.89 0.81-0.98) MVI. calibration decision analysis demonstrated that had high goodness-of-fit clinical benefits.The can effectively predict HCC within criteria serves as valuable biomarker facilitating individualized decision-making.