A preoperative model for predicting microvascular invasion in hepatocellular carcinoma based on laboratory findings and MRI features.

Microvascular invasion (MVI) is an important adverse prognostic factor in hepatocellular carcinoma (HCC), and accurate preoperative prediction of MVI may help optimize treatment planning. This study aimed to develop and validate a preoperative model based on laboratory indicators and MRI features for predicting MVI in patients with HCC. This retrospective study included 535 patients with HCC who underwent surgical resection at two hospitals. Patients from the Affiliated Xuzhou Hospital of Xuzhou Medical University were assigned to the training cohort (n = 388), and those from Yancheng First People's Hospital were assigned to the validation cohort (n = 147). Univariate logistic regression was performed to identify potential risk factors for MVI, and variables with statistical significance were entered into multivariate logistic regression to establish a predictive model. Model performance was assessed using the area under the receiver operating characteristic curve (AUC). Multivariate logistic regression identified AFP > 400 ng/mL, larger tumor size, and absence of a radiological capsule as independent predictors of MVI. A predictive model incorporating these three variables was constructed. The model showed acceptable discriminatory ability in both the training cohort (AUC = 0.782) and the validation cohort (AUC = 0.776). AFP > 400 ng/mL, larger tumor size, and absence of a radiological capsule were independent predictors of MVI in HCC. A preoperative model combining laboratory and MRI features may provide supportive information for estimating the risk of MVI and assisting clinical decision-making.