Objective To explore the impact of microvascular invasion (MVI) on the survival prognosis of patients after radical hepatectomy for hepatocellular carcinoma, to analyze its related risk factors and preoperative prediction methods, and to provide reference and support for the treatment of early postoperative recurrence. MethodsBy searching domestic and international medical literature databases, we screened studies related to MVI in hepatocellular carcinoma, focusing on the definition, grading, risk factors, preoperative prediction methods, and postoperative treatment strategies of MVI, and summarized the results of the existing studies. ResultsMVI was a well-established risk factor for the intrahepatic metastasis and early postoperative recurrence of hepatocellular carcinoma. Currently, various methods were employed to predict MVI, including laboratory indicators, imaging genomics, and genomics. The laboratory indicators used for prediction included alpha-fetoprotein, protein induced by vitamin K absence or antagonist-Ⅱ, hepatitis B virus, tumor diameter, vascular endothelial growth factor A, and circulating tumor cells. Imaging genomics involved preoperative MRI with irregular tumor shape and intra-voxel incoherent motion diffusion-weighted imaging D value < 1.16 × 10-3 mm2/S, CT enhancement imaging features with irregular tumor margins, multiple foci, and contrast-enhanced ultrasound portal venous and delayed phase scores. Genomics included the maximum variant allele frequency of circulating tumor DNA. In cases where MVI was detected after surgery, adjuvant therapy options had gained attention, such as transcatheter arterial chemoembolization, hepatic arterial infusion chemotherapy, targeted therapy, immunotherapy, radiation therapy, antiviral therapy, and local treatment combined with systemic treatment. ConclusionsThe study of MVI and its targeted treatment strategies are important for reducing the postoperative recurrence rate of hepatocellular carcinoma and improving patient survival. The preoperative prediction model and postoperative treatment plan should be optimized in the future to provide more effective treatment reference for patients.