Abstract: Objective　To explore the related risk factors of microvascular invasion (MVI) in hepatocellular carcinoma (HCC), and to construct a predictive model of MVI. Methods　The clinical data of 178 patients with HCC who underwent orthotopic liver transplantation for the first time were analyzed. Patients were divided into the MVI positive group (76 cases) and the MVI negative group (102 cases) according to postoperative pathological results. The preoperative general information, preoperative CT or MRI examination results, preoperative laboratory examination results, and pathological results of the patients were collected, and the patients were followed up. The recurrence-free survival (RFS) time was recorded for all patients. The diagnostic values of alpha-fetoprotein (AFP), aspartate aminotransferase to lymphocyte ratio (ALRI), tumor maximum diameter, γ-glutamyl transpeptidase to lymphocyte count ratio (GLR) and neutrophil to lymphocyte ratio (NLR) were analyzed through receiver operating characteristic (ROC) curve for the diagnosis of MVI, and the optimal cut-off value was determined. Multivariate Logistic regression was used to analyze the independent risk factors of MVI, and a predictive scoring model was established. The patients were divided into the MVI high-risk group (3-4 points, 67 cases) and the MVI low-risk group (0-2 points, 111 cases) according to the optimal cut-off value of the scoring model. The Kaplan-Meier survival curve was drawn to compare the RFS rate of the two groups. Results　Compared with the MVI-negative group, there were higher proportions of multiple tumors, tumor maximum diameter＞3.75 cm, preoperative AFP＞53.8 μg/L, NLR＞3, GLR＞85.84 and ALRI＞75.36 in the MVI-positive group (P＜0.05). The largest tumor diameter＞3.75 cm, preoperative AFP＞53.8 μg/L, GLR＞85.84, ALRI＞75.36 were independent risk factors for MVI in patients with hepatocellular carcinoma (P＜0.05). The scoring system composed of the above risk factors had good goodness of fit (χ2=2.553, P=0.862), area under the curve=0.787, P＜0.01, the optimal cut-off value was 2, the specificity was 0.814, and the sensitivity was 0.632. The results of survival analysis showed that the RFS rate of patients was significantly higher in the MVI low-risk group than that of patients in the MVI high-risk group (Log-rank χ2=37.584, P＜0.01). Conclusion　Preoperative GLR, ALRI, AFP and maximum tumor diameter are independent risk factors for MVI, and the predictive model composed of them has certain predictive value for the presence of MVI before surgery and postoperative recurrence.

Key words: liver neoplasms, gamma-glutamyltransferase, aspartate aminotransferases, liver transplantation, lymphocyte count, microvascular invasion, predictive model