ObjectiveTo assess value of preoperative clinical data and enhanced CT imaging features in predic-tion of microvascular invasion (MVI) and early recurrence (recurrence in one year) after curative resection for hepatoce-llular carcinoma (HCC). MethodsA retrospective analysis was conducted for 150 patients with HCC who underwent curative tumor resection in West China Hospital of Sichuan University from April 2014 to May 2015. The roles of preoperative CT characteristics and clinical data on MVI and early recurrence after curative tumor resection were evaluated by univariate and multivariate analyses. Resultscompared with HCC with no MVI and no early recurrence after curative resection, univariate analysis results showed that HCC with MVI and early recurrence had larger tumor size (P=0.002, P=0.005), a higher proportion of non-smooth tumor margin (P<0.001, P<0.001), and tumor multifocality (P=0.005, P=0.038), HCC with MVI had a higher proportion of incomplete tumor capsule (P=0.032), HCC with early recurrence had a higher proportion of incomplete and absence tumor capsules (P=0.038) and a faster washout on portal venous phase-the percentage attenuation ratio on the portal venous phase (P=0.049) and relative washout ratio on the portal venous phase (P=0.020) were higher. A multivariate logistic regression analysis results showed that non-smooth tumor margin (OR=7.075, P<0.001; OR=4.125, P<0.001) and tumor multifocality (OR=3.290, P=0.008; OR=2.354, P=0.047) were the independent predictors for MVI and early recurrence after curative tumor resection, HCC with early recurrence also had a faster washout on the portal venous phase (OR=1.023, P=0.017). ConclusionNon-smooth tumor margin and tumor multifocality are independent risk factors for MVI and early recurrence after curative tumor resection, and HCC with early recurrence has a faster washout on portal venous phase. Preoperative enhanced CT imaging could predict MVI and early recurrence after curative tumor resection and CT imaging findings are helpful to choose reasonable treatment and predict prognosis.