Construction of a microvascular invasion risk scoring model for hepatocellular carcinoma based on preoperative data / 中国现代医生

ABSTRACT

Objective Microvascular invasion(MVI)risk scoring model was established based on the preoperative data of hepatocellular carcinoma(HCC)patients.Methods The clinical data of 1153 HCC patients who underwent hepatectomy in Hangzhou First People's Hospital from January 2000 to December 2021 were retrospectively analyzed.Random sampling method was used to divide the samples into modeling group(n=864)and verification group(n=289)at a ratio of 31.The modeling group used Logistic regression analysis model to explore the independent risk factors of MVI and established a prediction model accordingly.Receiver operating characteristic(ROC)curve and correction curve were drawn to evaluate the predictive ability and performance of the risk model.Results The incidence of MVI was 24.1%(208/864)in modeling group.Multivariate Logistic regression analysis showed that alpha-fetoprotein(AFP)>160ng/ml and total tumor volume(TTV)>30cm3 were independent risk factors for MVI(P<0.05).The total score of risk scoring model was 6 points,0-1 was classified as low risk,2-3 was classified as medium risk,and 4-6 was classified as high risk.The model predicted that the area under the curve(AUC)of MVI was 0.714 in modeling group and 0.731 in verification group.The calibration diagram showed that the prediction model had good performance.Conclusion The MVI risk prediction model for HCC patients based on TTV and AFP is simple and easy to use,which is conducive to preoperative treatment decision-making and doctor-patient communication.