The Value Of Preoperative Imaging Nomogram Model For Predicting Ki 67 Expression Of Hepatocellular Carcinoma

Objective:To establish a nomogram model for preoperative prediction of Ki-67 expression level in hepatocellular carcinoma(HCC)based on contrast-enhanced CT or MRI features,and to provide a new screening tool for predicting the prognosis of HCC patients and formulating individualized treatment plans.Materials and methods:In this retrospective study,the data of 101 HCC patients who were pathologically confirmed and underwent immunohistochemistry in the Affiliated Hospital of North Sichuan Medical College(institution 1),including general clinical data,preoperative liver-enhanced CT or MRI imaging data,and postoperative pathological data,were used as training group.According to the state of postoperative pathological Ki-67,they were divided into high and low expression groups of Ki-67.SPSS22.0 and R4.1.2 were used for statistical analysis.Univariate and multivariate Logisticanalysis were used to screen out independent risk factors for high Ki-67 expression and establish a prediction model.The AUC was used to evaluate differentiation,the calibration curve was used to evaluate the calibration,and the Hosmer-Lemeshow test was used toevaluate the goodness-of-fit.The data of 45 HCC patients in The Second Affiliated Medical College of North Sichuan Medical College(institution 2)were collected as the test group for external validation of the prediction model.Results:Immunohistochemical results indicated that ki-67 expression was high in 60 cases(59.4%)and low in 41 cases(40.6%)of 101 HCC patients in the training group.In the test group,Ki-67 expression was high in 26 cases(57.8%)and low in 19 cases(42.2%)of 45 HCC patients.Multivariate analysis showed AFP(OR=5.963,95%CI:1.413-25.164,P=0.015),tumor pseudocapsule(OR=0.259,95%CI:0.068-0.990,P=0.048),tumor margin(OR=0.137,95%CI:0.031-0.594,P=0.008)and periumoral enhancement(OR=7.249,95%CI:1.626-32.314,P=0.009)were independent risk factors for the high level of Ki-67 expression in HCC.A nomogram model was constructed to predict the risk of high expression of Ki-67 in HCC,with AUC values of 0.911(95%CI:0.850-0.971)and 0.860(95%CI:0.794-0.971)in the training and validation groups,respectively.The sensitivity and specificity of the training group were 93.3%and 78.0%,respectively,and those of the validation group were 80.8%and 78.9%,respectively.The Bias-correctrd curve in the calibration curve fit well with the Ideal curve,and the nomogram predicted the risk of high Ki-67 expression was consistent with the actual situation.Conclusion:The nomogram model based on preoperative imaging can easily predict the expression level of Ki-67 in HCC with good predictive efficiency,providing a new non-invasive strategy for evaluating the cell proliferation status of HCC,and helping clinicians to judge the prognosis of HCC patients and guide the individualized treatment.