Objective To develop a nomogram including clinical and MRI-based radiomics characteristics to predict the expression of Ki67 in hepatocellular carcinoma(HCC)and validate its efficacy.Methods The clinical,imaging,and pathological data of 183 patients with HCC who underwent surgical resection in Nanjing Drum Tower Hospital from 2016 to 2020 were retrospectively analyzed.According to the results of immunohistochemistry,they were divided into positive and negative expression groups.For continuous variables,the student t-test and Mann-Whitney U test were used for comparison between the two groups.For categorical data,Chi-square or Fisher's exact tests were used for comparison between the two groups.LASSO regression were used to extract and screen radiomics features and calculate Radscore.Univariate and multi-variate logistic regression analyses were used to screen independent predictors of Ki67 expression and construct nomogram.ROC analysis and bootstrapping were used to verify the discrimination and calibration of the model.Decision curve was used for the clinical benefit analysis of the model.Results Compared with the Ki67-negative group,the Ki67-positive group was confirmed by lower average age,lower serum AFP level,lower tumor proton density fat fraction(PDFF),higher propor-tion of type Ⅱ and type Ⅲ shapes,higher Edmondson-Steiner grades and microvascular invasion probability(all P＜0.05).Univariate and multivariate Logistic regression analysis showed that serum AFP level,tumor shape,PDFF,and Rad-score were independent predictors of Ki67 expression.Based on the above parameters,a nomogram prediction model was es-tablished.With ROC analysis and internal validation,the model confirmed good discrimination(area under the curve=0.856,sensitivity 75％,specificity 87％)and calibration(mean absolute error 0.016).Net benefits were obtained as the threshold probabilities ranging from 0.05 to 1.0.Conclusion The nomogram model incorporating clinical and MRI-based radiomics features could effectively predict the Ki67 level in HCC tissues before surgery,which is conducive to individual-ized treatment.
维普
万方数据
