基于增强CT的细胞外体积分数联合常规检验标志物在肝癌分级中的预测价值

Prediction value of extracellular volume fraction combined with routine testing ofmarkers in the classification of hepatocellular carcinoma based on enhanced CT

目的 探讨细胞外体积分数及常规检验标志物与肝细胞肝癌病理分级的相关性,分析联合预测肝细胞肝癌病理分级的价值。方法 回顾性收集2019年1月~2023年10月于蚌埠医科大学第一附属医院就诊、经病理证实为肝细胞肝癌的患者170例,根据病理分级分为高分化组(n=49)和中低分化组(n=121),提取平扫期及平衡期病灶实质增强CT图像计算细胞外体积(ECV)值。使用单因素及多因素Logistic回归分析影像及临床特征;采用Spearman分析其与肝细胞肝癌病理分级相关性;通过绘制ROC曲线计算曲线下面积(AUC)评估单一模型及联合模型诊断效能。采用Kappa检验验证Logistic回归模型。采用R语言生成列线图使预测结果可视化。结果 两组间ECV、中性粒细胞计数与淋巴细胞计数比值(NLR)及甲胎蛋白的差异有统计学意义(P<0.01);与分级相关的因素有NLR(r=0.381)、血小板计数与淋巴细胞计数比值(r=0.338)、系统免疫炎症指数(r=0.262)、ECV(r=-0.545)、对比增强值(r=-0.349);联合预测模型AUC为0.919,优于各单一模型(ECV、NLR及甲胎蛋白的AUC分别为0.846、0.743、0.635)。结论 基于增强CT的细胞外体积分数联合常规检验标志物对于肝细胞肝癌病理分级有良好的预测价值。

Objective To explore the correlation between extracellular volume fraction(ECV),routine testing of markers and pathological grading of hepatocellular carcinoma(HCC),to analyze the value of joint prediction of pathological grading of HCC.Methods A total of 170 patients with HCC confirmed by pathology in the First Affiliated Hospital of Bengbu Medical University from January 2019 to October 2023 were retrospectively collected,and divided into well-differentiated group(n=49)and moderately poorly-differentiated group(n=121)according to the pathological classification.ECV was calculated by extracting the CT images of the lesions in plain scan period and equilibrium period.The image and clinical features were analyzed by univariate and multivariate Logistic regression.Spearman was used to analyze its correlation with pathological grading of HCC.ROC curve was drawn to calculate the AUC to evaluate the diagnostic efficiency of single model and joint model.Kappa test was used to verify the Logistic regression model.Using R language to generate nomograms to visualize the prediction results.Results The differences in ECV,the ratio of neutrophil count to lymphocyte count(NLR)and alpha-fetoprotein between the two groups were statistically significant(P<0.01).The factors related to classification are NLR(r=0.381),platelet-to-lymphocyte ratio(r=0.338),system immune-inflammation index(r=0.262),ECV(r=-0.545)and contrast enhances(r=-0.349).The AUC of the joint prediction model was 0.919,which was better than each single model(AUC in ECV,NLR and alpha-fetoprotein were 0.846,0.743 and 0.635 respectively).Conclusion Extracellular volume fraction of enhanced CT combined with routine test markers has good diagnostic value for pathological grading of HCC.