建立临床-影像学联合列线图模型早期识别TACE抵抗肝细胞癌

Establishing combined clinical-imaging nomogram model for early identification of TACE-refractory hepatocellular carcinoma

目的建立预测TACE抵抗肝细胞癌(HCC)的临床-影像学联合列线图模型,评价其早期识别TACE抵抗HCC的价值。方法回顾性分析218例接受TACE的HCC患者临床资料,以其中137例作为训练集,81例作为验证集。采用Cox风险回归模型,基于训练集临床及影像学资料筛选预测TACE抵抗HCC的独立因素;建立列线图模型,评价其判断训练集与验证集中TACE抵抗低、高危HCC患者中位总生存期(OS)的差异;并以校准曲线及决策曲线验证列线图模型的校准度及临床价值。结果回归分析显示巴塞罗那临床肝癌(BCLC)分期、肿瘤最大径、边界及病灶数目是影响HCC患者中位OS的独立因素。列线图模型判断的训练集及验证集TACE抵抗低、高危HCC患者中位OS差异均有统计学意义(P均<0.05)。校准曲线及决策曲线显示临床-影像学联合列线图模型校准度良好,可使患者净获益。结论临床-影像学联合列线图预测模型可用于早期判别TACE抵抗HCC。

Objective To establish a combined clinical-imaging nomogram for early identification of TACE-refractory hepatocellular carcinoma(HCC).Methods Data of 218 patients with HCC who underwent TACE were retrospectively analyzed,among them 137 patients were divided into the training set,while the other 81 were divided into the validation set.Cox regression model was used to identify clinical-imaging risk factors of TACE-refractory HCC based on the clinical and imaging data of the training set,and the predictive nomogram was established.The median overall survival(OS)were compared between HCC patients with low and high TACE refractoriness in both sets.The calibration curve and decision curve analysis were used to observe the calibration degree and clinical value of the clinical-imaging combined nomogram.Results Regression analysis showed that Barcelona clinic liver cancer(BCLC)stage,the maximum tumor diameter,irregular tumor margin and tumor number were all independent predictors of median OS of HCC patients.The median OS of patients with low or high risk of TACE refractoriness were significantly different in both training set and validation set(both P<0.05).The calibration curve and the decision curve analysis showed that the combined clinical-imaging nomogram model had good calibration degree,and could bring net benefits to HCC patients.Conclusion The combined clinical-imaging nomogram model could be used to early identify TACE-refractory HCC.