肝癌患者经肝动脉化疗栓塞术后疼痛列线图预测模型的构建及验证

Construction and verification of a nomogram model for predicting pain after transcatheter arterial chemoembolization in patients with hepatocellular carcinoma

目的构建肝癌患者经肝动脉化疗栓塞(TACE)术后中重度疼痛列线图预测模型并进行内外部验证。方法回顾性收集2017年1月至2018年4月南方医科大学南方医院行TACE手术的肝癌患者264例作为建模集,采用数字疼痛评分法术后对患者进行疼痛评估,根据术后24 h内是否发生中重度疼痛分为疼痛组96例和非疼痛组168例,将单因素分析差异有统计学意义的变量纳入多因素Logistic回归模型确定风险预测因子,构建列线图预测模型并对模型进行内部验证。另收集2020年1—6月该院行TACE手术的肝癌患者87例作为验证集对模型进行外部验证。结果264例肝癌行TACE手术患者术后24 h内96例占36.36%发生中重度疼痛,术后24 h内疼痛组患者肌注吗啡用量共计1015 mg,平均每例患者10.57 mg。多因素Logistic回归分析显示术前疼痛、距肝包膜距离≤2 cm、凝血酶原活动度高、碘油剂量>10 ml、栓塞肿瘤多个是TACE术后中重度疼痛发生的独立危险因素(P<0.05)。而年龄>50岁是TACE术后中重度疼痛发生的保护因素(P<0.05)。根据上述指标建立列线图预测模型,该模型ROC曲线下面积(AUC)为0.799(95%CI 0.745~0.853);模型内部验证和外部验证的AUC分别为0.780和0.788。校准曲线显示列线图预测模型预测概率与实际概率之间的一致性良好。结论构建的肝癌患者TACE术后中重度疼痛列线图预测模型具有良好的区分度、准确度,对预测TACE术后中重度疼痛高风险人群、制订针对性的预防策略具有一定的指导意义。

Objective To establish a predictive model of moderate to severe pain in patients with hepatocellular carcinoma(HCC)undergoing transcatheter arterial chemoembolization(TACE).Methods 264 patients with HCC who underwent TACE operation in Southern Medical University from January 2017 to April 2018 were selected as the modeling set.The pain was assessed by numeric rating scales.The patients were divided into pain group(n=96)and non-pain group(n=168)according to whether moderate to severe pain occurred within 24 hours after the operation.Binary Logistic regression analysis were performed for variables that were statistically significant in the univariate analyses.The predictive nomogram was constructed and the internal validation was performed.In addition,87 patients with HCC who underwent TACE operation from January 2020 to June 2020 were selected as the validation set for external validation.Results In the modeling set,96 patients(36.36%)had moderate to severe pain within 24 hours after TACE operation in 264 patients with HCC,and the dosage of morphine intramuscularly injected within 24 hours was 1015 mg,with an average of 10.57 mg per patient.Multivariate Logistic regression analysis showed that preoperative pain,the distance between the tumor and capsule≤2 cm,high prothrombin activity,dosage of lipiodol>10 ml,and several thromboembolic tumors were independent risk factors for moderate to severe pain after TACE(P<0.05).Age>50 was the protective factor of moderate to severe pain after TACE(P<0.05).The area under ROC curve was 0.799(95%CI:0.745-0.853)in the modeling set.The area under Roc curve for internal validation and external validation were 0.780 and 0.788,respectively.The calibration curves showed satisfactory agreements between the model predicted probability and the actually observed probability.Conclusion The predictive model of moderate to severe pain after TACE was established in this study has good differentiation and accuracy,it has certain guiding significance for predicting the high-risk group of moderate to severe pain after TACE operation and formulating the targeted prevention strategy.