HER2+乳腺癌患者蒽环类药物相关急性心脏毒性危险因素分析及预测模型构建

Risk Factors Analysis and Predictive Model Construction of Acute Cardiotoxicity in HER2+Breast Cancer Patients Induced by Anthracyclines

目的本研究旨在探讨人类表皮生长因子受体2(humman epidermal growth factor receptor 2,HER2)阳性(+)乳腺癌患者蒽环类药物相关急性心脏毒性(anthracycline-induced cardiotoxicity,ACT)的危险因素,并构建其诺曼图(nomogram)预测模型,以期为此类患者的预测、诊断与预防提供理论依据。方法回顾性分析我院2016年1月至2021年12月收治的HER2+乳腺癌患者319例,根据蒽环类药物化疗后是否发生急性ACT将其分成急性ACT组(n=149)与无急性ACT组(n=170)。采用多因素Logistic回归分析探讨HER2+乳腺癌患者发生急性ACT的影响因素,并进一步用R软件建立nomogram预测模型。结果两组患者在年龄、身体质量指数(body mass index,BMI)、高血压病史、高血脂病史、糖尿病病史、联合曲妥珠单抗、联用右丙亚胺、总胆固醇(TC)、低密度脂蛋白胆固醇(LDL-H)、肌酸激酶同工酶(CK-MB)及α-羟丁氨酸脱氢酶(α-HBDH)方面比较,差异有统计学意义(P<0.05)。多因素Logistic回归分析显示,BMI、糖尿病病史、联合曲妥珠单抗、联合右丙亚胺、LDL-H、CK-MB是HER2+乳腺癌患者发生急性ACT的危险因素(P<0.05)。以上述危险因素为指标构建nomogram预测模型,一致性指数(C-Index)为0.870[95%CI(0.831,0.909)],Hosmer-Lemeshow拟合优度检验结果显示,该预测模型的拟合效果良好(χ^(2)=11.180,P=0.192)。结论BMI、糖尿病病史、联合曲妥珠单抗、联合右丙亚胺、LDL-H、CK-MB是HER2+乳腺癌患者发生急性ACT的危险因素,基于上述危险因素构建的nomogram预测模型具有良好的区分度及准确性,可有效识别高危患者,对临床有一定的指导意义。

OBJECTIVE To explore the risk factors of acute cardiotoxicity(ACT)in humman epidermal growth factor receptor 2+(HER2+)breast cancer patients caused by anthracyclines,and construct its nomogram prediction model,in order to provide theoretical basis for the prediction,diagnosis and prevention of such patients.METHODS A total of 319 patients with HER2+breast cancer treated in our hospital from January 2016 to December 2021 were included and divided into ACT group(n=149)and no ACT group(n=170)according to whether ACT occurred after anthracycline chemotherapy.The clinical data of all patients were collected,and the influencing factors of ACT in HER2+breast cancer patients were investigated by multivariate logistic regression analysis.The risk factors were introduced into R 4.2.0 software.RESULTS Significant differences between the two groups were shown in the age,BMI,history of hypertension,history of hyperlipidemia,history of diabetes,combination of trastuzumab,combination of cardioprotective drugs dextropropimine,total cholesterol(TC)、low density lipoprotein-cholesterol(LDL-H),creatine kinase-myocardial band isoenzyme(CK-MB),α-hydroxybutyrate dehydrogenase(α-HBDH)(P<0.05).The results of multivariate logistic regression analysis showed that BMI,history of diabetes,trastuzumab combined with dextropropamine,LDL-H and CK-MB were the risk factors of ACT in HER2+breast cancer patients(P<0.05).Nomogram prediction model was constructed with the above risk factors as indicators.The area under ROC curve was 0.870(95%CI:0.831,0.909),and the calibration curve of the prediction model was basically consistent with the actual curve.Hosmer lemeshow goodness of fit test showed that the prediction model had a good fitting effect(χ^(2)=11.180,P=0.192).CONCLUSION BMI,history of diabetes,combined trastuzumab,combination with dextropropimine,LDL-H and CK-MB are the risk factors of ACT in patients with HER2+breast cancer.The nomogram prediction model has good differentiation ability and accuracy and can effectively identify the high-risk patients.