乳腺癌化疗患者核素门控心肌灌注纹理特征及其对心肌损害的早期预测价值

Characteristics of radionuclide gated myocardial perfusion texture in patients with breast cancer undergoing chemotherapy and its early predictive value for myocardial damage

目的分析乳腺癌化疗患者核素门控心肌灌注纹理特征及其在早期预警蒽环类药物化疗后心肌损害的价值。方法纳入2021年1月—2022年2月新疆医科大学第一附属医院核医学科接受蒽环类药物规律化疗的乳腺癌患者122例,于化疗1疗程结束后采用SPECT和超声心动图检测患者心功能和心肌细胞活力状态,通过纹理分析提取心肌的40个纹理特征;随访1年,根据心脏毒性评价指标评估患者化疗后心肌损害情况,据此分为心肌损害组和无心肌损害组。比较2组患者临床资料,采用Logistic回归分析乳腺癌化疗后心肌损害的独立影响因素;使用R语言软件绘制列线图模型,绘制校正曲线、ROC曲线评估模型在早期预测乳腺癌化疗后心肌损害的内部效能。结果随访1年,失访7例,剩余115例患者中出现35例不同程度的心脏毒性(心肌损害组),余80例为无心肌损害组。心肌损害组化疗1疗程后SPECT指标中的相位标准差(SD)、相位直方图带宽(BW)高于无心肌损害组(t/P=2.418/0.017、2.304/0.023);纹理分析中心肌损害组Energy、Total energy、Contrast值高于无心肌损害组(t=8.003、6.178、4.911,P均<0.001);多因素Logistic回归分析显示,SD、BW、Energy、Contrast升高为乳腺癌化疗后心肌损害的危险因素[OR(95%CI)=1.480(1.027~2.134)、1.615(1.191~2.191)、5.953(2.247~15.766)、1.041(1.018~1.065)];以独立相关因素构建乳腺癌化疗后心肌损害的早期预测列线图模型得出C-指数为0.915(95%CI 0.889~0.984),校正曲线与理想曲线走形接近;列线图模型的ROC曲线分析显示,早期评估乳腺癌化疗后心肌损害的AUC为0.924,敏感度为0.886、特异度为0.838。结论早期核素门控成像心肌灌注的纹理特征能够预测乳腺癌蒽环类药物化疗后心肌损害,对于指导临床早期进行心肌保护,调整治疗方案具有重要意义。

Objective To analyze the characteristics of radionuclide gated myocardial perfusion texture in patients with breast cancer undergoing chemotherapy and its value in early warning of myocardial damage after anthracycline chemotherapy.Methods One hundred and twenty-two patients with breast cancer who received regular chemotherapy with anthracyclines in the Nuclear Medicine Department of the First Affiliated Hospital of Xinjiang Medical University from January 2021 to February 2022 were included.After one course of chemotherapy,SPECT and echocardiography were used to detect the patients'cardiac function and myocardial cell viability,and 40 texture features of myocardium were extracted through texture analysis;Follow up for 1 year,evaluate the myocardial damage of patients after chemotherapy based on cardiac toxicity evaluation indicators,and divide them into myocardial damage group and no myocardial damage group based on this.The clinical data of the two groups were compared,and the independent influencing factors of myocardial damage after chemotherapy for breast cancer were analyzed by Logistic regression;Use R language software to draw nomogram model,draw correction curve and ROC curve to evaluate the internal efficacy of the model in early prediction of myocardial damage after chemotherapy for breast cancer.Results Following up for one year,7 cases were lost,and out of the remaining 115 patients,35 showed varying degrees of cardiac toxicity,namely 35 in the myocardial injury group and 80 in the non-myocardial injury group.The phase standard deviation(SD)and phase histogram bandwidth(BW)of SPECT indicators in the myocardial injury group after one course of chemotherapy were higher than those in the non-myocardial injury group(t/P=2.418/0.017,2.304/0.023).The energy,total energy,and contrast values of the myocardial damage group in texture analysis were higher than those of the non-myocardial damage group(t=8.003,6.178,4.911,P<0.001).Multivariate logistic regression analysis showed that increased SD,BW,Energy and Contrast were risk factors for myocardial damage after chemotherapy in breast cancer[OR(95%CI)=1.480(1.027-2.134),1.615(1.191-2.191),5.953(2.247-15.766),1.041(1.018-1.065)];The nomogram model for early prediction of myocardial damage after chemotherapy for breast cancer was constructed based on independent related factors.The C-index was 0.915(95%CI 0.889-0.984),and the correction curve was close to the ideal curve;ROC curve analysis of nomogram model showed that AUC for early assessment of myocardial damage after chemotherapy for breast cancer was 0.924,sensitivity was 0.886,and specificity was 0.838.Conclusion The texture features of myocardial perfusion in early radionuclide gated imaging can predict myocardial damage after anthracycline chemotherapy for breast cancer,which is of great significance for guiding clinical early myocardial protection and adjusting treatment plans.