基于剪切波弹性成像及超微血管成像影像学参数构建乳腺癌新辅助化疗疗效预测模型的预测效能

Predictive efficiency of a model of neoadjuvant chemotherapy for breast cancer based on parameters of shear wave elastography and superb microvascular imaging

目的:探讨基于剪切波弹性成像(SWE)及超微血管成像(SMI)影像学参数构建的乳腺癌新辅助化疗疗效预测模型的预测效能。方法:收集160例乳腺癌患者,其中训练集120例,验证集40例。根据化疗效果分为完全缓解(pCR)组和非pCR组。收集患者相关资料,采用单因素和多因素二元logistic回归分析筛选乳腺癌患者新辅助化疗后非pCR的独立影响因素,通过RStudio 4.2.1构建列线图风险模型并验证。结果:单因素分析显示,2组雌激素受体(ER)、孕激素受体(PR)、人类表皮生长因子受体-2(HER-2)、内部回声、钙化、边缘毛刺、最大径、动脉收缩期血流峰值速度(PSV)、RI、双侧横向剪切波速度比值(SWVr)及Adler血流分级差异均有统计学意义(均P<0.05)。多因素分析:临床特征分析显示,ER、PR、HER-2均是非pCR的独立影响因素(均P<0.05);影像学特征分析显示,PSV、SWVr均是非pCR的独立影响因素(均P<0.05);联合分析显示,ER、PR、HER-2、PSV、SWVr均是非pCR的独立影响因素(均P<0.05)。训练集临床特征模型、影像学特征模型、联合模型,验证集临床特征模型、影像学特征模型、联合模型的AUC分别为0.770(95%CI 0.681~0.858)、0.960(95%CI 0.915~1.000)、0.969(95%CI 0.926~1.000)、0.745(95%CI 0.634~0.855)、0.980(95%CI 0.947~1.000)、0.948(95%CI 0.871~1.000);敏感度分别为81.10%、93.20%、94.60%、82.20%、100.00%、98.60%;特异度分别为65.20%、91.30%、93.50%、88.00%、85.20%、88.00%。结论:乳腺癌患者新辅助化疗后非pCR受ER、PR、HER-2、PSV、SWVr的共同影响,根据上述因素建立预测模型具有较好的预测效能。

Objebtive:To explore the predictive efficiency of the model of neoadjuvant chemotherapy for breast cancer based on shear wave elastography(SWE)and superb microvascular imaging(SMI)parameters.Methods:A total of 160 patients with breast cancer were collected and divided into the training cohort(120 cases)and the validation cohort(40 cases).After neoadjuvant chemotherapy,the patients were divided into the pathological complete remission(pCR)group and non-pCR group according to the effect of chemotherapy.Univariate and multivariate logistic regression analysis were performed on the related factors,and a nomogram risk model was constructed and verified.Results:Univariate analysis showed that ER,PR,HER-2,internal echo,calcification,edge burr,the maximum diameter,PSV,RI,SWVr and Adler grade of blood flow were significantly different between the two groups(all P<0.05).Multivariate analysis showed that ER,PR,HER-2,PSV and SWVr were independent factors affecting the occurrence of non-pCR(all P<0.05).The AUCs of the clinical model,the imaging features model and the combined model of the training cohort were 0.770(95%CI 0.681~0.858),0.960(95%CI 0.915~1.000)and 0.969(95%CI 0.926~1.000),the sensitivities were 81.10%,93.20%and 94.60%,and the specificities were 65.20%,91.30%and 93.50%.The AUCs of the clinical model,the imaging features model and the combined model of the validation cohort were 0.745(95%CI 0.634~0.855),0.980(95%CI 0.947~1.000)and 0.948(95%CI 0.871~1.000),the sensitivities were 82.20%,100.00%and 98.60%,and the specificities were 88.00%,85.20%and 88.00%.Conclusions:The occurrence of non-pCR in breast cancer patients after neoadjuvant chemotherapy is affected by ER,PR,HER-2,PSV and SWVr.According to the above factors,the predictive model has a good predictive efficiency.