基于临床-病理-超声特征构建评估乳腺癌新辅助化疗疗效的列线图模型

Construction of nomogram model for evaluating neoadjuvant chemotherapy efficacy of breast cancer based on clinical pathological ultrasonic characteristics

目的 探讨基于临床-病理-超声特征构建评估乳腺癌(BC)新辅助化疗(NAC)疗效,构建并验证该风险预测列线图模型.方法 回顾性分析2021年6月-2023年3月赣州市肿瘤医院100例采用NAC的BC患者相关资料,按照7∶3比例分为建模集(n=70例)和验证集(n=30例).收集患者基线资料、临床资料及超声特征资料,根据NAC疗效状况分为有效组和无效组.通过单因素分析和多因素分析筛选影响BC患者NAC疗效的独立影响因素后构建预测模型.采用R软件绘制列线图和模型校准图,绘制ROC曲线评估风险预测模型的预测价值,使用H-L检验判断模型的拟合优度.结果 多因素结果显示,患者病灶大小、侧方声影、血流分级、淋巴结转移、分化程度、ER、PR和HER-2为评估NAC疗效的独立影响因素(P<0.05).建模集中临床-病理-超声特征构建的联合模型AUC为0.967(95%CI:0.930-1.000);验证集ROC曲线联合模型AUC为0.955(95%CI:0.913-0.997).二者联合模型对比各自单一构建模型均具有更高的临床预测价值.联合模型校准图校准曲线贴近标准曲线,提示该模型一致性较好.经H-L拟合优度检验结果分别显示为:χ^(2)=5.042,P=0.655(建模集),χ^(2)=4.392,P=0.355(验证集),拟合度较好,预测价值高.结论 基于临床-病理-超声特征构建的联合模型对BC患者NAC疗效相较于单一构建模型具有更高的临床评估效能和预测价值.

Objective To evaluate the efficacy of neoadjuvant chemotherapy(NAC)for breast cancer(BC)based on clinical pathological ultrasonic characteristics,and to construct and validate the risk prediction nomogram model.Methods A retrospective analysis was performed on the data of 100 BC patients who used NAC in our hospital from June 2021 to March 2023,and divided into model set(n=70)and validation set(n=30)according to the 7:3 ratio.Baseline data,clinical data and ultrasound characteristics of patients were collected,and they were divided into effective group and ineffective group according to the efficacy status of NAC.A predictive model was constructed by screening the independent influencing factors affecting the efficacy of NAC in BC patients through univariate analysis and multivariate analysis.R software draws nomogram and model calibration plot,ROC evaluates the predictive value of the risk prediction model,and H-L tests to determine the goodness-of-fit of the model.Results Multivariate results showed that lesion size,lateral sound and opacity,blood flow grade,lymph node metastasis,degree of dfferentiation,ER,PR and HER-2 were independent influencing factors in assessing the efficacy of NAC(P<0.05).The AUC of the combined model constructed by clinical-pathological-ultrasound features in the modeling set was 0.967(95%CI:0.930-1.000),and the AUC of the combined model of the validation set ROC curve was 0.955(95%CI:0.913-0.997).The combined model of the two has higher clinical prediction value than the single construction model of each.The calibration curve of the joint model calibration plot is close to the standard curve,indicating that the model has good consistency.The results of H-L goodness-of-fit test were shown as follows:χ^(2)=5.042,P=0.655(modeling set),χ^(2)=4.392,P=0.355(validation set),with good fit and high prediction value.Conclusion The combined model based on clinical pathology ultrasound features has higher clinical evaluation efficacy and predictive value for NAC efficacy in BC patients compared to a single constructed model.