磁共振联合临床病理特征对乳腺癌新辅助化疗疗效的早期预测

Early Prediction of Neoadjuvant Chemotherapy Efficacy in Breast Cancer by Dynamic Contrast-Enhanced Magnetic Resonance Imaging Combined with Clinicopathological Features

目的通过动态增强磁共振成像联合临床病理特征确定乳腺癌新辅助化疗疗效相关指标,同时评估2周期新辅助化疗后肿瘤减少率(TRR)和早期强化率(EER)降低的最佳阈值,并据此建立模型早期预测新辅助化疗后的病理完全缓解(pCR)。方法回顾性收集2021年2月至2022年12月在我院接受新辅助化疗的浸润性乳腺癌患者。按疗效分为pCR组和Non-pCR组,比较两组患者的临床病理特征,采用多因素Logistic回归分析确定pCR的独立预测因子,并基于这些指标建立预测病理反应的列线图模型。结果共收集到符合入组条件的患者267例,其中pCR 107例,Non-pCR 160例。TRR的最佳阈值为50.8%,EER降低的最佳阈值为66.7%。多因素分析显示,临床T分期、分子亚型、TRR、EER差值是pCR的独立影响因素(均P<0.05),列线图模型受试者工作特征曲线下面积(AUC)为0.926,敏感度96.3%,特异度80.0%。经Bootstrap内部验证得出AUC值为0.905,校准曲线也显示预测概率与实际概率吻合较好,表明该模型有较好的预测能力。结论2周期新辅助化疗后,最佳TRR为50.8%,EER差值为66.7%。临床T分期、分子亚型、TRR、EER差值是pCR的独立影响因素,以此建立的乳腺癌疗效预测模型具有较好的预测能力。

Objective To identify effectiveness-related indicators by dynamic contrast-enhanced magnetic resonance imaging(DCE-MRI)combined with clinicopathological features and evaluate the optimal threshold for tumor reduction rate(TRR)and early phase enhancement rate(EER)reduction after two cycles of neoadjuvant chemotherapy(NAC),and a comprehensive model was established to predict pathological complete response(pCR)after NAC at an early stage.Methods Patients with invasive breast cancer who underwent NAC at the Affiliated Hospital of Southwest Medical University from February 2021 to December 2022 were retrospectively collected.According to the efficacy,it is divided into pCR group and Non-pCR group,the clinicopathological features of the two groups were compared,multivariate Logistic regression analysis was used to determine the independent predictors of pCR,and a nomogram model for predicting pathological responses was established based on these indicators.Results A total of 267 patients were collected,including 107 pCR and 160 Non-pCR.The optimal threshold for TRR was 50.8%and for EER reduction was 66.7%.Multivariate analysis showed that clinical T stage,molecular subtype,TRR,and EER differences were independent influencing factors of pCR(all P<0.05),and the area under the receiver operating characteristic curve(AUC)of the subjects in the nomogram model established based on these factors was 0.926,the sensitivity was 96.3%,and the specificity was 80.0%.After Bootstrap's internal verification,the AUC value is 0.905,and the calibration curve also shows that the predicted probability matches the actual probability well,indicating that the model has better predictive ability.ConclusionThe optimal TRR is 50.8%and the EER difference is 66.7%after two cycles of NAC.Clinical T stage,molecular subtype,TRR,and EER difference are independent influencing factors of pCR,and the prediction model of breast cancer efficacy established by this method has good predictive ability.