初始淋巴结阳性乳腺癌新辅助治疗后的病理 完全缓解对腋窝手术的指导意义

The Significance of Pathological Complete Response after Neoadjuvant Therapy in Guiding Axillary Surgery for Initially Lymph Node-Positive Breast Cancer

摘要: 目的：探讨初始腋窝淋巴结阳性(clinically Node-positive, cN+)乳腺癌患者对新辅助治疗(Neoadjuvant Therapy, NAT)的病理反应差异，分析影响腋窝病理完全缓解(axillary pathological Complete Response, apCR)的相关因素，并构建ROC预测模型，为NAT后腋窝手术降级决策提供参考。方法：回顾性分析2018年10月至2025年4月于烟台毓璜顶医院乳腺外科接受NAT的481例乳腺癌患者，其中初始cN+患者237例。收集临床病理资料，对NAT后的乳腺及腋窝病理完全缓解率(breast pCR, bpCR; axillary pCR, apCR)的影响因素进行单因素及二元多因素Logistic回归分析。并基于独立预测因素构建apCR的受试者工作特征(ROC)曲线预测模型。结果：初始临床淋巴结阳性(clinically Node-positive, cN+)患者中，bpCR率为30.8% (73/237)，apCR率为46.0% (109/237)。不同分子分型间bpCR和apCR率差异显著：HER2过表达型bpCR率最高(75.5%)，Luminal A型最低(0.0%)；apCR率以HER2过表达型(79.2%)和Luminal B HER2+型(70.2%)最高，Luminal B HER2−型(19.0%)和Luminal A型(18.2%)最低。多因素分析显示，PR阴性(OR = 0.227, 95% CI: 0.086~0.603)是bpCR的独立预测因素；bpCR (OR = 9.997, 95% CI: 3.891~25.688)是apCR的独立预测因素。基于bpCR、HER2表达及ER状态构建的apCR预测模型，曲线下面积(AUC)为0.824 (95% CI: 0.796~0.899)，敏感度78.0%，特异度82.0%。结论：初始cN+乳腺癌患者中，分子分型与NAT病理反应密切相关，HER2过表达型apCR率最高。bpCR是apCR最强有力的独立预测因素。基于bpCR、HER2表达及ER状态构建的预测模型对apCR具有较好的预测能力。但bpCR状态需通过乳腺手术后的病理结果才能确定，该模型在严格意义上难以用于术前腋窝评估，仅可作为术中或术后评估腋窝手术降级策略可行性的辅助工具。其临床应用仍需在结合靶向腋窝清扫(TAD)等技术的前提下，通过前瞻性研究进一步验证。

Abstract: Objective: To investigate pathological response heterogeneity in breast cancer patients with clinically Node-positive (cN+) after neoadjuvant therapy (NAT), identify factors influencing axillary pathological complete response (apCR), and construct a receiver operating characteristic (ROC) prediction model to inform axillary surgery de-escalation decisions. Methods: A retrospective cohort study was conducted on 481 breast cancer patients who received NAT at the Department of Breast Surgery, Yantai Yuhuangding Hospital, from October 2018 to April 2025. Among these, 237 patients with baseline cN+ status were included. Clinical and pathological data were collected. Univariate and binary multivariate Logistic regression analyses were performed to explore predictors of breast pathological complete response (bpCR) and apCR post-NAT. A ROC curve model for apCR was constructed using independent predictive factors. Results: In the cN+ cohort, the bpCR rate was 30.8% (73/237) and the apCR rate was 46.0% (109/237). Significant differences in bpCR and apCR rates were observed across molecular subtypes: The HER2-overexpressing subtype had the highest bpCR rate (75.5%), while the Luminal A subtype had the lowest (0.0%). apCR rates were highest in the HER2-overexpressing (79.2%) and Luminal B HER2-positive (70.2%) subtypes, and lowest in the Luminal B HER2-negative (19.0%) and Luminal A (18.2%) subtypes. Multivariate regression revealed: Negative progesterone receptor (PR) status (OR = 0.227, 95% CI: 0.086~0.603) was an independent predictor of bpCR. bpCR (OR = 9.997, 95% CI: 3.891~25.688) was the strongest independent predictor of apCR. The apCR prediction model, incorporating bpCR, HER2 expression, and estrogen receptor (ER) status, had an area under the curve (AUC) of 0.824 (95% CI: 0.796~0.899), with a sensitivity of 78.0% and specificity of 82.0%. Conclusion: Molecular subtype is closely associated with NAT-induced pathological response in initially cN+ breast cancer patients, with the HER2-overexpressing subtype exhibiting the highest apCR rate. bpCR is the most robust independent predictor of apCR. The model integrating bpCR, HER2 expression, and ER status shows good apCR predictive performance, providing a reference for axillary surgery de-escalation in post-NAT cN+ patients. However, the status of bpCR can only be confirmed based on postoperative pathological findings of breast tissue. Strictly speaking, this model is not applicable to preoperative axillary assessment; instead, it can only serve as an auxiliary tool for evaluating the feasibility of axillary surgery downgrading strategies during or after the operation. Its clinical application requires further validation through prospective studies combined with techniques, such as targeted axillary dissection (TAD).