From the Departments of Radiology (R.K., J.M.C., S.H.L., S.Y.K., E.S.K., N.C., W.K.M.) and Breast Surgery (H.B.L.), Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul 03080, Republic of Korea; and Department of Healthcare Center, Seoul Metropolitan Government-Seoul National University Boramae Medical Center, Seoul, Republic of Korea (R.K.).
Radiology. 2019 Oct;293(1):49-57. doi: 10.1148/radiol.2019190014. Epub 2019 Aug 13.
Background In patients who are expected to achieve axillary pathologic complete response (pCR) after neoadjuvant chemotherapy (NAC), omission of axillary lymph node (LN) dissection could prevent morbidity and complications. Purpose To develop a clinical model to predict residual axillary LN metastasis in patients with clinically node-positive breast cancer after NAC by using MRI and US. Materials and Methods In this retrospective study, women with clinically node-positive breast cancer who were treated with NAC following surgery between January 2015 and September 2017 were included. The patients were randomly assigned to a test and validation set (7:3 ratio). Univariable and multivariable logistic regression analyses were performed to evaluate the independent predictors of residual axillary LN metastasis in the test set. A prediction risk score was developed based on the odds ratios from the multivariable analysis and validated in both sets. Results A total of 408 women were included (mean age ± standard deviation, 47.9 years ± 9.6). The axillary pCR rate was 56.6% (231 of 408). Independent predictors of residual axillary LN metastasis were clinical stage N2 or N3, presence of axillary lymphadenopathy at US after NAC, tumor size reduction less than 50% at MRI, Ki-67 negativity, hormone receptor positivity, and human epidermal growth factor receptor 2 negativity (all, < .05). In a model using these predictors, the area under the receiver operating characteristic curve in the test and validation sets was 0.84 (95% confidence interval: 0.79, 0.88) and 0.78 (95% confidence interval: 0.70, 0.87), respectively. When the patients had a simplified risk score of 1, the false-negative rates ranged between 5%-10%. Conclusion A prediction model incorporating nodal status stage, US finding, MRI response, and molecular receptor status shows good diagnostic performance for residual axillary lymph node metastasis after neoadjuvant chemotherapy in patients with clinically node-positive breast cancer. © RSNA, 2019 See also the editorial by Whitman in this issue.
背景 在接受新辅助化疗 (NAC) 后预计达到腋窝病理完全缓解 (pCR) 的患者中,省略腋窝淋巴结 (LN) 清扫术可以预防发病率和并发症。目的 通过 MRI 和 US ,为接受 NAC 治疗后的临床阳性乳腺癌患者开发一种预测 LN 残留的临床模型。材料与方法 在这项回顾性研究中,纳入了 2015 年 1 月至 2017 年 9 月期间接受 NAC 联合手术治疗的临床阳性乳腺癌患者。将患者随机分配到测试和验证集(7:3 比例)。在测试集中进行单变量和多变量逻辑回归分析,以评估 LN 残留的独立预测因素。根据多变量分析的优势比,建立预测风险评分,并在两个数据集进行验证。结果 共纳入 408 例患者(平均年龄 ± 标准差,47.9 岁 ± 9.6)。腋窝 pCR 率为 56.6%(231/408)。LN 残留的独立预测因素包括临床分期 N2 或 N3、NAC 后 US 检查有腋窝淋巴结肿大、MRI 上肿瘤大小缩小小于 50%、Ki-67 阴性、激素受体阳性和人表皮生长因子受体 2 阴性(均 <.05)。在使用这些预测因素的模型中,测试和验证集的受试者工作特征曲线下面积分别为 0.84(95%置信区间:0.79,0.88)和 0.78(95%置信区间:0.70,0.87)。当患者简化风险评分达到 1 时,假阴性率在 5%-10%之间。结论 纳入淋巴结状态分期、US 发现、MRI 反应和分子受体状态的预测模型,对临床阳性乳腺癌患者接受 NAC 治疗后 LN 残留具有良好的诊断性能。版权所有©2019,RSNA。本期另见 Whitman 编辑述评。