Zhang Kai, Zhu Qian, Sheng Danli, Li Jiawei, Chang Cai
Department of Medical Ultrasound, Fudan University Shanghai Cancer Center, Shanghai, People's Republic of China.
Cancer Manag Res. 2020 Feb 10;12:965-972. doi: 10.2147/CMAR.S239921. eCollection 2020.
Few models with good discriminative power have been introduced to predict the risk of non-sentinel lymph node (non-SLN) metastasis in breast cancer after neoadjuvant chemotherapy (NAC). We aimed to develop a new and simple model for predicting the probability of non-SLN metastasis in breast cancer and facilitate the selection of patients who could avoid unnecessary axillary lymph node dissection following NAC.
A total of 298 patients diagnosed with invasive breast cancer, who underwent SLN biopsy after completing NAC and subsequently breast surgery, were included and classified into the training set (n=228) and testing set (n=70). Univariate and multivariate analyses were used to select factors that could be determined prior to breast surgery and significantly correlated with non-SLN metastasis in the training set. A logistic regression model was developed based on these factors and validated in the testing set.
Nodal status before NAC, post-NAC axillary ultrasound status, SLN number, and SLN metastasis number were independent predictors of non-SLN metastases in breast cancer after NAC. A predictive model based on these factors yielded an area under the curve of 0.838 (95% confidence interval: 0.774-0.902, < 0.001) in the training set. When applied to the testing set, this model yielded an area under the curve of 0.808 (95% confidence interval: 0.609-1.000, = 0.003).
A new and simple model, which incorporated factors that could be determined prior to breast surgery, was developed to predict non-SLN metastasis in invasive breast cancer following NAC. Although this model performed excellently in internal validation, it requires external validation before it can be widely utilized in the clinical setting.
很少有具有良好鉴别能力的模型被用于预测新辅助化疗（NAC）后乳腺癌非前哨淋巴结（non-SLN）转移的风险。我们旨在开发一种新的、简单的模型来预测乳腺癌非前哨淋巴结转移的概率，并有助于选择那些在NAC后可避免不必要腋窝淋巴结清扫的患者。
纳入298例诊断为浸润性乳腺癌的患者，这些患者在完成NAC后接受了前哨淋巴结活检，随后进行了乳房手术，并被分为训练集（n = 228）和测试集（n = 70）。采用单因素和多因素分析来选择在乳房手术前可确定且与训练集中非前哨淋巴结转移显著相关的因素。基于这些因素建立了逻辑回归模型，并在测试集中进行验证。
NAC前的淋巴结状态、NAC后的腋窝超声状态、前哨淋巴结数量和前哨淋巴结转移数量是NAC后乳腺癌非前哨淋巴结转移的独立预测因素。基于这些因素的预测模型在训练集中的曲线下面积为0.838（95%置信区间：0.774 - 0.902，P < 0.001）。当应用于测试集时，该模型的曲线下面积为0.808（95%置信区间：0.609 - 1.000，P = 0.003）。
开发了一种新的、简单的模型，该模型纳入了乳房手术前可确定的因素，以预测NAC后浸润性乳腺癌的非前哨淋巴结转移。尽管该模型在内部验证中表现出色，但在广泛应用于临床之前需要进行外部验证。
