Gao Lu-Ying, Ran Hai-Tao, Deng You-Bin, Luo Bao-Ming, Zhou Ping, Chen Wu, Zhang Yu-Hong, Li Jian-Chu, Wang Hong-Yan, Jiang Yu-Xin
Department of Ultrasound, State Key Laboratory of Complex Severe and Rare Diseases, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Department of Ultrasound, The Second Affiliated Hospital of Chongqing Medical University, Chongqing, China.
Asia Pac J Clin Oncol. 2023 Apr;19(2):e71-e79. doi: 10.1111/ajco.13781. Epub 2022 May 20.
We aim to assess the performance of the Gail model and the fifth edition of ultrasound BI-RADS (Breast Imaging Reporting and Data System) in breast cancer for predicting axillary lymph node metastasis (ALNM).
We prospectively studied 958 female patients with breast cancer between 2018 and 2019 from 35 hospitals in China. Based on B-mode, color Doppler, and elastography, radiologists classified the degree of suspicion based on the fifth edition of BI-RADS. Individual breast cancer risk was assessed with the Gail model. The association between the US BI-RADS category and the Gail model in terms of ALNM was analyzed.
We found that US BI-RADS category was significantly and independently associated with ALNM (P < 0.001). The sensitivity, specificity, and accuracy of BI-RADS category 5 for predicting ALNM were 63.6%, 71.6%, and 68.6%, respectively. Combining the Gail model with the BI-RADS category showed a significantly higher sensitivity than using the BI-RADS category alone (67.8% vs. 63.6%, P < 0.001). The diagnostic accuracy of the BI-RADS category combined with the Gail model was better than that of the Gail model alone (area under the curve: 0.71 vs. 0.50, P < 0.001).
Based on the conventional ultrasound and elastography, the fifth edition of ultrasound BI-RADS category could be used to predict the ALNM of breast cancer. ALNM was likely to occur in patients with BI-RADS category 5. The Gail model could improve the diagnostic sensitivity of the US BI-RADS category for predicting ALNM in breast cancer.
我们旨在评估盖尔模型和超声乳腺影像报告和数据系统（BI-RADS）第五版在预测乳腺癌腋窝淋巴结转移（ALNM）方面的性能。
我们对2018年至2019年间来自中国35家医院的958例女性乳腺癌患者进行了前瞻性研究。放射科医生基于B超、彩色多普勒和弹性成像，根据BI-RADS第五版对可疑程度进行分类。使用盖尔模型评估个体患乳腺癌的风险。分析了美国BI-RADS分类与盖尔模型在ALNM方面的关联。
我们发现美国BI-RADS分类与ALNM显著且独立相关（P<0.001）。BI-RADS 5类预测ALNM的敏感性、特异性和准确性分别为63.6%、71.6%和68.6%。将盖尔模型与BI-RADS分类相结合显示出比单独使用BI-RADS分类更高的敏感性（67.8%对63.6%，P<0.001）。BI-RADS分类与盖尔模型相结合的诊断准确性优于单独使用盖尔模型（曲线下面积：0.71对0.50，P<0.001）。
基于传统超声和弹性成像，超声BI-RADS第五版分类可用于预测乳腺癌的ALNM。BI-RADS 5类患者可能发生ALNM。盖尔模型可提高美国BI-RADS分类预测乳腺癌ALNM的诊断敏感性。
