本多中心研究旨在探讨基于术前乳腺癌增强钼靶摄影（CEM）中肿瘤内和周围区域的放射组学预测腋窝淋巴结（ALN）转移的可行性。共回顾性招募两个中心809例术前乳腺癌CEM图像患者。采用最小绝对值收缩和选择算子（LASSO）回归选择从肿瘤区域和周围5 mm和10 mm的区域中提取的CEM图像的放射组学特征，并构建放射组学签名。然后构建包括最佳放射组学签名和临床病理因素的评分卡。使用AUC评估评分卡的性能并与乳腺放射科医师进行直接比较。在内部测试集中，当周围区域增加时，周围瘤签名的AUC值下降，而签名tumor + 10mm表现最佳，AUC为0.712。包含signaturetumor + 10mm、肿瘤直径、孕酮受体（PR）、人表皮生长因子受体2（HER-2）和CEM报道的淋巴结状态的评分卡在内、外部测试集中最大的AUC分别为0.753和0.732。此外，评分卡优于放射科医师并改善了放射科医师的诊断性能。基于CEM肿瘤内和周围区域的评分卡可能提供一种非侵入性辅助工具，以指导乳腺癌ALN转移的治疗策略。版权所有©2023 Elsevier Inc.

This multicenter study aimed to explore the feasibility of radiomics based on intra- and peritumoral regions on preoperative breast cancer contrast-enhanced mammography (CEM) to predict axillary lymph node (ALN) metastasis.A total of 809 patients with preoperative breast cancer CEM images from two centers were retrospectively recruited. Least absolute shrinkage and selection operator (LASSO) regression was used to select radiomics features extracted from CEM images in regions of the tumor and peritumoral area of five and ten mm as well as construct radiomics signature. A nomogram, including the optimal radiomics signature and clinicopathological factors, was then constructed. Nomogram performance was evaluated using AUC and compared with breast radiologists directly.In the internal testing set, AUCs of peritumoral signatures decreased when the peritumoral area increased and signaturetumor + 10mm demonstrated the best performance with an AUC of 0.712. The nomogram incorporating signaturetumor + 10mm, tumor diameter, progesterone receptor (PR), human epidermal growth factor receptor 2 (HER-2), and CEM-reported lymph node status yielded maximum AUCs of 0.753 and 0.732 in internal and external testing sets, respectively. Moreover, the nomogram outperformed radiologists and improved diagnostic performance of radiologists.The nomogram based on CEM intra- and peritumoral regions may provide a noninvasive auxiliary tool to guide treatment strategy of ALN metastasis in breast cancer.Copyright © 2023. Published by Elsevier Inc.