随着设备的进步以及技术的不断发展近年来新兴的对比增强乳腺X线摄影contrast-enhanced mammography CEM技术是将FFDM与对比增强相结合利用碘的K边缘效应通过静脉注射对比剂后快速进行高、低能双能曝光再经后处理后获得低能图和减影图的一种新兴的乳腺成像技术10。CEM在低能图上提供了与FFDM类似的病灶的形态学信息及钙化细节显示同时也可以在减影图上消除乳腺腺体的重叠效应对病灶

With the advancement of technology and equipment, contrast-enhanced mammography (CEM) has emerged as a new breast imaging technique that combines full-field digital mammography (FFDM) with contrast-enhancement. By using the K-edge effect of iodine and performing fast high- and low-energy double exposure after intravenous injection of contrast agent, CEM can obtain low-energy and subtracted images through post-processing, providing similar morphological and calcification details as FFDM on the low-energy image, while eliminating the overlapping effect of breast gland on the subtracted image, making the determination of lesion size more similar to pathological findings. In this article, we propose a novel network XXUNet for breast CEM image in the cranio-caudal (CC) and mediolateral oblique (MLO) views. We designed a CNN-Transformer Hybrid as Encoder to simultaneously obtain long-distance and local information on the CEM image. As shown in Figure 1, our XXUNet can not only extract the complete feature map of a single view, but also interact with the information of the same lesion from different views, making the segmentation of a specific lesion more clinically significant and efficient. As shown in Figure 1, each patient's chest is imaged from two different angles, CC and MLO views, respectively. The complementary information from the two views can help doctors understand the 3D shape of the lesion more comprehensively, as the lesion itself is 3D and doctors need to view both views multiple times during clinical practice