乳腺癌是当前危害女性健康最主要的恶性肿瘤，近年来发病率不断攀升。近年来，依托于多组学信息的肿瘤精准诊疗模式的创立，为乳腺癌诊疗效果的进一步提升带来了新的希望。本项目拟利用人工智能技术，基于多组学，多模态的乳腺癌影像、病理、基因等数据，深度挖掘影像和病理图像的信息，筛选乳腺癌关键组学靶点，开发人工智能辅助肿瘤诊疗的相关技术，建立人工智能辅助的精准临床诊断系统和病理诊断系统、精准风险评估和疗效评估系统，从而实现肿瘤病理、风险和疗效的可视化。通过项目建设，建立乳腺癌精准诊疗决策新理论、新模式、新技术、新标准，并开展相应的临床应用，从而进一步提高乳腺癌早期诊断率，提高治疗效果，减少不必要的治疗毒副反应和医疗资源耗费，同时奠定人工智能赋能医疗的技术基础，促进相关产业发展。

Breast cancer is the most important cancer endangering women's health, and its incidence has been rising in recent years. In recent years, the establishment of a precise diagnosis and treatment model for breast cancer based on multi-omics data has brought new hope for the improvement of the diagnosis and treatment effect of breast cancer. This project aims to screening key histological targets of breast cancer, developing the relevant technology of artificial intelligence (AI)-assisted cancer diagnosis and treatment, establishing AI-assisted precise clinical and pathological diagnosis system, and precise risk assessment and treatment, using multi-omics, multi-modal breast cancer data (imaging, pathology and bioinformatics, etc.) and deep learning algorithms and other data, in-depth mining of image and pathological image information. Effectiveness evaluation system is to be developed to achieve the visualization of tumor pathology, risk and treatment efficacy. This project will develop AI-assisted systems based on distributed storage and high-performance computing framework, constructing an accurate diagnosis and treatment decision system and decision model based on online learning algorithms for incremental multimodal data, and launching clinical trials to validate and verify the systems. Based on the new technologies and systems stated above, new theories and models will be proposed and novel technologies and standards for precision diagnosis and treatment of breast cancer will be set up, so as to improve the early diagnosis of breast cancer, improve the treatment effect, reduce side effects and medical resource consumption, lay the technical foundation of AI-assisted medical treatment.