在本篇综述中，我们评估了利用人工智能技术来降低局部和区域性乳腺癌治疗风险的潜力和最近的进展，特别关注新辅助系统治疗（NAST）后的手术治疗。 NAST的增加使用和功效使得具有病理完全反应(pCR)的患者的最佳局部和区域治疗成为临床相关的知识空白。有假设认为，具有pCR的患者没有受益于治疗性手术，因为NAST已经消灭了所有的肿瘤。然而，目前还不清楚如何可靠地排除NAST后的残余癌细胞，以确定适合忽略乳腺癌手术的患者。评估成像和微创活检排除残余癌细胞的临床试验的证据表明，存在高风险错过残余癌细胞。最近，基于人工智能算法显示了可靠地排除NAST后的残余癌细胞的良好结果。这个例子展示了基于人工智能算法进一步减少和个性化局部和区域性乳腺癌治疗的巨大潜力。版权所有©2023 Elsevier Ltd.出版。

In this review, we evaluate the potential and recent advancements in using artificial intelligence techniques to de-escalate loco-regional breast cancer therapy, with a special focus on surgical treatment after neoadjuvant systemic treatment (NAST). The increasing use and efficacy of NAST make the optimal loco-regional management of patients with pathologic complete response (pCR) a clinically relevant knowledge gap. It is hypothesized that patients with pCR do not benefit from therapeutic surgery because all tumor has already been eradicated by NAST. It is unclear, however, how residual cancer after NAST can be reliably excluded prior to surgery to identify patients eligible for omitting breast cancer surgery. Evidence from clinical trials evaluating the potential of imaging and minimally-invasive biopsies to exclude residual cancer suggests that there is a high risk of missing residual cancer. More recently, AI-based algorithms have shown promising results to reliably exclude residual cancer after NAST. This example illustrates the great potential of AI-based algorithms to further de-escalate and individualize loco-regional breast cancer treatment.Copyright © 2023. Published by Elsevier Ltd.