Abstract
Breast cancer (BC), specifically HER2-positives subtype, has a poor prognosis. Nevertheless, the development of anti-HER2 therapy yielded satisfactory outcomes. Therefore, evaluating patient HER2 status and ascertaining responsiveness to anti-HER2 therapy is crucial. The advent of deep learning has propelled the artificial intelligence (AI) revolution, leading to an increased applicability of AI in predictive models. In the field of medicine, AI is an emerging modality that is gaining momentum for facilitating cancer diagnosis and treatment, particularly in the effective management of breast cancer. This study aims to provide a comprehensive review of current diagnostic and predictive models that utilize data obtained from histopathological slides, radiomics, and HER2 binding sites. Advancements and practical applications of these models were also evaluated. Additionally, we examined existing obstacles that AI encounters for anti-HER2 therapy. We also proposed future directions for integrating AI in assessing and managing anti-HER2 therapy. The findings of this study offer valuable insights into the evaluation of AI-based anti-HER2 therapy, emphasizing key concepts and obstacles that, if addressed, could facilitate the integration of AI-assisted anti-HER2 therapy. The integration of AI has the potential to enhance the precision and customization of screening and treatment protocols for HER2+ breast cancer.