Integrating Explainable AI: Breakthroughs in Medical Diagnosis and Surgery

Ana Henriques, Henrique Parola, Raquel Gonçalves, Manuel Rodrigues

Abstract: The challenge of explaining the results generated by artificial intelligence (AI) is a significant obstacle to their widespread acceptance, which is why increased attention has been paid to the explainability in AI (XAI) in recent years. Given its impact on the medical sector, the survey seeks to demonstrate the important role of XAI in ensuring the reliability and accountability of AI in the domains of diagnosis and surgery. Therefore, we conduct an in-depth look at the applications and challenges of XAI in these areas by reviewing articles published between 2022 and 2023. The survey aims to explore the categorization of XAI techniques, establish their taxonomy, address trade-offs between model performance and interpretability and emphasize the importance of achieving a balance in practical applications. The findings of this study confirm the potential of XAI in medicine as a promising avenue for exploration, providing guidance for the development of medical XAI applications.