Unveiling the Algorithm: The Role of Explainable Artificial Intelligence in Modern Surgery

Area of Science:

• Surgical innovation
• Medical Artificial Intelligence
• Explainable AI

Background:

• Artificial Intelligence (AI) is transforming surgical care through enhanced diagnosis, prediction, and decision-making.
• Challenges include AI model opacity, bias, and data sharing, hindering clinical trust and regulatory approval.
• Explainable Artificial Intelligence (XAI) is emerging to address these gaps, facilitating AI adoption in data-driven surgery.

Purpose of the Study:

• To review current applications of XAI in surgical contexts.
• To highlight the absence of XAI in Generative AI tools like ChatGPT.
• To emphasize XAI as a prerequisite for responsible surgical AI innovation.

Main Methods:

• Literature review of XAI applications in surgery.
• Analysis of XAI's role in preoperative assessment, surgical planning, intraoperative guidance, and postoperative monitoring.
• Identification of challenges including model bias, overfitting, and user interface design.

Main Results:

• XAI applications are reviewed across the surgical workflow.
• Generative AI tools currently lack XAI mechanisms.
• Key challenges for XAI integration include bias, overfitting, and UI design.

Conclusions:

• XAI is essential for surgeons to interpret, validate, and trust AI tools in surgery.
• Responsible innovation in surgical AI necessitates XAI implementation.
• Overcoming challenges in bias, overfitting, and design is key for integrating XAI into surgical practice for transparent and human-centered AI.