Abstract
AIM: To map and critically appraise the current literature on Artificial Intelligence (AI) applications in emergency general surgery, with a focus on clinical decision-support tools for preoperative risk stratification and intraoperative assistance, and to identify ethical, structural, and regulatory barriers to implementation.
METHODS: A scoping review was conducted within the ARIES project, following established methodological frameworks. Relevant studies evaluating AI-based tools in emergency surgical settings were systematically identified and analyzed.
RESULTS: The literature describes AI applications mainly in two domains: preoperative decision support, including risk prediction and diagnostic or triage models for acute abdominal and traumatic conditions, and intraoperative assistance, largely focused on computer vision-based systems for anatomical recognition, safety guidance, and navigation in minimally invasive emergency procedures. Additional contributions address training and telementoring platforms, as well as cross-cutting ethical, legal, and regulatory considerations relevant to AI adoption in emergency surgical care.
CONCLUSIONS: AI has the potential to complement emergency surgeons' clinical judgment, but its routine adoption in emergency surgical practice remains limited. Addressing methodological, ethical, and regulatory challenges, together with the development of robust data infrastructures and targeted training pathways, is essential to support safe, effective, and equitable implementation in acute care settings. In addition, the lack of dedicated investment and sustainable funding models for large-scale clinical implementation and prospective evaluation represents a critical barrier to the translation of AI from research into routine emergency surgical practice.