The Operative Role of Artificial Intelligence in Vascular Surgery: A Systematic Review of Literature

Abstract

This study aims to systematically review existing literature on the impact of artificial intelligence (AI) on operative workflow and safety in vascular surgery. This systematic review was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines and registered with the International Prospective Register of Systematic Reviews (CRD420251004635). A comprehensive literature search was conducted across PubMed, Embase (via Ovid), Scopus, Google Scholar, and Science Direct for studies published between January 2005 and February 2025. The search strategy used for PubMed, Embase (via Ovid), Scopus, and Google Scholar was: ("artificial intelligence" OR "AI" OR "machine learning" OR "deep learning" OR "neural network") AND ("vascular surgery" OR "endovascular" OR "EVAR" OR "TEVAR" OR "aneurysm repair" OR "carotid endarterectomy" OR "arteriovenous fistula" OR "bypass" OR "inferior vena cava") AND ("segmentation" OR "prediction" OR "risk model" OR "imaging" OR "perioperative" OR "intraoperative" OR "operative planning" OR "workflow" OR "patient safety). The strategy for Science Direct was: ("artificial intelligence" OR "machine learning" OR "deep learning") AND ("vascular surgery" OR "endovascular") AND ("operative role" OR "workflow" OR "patient safety"), as limited to eight words. The initial search identified 817 studies, and after screening, studies that did not meet the criteria were excluded, leaving eight relevant studies. Six were retrospective studies, one prospective study and one hybrid (retrospective-prospective cohort) study. The exclusion criteria included irrelevant titles, duplicate papers, abstracts, themes and non-English papers. The review demonstrated predominant utilisation of AI in the preoperative phase for risk prediction, decision support, anatomical assessment, and operative planning. Intraoperatively, AI applications encompassed real-time risk updates and intraoperative guidance based on preoperative computed tomography, while postoperative models enhanced surveillance following endovascular procedures. Across the reviewed studies, conventional methods were often outperformed by AI models in predictive accuracy, workflow efficiency and safety. AI application shows potential in improving operative workflow and safety in vascular surgery through enhanced decision support, risk prediction, and process automation. However the performance should be interpreted cautiously given current evidence limitations. Overall, the evidence remains limited by small, retrospective, and heterogeneous studies with potential bias, highlighting the need for large-scale prospective validation before routine clinical adoption.

Ähnliche Arbeiten