Artificial intelligence in vascular surgery demonstrates strong performance in diagnostic imaging, with deep learning algorithms achieving sensitivities exceeding 85% and AUC values above 0.90.
AI demonstrates strong potential in vascular surgery for diagnostic imaging, risk prediction, and intraoperative guidance, though widespread adoption requires robust prospective validation and transparent model development.
Artificial intelligence (AI) is rapidly reshaping modern medicine, with expanding applications in vascular surgery ranging from diagnostic imaging and risk prediction to intraoperative guidance and postoperative surveillance. Despite the growing volume of research in this field, real-world clinical implementation remains limited, highlighting the need to critically evaluate current evidence, identify existing limitations, and outline future directions for successful integration into clinical practice. This narrative review summarizes recent evidence from PubMed, Scopus, and Web of Science, focusing on clinically relevant applications of machine learning and deep learning in vascular surgery. Current evidence demonstrates strong performance across multiple domains. In diagnostic imaging, deep learning algorithms have achieved high accuracy in detecting aneurysms, grading carotid plaque characteristics, and identifying peripheral arterial disease, with several studies reporting diagnostic sensitivities and specificities exceeding 85% and predictive models achieving area-under-the-curve values above 0.90. Predictive analytics have also shown promise for perioperative risk stratification and long-term outcome prediction. Intraoperative applications, including augmented reality and robotic assistance, are emerging as valuable tools to improve surgical precision and streamline workflows. Postoperative surveillance systems supported by AI enable automated detection of complications such as endoleaks and restenosis, facilitating earlier intervention and more personalized follow-up strategies. However, significant barriers continue to hinder widespread adoption, including concerns regarding dataset heterogeneity, external validation, algorithm transparency, ethical considerations, and integration into routine clinical workflows. Overall, AI has the potential to substantially improve surgical accuracy, efficiency, and patient outcomes in vascular surgery, but future progress will require robust prospective validation studies, transparent model development, and clinician-centered implementation strategies.
Bozzani et al. (Thu,) conducted a review in Vascular surgery. Artificial intelligence was evaluated. Artificial intelligence in vascular surgery demonstrates strong performance in diagnostic imaging, with deep learning algorithms achieving sensitivities exceeding 85% and AUC values above 0.90.