Artificial Intelligence in Vascular and Endovascular Surgery: From Prediction to Surgeon-Led Clinical Implementation

Artificial intelligence (AI) has shown promising performance in vascular imaging, risk prediction, procedural support, surveillance, and education, but its translation into routine surgical care remains limited. This narrative review addresses a focused question: what evidence supports the transition of vascular AI from technical prediction to surgeon-led clinical implementation? A targeted PubMed search through July 24, 2026 identified international and Japan-based studies across the perioperative pathway. Deep-learning systems can segment aortic anatomy, characterize aneurysm or dissection, and support post-EVAR imaging. Machine-learning models may stratify mortality and treatment risk, whereas computer vision, image fusion, wearable monitoring, and large language models extend potential use into procedures, follow-up, and education. Japanese reports demonstrate locally relevant innovation in aortic imaging, EVAR prognosis, postoperative monitoring, and specialty education; however, most evidence remains retrospective, single-center, or developmental. Selected surgeon-led workflows, including quantitative spinal cord imaging after aortic repair, are presented with explicit data provenance as illustrative demonstrations, not proof of clinical efficacy. Progress requires external validation, workflow and human-factors evaluation, prospective impact studies, transparent reporting, and post-deployment monitoring. AI should augment vascular specialists, with surgeons retaining responsibility for clinical questions, verification, interpretation, and patient care.