AI Guided Operative Planning and Navigation in Vascular Surgery: A Systematic Review
Keywords:
Artificial intelligence; Vascular surgery; Endovascular navigation; Operative planning; Reinforcement learning; Deep learning; Endovascular aneurysm repair (EVAR); Surgical augmented intelligence; Autonomous catheter navigation; Hybrid operating room.Abstract
Background: Artificial intelligence (AI) is rapidly advancing in vascular surgery, with applications in operative planning and navigation. AI‑based systems may enhance preoperative mapping, intraoperative guidance, and autonomous device control, potentially improving procedural precision and patient outcomes.
Objectives: To systematically review current evidence on AI‑enabled operative planning and navigation in vascular surgery, assess technology readiness, and identify future research priorities.
Methods: A systematic search of PubMed, IEEE Xplore, and arXiv (to June 2025) identified studies evaluating AI in planning or navigation during vascular or endovascular interventions. Eligible studies included deep learning, machine learning, or reinforcement learning approaches validated in simulation, phantom, or clinical environments. Data were synthesized narratively and grouped into planning/augmented guidance versus autonomous navigation systems.
Results: Twenty-four studies met inclusion. AI‑augmented planning tools, particularly deep learning–based overlays for endovascular aneurysm repair (EVAR), demonstrated reductions in fluoroscopy time, contrast use, and radiation exposure in early clinical studies. Autonomous navigation systems using reinforcement learning achieved >95% success in simulated catheter and guidewire navigation but lacked patient-level validation. Technology readiness levels remain low (TRL≈3 for autonomous navigation).
Conclusions: AI in operative planning shows promising clinical translation, especially in EVAR, while autonomous navigation is largely experimental. Future research should focus on multicentre validation, semi‑autonomous human–machine collaboration, and regulatory/ethical frameworks to ensure safe integration into clinical workflows.
