Abstract
Background:
Vascular and interventional radiology (VIR) procedures require high precision under fluoroscopic guidance, yet procedural inefficiencies persist due to anatomical complexity, navigation uncertainty, and operator-dependent variability.
Objective:
To review current applications of artificial intelligence (AI) in VIR and propose a conceptual framework for integrating AI across the procedural workflow.
Methods:
A structured narrative review was conducted using PubMed, Scopus, and IEEE Xplore databases (2018–2025). Keywords included “artificial intelligence,” “interventional radiology,” “embolization,” “navigation,” and “dose optimization.” Studies focusing on AI applications in procedural planning, intra-procedural guidance, and outcome prediction were included.
Results:
Current AI applications are largely task-specific, including bile-duct segmentation, tumor feeder detection, and radiation dose optimization. However, these systems operate in isolation and lack integration into a unified procedural workflow.
Conclusion:
We propose a clinically adaptive AI framework integrating pre-procedural planning, intra-procedural navigation, and post-procedural optimization. Such systems have the potential to enhance procedural efficiency, reduce complications, and standardize outcomes in VIR.