Clinical utility of artificial intelligence models in radiology: a systemic scoping review of diagnostic and endovascular applications.
Authors
Affiliations (3)
Affiliations (3)
- University of Texas Health Sciences Center at Houston, Houston, TX, USA.
- University of Missouri Kansas City School of Medicine, Kansas City, MO, USA.
- Division of Vascular and Interventional Radiology, Department of Radiology, The Ohio State University Wexner Medicine Center, 395 W 12Th Ave, 4Th Floor Faculty Office Tower, Columbus, OH, 43210, USA. [email protected].
Abstract
To systematically scope the clinical integration of artificial intelligence (AI) in diagnostic and interventional radiology. This integration encompasses various components of AI forms such as deep learning, convolutional neural networks, natural language processing, and machine learning. A Preferred Reporting Items for Systematic Reviews and Meta-Analysis Extension for Scoping Reviews (PRISMA-ScR) was employed to evaluate current primary and translation literature on the utility of AI in diagnostic and interventional radiology in broad disease categories. Following the review for inclusion criteria, a total of 23 peer-reviewed research articles were selected for review. Notably, most studies were found to focus on diagnostic and interventional radiology and oncologic diseases, including lung, hepatocellular, colorectal, prostate, pancreatic, breast, and blood cancers. Radiologists have an advantageous role with the integration of these tools in clinical practice. This may include disease prediction models, catheter navigation, and image reconstruction. Utilization of these AI tools can help improve and further expose of the capabilities of diagnostic and interventional radiology to patients worldwide. From a disease standpoint, this review found most of the clinical literature has implemented AI tools for diagnostic and interventional radiology in oncology, followed by vascular diseases. Careful navigation is necessary to address the current logistical challenges, educational demands, and ethical dilemmas to ensure the safe and effective incorporation of these technologies into clinical radiologic settings.