Artificial intelligence-assisted analysis of musculoskeletal imaging-A narrative review of the current state of machine learning models
Artikel i vetenskaplig tidskrift, 2025

Artificial intelligence (AI) has emerged as a powerful tool in medical imaging, including musculoskeletal radiology [44, 49]. By leveraging advanced algorithms and big data, AI enables the automation of image analysis, offering the potential to improve diagnostic accuracy, reproducibility and efficiency [9, 15, 24]. This is especially valuable in improving patient flow and reducing turnover times in musculoskeletal imaging, where the interpretation of radiographs, computed tomography (CT), and magnetic resonance imaging (MRI) scans, can be intricate and time-consuming.

Considering increasing imaging volumes and a relative shortage of subspecialized musculoskeletal radiologists, AI offers the potential for a scalable solution, democratising accurate diagnosis and therefore quality of care. By triaging imaging examinations and assisting with image interpretation, AI-based tools furthermore reduce reporting times. The integration of AI into clinical practice is not intended to replace radiologists but to augment their capabilities, allowing them to manage increasing workloads effectively [12].

While AI has demonstrated significant potential in musculoskeletal imaging, its clinical adoption requires validation, ethical oversight, and consideration of its limitations. This review aims to explore these aspects to provide a balanced perspective on the future of AI in musculoskeletal radiology including exploration of the types of AI models utilised, ethical considerations, technical challenges, and current solutions.

deep learning

AI in musculoskeletal imaging

clinical integration

image analysis

computer vision

Författare

Felix C. Oettl

Universität Zürich

Hosp Special Surg

Balint Zsidai

Göteborgs universitet

Jacob F. Oeding

Göteborgs universitet

Michael T. Hirschmann

Universität Basel

Kantonsspital Baselland

Robert Feldt

Chalmers, Data- och informationsteknik, Software Engineering

David Fendrich

Tenfifty AB

Matthew J. Kraeutler

Göteborgs universitet

Philipp W. Winkler

Johannes Kepler Univ Linz

Pawel Szaro

Göteborgs universitet

Kristian Samuelsson

Göteborgs universitet

Knee Surgery, Sports Traumatology, Arthroscopy

0942-2056 (ISSN) 1433-7347 (eISSN)

Vol. In Press

Ämneskategorier (SSIF 2025)

Radiologi och bildbehandling

DOI

10.1002/ksa.12702

PubMed

40450562

Mer information

Senast uppdaterat

2025-06-16