Mishra, Arya Suman, Meacher, Hannah, Khan, Muhammad Murtaza, Gaddoura, Zaina G, Hough, Matthew, Saeed, Abu, Heaver, Catriona, Bing, Andrew, Marquis, Chris, Hill, Simon, Makwana, Nilesh, Kyriacou, Theocharis ORCID: https://orcid.org/0000-0002-5211-3686 and Cool, Paul (2026) Artificial Intelligence to Evaluate Radiographic Outcome following Hallux Valgus Surgery. Journal of Orthopedic Research and Therapy, 11 (4).

Preview

Text mishra_etal_2026.pdf - Published Version Available under License Creative Commons Attribution. Download (1MB) | Preview

Abstract

Purpose: Angular measurements in Hallux valgus are commonly used, but the methods described vary, particularly for postoperative films. This study aims to assess this variation, evaluate a new artificial-intelligence-based computer method, and determine whether it can serve as an objective outcome measure. Methods: Consecutive patients who had surgery for hallux valgus over six months were included. We collected information on age, gender, and type of surgery. Hallux valgus (HV), intermetatarsal (IM) and distal metatarsal articular (DMA) angles were measured on pre- and post-operative radiographs by seven observers of varying degrees of experience. We also read the images into the computer, and after users identified five bony landmarks, we segmented them with Segment Anything (SAM)® software. We then used the first eigenvector to define the long axis of each bone and calculated angles. Results: In both pre-operative and post-operative radiographs, the computer method showed less variation than human observers for all angles. For post-operative radiographs, the ICC was rated excellent for the HV angle and good for the IM angle. The computer method was considerably faster than human observers. Conclusion: The described computer method is reproducible with less variation and bias than human observers. In addition, it can be used on postoperative radiographs as an objective outcome measure.

Keywords: Hallux valgus, Angle measurement, Observer error, Artificial intelligence, Outcome measure.

Item Type:	Article
Status:	Published
DOI:	10.29011/2575-8241.001429
Subjects:	R Medicine > R Medicine (General)
School/Department:	York Business School
Institutes:	Institute for Health and Care Improvement
URI:	https://ray.yorksj.ac.uk/id/eprint/15924

University Staff: Request a correction | RaY Editors: Update this record