Back to all papers

Evaluating the Accuracy of an Artificial Intelligence Solution Program for Identifying Anatomical Landmarks in Orthodontics.

October 31, 2026pubmed logopapers

Authors

Cho HR,Kim J,Kim JY

Affiliations (3)

  • Department of Medical Artificial Intelligence, Konyang University, Daejeon, Korea.
  • Department of Nursing, Woosuk University, Wanju, Korea.
  • Department of Biomedical Informatics, Konyang University College of Medicine, Daejeon, Korea. [email protected].

Abstract

This study quantitatively evaluated the accuracy and clinical reliability of EzOrtho, deep learning software designed to identify anatomical landmarks automatically for orthodontic diagnosis. Real-world clinical data were used to assess its applicability in practice. We analyzed 23 lateral cephalometric radiographs. EzOrtho automatically identified 36 anatomical landmarks per radiograph, yielding 828 data points for comparison with ground truth established by experienced orthodontists using a standardized annotation protocol. Accuracy was quantified using mean radial error (MRE), and clinical reliability was assessed using the successful detection rate (SDR), defined as the percentage of all points within the clinically acceptable threshold of 2.0 mm. The overall MRE was 0.78 mm, and the SDR at the 2.0 mm threshold was 86%. Although most landmarks were detected precisely, the landmark-specific analysis identified higher error rates at a few points in anatomically complex regions, indicating areas for improvement. The findings suggest that EzOrtho may offer clinical utility and reliability for automatic cephalometric landmark detection using real clinical data. Its performance indicates the potential to improve efficiency and consistency relative to manual tracing. However, the study's limitations, including its small sample size, mean that these results remain preliminary. Further validation is needed to establish its role as a decision-support tool in orthodontic practice.

Topics

Journal Article

Ready to Sharpen Your Edge?

Subscribe to join 11k+ peers who rely on RadAISlice. Get the essential weekly briefing that empowers you to navigate the future of radiology.

We respect your privacy. Unsubscribe at any time.