Artificial-Intelligence-Based Cephalometric Landmark Detection in Lateral Cephalograms.
Authors
Affiliations (4)
Affiliations (4)
- Department of Orthodontics, School of Dentistry, Kanagawa Dental University, Yokosuka 238-8580, Japan.
- Life Science Center for Survival Dynamics, Tsukuba Advanced Research Alliance, 1-1-1 Tennodai, Tsukuba 305-8577, Japan.
- Laboratory of Data-Driven Biology, Department of Integrative Cellular Informatics, Center for Neurological Diseases and Cancer, Nagoya University Graduate School of Medicine, Tsurumai 65, Showa-ku, Nagoya 466-8550, Japan.
- EDIAND Inc., Tokyo 135-0062, Japan.
Abstract
<b>Background/Objectives:</b> Accurate landmark identification underpins reliable cephalometric analysis. This study evaluated a ResNet50-based, single-stage regression convolutional neural network for direct automatic localization of 15 landmarks on lateral cephalograms. <b>Methods</b>: This retrospective study included 669 lateral cephalograms (669 patients): 619 for training and 50 randomly selected for an internal holdout test set. One of five orthodontists annotated each cephalogram, and coordinates served as the reference standard. One orthodontist independently re-annotated all test images after more than 2 weeks to assess reproducibility. Model performance was evaluated using Euclidean localization errors and success detection rates (SDR). <b>Results</b>: Across 750 landmark predictions, mean localization error was 1.25 ± 1.39 mm (95% confidence interval, 1.15-1.35 mm) and median error was 0.72 mm. SDRs within 1.0, 2.0, and 4.0 mm were 64.5%, 81.5%, and 94.4%, respectively. A statistically significant overall difference was observed among the 15 landmarks (Friedman χ<sup>2</sup>(14) = 28.84, <i>p</i> = 0.011). Point A had the numerically largest mean error (1.66 mm) and the mandibular central incisor the smallest (1.04 mm). The mean difference between original and repeated annotations was 0.52 ± 0.92 mm. <b>Conclusions</b>: In this single-center internal holdout test set, mean localization error was below the 2.0 mm benchmark. However, 18.5% of predictions exceeded 2.0 mm, and external validity remains unconfirmed. Artificial-intelligence-generated landmarks should be verified by orthodontists, and external validation is required before broader clinical use.