Fully Automatic System for Accurate Localisation and Analysis of Cephalometric Landmarks in Lateral Cephalograms

Abstract

Cephalometric tracing is a standard analysis tool for orthodontic diagnosis and treatment planning. The aim of this study was to develop and validate a fully automatic landmark annotation (FALA) system for finding cephalometric landmarks in lateral cephalograms and its application to the classification of skeletal malformations. Digital cephalograms of 400 subjects (age range: 7-76 years) were available. All cephalograms had been manually traced by two experienced orthodontists with 19 cephalometric landmarks, and eight clinical parameters had been calculated for each subject. A FALA system to locate the 19 landmarks in lateral cephalograms was developed. The system was evaluated via comparison to the manual tracings, and the automatically located landmarks were used for classification of the clinical parameters. The system achieved an average point-to-point error of 1.2 mm, and 84.7% of landmarks were located within the clinically accepted precision range of 2.0 mm. The automatic landmark localisation performance was within the inter-observer variability between two clinical experts. The automatic classification achieved an average classification accuracy of 83.4% which was comparable to an experienced orthodontist. The FALA system rapidly and accurately locates and analyses cephalometric landmarks in lateral cephalograms, and has the potential to significantly improve the clinical work flow in orthodontic treatment.

Figure 1. Cephalogram annotation example showing the 19 landmark positions used in this study.

A description of all landmarks is given in Supplementary Table S2.

Figure 2. Schematic overview of the FALA system as described in refs and , applied to lateral cephalograms.

Figure 3. Cumulative distribution of image-specific point-to-point results for parameter optimisation of the FALA system: (left) comparing different combinations of patch size (ps), training range (tr) and search range (sr); and (right) showing the difference in performance between the original and the improved parameter settings.

Error bars show the 95% confidence interval across the four cross-validation experiments.

Figure 4. Cumulative distribution of image-specific point-to-point results of the FALA system tested on 400 images: (left) demonstrating the impact of the quality of the training data on performance; and (right) demonstrating that the runtime-improved system does not lead to any decrease in performance.

Error bars show the 95% confidence interval across the four cross-validation experiments.

Figure 5. Fully automatic annotation results of the runtime-improved doctor2-trained FALA system (sorted by PEI percentiles): (left) 50%ile/median: PEI = 1.1 mm; (middle) 95%ile: PEI = 1.8 mm; and (right) 99%ile: PEI = 2.4 mm.