Determination of the accuracy of semi-automatic and fully automatic 2d lateral cephalometric analysis programs

Abstract

AIM: To evaluate the accuracy of current semi-automatic and fully automatic 2D lateral cephalometric analysis programs. MATERIALS AND METHODS: 60 lateral cephalometric radiographs were randomly selected and grouped based their skeletal malocclusions to form 3 equal groups of 20 Class I, 20 Class II and 20 Class III. These radiographs were then analyzed via traditional hand-based analysis. The values obtained from this method of analysis were compared to 4 subsequent methods of analysis. These consisted of semi-automatic analysis using Dolphin Imaging software, semi-automatic analysis using Kodak Orthodontic Imaging software, fully automatic analysis using Kodak Orthodontic Imaging software and fully automatic analysis combined with limited landmark changes using Kodak Orthodontic Imaging software. RESULTS: ICC tests were completed to compare the gold standard hand-based analysis to the 4 subsequent methods. The values obtained from semi-automatic Dolphin and Kodak Orthodontic Imaging software were found to be comparable to hand-based analysis. Whereas, the values obtained from the fully automatic mode of Kodak Orthodontic Imaging software were not found to be comparable to hand-based analysis. CONCLUSIONS: Digital cephalometric programs can be used as an accurate method when performing lateral cephalometric analyses. The fully automatic mode of these programs should only be used as a support to diagnosis and not as a diagnostic tool.