Comparison of cephalometric analyses with artificial intelligence algorithm with digital two and three dimensional cephalometric analyses in orthognatic surgery patients

The purpose of this prospective study is to compare fully automated cephalometric analyses based on artificial intelligence technology from 2D images obtained from cephalometric images taken before surgery in patients requiring bimaxillary orthognathic surgery with computer-aided digital cephalometric analyses and cephalometric measurements of 3D CT reconstructions and to evaluate the adequacy of AI-based cephalometric analyses. Fifty patients who applied to the Department of Orthodontics at Bezmialem Vakıf University Faculty of Dentistry and were treated with orthognathic surgery based on the results of orthodontic analyses were included in this study. The CT images of 23 male and 27 female patients were included. The average age of the male patients was 25.30 ± 5.83, and the average age of the female patients was 23.96 ± 5.61. The CT scans taken from the patients were performed by a single technician in accordance with the standards at the Department of Radiology, Bezmialem Vakıf University Faculty of Medicine. Anatomical points in the CT images of the patients were recorded and marked in the NemoCeph 3D (SurgeryFAB) program, which reconstructs 3D images into 2D and cephalometric images. These points were also automatically marked using artificial intelligence in the Webceph, OrthoDx, and Audox programs. The results obtained in our study showed that orthodontists could use the linear and proportional measurements obtained from 3D CT in a manner similar to the methods used in the analysis of 2D cephalogram images created by the superimposition of CT images obtained with parallel rays that do not cause magnification. It was observed that fully automated cephalometric analyses made erroneous cephalometric measurements in orthognathic surgery patients.