Comparison of computer aided cephalometric analysis methods in which cephalometric reference points are determined by physician and artificial intelligence in patients with cleft lip and palate

Aim: The aim of this study is to compare the measurements of cleft lip and palate patients performed by artificial intelligence supported automatic cephalometric analysis tool OrthoDx and manual landmark identification based Dophin software and thereby analyze the applicability of OrthoDx for cleft lip and palate patients. Material and Method: The lateral cephalometric radiographs of 188 cleft lip and palate patients with a mean age of 9,176±4,726 years selected from the archives of Hacettepe University Faculty of Dentistry Department of Orthodontics were included in the study. 20 angular and 8 linear cephalometric parameters were used. The measurements made by the researcher using the Dolphin program were accepted as reference measurements and were compared with the artificial intelligence supported OrthoDx program measurements with and without intervention. Results: As a result of the comparison, the measurements made by the OrthoDx program; Intraclass correlation coefficient was observed to be low in the values of Face Depth, Face Axis, Facial Taper. ANB angle, FMA angle, Mandibular plane angle and L1-NB distance values were very high in agreement and no statistically significant difference was found. Conclusion: According to our analysis, in the current version of the artificial intelligence-based OrthoDx program, it is a promising tool for cephalometric landmark identification in patients with cleft lip and palate, provided that it is used under the supervision of an orthodontist .