A Comparative Study of Three Different Methods for Cephalometric Analysis

Aim: The aim of the research was to compare and assess the accuracy and reliability of fully automatic (artificial intelligence [AI]), conventional, and digital cephalometric analysis. Materials and Methods: Fifty-two lateral cephalometric radiographs of patients aged 15–30 years with skeletal class I malocclusion were analyzed using three methods, i.e., manual tracing, digital tracing with NemoCeph NX 2009 software, and fully automatic tracing using Ceph.io software. Out of which, manual tracings were repeated after 4 weeks to assess intraexaminer reliability. Results: Almost all the variable linear parameters have showed statistically significant difference between the two different comparisons except for lower anterior facial height who showed statistically nonsignificant difference. The fully automatic cephalometric analysis demonstrated higher accuracy and reliability in landmark identification and measurements compared to the conventional digital and manual tracing methods and the digital tracing method, while more accurate than manual tracing. Conclusion: More studies need to be conducted with regard to AI methods as it will bring significant changes in diagnosis, measurement procedures, and formulating a definitive proper treatment planning.