How reliable are AI-Assisted cephalometric programs in assessing measurements involving bilateral landmarks?

Abstract Objective Recent advancements in technology have promoted artificial intelligence (AI) to automatically detect landmarks on lateral cephalograms. This study aimed to evaluate the reliability of an AI-assisted cephalometric analysis program (Webceph, Gyeonggi-do, Republic of Korea) in the assessment of cephalometric measurements involving bilateral landmarks. Materials and methods Fifty-one high-quality cephalograms were used and inclusion/exclusion criteria were applied. Two researchers manually traced the cephalograms after which an AI-assisted cephalometric analysis program (Webceph) was applied. Both intra-and inter-operator reliability were tested. Independent Sample t-tests were used to compare the means of measurements. Results The inter- and intra-class correlation coefficients were 0.80 which indicated ‘good’ reliability. Statistically significant differences were found in the gonial angle and effective mandibular length measurements ( p <0.05), but not in the articular angle and FMA angle ( p >0.05). The results suggest that, while AI-assisted programs provide reliable measurements, differences in certain measurements may be attributed to inherent AI algorithm limitations. Clinicians should verify and, if needed, correct bilateral landmark locations after the initial AI digitisation. Conclusion AI-driven cephalometric analysis holds promise for improving diagnostic efficiency and precision. However, limitations and further AI advancements require consideration to ensure appropriate clinical use.