This study evaluated the reliability of artificial intelligence (AI)-based cephalometric measurements of overjet (OJ) and overbite (OB) and their agreement with manual assessment and investigated factors that may influence measurement outcomes. This retrospective cross-sectional study included 200 orthodontic patients (125 females, 75 males; mean age 20.12 years) selected from 440 records. Manual OJ and OB measurements were obtained twice from dental models using digital calipers and served as the reference standard. AI-based measurements were generated from lateral cephalometric radiographs using WebCeph™. Reliability was assessed using intraclass correlation coefficients (ICCs), while agreement between methods was evaluated using paired t-tests and Bland–Altman analysis. AI-based measurements showed significantly higher OJ and slightly lower OB values than manual measurements, with mean differences of 0.89 mm for OJ and −0.44 mm for OB (p 0.90). Bland–Altman analysis demonstrated that most measurement differences were within the 95% limits of agreement despite a small systematic bias. These findings suggest that AI-assisted cephalometric analysis provides a rapid and reliable approach for routine orthodontic assessment; however, individual discrepancies may be clinically relevant when high measurement precision is required. Therefore, AI should be considered a useful adjunct to conventional assessment, whereas manual measurements remain the reference standard when maximum measurement accuracy is essential.
Ganjali et al. (2026) studied this question.