Comparative Evaluation of Craniofacial Growth Patterns Using AI-Driven Longitudinal Cephalometric Superimpositions in Adolescents Article Swipe
YOU?
·
· 2025
· Open Access
·
· DOI: https://doi.org/10.4103/jpbs.jpbs_1346_25
· OA: W4417451143
Background: Accurate analysis of craniofacial growth is essential for individualized orthodontic treatment planning. Traditional cephalometric methods often rely on manual tracing, which can be time-consuming and subject to interobserver variability. Recent advances in artificial intelligence (AI) offer promising tools for precise and reproducible assessment of growth changes. Methods: A longitudinal cohort of 90 adolescents (aged between 11 and 16 years) was analyzed, comprising 30 each from horizontal, vertical, and average growth pattern groups, determined by initial cephalometric assessment. AI algorithms were used to superimpose lateral cephalograms taken at baseline and after 18 months. Linear and angular measurements, including S-N, ANB, mandibular length (Co-Gn), and lower anterior facial height (ANS-Me), were computed. Statistical analysis involved ANOVA and post hoc Tukey’s test ( P < 0.05 considered significant). Results: The AI-based analysis revealed significant differences in mandibular length increase among the groups (horizontal: 5.2 ± 1.3 mm, vertical: 3.1 ± 1.2 mm, average: 4.1 ± 1.1 mm; P = 0.003). Vertical growers exhibited the highest increase in lower anterior facial height (6.7 ± 1.5 mm), compared to horizontal (3.5 ± 1.0 mm) and average growers (4.2 ± 1.2 mm) ( P < 0.001). The AI superimpositions demonstrated high reproducibility (intra-class correlation > 0.94). Conclusion: AI-driven cephalometric superimposition provides a reliable and efficient method for assessing craniofacial growth in adolescents. Vertical growers show greater vertical skeletal development, while horizontal growers exhibit more prominent mandibular lengthening. AI integration into routine cephalometric evaluation can significantly enhance diagnostic accuracy and growth prediction.
