Related Experiment Video
Updated: Aug 31, 2026
Measuring Maxillary Posterior Tooth Movement: A Model Assessment using Palatal and Dental Superimposition
Published on: February 23, 2024
The Role of Artificial Intelligence Models in Predicting Post-Prosthetic Facial Esthetics in Edentulous Patients:
Rania Moussa1, Mostafa I Fayad1, Omar M Alaryani1
1Department of Substitutive Dental Sciences, College of Dentistry, Taibah University, Madina, Saudi Arabia.
Objectives:
Artificial intelligence (AI) is increasingly used in esthetic dentistry; however, its accuracy in predicting post-prosthetic facial outcomes in edentulous patients remains unclear. This study aimed to evaluate the ability of AI models to simulate post-denture facial esthetics compared with actual clinical outcomes.
Materials And Methods:
A prospective within-subject observational study was conducted on 14 completely edentulous patients receiving new complete dentures. Standardized facial photographs were obtained pretreatment (T0) and posttreatment (T1). AI-generated images were produced from T0 photographs using Gemini (G-P) and FaceApp (FA-P). Outcomes included patient preference, expert esthetic evaluation, and quantitative facial anthropometric measurements.
Statistical Analysis:
Nonnormally distributed paired facial measurements were compared using the Wilcoxon signed-rank test. Differences among image types and expert panel scores were analyzed using the Friedman test (two-way analysis of variance by ranks), with post hoc Wilcoxon signed-rank tests where applicable. Participant preferences were assessed using the chi-square goodness-of-fit test. Statistical significance was set at p ≤0.05.
Results:
No significant differences were found in patient preferences (p = 0.247) and in expert evaluations between actual and AI-generated images (p = 0.316), with high inter-rater reliability (intraclass correlation coefficient >0.85). Anthropometric analysis revealed significant differences across most variables (p < 0.001) in oral commissure width, upper and lower lip height, and lower facial height. AI-generated images showed partial agreement but significant deviations in magnitude, particularly in FA-P, which were exaggerated. Bilateral comparison of commissure height showed no significant differences (p > 0.05), indicating preserved facial symmetry across all conditions.
Conclusions:
AI-generated simulations provide esthetically comparable representations to actual outcomes; however, they lack quantitative accuracy in reproducing facial anthropometric changes. While AI may serve as a useful adjunct for patient communication and visualization, it cannot replace clinically driven prosthodontic assessment.
