Comparative Analysis of Full Crown Morphology Designed by Artificial Intelligence and Dental Technicians.

Abstract

Objectives: To compare the morphological characteristics of full crowns designed by an experienced dental technician and a data-driven artificial intelligence (AI) system, with particular attention to their similarity to natural teeth.

Methods: Twelve digital casts, including mandibular right second premolars, were collected, three-dimensional (3D) printed, and prepared by an experienced dentist. For each case, two crowns were independently designed by an experienced dental technician (ET) and the AI system. Morphological characteristics, including buccal and lingual cusp angle, buccal-lingual diameter, mesial-distal diameter, functional cusp wear facets, 3D similarity, and occlusal contact point counts, were compared among natural teeth, ET crowns, and AI crowns. Functional cusp wear and occlusal contacts were evaluated qualitatively, based on morphological observation and visual analysis under virtual articulating paper conditions. Statistical analysis was performed using one-way analysis of variance with least significant difference post-hoc testing, with the significance level set at α = 0.05.

Results: Both AI crowns (56.2°±5.1°) and ET crowns (57.5°±6.3°) exhibited larger buccal cusp angles than natural teeth (45.7°±4.9°) (p < 0.05), while no significant differences were found in lingual cusp angles among the groups (p = 0.434, 0.607, 0.787). The buccal-lingual diameter of AI crowns closely matched natural teeth (p = 0.094), whereas ET crowns were significantly smaller than both AI crowns (p = 0.034) and natural teeth (p = 0.044). The mesial-distal diameter of AI crowns also showed no significant difference from natural teeth (p = 0.223), while ET crowns differed significantly from both groups (p = 0.047, 0.002). Although AI crowns exhibited slightly lower 3D discrepancy (0.43 ± 0.07) than ET crowns (0.48 ± 0.06), the difference was not statistically significant (p = 0.089). AI did not accurately reproduce functional cusp wear facets, whereas technician designs tended to overcompensate. Occlusal contact counts showed no significant differences among groups.

Conclusions: AI systems can reproduce overall crown morphology to a degree comparable with experienced technicians but remain limited in replicating functional wear facets.

Clinical implications: AI offers a practical and efficient tool for crown design, with potential to improve consistency and reduce manual workload in prosthodontics. However, further refinement is required before these systems can fully replicate functional morphological details for widespread clinical application.