In vitro evaluation of automated single-unit crown design using artificial intelligence: Design time, marginal...

To compare an artificial intelligence-based design workflow with a manual CAD workflow in terms of design time, vertical marginal misfit, dental morphology, and interproximal and occlusal contacts during single-unit crown planning.

Standardized preparations for single-unit crowns were performed on a typodont (teeth 22, 24, and 26). After digital scanning, single crowns were designed (n = 60, 20 per tooth) using a manual CAD workflow (Exocad DentalCAD; exocad GmbH) (C-E) or an artificial intelligence-based design system workflow (Medit Link v2.4.4; Medit Corp.) (E-AI), according to randomized tooth allocation. Design time (in seconds), vertical marginal misfit (mm), and dental morphology (mm) were assessed digitally, followed by a physical evaluation of occlusal and interproximal contacts after 3D printing of the single-unit crowns. Data were analyzed using mixed-effects models, nonparametric tests, and chi-square tests for categorical outcomes, followed by adjusted residual analysis when appropriate, with a significance level of 5%.

The E-AI workflow significantly reduced design time compared with C-E (p < 0.001), reducing planning time by approximately 8-15 min, depending on tooth type. Marginal discrepancies exceeding clinical thresholds were observed, as well as morphological deviations at the distal crest of single crowns for incisors (p = 0.020), with no differences observed for molars and premolars (p > 0.05). A significant difference was observed only for mesial interproximal contacts (p = 0.037), whereas distal interproximal (p = 0.136) and occlusal contacts (p = 0.247) did not differ between workflows. Clinically poor occlusal contacts predominated in both groups.

Although the artificial intelligence-based design workflow markedly reduced design time, it did not improve key clinical parameters, as limitations related to marginal fit and occlusal contact quality remained in both workflows.

The use of artificial intelligence-based design systems in prosthetic planning can enhance clinical efficiency by streamlining design steps and reducing operator dependence. This technology has significant potential for educational settings and clinical practice by enabling faster prosthetic planning, while still requiring professional oversight for occlusal refinement and marginal adjustment.