Comparison of automated grading of digital orthodontic models and hand grading of 3-dimensionally printed models.

Abstract

Emerging workflows in orthodontics enable automated analysis of digital models and production of physical study models from digital files for the evaluation of treatment outcomes. The objective of this study was to compare the automated assessment of digital orthodontic models and the hand grading of 3D-printed models with the use of the American Board of Orthodontics cast-radiograph evaluation (ABO CRE) system. Plaster models from 15 cases were scanned with the use of a desktop model scanner to create digital models from which physical models were produced with the use of a stereolithography-based 3D printer. All digital models from each case were graded with the use of an automated software tool (SureSmile), and 3D-printed models were scored by hand with the use of the ABO CRE grading system. All hand-graded models were scored a second time at least 2weeks later. SureSmile gave statistically significantly higher scores to alignment and rotations (P<0.001), overjet (P<0.001), occlusal contacts (P<0.001), and total score (P<0.001). Hand grading scored higher in buccolingual inclination (P<0.001). No significant differences were found in marginal ridges, occlusal relationships, and interproximal contacts. Scores assessed in an automated manner by SureSmile are generally significantly greater than those assessed by hand grading.