Semi-automated evaluation of the American Board of Orthodontics electronic cast-radiograph evaluation for assessing orthodontic treatment outcome

AIM: This study aimed to evaluate the reliability and agreement of a semi-automated evaluation of the electronic cast-radiograph evaluation (E-CRE) model score, as developed by the American Board of Orthodontics (ABO), compared to manual measurements. METHODS: ™ Lab (version 3.5, Image Instruments GmbH, Chemnitz, Germany), while the plaster models were scored manually using the ABO measurement gauge. In addition, 20 models were scored manually and digitally by a second observer to assess interobserver reliability. RESULTS: The interobserver reliability was excellent for digital measurements (intraclass correlation coefficient ICC 0.94-1.00) and higher compared to those for manual measurements (ICC 0.90-0.97). The reliability between digital and manual measurements was very high, with correlation coefficients ranging from 0.95 to 0.99. However, the digital evaluation produced slightly higher scores for alignment than the manual measurement (mean difference: +0.93). CONCLUSION: ™ is highly reliable and involves less interpersonal variance than manual grading. However, scores measured digitally were on average 0.79 point higher than those measured manually (scores below 30 are considered as acceptable treatment outcome).