Validity, Reliability and Reproducibility of Artificial Intelligence-Based Methods in Tooth Widths, Bolton Ratios and...

Area of Science:

• Dental measurements
• Artificial intelligence in orthodontics
• Biometrics

Background:

• Accurate dental measurements are crucial for orthodontic diagnosis and treatment planning.
• Traditional manual methods can be time-consuming and prone to human error.
• AI-based tools offer potential for automated and efficient analysis of dental models and images.

Purpose of the Study:

• To compare the validity, reliability, and reproducibility of manual, fully automated AI, and semi-automated AI methods for dental measurements.
• To assess the accuracy of AI in measuring tooth widths, calculating Bolton ratios, and performing space analysis.
• To determine the optimal method for consistent and accurate dental assessments.

Main Methods:

• Analysis of 102 plaster models and 102 occlusal photographs using manual, fully automated AI, and semi-automated AI.
• Measurement of mesiodistal tooth widths, calculation of Bolton ratios, and space analysis in both arches.
• Statistical evaluation of validity (ANOVA), reliability (Pearson's correlation), and reproducibility (ICC).

Main Results:

• Significant differences in mean tooth widths among methods; semi-automated AI yielded highest values.
• No significant differences in Bolton ratios, but AI methods showed increased variability in individual measurements.
• Fully automated AI demonstrated highest intra-examiner reliability for tooth width, Bolton ratios, and space analysis.
• All methods showed excellent reproducibility for tooth width; manual method superior for Bolton ratio reproducibility.

Conclusions:

• AI methods offer potential for dental measurements, with fully automated AI showing high reliability.
• Manual methods remain superior for reproducibility in specific derived measurements like Bolton ratios.
• Careful consideration of AI method variability is needed for accurate orthodontic assessments.