⥠Quick Summary
This study explored the use of machine learning (ML) models to classify cervical vertebral maturation stages (CVMS) using consensus-based datasets. The Support Vector Machine model achieved the highest accuracy of 77.4%, highlighting the importance of dataset quality in ML applications.
ð Key Details
- ð Dataset: 1380 lateral cephalograms rated by three clinicians
- ð§Đ Features used: Landmarks annotation of cervical vertebrae
- âïļ Technology: Various ML models including Support Vector Machine and Multi-Layer Perceptron
- ð Performance: Highest accuracy of 77.4% with the Support Vector Machine model
ð Key Takeaways
- ð Consensus-based datasets significantly improve model accuracy in CVMS classification.
- ðĄ Machine learning can effectively classify cervical vertebral maturation stages.
- ð The Support Vector Machine model outperformed others with an accuracy of 77.4%.
- ð Individual ratings yielded lower accuracies, ranging from 60.4% to 67.9%.
- ð Feature selection was crucial for identifying significant characteristics in the datasets.
- ð The study emphasizes the importance of high inter- and intra-observer agreement in dataset quality.
- ð§ Robust generalization capabilities were observed in consensus-based training models.
- ð Published in: Progress in Orthodontics, 2024.
ð Background
The classification of cervical vertebral maturation stages is essential in orthodontics and dental practice, as it aids in treatment planning and predicting growth patterns. Traditional methods of assessment can be subjective and vary between clinicians. The integration of machine learning offers a promising avenue for enhancing the accuracy and consistency of these classifications.
ðïļ Study
This study involved three clinicians who independently rated cervical vertebral maturation stages on a total of 1380 lateral cephalograms. The researchers created five datasets, including two consensus-based datasets (Complete Agreement and Majority Voting) and three datasets based on individual ratings. A feature selection process was applied to identify key landmarks and patient information, which were then utilized to train various ML models.
ð Results
The results indicated that the features deemed significant in the consensus-based datasets aligned with established CVMS guidelines. The Support Vector Machine model trained on the Complete Agreement dataset achieved the highest accuracy of 77.4%, while the Multi-Layer Perceptron model on the Majority Voting dataset reached 69.6%. Models based on individual ratings demonstrated lower accuracies, ranging from 60.4% to 67.9%.
ð Impact and Implications
The findings of this study underscore the potential of machine learning in enhancing the classification of cervical vertebral maturation stages. By utilizing consensus-based datasets, clinicians can achieve higher accuracy and reliability in their assessments. This advancement could lead to improved treatment planning and outcomes in orthodontics, ultimately benefiting patient care.
ðŪ Conclusion
This research highlights the significant role of dataset quality in the performance of machine learning models for CVMS classification. The superior accuracy of models trained on consensus-based datasets demonstrates the importance of collaborative evaluations in clinical settings. As machine learning continues to evolve, its integration into orthodontic practices could revolutionize how clinicians assess and plan treatments.
ðŽ Your comments
What are your thoughts on the application of machine learning in orthodontics? We would love to hear your insights! ðŽ Share your comments below or connect with us on social media:
Classification of cervical vertebral maturation stages with machine learning models: leveraging datasets with high inter- and intra-observer agreement.
Abstract
OBJECTIVES: This study aimed to assess the accuracy of machine learning (ML) models with feature selection technique in classifying cervical vertebral maturation stages (CVMS). Consensus-based datasets were used for models training and evaluation for their model generalization capabilities on unseen datasets.
METHODS: Three clinicians independently rated CVMS on 1380 lateral cephalograms, resulting in the creation of five datasets: two consensus-based datasets (Complete Agreement and Majority Voting), and three datasets based on a single rater’s evaluations. Additionally, landmarks annotation of the second to fourth cervical vertebrae and patients’ information underwent a feature selection process. These datasets were used to train various ML models and identify the top-performing model for each dataset. These models were subsequently tested on their generalization capabilities.
RESULTS: Features that considered significant in the consensus-based datasets were consistent with a CVMS guideline. The Support Vector Machine model on the Complete Agreement dataset achieved the highest accuracy (77.4%), followed by the Multi-Layer Perceptron model on the Majority Voting dataset (69.6%). Models from individual ratings showed lower accuracies (60.4-67.9%). The consensus-based training models also exhibited lower coefficient of variation (CV), indicating superior generalization capability compared to models from single raters.
CONCLUSION: ML models trained on consensus-based datasets for CVMS classification exhibited the highest accuracy, with significant features consistent with the original CVMS guidelines. These models also showed robust generalization capabilities, underscoring the importance of dataset quality.
Author: [‘Kanchanapiboon P’, ‘Tunksook P’, ‘Tunksook P’, ‘Ritthipravat P’, ‘Boonpratham S’, ‘Satravaha Y’, ‘Chaweewannakorn C’, ‘Peanchitlertkajorn S’]
Journal: Prog Orthod
Citation: Kanchanapiboon P, et al. Classification of cervical vertebral maturation stages with machine learning models: leveraging datasets with high inter- and intra-observer agreement. Classification of cervical vertebral maturation stages with machine learning models: leveraging datasets with high inter- and intra-observer agreement. 2024; 25:35. doi: 10.1186/s40510-024-00535-1