Abstract
Objective
To develop an artificial intelligence (AI)-based algorithm for the assessment and comparison of skeletal maturation in patients with and without cleft lip and/or palate and to detect the presence of cervical vertebral anomalies (CVAs).
Design
Retrospective cohort study.
Setting
A university orthodontic clinic and comprehensive cleft care centers.
Patients/Participants
In total, 1080 cephalograms of patients with and without unilateral cleft lip and palate (UCLP) aged 6 to 18 years, without any associated syndromes, congenital disorders, or history of trauma or illness, were collected. About 960 cephalograms were assessed in the study upon elimination of poor-quality lateral cephalograms.
Methods
The MobileNet architecture using TensorFlow framework was employed to develop 2 convolutional neural network (CNN)-based AI models for automated assessment of skeletal age and detection of CVAs. Inter-rater reliability for manual cervical vertebral maturation (CVM) staging was assessed using Cohen's kappa coefficient, and intraclass correlation coefficient (ICC) was calculated. The results of each model were separately analyzed using chi-square test, and the statistical significance was tested at 5% level.
Results
The CNN-based AI model yielded an average accuracy rate of 74.5%, with an accuracy of up to 88% for detecting skeletal maturity and an accuracy rate of 83% for detecting CVAs.
Conclusions
It can be concluded that CVM methods help detect skeletal maturity objectively in patients with UCLP and have shown delayed skeletal growth compared to patients without UCLP. CVAs were found to be more prevalent in patients with UCLP than in their non-cleft counterparts, with these findings facilitated by utilizing a novel AI algorithm.
Keywords
Get full access to this article
View all access options for this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
