METHODS: Radiographs of OKCs and ameloblastomas were retrospectively reviewed. Location, border, shape, association with impacted tooth, tooth displacement, root resorption, and bone expansion were evaluated. Chi-squared or Fisher's exact tests were used for statistical analysis. A p value
DESIGN: Retrospective assessment using the Peer Assessment Rating (PAR) index.
SETTING: Consecutive patients treated by one consultant orthodontist at a tertiary care cleft center.
PARTICIPANTS: One hundred twenty-seven patients with either complete unilateral cleft lip and palate (UCLP) or bilateral cleft lip and palate (BCLP) consecutively treated with fixed appliances.
INTERVENTION: Fixed orthodontic appliance treatment and orthognathic surgery when required.
OUTCOMES: The PAR index assessment was carried out by a calibrated-independent assessor. Treatment duration, the number of patient visits, and data on dental anomalies were drawn from patient records and radiographs.
RESULTS: One hundred two patients' study models were assessed after exclusions. Mean start PAR score for UCLP (n = 71) was 43.9 (95% CI, 41.2-46.6, SD 11.5), with a mean score reduction of 84.3% (95% CI, 81.9-86.7, SD 10.1). The UCLP mean treatment time was 23.7 months with 20.1 appointments. Mean start PAR score for BCLP (n = 31) was 43.4 (95% CI, 39.2-47.6, SD 11.4), with a mean score reduction of 80.9% (95% CI, 76.3-85.5, SD 12.5). The BCLP mean treatment time was 27.8 months with 20.5 appointments.
CONCLUSION: These results compare well with other outcome reports, including those for patients without a cleft, and reflect the standard of care provided by an experienced cleft orthodontist. As with high-volume surgeons, orthodontic treatment for this high need group is favorable when provided by a high-volume orthodontist. These findings may be used for comparative audit with similar units providing cleft care.
Methods: We collected 3794 corneal images from 542 eyes of 280 subjects and developed seven deep learning models based on anterior and posterior eccentricity, anterior and posterior elevation, anterior and posterior sagittal curvature, and corneal thickness maps to extract deep corneal features. An independent subset with 1050 images collected from 150 eyes of 85 subjects from a separate center was used to validate models. We developed a hybrid deep learning model to detect KCN. We visualized deep features of corneal parameters to assess the quality of learning subjectively and computed area under the receiver operating characteristic curve (AUC), confusion matrices, accuracy, and F1 score to evaluate models objectively.
Results: In the development dataset, 204 eyes were normal, 123 eyes were suspected KCN, and 215 eyes had KCN. In the independent validation dataset, 50 eyes were normal, 50 eyes were suspected KCN, and 50 eyes were KCN. Images were annotated by three corneal specialists. The AUC of the models for the two-class and three-class problems based on the development set were 0.99 and 0.93, respectively.
Conclusions: The hybrid deep learning model achieved high accuracy in identifying KCN based on corneal maps and provided a time-efficient framework with low computational complexity.
Translational Relevance: Deep learning can detect KCN from non-invasive corneal images with high accuracy, suggesting potential application in research and clinical practice to identify KCN.