AI-Dentify: deep learning for proximal caries detection on bitewing x-ray - HUNT4 Oral Health Study.

BACKGROUND: Dental caries diagnosis requires the manual inspection of diagnostic bitewing images of the patient, followed by a visual inspection and probing of the identified dental pieces with potential lesions. Yet the use of artificial intelligence, and in particular deep-learning, has the potential to aid in the diagnosis by providing a quick and informative analysis of the bitewing images. METHODS: A dataset of 13,887 bitewings from the HUNT4 Oral Health Study were annotated individually by six different experts, and used to train three different object detection deep-learning architectures: RetinaNet (ResNet50), YOLOv5 (M size), and EfficientDet (D0 and D1 sizes). A consensus dataset of 197 images, annotated jointly by the same six dental clinicians, was used for evaluation. A five-fold cross validation scheme was used to evaluate the performance of the AI models. RESULTS: The trained models show an increase in average precision and F1-score, and decrease of false negative rate, with respect to the dental clinicians. When compared against the dental clinicians, the YOLOv5 model shows the largest improvement, reporting 0.647 mean average precision, 0.548 mean F1-score, and 0.149 mean false negative rate. Whereas the best annotators on each of these metrics reported 0.299, 0.495, and 0.164 respectively. CONCLUSION: Deep-learning models have shown the potential to assist dental professionals in the diagnosis of caries. Yet, the task remains challenging due to the artifacts natural to the bitewing images.

The validity of self-reported number of teeth and edentulousness among Norwegian older adults, the HUNT Study.

BACKGROUND: Number of teeth is an established indicator of oral health and is commonly self-reported in epidemiological studies due to the costly and labor-intensive nature of clinical examinations. Although previous studies have found self-reported number of teeth to be a reasonably accurate measure, its accuracy among older adults ≥ 70 years is less explored. The aim of this study was to assess the validity of self-reported number of teeth and edentulousness in older adults and to investigate factors that may affect the accuracy of self-reports. METHODS: This study included two different samples of older adults ≥ 70 years drawn from the fourth wave of the Trøndelag Health Study (the HUNT Study), Norway. Sample 1 (n = 586) was used to evaluate the validity of self-reported number of teeth and sample 2 (n = 518) was used to evaluate self-reported edentulousness. Information on number of teeth and background variables (education, smoking, cognitive function, and self-perceived general and oral health) were self-reported in questionnaires, while clinical oral health examinations assessed number of teeth, number of teeth restored or replaced by fixed prosthodontics and edentulousness. Spearman and Pearson correlation coefficients, Bland-Altman plot, chi-square test and kappa statistics were used to assess the agreement between self-reported and clinically recorded number of teeth. RESULTS: The mean difference between self-reported and clinically recorded number of teeth was low (- 0.22 teeth), and more than 70% of the participants reported their number of teeth within an error of two teeth. Correlations between self-reports and clinical examinations were high for the total sample (0.86 (Spearman) and 0.91 (Pearson)). However, a lower correlation was found among participants with dementia (0.74 (Spearman) and 0.85 (Pearson)), participants having ≥ 20 teeth (0.76 (Spearman) and 0.67 (Pearson)), and participants with ≥ 5 teeth restored or replaced by fixed prosthodontics (0.75 (Spearman) and 0.77 (Pearson)). Self-reports of having teeth or being edentulous were correct in 96.3% of the cases (kappa value 0.93, p value < 0.001). CONCLUSIONS: Among older Norwegian adults, self-reported number of teeth agreed closely with clinical tooth counts and nearly all the edentulous participants correctly reported having no teeth.