Clustering effects of oral conditions based on clinical and radiographic examinations

Artificial intelligence in endodontics: Fundamental principles, workflow, and tasks

The integration of artificial intelligence (AI) in healthcare has seen significant advancements, particularly in areas requiring image interpretation. Endodontics, a specialty within dentistry, stands to benefit immensely from AI applications, especially in interpreting radiographic images. However, there is a knowledge gap among endodontists regarding the fundamentals of machine learning and deep learning, hindering the full utilization of AI in this field. This narrative review aims to: (A) elaborate on the basic principles of machine learning and deep learning and present the basics of neural network architectures; (B) explain the workflow for developing AI solutions, from data collection through clinical integration; (C) discuss specific AI tasks and applications relevant to endodontic diagnosis and treatment. The article shows that AI offers diverse practical applications in endodontics. Computer vision methods help analyse images while natural language processing extracts insights from text. With robust validation, these techniques can enhance diagnosis, treatment planning, education, and patient care. In conclusion, AI holds significant potential to benefit endodontic research, practice, and education. Successful integration requires an evolving partnership between clinicians, computer scientists, and industry.

Artificial intelligence for caries detection: Randomized trial

OBJECTIVES

We aimed to assess the impact of an artificial intelligence (AI)-based diagnostic-support software for proximal caries detection on bitewing radiographs.

METHODS

A cluster-randomized cross-over controlled trial was conducted. A commercially available software employing a fully convolutional neural network for caries detection (dentalXrai Pro, dentalXrai Ltd.) was randomly employed by 22 dentists, supporting their caries detection on 20 bitewings randomly chosen from a pool of 140 bitewings, with 10 bitewings randomly being supported by AI and 10 not. The reference test had been established by 4 + 1 independent experts in a pixelwise fashion. Caries was subgrouped as enamel, early dentin and advanced dentin caries, and accuracy and treatment decisions for each caries lesion assessed.

RESULTS

Dentists with AI showed a significantly higher mean (95% CI) area under the Receiver-Operating-Characteristics curve (0.89; 0.87-0.90) than those without AI (0.85; 0.83-0.86; p<0.05), mainly as their sensitivity was significantly higher (0.81; 0.74-0.87 compared with 0.72; 0.64-0.79; p<0.05) while the specificity was not significantly affected (p>0.05). This increase in sensitivity was found for enamel, but not early or advanced dentin lesions. Higher sensitivity came with an increase in non-invasive, but also invasive treatment decisions (p<0.05).

CONCLUSION

AI can increase dentists' diagnostic accuracy but may also increase invasive treatment decisions.

CLINICAL SIGNIFICANCE

AI can increase dentists' diagnostic accuracy, mainly via increasing their sensitivity for detecting enamel lesions, but may also increase invasive therapy decisions. Differences in the effects of AI for different dentists should be explored, and dentists should be guided as to which therapy to choose when detecting caries lesions using AI support.

Impact of Image Context on Deep Learning for Classification of Teeth on Radiographs

Objectives: We aimed to assess the impact of image context information on the accuracy of deep learning models for tooth classification on panoramic dental radiographs. Methods: Our dataset contained 5008 panoramic radiographs with a mean number of 25.2 teeth per image. Teeth were segmented bounding-box-wise and classified by one expert; this was validated by another expert. Tooth segments were cropped allowing for different context; the baseline size was 100% of each box and was scaled up to capture 150%, 200%, 250% and 300% to increase context. On each of the five generated datasets, ResNet-34 classification models were trained using the Adam optimizer with a learning rate of 0.001 over 25 epochs with a batch size of 16. A total of 20% of the data was used for testing; in subgroup analyses, models were tested only on specific tooth types. Feature visualization using gradient-weighted class activation mapping (Grad-CAM) was employed to visualize salient areas. Results: F1-scores increased monotonically from 0.77 in the base-case (100%) to 0.93 on the largest segments (300%; p = 0.0083; Mann-Kendall-test). Gains in accuracy were limited between 200% and 300%. This behavior was found for all tooth types except canines, where accuracy was much higher even for smaller segments and increasing context yielded only minimal gains. With increasing context salient areas were more widely distributed over each segment; at maximum segment size, the models assessed minimum 3-4 teeth as well as the interdental or inter-arch space to come to a classification. Conclusions: Context matters; classification accuracy increased significantly with increasing context.

Generalizability of Deep Learning Models for Caries Detection in Near-Infrared Light Transillumination Images

OBJECTIVES

The present study aimed to train deep convolutional neural networks (CNNs) to detect caries lesions on Near-Infrared Light Transillumination (NILT) imagery obtained either in vitro or in vivo and to assess the models' generalizability.

METHODS

In vitro, 226 extracted posterior permanent human teeth were mounted in a diagnostic model in a dummy head. Then, NILT images were generated (DIAGNOcam, KaVo, Biberach), and images were segmented tooth-wise. In vivo, 1319 teeth from 56 patients were obtained and segmented similarly. Proximal caries lesions were annotated pixel-wise by three experienced dentists, reviewed by a fourth dentist, and then transformed into binary labels. We trained ResNet classification models on both in vivo and in vitro datasets and used 10-fold cross-validation for estimating the performance and generalizability of the models. We used GradCAM to increase explainability.

RESULTS

The tooth-level prevalence of caries lesions was 41% in vitro and 49% in vivo, respectively. Models trained and tested on in vivo data performed significantly better (mean ± SD accuracy: 0.78 ± 0.04) than those trained and tested on in vitro data (accuracy: 0.64 ± 0.15; p < 0.05). When tested in vitro, the models trained in vivo showed significantly lower accuracy (0.70 ± 0.01; p < 0.01). Similarly, when tested in vivo, models trained in vitro showed significantly lower accuracy (0.61 ± 0.04; p < 0.05). In both cases, this was due to decreases in sensitivity (by -27% for models trained in vivo and -10% for models trained in vitro).

CONCLUSIONS

Using in vitro setups for generating NILT imagery and training CNNs comes with low accuracy and generalizability.

CLINICAL SIGNIFICANCE

Studies employing in vitro imagery for developing deep learning models should be critically appraised for their generalizability. Applicable deep learning models for assessing NILT imagery should be trained on in vivo data.