By Danielle 
A review of 21 studies conducted between 2014 and 2024 suggests that artificial intelligence (AI), particularly deep learning (DL) algorithms, shows strong potential in detecting and predicting early childhood caries (ECC).
Published July 26 in Nature, the study found that DL algorithms—models based on complex neural networks that mimic how the human brain detects patterns in large, unstructured datasets—can detect ECC with an accuracy ranging from 78 to 86 per cent. Reported sensitivity ranged from 67 to 96 per cent, while specificity varied from 81 to 99 per cent.
Read related link: University of Texas researchers training AI to predict dental composite performance
For ECC prediction, the studies reported an accuracy range of 60 to 100 per cent, sensitivity from 20 to 100 per cent, and specificity from 54 to 94 per cent. The pooled sensitivity and specificity across all studies were 80 and 81 per cent, respectively, with 95 per cent confidence intervals—indicating statistically significant effects.
Despite the promising results, the authors advised that further research is needed to improve the technology and determine its clinical application in paediatric dentistry.
Read related article: Reshaping Dental Care with Artificial Intelligence
Listen: Episode 2: Integrating Technology and Training Staff