Detection of caries around restorations on bitewings using deep learning

• DOI: 10.1016/j.jdent.2024.104886
• Publication Date: 2024-02-09T08:20:10Z
• Authors: Eduardo Trota Chaves; Shankeeth Vinayah...; Niels van Nistelr...; Tong Xi; Vitor Henrique Di...; Tabea Flügge; Hadi Saker; Alexander Kim; Giana da Silveira...; Bas Loomans; Marie-Charlotte H...; Fausto Medeiros M...; Maximiliano Sergi...
• Keywords: Clinical Practice

Objective: Secondary caries lesions adjacent to restorations, a leading cause of restoration failure, require accurate diagnostic methods ensure an optimal treatment outcome. Traditional strategies rely on visual inspection complemented by radiographs. Recent advancements in artificial intelligence (AI), particularly deep learning, provide potential improvements detection. This study aimed develop convolutional neural network (CNN)-based algorithm for detecting primary and secondary around restorations using bitewings. Methods: Clinical data from 7 general dental practices the Netherlands, comprising 425 bitewings 383 patients, were utilized. The used Mask-RCNN architecture, instance, segmentation, supported Swin Transformer backbone. After augmentation, model training was performed through ten-fold cross-validation. accuracy evaluated calculating area under Free-Response Receiver Operating Characteristics curve, sensitivity, precision, F1 scores. Results: achieved areas FROC curves 0.806 0.804, F1-scores 0.689 0.719 detection, respectively. Conclusion: An CNN-based automated system developed detect bitewings, highlighting significant advancement diagnostics. significance: that integrates detection both will permit development systems aid clinicians their daily clinical practice.