Accuracy of digital image processing in radiographic early caries detection

Tooth surface with early caries lesion can remineralize into sound tooth surface with early intervention. However, early caries lesion is difficult to detect and frequently goes unnoticed during clinical or radiographic examinations. Image quality enhancement through imaging tools such as filters has the potential to increase the accuracy of early caries lesion detection. To compare the accuracy of original radiographic images to enhanced images with filters on digital dental x-ray for detection of early caries lesions. Five working models were utilized, containing 42 sound tooth surfaces, and 36 surfaces with early caries lesions. Radiographic images were takenusing Vistascan storage phosphor plate system (Dürr Dental, Germany) and enhanced with Fine, Caries 1, Caries 2, and High Diagnostic (HD) filters. Original and enhanced radiographic images were assessed by three observers to record the presence or absence of early caries lesions. Sensitivity, specificity, and overall accuracy were calculated. Sensitivity of HD and Fine filter images were significantly higher (p < 0.05), but Caries 1 and Caries 2 images did not differ significantly from the original images (p > 0.05). There was no significant difference between the specificity of original filter images and filter enhanced images. The overall accuracy of original filter images and filter enhanced images increased significantly only on HD filter (p < 0.05). HD filter had significantly higher overall accuracy than the original images.In addition, it showed the highest sensitivity, specificity, and overall accuracy. Therefore, some filters may be used to enhance early caries lesion detection.