Objectives: The objective of this study is to assess the potential of DIAGNOdent readings in detecting carious lesions under composite resin restorations in comparison with digital radiography.
Materials and methods: One hundred extracted third molars were visually selected: 50 with and 50 without carious lesions. Class I occlusal cavities were prepared. In the carious group, caries were left on the floor of cavities before composite resin restoration. The occlusal surface relevant to the caries position in the carious group and a counterfeit point in the non-carious group were marked for DIAGNOdent reading. Teeth were fixed in a plaster box. Digital radiographs were taken and examined by four observers. Definitive diagnosis of caries was based on a histological assessment. Diagnostic accuracy for each method was expressed by the area under the receiver operating characteristic (ROC) curve. Differences between the areas under the ROC curves were assessed using the McNemar test.
Results: The respective sensitivity (95 % confidence interval (%CI)) and specificity (95 %CI) of DIAGNOdent were 0.74 (0.66-0.83) and 0.84 (0.76-0.92). The respective average diagnostic performance values for digital radiographs assessed by four observers were 0.54 (0.36-0.72) and 0.77 (0.65-0.86). The respective ROC values for DIAGNOdent and digital radiographs based on four observers were 0.79 and 0.65. Cohen's kappa statistic revealed a moderate to substantial agreement among interobserver reliabilities (k = 0.60-0.77).
Conclusions: There were no statistically significant differences between DIAGNOdent and digital radiographs for the detection of dental caries under composite restoration (p value > 0.05).
Clinical relevance: DIAGNOdent was developed to detect caries on occlusal surfaces and was tested for caries adjacent to filling materials. This study demonstrated the use of DIAGNOdent in detecting caries under old composite restorations.