Validity of caries detection on occlusal surfaces and treatment decisions based on results from multiple caries-detection methods

The aim of this in vitro study was to evaluate whether having results available from multiple detection methods influences dentist's treatment decisions for incipient caries lesions on occlusal surfaces. The occlusal surface of 96 extracted permanent molars without frank cavitation was examined by three examiners initially by visual examination alone, following which they chose one of three treatment options: (i) no treatment, (ii) preventive or non-invasive treatment (sealants), and (iii) invasive treatment. Four weeks later the examiners again selected one of the three treatment options for the surfaces, but this time were able to refer to the results from additional caries-detection methods [bitewing radiographs, electric conductance measurement (ECM), quantitative light fluorescence (QLF), and DIAGNOdent] that had been performed in the interim time. Stereomicroscopy was used to evaluate sensitivity, specificity, accuracy, and area under the Receiver Operating Characteristic (ROC) curve (AUC) for the detection methods at the D1 diagnostic threshold. Slight improvement was obtained in the percentage of sites correctly diagnosed, and in the AUC, when referring to the results obtained from all detection methods compared with visual examination alone. However, a drastic effect on the selection of treatment options was observed by having results available from multiple methods, with the choice of invasive treatment increasing substantially. In conclusion, having data available from multiple methods did not improve the accuracy of examiners in detecting early occlusal caries lesions, but it had a great influence on the number of surfaces indicated for operative treatment. The potential decrease in overall specificity while using multiple methods of detection may be of concern in populations with a low prevalence of occlusal caries lesions.