Abstract
Background/Aims: This study aimed to compare and contrast in vitro six methods to determine the most accurate method for detecting approximal carious lesions in primary molars. Methods: Extracted primary molars (n = 140) were stored in 0.02% chlorhexidine solution and mounted in light-cured resin in pairs. The six carious lesion detection methods used by the three examiners to assess approximal carious lesions were visual inspection, digital radiography, two transillumination lights (SDI and NSK), and two laser fluorescence instruments (CDD and DDP). Five damaged teeth were discarded. The teeth (n = 135) were sectioned, serially ground, and examined under light microscopy using Downer’s histological (HST) criteria as the gold standard. Intra- and inter-examiner reliability, agreement with HST, specificity, sensitivity, receiver operating characteristic (ROC) curves, and areas under the curve were calculated. Results: This study found visual inspection to be the most accurate method when validated by histology. Transillumination with NSK light had the highest specificity, and digital radiography had the highest sensitivity for detecting enamel and/or dentinal carious lesions. Combining specificity and sensitivity into the area under ROC curves, enamel plus dentinal lesions were detected most accurately by visual inspection followed by digital radiography; dentinal lesions were detected most accurately by digital radiography followed by visual inspection. Conclusions: None of the four newly developed methods can be recommended as suitable replacements for visual inspection and digital radiography in detecting carious lesions on approximal surfaces of primary molars, and further developmental work is needed.