Assessment of two methods for detecting carious dentin: an in vitro study.

Abstract

Background: The objective of this study was to compare, in vitro, two dentinal caries lesion detector methods, Caries Finder and BlueCheck, to determine if they were substantially equivalent in their ability to aid visualization of demineralized dentin and to also to compare their performance compared to the traditional visual/tactile method of dentinal caries lesion detection in vitro.

Methods: Sixty-five extracted human teeth containing lesions rated as ICDAS 4,5 or 6 were chosen and then randomly assigned to two groups. Specimens were then evaluated in standard operatory conditions by three evaluators using the traditional visual and tactile method, the Caries Finder method, and the BlueCheck method of detection. The study employed a parallel, randomized controlled study design. To test the equivalence claim, a "two-one sided test" (TOST) approach was utilized.

Results: As compared to the traditional method, the Caries Finder method had a 0.9742 accuracy, 95% confidence interval [0.9578, 0.9855], 94.80% sensitivity, 98.53% specificity, 96.47% positive predictive value, 97.82% negative predictive value, 0.938 Kappa value, p < 2.2e-16). The BlueCheck method had a 0.9821 accuracy, 95% confidence interval [0.9682, 0.9910], 96.02% sensitivity, 99.09% specificity, 97.69% positive predictive value, 98.42% negative predictive value, 0.956 Kappa value, p < 2.2e-16). Inter-rater reliability and intra-rater reliability ratings were good to excellent.

Conclusions: The results of this study support the conclusion that the Caries Finder and BlueCheck methods compare favorably with the traditional method of carious dentin detection. Caries Finder and BlueCheck detection methods were found to have comparable performance in their ability to differentiate carious dentin from healthy tooth structure in vitro; however further in vivo validation is required to confirm clinical equivalence. Both show good to excellent inter-rater and intra-rater reliability.