Intra-examiner agreement of early caries detection methods

Methods: Molars and premolars of 119 subjects, age 8-13, were examined every 6 months. Presented here are the results for intra-examiner agreement after three visits. Several early caries detection methods were used for detection of carious lesions, including non-cavitated lesions. Clinical visual examination (VE) was performed using the ICDAS criteria, followed by QLF (quantitative light-induced fluorescence), fiber-optic transillumination (FOTI), digital fiber-optic transillumination (DiFOTI), and Diagnodent (LF). Different examiners performed examinations for each method independently.

Results: There were no differences found between the three visits, so the data were pooled for each method. Visual examination was repeated on 1466 surfaces. The number of surfaces varied slightly between methods, as neither Diagnodent nor QLF examinations were done on approximal surfaces. Agreement for methods using an ordinal scale was for VE 0.79 (Weighted Kappa), for FOTI 0.84 (Weighted Kappa), and for DiFOTI 0.56 (Weighted Kappa). During the QLF examination, a decision was made on presence or absence of caries, and the dichotomous decision showed a repeatability of 0.69 (Kappa). Agreement for Diagnodent (continuous scale) was assessed using Intra-Class Correlation Coefficient (ICC), and gave an ICC of 0.75. Agreement for each method for occlusal surfaces was found to be similar as for all surfaces, but the lowest agreement was found for approximal surfaces for VE, FOTI, and DiFOTI.

Conclusions: All methods showed substantial intra-examiner agreement in their caries scoring in posterior permanent teeth.

Supported by NIH/NIDCR P01 DE013540.