ABSTRACT
OBJECTIVE: Histology is frequently used as a gold standard to validate caries detection devices. Poor assessment consistency could lead to apparent changes in diagnostic accuracy. In multi-center, multi-examiner studies electronic transfer of information would be convenient, provided there is no deteriation in quality. This study tested the hypothesis that examiner reproducibility in the assessment of caries lesion depth when viewing photographic images of histological sections on a computer monitor, is comparable with viewing the same sections under a microscope using two histological classification systems. METHODS: 166 investigation sites (96 teeth) were selected for visual examination (ICDAS-II) and sections made using a novel technique which reduced risk of section damage and allowed accurate allocation of section to each investigation site. Digital images of the sections were produced and four examiners viewed the sections under a microscope and on a separate occasion corresponding digital images on a computer monitor. Presence and extent of caries was scored according to two histological classification systems (Downer, ERK). RESULTS: The inter- and intra-examiner reproducibility for both histological classification systems and both examination techniques was substantial to almost perfect (weighted kappa=0.63-0.90). Comparing the kappa values between microscopy and viewing digital images, there was no effect or only a small effect between both examination techniques (effect size 0.00-0.28). There was also a strong relationship between the two viewing techniques (r(s)=0.748-0.844). CONCLUSIONS: Viewing digital images of tooth sections produces results comparable to viewing images directly under a microscope and therefore has potential benefits for multi-centre studies.
