Method: A sample of fifty extracted molars and premolars was independently evaluated by 3 examiners. Assessment of histological section using stereomicroscope was used to validate lesion depth. Un-weighted kappa coefficient was used to assess inter- and intra- examiner reproducibility. Data obtained from the visual examination were compared to the histological classification for each investigation site using the Spearman correlation coefficient.
Result: The intra-examiner reproducibility was calculated by reexamination of 10% of the sample, showing 0.92 for Nyvad and 1.00 for both ICDAS-II and WHO criteria. Inter-examiner kappa values were 0.83 and 0.82 for Nyvad and OMS criteria, respectively and 0.65 for ICDAS-II. Among the visual examination discordant scores, the majority was related to healthy sites and pre-cavity lesions: 25% of Nyvad discordant scores and 32% of ICDAS-II. There were also discordant sites related to lesions depth: 26% of ICDAS-II discordant scores. The major correlation between visual (ICDAS-II) and histological examination was found for cavitated lesion characteristic (D3) and the best value was observed for Examiner 3 (0.95) and Examiner 1 (0.67).
Conclusion: The ICDAS-II and Nyvad presented good reproducibility and correlation with histological assessment at varying stages of the disease process.