Variables affecting the inter- and intra-examiner reliability of ICDAS for occlusal caries diagnosis in permanent...

The aims of this study were to assess the validity and reproducibility of the International Caries Detection and Assessment System (ICDAS) in detecting occlusal caries in permanent molars and to investigate variables that can affect the reproducibility of ICDAS.

Five experienced, trained, and calibrated investigators examined the occlusal surfaces of 142 extracted permanent molars, ranging from sound to cavitated dentine caries. All visual examinations were conducted under standard conditions. Examinations were repeated after a period of 1 and 3 months from the initial examination. After all examinations were completed, teeth were sectioned and histologic validation was undertaken. Maximum kappa statistics were used to assess bias among examiners and linear weighted kappa was used to assess inter- and intra-examiners' reproducibility. Sensitivity, specificity, accuracy and area under the receiver operating characteristic curve were obtained at D3 threshold.

Differences in the marginal homogeneity were observed for the general dentist. Kappa for intra-examiner reproducibility ranged between 0.66 and 0.81 and the inter-examiner reproducibility ranged between 0.42 and 0.75. ICDAS score of ≥3 presented the best Spearman's correlation coefficient (0.69) at Downer's histologic classification D3. Corresponding sensitivity and specificity values at that level were 0.73 and 0.95, respectively. While variables like the interval of time between teeth examinations did not affect the intra-examiner reproducibility, examiners' specialty influenced inter-examiner's reproducibility over time.

ICDAS presented acceptable intra-examiners reproducibility in detecting occlusal caries over time. However, there was a clear effect for examiner's specialty and clinical experience on the inter-examiner's reproducibility of the ICDAS system at different time intervals.