Comparison of ICDAS, CAST, Nyvad's Criteria, and WHO-DMFT for Caries Detection in a Sample of Italian Schoolchildren.

Caries measurement methods vary considerably in terms of the stages of lesion considered making the comparison problematic among different surveys. In this cross-sectional study, four caries measurement methods, the WHO-DMFT, the International Caries Detection and Assessment System (ICDAS), the Caries Assessment Spectrum and Treatment (CAST), and the Nyvad Criteria were tested in a sample of children. Five-hundred 12-year old children (236 males and 264 females) were examined four times by four calibrated examiners. The calibration process showed that Cohen's Kappa exceeded the criterion of K = 0.75 and K = 0.80 for inter/intra-examiner agreement, respectively. In the survey, the total number of misclassification errors for the four methods amounted to 312 observations (67.94% regarding enamel lesions). The greatest difference among methods was shown by number of sound teeth (p < 0.01): WHO-DMFT n = 9505, 74.14%; ICDAS n = 2628, 20.49%; CAST n = 5053, 39.41%; and Nyvad Criteria n = 4117, 32.11%. At the level of dentinal Distinct/Active Cavity lesions, no statistically significant difference was observed (p = 0.40) between ICDAS (n = 1373, 10.71%), CAST (n = 1371, 0.69%), and Nyvad Criteria (n = 1720, 13.41%). In the severe caries levels, all methods were partially in agreement, while no accordance was found for the initial (enamel) lesions. A common language in caries detection is critical when different studies are compared.