Histologic verification of the reliability of clinical caries diagnoses

: The reliability of different investigators' assessment distinguishing carious cavities from caries-free ones was examined in an in vitro study. In 23 recently extracted human teeth caries was completely removed by the principal investigator, while in 17 teeth caries was only incompletely removed. On two occasions at an interval of 2 weeks, 5 investigators established the diagnosis "carious" or "caries-free" after clinical inspection using probe and mirror. Decalcified thin sections were examined in order to determine objectively if a tooth was caries-free or not. On an average, 8 out of 10 carious and 6 out of 10 caries-free cavities were correctly diagnosed. An analysis of the erroneous diagnoses showed that it was deep interproximal lesions that went unnoticed most frequently. Lesions at the enamel-dentin junction were also overlooked. False positive diagnoses were three times more frequent than false negative results. The reliability of caries detection (sensitivity 83.5%) was far higher than reported in the literature.