Clinical and historical predictors of dental caries on radiographs.

Abstract

This report evaluates the efficacy of the clinical predictors of caries proposed in the USA FDA guidelines for prescribing dental radiographs. The clinical findings that best associate with the presence of any caries where a history of pain, a defective restoration, unusual calcification, and an abutment tooth for a fixed or removable prosthesis. There is a group of measures of periodontal disease that were also weakly associated with the presence of caries. The best predictors of caries extending into the dentin were the presence of clinically defective restorations, a history of pain, and signs of periodontal disease. Proximal lesions are likely to appear on teeth with defective restorations, unusual calcification or large or deep restorations. The best predictors of radiographic root caries are periodontal findings such as furcation involvement, increased mobility, a history of periodontal therapy and gingival recession. While the specificities for these findings were generally high, the sensitivities and positive predictive values were usually under 50%, and often much lower. Thus these clinical findings cannot successfully be used as exclusive criteria for ordering radiographs for caries detection. Because caries is found fairly frequently, and because we are unable to identify tooth-specific criteria with clinically useful sensitivity and specificity values, these data support the FDA panel recommendation of bitewing examinations for all new patients and at periodic intervals for recall patients.