Visual-tactile versus radiographic caries detection agreement in caries-active adults

Abstract

The study aims to determine a) the extent of agreement between visual-tactile caries examination (VTE) and radiographic examination (RE) in detecting presumptive caries lesions on occlusal and proximal surfaces of posterior teeth of UNC Xylitol for Adult Caries Trial participants; and b) the additional caries diagnostic yield obtained by adding RE to VTE. Data consisted of surface-level visual-tactile and radiographic classification of disease (cavitated and noncavitated caries lesions) or nondisease (sound surfaces). Participants (n = 114, adults with ≥12 erupted teeth and 1-10 caries lesions) received baseline VTE by a trained and calibrated examiner, and had interproximal radiographs obtained within 7 months before or after the VTE. Radiographs were assessed independently by two trained and calibrated examiners masked with respect to VTE results. The diagnostic threshold was surface-level disease/nondisease status. Kappa statistics provided an estimate of VTE-RE agreement on diseased surfaces. The additional diagnostic yield of the RE over VTE was calculated as the additional lesions detected radiographically as a percentage of the total number of lesions detected by VTE. Four-hundred ninety-four (51 occlusal, 433 proximal) lesions were detected; of these, 81 (2 occlusal, 79 proximal) lesions were detected by both VTE and RE. Kappa statistics were 0.18 (all surfaces), 0.04 (occlusal), and 0.18 (proximal). The additional diagnostic yield was 69 percent (all surfaces), 55 percent (occlusal), and 71 percent (proximal). There is poor agreement between VTE and RE to detect caries in posterior teeth of caries-active adults. However, an RE performed within 7 months of a VTE adds caries diagnostic yield in a clinical trial, especially on proximal surfaces.