Accuracy of radiographic detection of residual caries in connection with tunnel restorations.

The aim of this study was to evaluate the diagnostic accuracy of radiographic examination for the detection of residual caries after tunnel preparation and filling with glass ionomer cement. Further, two different radiographic modalities were compared. Forty-five extracted human premolars and molars with clinical signs of approximal caries were selected for the study. The teeth were radiographed in order to identify surfaces with dentinal lesions suitable for tunnel preparation. Seventeen of the teeth had lesions that were as a minimum through the enamel and as a maximum into the outer half of the dentin (14 teeth = 1 lesion, 3 teeth = 2 lesions). Seventeen surgically removed third molars were thereafter included. These teeth were sound with respect to caries. The teeth were arranged in blocks with approximal contacts and radiographed using two image receptors: Ektaspeed Plus film (E) and the Digora (D) digital storage phosphor plate system. The 20 carious surfaces and 20 surfaces of the impacted teeth were prepared following the guidelines for the class II tunnel technique and filled with glass ionomer cement. The radiographic examination was repeated with both modalities after filling, and the approximal surfaces scored on the radiographs by 5 observers using the criteria: 0 = no filling, 1 = filling with residual caries, 2 = filling with no adjacent caries. All teeth except the unerupted third molars were sectioned, and the sections examined under the microscope. Residual caries was observed in 8 surfaces. On average, sensitivity was 0.25 for E and 0.32 for D, specificity 0.83 for E and 0.76 for D, positive predictive value 0.38 for E and 0.32 for D, and negative predictive value 0.75 for E and 0.77 for D. The differences between the two modalities were not statistically significant (p > 0.3). Based on the relatively small number of lesions in this study, it is suggested that radiography immediately after tunnel restoration cannot be recommended, as the diagnostic benefit from this examination seems to be minuscule.