BACKGROUND: Vertical root fractures (VRFs) can only be detected by radiography. In recent years, direct digital dental radiography (DDR) has become a substitute to film-based radiography. Purpose of this study was to compare accuracy and reliability of charge couple device (CCD)-based direct digital radiography with conventional radiography (CR) in VRF diagnosis. METHODS AND MATERIALS: In this in vitro study, 230 extracted single-rooted human teeth were endodontically instrumented. VRFs were performed experimentally in half of the samples. Each tooth was imaged using the paralleling technique with E-speed film and a CCD-based digital image receptor. Two oral radiologists interpreted the images and repeated the procedure a month later with half of the samples. The accuracy, sensitivity, and specificity of each technique were determined. The reliability and degree of agreement were also determined by the means of Cohen's Kappa analysis. χ² test was used to compare two observers' diagnosis, considering the location of fracture line. RESULTS: The accuracy of CR and DDR was 65% and 70% retrospectively. The sensitivity was 60% for CR, 61% for DDR and the specificity was 70% for CR, 78% for DDR, but the differences were not significant. The interobserver reliability was moderate for CR (K = 0.366, 95% CI) and fair for DDR (K = 0.538, 95% CI). CONCLUSION: No significant difference was seen between the two techniques. The specificity of DDR was slightly better than CR, and their accuracy and sensitivity showed small differences.
BACKGROUND:Vertical root fractures (VRFs) can only be detected by radiography. In recent years, direct digital dental radiography (DDR) has become a substitute to film-based radiography. Purpose of this study was to compare accuracy and reliability of charge couple device (CCD)-based direct digital radiography with conventional radiography (CR) in VRF diagnosis. METHODS AND MATERIALS: In this in vitro study, 230 extracted single-rooted human teeth were endodontically instrumented. VRFs were performed experimentally in half of the samples. Each tooth was imaged using the paralleling technique with E-speed film and a CCD-based digital image receptor. Two oral radiologists interpreted the images and repeated the procedure a month later with half of the samples. The accuracy, sensitivity, and specificity of each technique were determined. The reliability and degree of agreement were also determined by the means of Cohen's Kappa analysis. χ² test was used to compare two observers' diagnosis, considering the location of fracture line. RESULTS: The accuracy of CR and DDR was 65% and 70% retrospectively. The sensitivity was 60% for CR, 61% for DDR and the specificity was 70% for CR, 78% for DDR, but the differences were not significant. The interobserver reliability was moderate for CR (K = 0.366, 95% CI) and fair for DDR (K = 0.538, 95% CI). CONCLUSION: No significant difference was seen between the two techniques. The specificity of DDR was slightly better than CR, and their accuracy and sensitivity showed small differences.