Methods: 44 extracted single rooted premolars mounted in a dry human mandible for assessment of longitudinal root fractures were endodontically prepared and divided into 2 groups. The teeth in group A were artificially fractured and those in group B were not. Longitudinal fractures were induced in 22 of 44 premolars. For horizontal root fracture, 44 maxillary central incisors were used that in 22 of which, fractures were created in horizontal plane by a mechanical force with hammer and fragments were relocated. They were placed in the respective empty maxillary anterior sockets of a human dry skull. Intraoral radiographs were obtained by a PSP sensors (Digora Optime). Cone beam CT images were taken with a unit (Promax 3D, Planmeca Helsinki Finland). 4 observers evaluated the CBCT scans and PSP images.
Results: Sensitivity and specificity for longitudinal root fracture detection of CBCT were 0.48, 0.92 and for PSP were 0.29, 0.87, respectively and in horizontal root fracture were 0.97, 0.73 for CBCT and 0.89, 0.85 for PSP. The results showed an overall higher accuracy for CBCT (0.70) scans higher than PSP (0.59) for detecting longitudinal root fractures and slightly overall higher accuracy for PSP (0.87) scans than CBCT (0.85) for detecting horizontal root fractures.
Conclusions: CBCT outperformed the digital radiography in detecting simulated longitudinal root fractures but not in horizontal root fractures.