Detection of artificially induced vertical root fractures of different widths by cone beam computed tomography...

To compare the diagnostic accuracy of cone beam computed tomography (CBCT) for the detection of artificially induced vertical root fractures (VRFs) of different widths in vitro and in vivo.

Vertical root fractures were induced in 25 extracted nonendodontically treated single-rooted human teeth (maxillary first premolars, maxillary canines and mandibular incisors). Twenty teeth without VRFs served as a control group. CBCT scanning (3D Accuitomo 170) was performed in vitro and in vivo. For the in vivo scanning, teeth were autoclaved, embedded into bite plates, placed in sterile plastic bags and then inserted into the mouths of volunteers. Teeth with VRFs were sectioned into axial slices and examined using a stereomicroscope to measure the widths of the VRFs. Five observers assessed the presence of VRFs using axial CBCT. Values for sensitivity, specificity, accuracy and interexaminer agreement were calculated.

The accuracy, specificity and sensitivity of CBCT were significantly higher in vitro than in vivo for VRFs with widths 50-150 μm (P < 0.05). The sensitivity and accuracy of CBCT were significantly higher for the detection of VRFs with widths greater than 150 μm in vivo and in vitro (P < 0.05). The accuracy of CBCT in vivo was 0.29 and 0.8 for fracture widths ranging from 50 to 150 μm and wider than 150 μm, respectively. No significant differences in CBCT specificity were found between VRF widths both in vitro and in vivo. The interexaminer reliability of the raters revealed a kappa value of 0.72, demonstrating substantial agreement.

The detectability of VRFs by CBCT in vitro and in vivo was dependent upon fracture width. The accuracy of CBCT in detecting VRFs of 50-300 μm width in vivo was significantly lower compared to the in vitro accuracy.