DIAGNOSTIC ACCURACY OF CONE BEAM COMPUTED TOMOGRAPHY USING DIFFERENT VOXEL SIZES IN DETECTION OF VERTICAL ROOT FRACTURES OF TEETH WITH METALLIC POSTS

Islam S. Shaker 1 , Nashwa S. Mohamed 2 and Ahmed M. Abdelsamad 3 . 1. Assistant lecturer of Oral Radiology, Faculty of Oral and Dental Medicine, Misr International University. 2. Professor of Oral and Maxillofacial Radiology, Faculty of Dentistry, Cairo University. Professor of Oral Radiology, Faculty of Oral and Dental Medicine, Misr International University. 3. Associate Professor of Oral and Maxillofacial Radiology, Faculty of Dentistry, Cairo University. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History


1054
CBCT has been used for proper diagnosis of root fractures to overcome the inherent disadvantages of anatomic superimposition. CBCT has allowed dental practitioners to visualize teeth three-dimensionally and with high spatial resolution (Long et al.

2014).
While recent studies have demonstrated the superiority of CBCT in detecting root fractures, there are some limitations to proper imaging when high-density materials such as gutta-percha and intra-canal metal posts are present. These materials may create artefacts that impair the quality of CBCT images. Beam hardening and streak artefacts can be superimposed on the root, creating difficulties for image assessment and may even mimic root fractures

Study Design:
This study was planned to assess the accuracy of Cone Beam Computed Tomography versus intra-oral digital periapical radiography in detection of root fractures of teeth with metallic post. Our study is a cross sectional prospective (data collection was planned before the index test and reference standard test were performed) study with a blind comparison to a gold standard.

Teeth Selection, Preparation & Grouping:
Teeth selection: 120 Single rooted extracted human teeth were obtained from Oral Surgery Department in the outpatient clinic in Misr International University. Selection of the teeth was based on fulfilling the eligibility criteria which were: 1. Maxillary or mandibular single rooted teeth. 2. Absence of root fracture on clinical examination.

Teeth Preparation & Grouping:
All teeth included in our study were endodonticaly treated then the gutta percha in the cervical part of the root was removed using gates glidden drills. The roots of teeth were covered by wax.
A Rectangular acrylic block was prepared in a rectangular plastic box with sufficient height to hold teeth in a position simulating patient radiographic imaging. Then eight holes were made in each block to support the teeth included in the study ( Figure.1).
Each Eight teeth were embedded in each acrylic block with their palatal surfaces directed outside towards the block borders.
The canals were prepared using standardized drills and the final standardized metallic posts were inserted ( Figure.2).
After post insertion the teeth roots were divided equally and randomly into 2 groups: In Group 1 Induced vertical root fractures were created by applying excessive forces to the roots through excessive rotation of posts with larger sizes (Figure.3). In Group 2 The teeth having intact roots with the posts inserted.

Imaging of Teeth:
Each tooth was coded by special code (Known only by the researcher) and placed randomly in acrylic blocks to be imaged 2 times as follows: CBCT with Standard resolution:-Each block was sectioned into two halves each containing four teeth to be adjusted to the CBCT field of view . The teeth were imaged using CBCT machine (Scanora3D, Soredex, Finland) after adjusting the control panel for image capturing with standard resolution. Parameters were adjusted at KVp=90 mA=10 Voxel size= 133µm FOV=5×5 cm and exposure time=6.1 s.

CBCT with high resolution:-
The exposure parameters used were the same as the first protocol but with different voxel size (voxel size of 85 µm). All CBCT images were analyzed using software (On Demand 3D) viewer in axial, coronal, sagital and 3D views to detect presence or absence of root fracture (Figure.4).
The outcome was binary for the presence or absence of fracture lines.

Blinding:-
The images were evaluated by 2 radiologists with an experience ranging from 10-25 years for presence or absence of root fractures.
Two radiologists were blinded during periapical radiographs and CBCT images assessment as tooth preparation and coding was done by a third researcher. Moreover, the statistician was blinded.
Statistical Analysis:-Data were presented as frequencies and percentages. ROC (Receiver Operating Characteristic) curve was constructed to evaluate diagnostic accuracy measures of the different modalities. Areas under the ROC curve (AUC) of the four modalities were compared using z-statistic. Sensitivity, Specificity, Positive, Negative predictive values and Diagnostic accuracy measures were calculated.
The significance level was set at P ≤ 0.05. ROC curve analysis was performed with MedCalc Version 11.3 for Windows (MedCalc Software bvba).

Results:-
Frequencies and percentages of fracture detection by different modalities were calculated.

Discussion:-
Accurate diagnosis of VRF depends on a careful clinical examination, complete evaluation of the case, and on an imaging examination which assesses the integrity of the bone and of the dental structure Detection of VRF is not only influenced by the type of imaging examination, either conventional radiography or CBCT, but also by the presence of material in the root canal, such as MPs, filling material, or remaining restorative material, So in our study we used CBCT with different voxel sizes to detect VRF in the presence of metallic posts. Induced vertical root fractures were created by applying excessive forces to the roots through excessive rotation of posts to try to mimic the shape of fracture line in clinical situation.

1056
Our results showed that the highest diagnostic accuracy was found with CBCT1 (76.7%) followed by CBCT2 (75.9%).
These results are in line with Da Silveira et al. 2013 who reported that different voxel sizes has no statistically significant difference in accuracy of detection of vertical root fractures.

Conclusion:-
Using voxel size of 133µm is sufficient to accurately detect vertical root fracture in presence of metallic posts.    Table 1:-Sensitivity, specificity, predictive values, diagnostic accuracy, Area Under the ROC curve (AUC) and 95% confidence interval (95% CI) for the detection of fracture by different modalities.