|Year : 2022 | Volume
| Issue : 3 | Page : 309-313
Comparison of Cone Beam Computed Tomography Performance at Different Voxel Sizes in the Evaluation of Mandibular Canal – An In vitro Study
Rini Joy1, A Kannan1, Krithika C Lakshmi1, D K S. Lakshminrusimhan2, Anitha Roy3
1 Department of Oral Medicine and Radiology, SRM Dental College and Hospital, Ramapuram, Chennai, Tamil Nadu, India
2 Department of Oral Medicine and Radiology, RVS Dental College and Hospital, Coimbatore, Tamil Nadu, India
3 Department of Pharmacology, Saveetha Dental College and Hospitals, Saveetha Institute of Medical and Technical Sciences, Chennai, Tamil Nadu, India
|Date of Submission||28-Aug-2021|
|Date of Decision||15-Aug-2022|
|Date of Acceptance||15-Aug-2022|
|Date of Web Publication||26-Sep-2022|
Department of Oral Medicine and Radiology, SRM Dental College and Hospital, Ramapuram Campus, Bharathi Salai, Chennai - 600 089, Tamil Nadu
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Background and Aim: Even though cone beam computed tomography (CBCT) imaging is a propitious tool to track the mandibular/ inferior alveolar canal (MC) course, documentation regarding indefectible CBCT imaging strategy for MC localization is sparse. This study aimed to appraise CBCT's functioning at specified voxel sizes for deducing an explicit voxel size setting that simplifies MC evaluation in individual imaging sections. Materials and Methods: CBCT scan of 12 dry human mandibles was produced at 0.5, 0.3, 0.25, and 0.2 mm voxel settings. Two specialists/assessors explored the generated images in coronal (buccolingual) and curved (anterior-posterior) sections. Statistical Analyses: Mann–Whitney U, Kruskal–Wallis, post-hoc Tukey HSD, and Kappa statistics. Results: All voxel specifications revealed appreciable statistical variance; coronal sections comparatively excelled in serving the study objective. Statistical authentication was spotted among voxel sizes 0.5–0.2 mm and 0.5–0.25 mm (both assessors, either section), 0.3–0.2 mm (first specialist, curved), and 0.3–0.5 mm (second specialist, curved). Inter-observer agreement was excellent for voxels 0.25 mm (coronal) and 0.2 mm (curved). Conclusion: Precise voxel setting to trace MC is 0.3 mm and the suitable imaging section is coronal.
Keywords: Cone-beam computed tomography, mandibular canal, radiation exposure, voxel size
|How to cite this article:|
Joy R, Kannan A, Lakshmi KC, Lakshminrusimhan D K, Roy A. Comparison of Cone Beam Computed Tomography Performance at Different Voxel Sizes in the Evaluation of Mandibular Canal – An In vitro Study. J Indian Acad Oral Med Radiol 2022;34:309-13
|How to cite this URL:|
Joy R, Kannan A, Lakshmi KC, Lakshminrusimhan D K, Roy A. Comparison of Cone Beam Computed Tomography Performance at Different Voxel Sizes in the Evaluation of Mandibular Canal – An In vitro Study. J Indian Acad Oral Med Radiol [serial online] 2022 [cited 2022 Dec 1];34:309-13. Available from: http://www.jiaomr.in/text.asp?2022/34/3/309/356955
| Introduction|| |
The mandibular canal (MC) is a variable anatomical structure, differing in cortication, shape, or position, thus, making its visualization a challenging clinical task. MC localization is crucial because it has plenty of dental applications.
For tracking MC course, cone beam computed tomography (CBCT) has been extensively studied and employed over time, as it is more advantageous than other imaging modalities.,,,,
Each CBCT equipment permits one to tailor the scanning protocols, such as the number of projections, voxel size, the field of view (FOV), and other settings, based on individual diagnostic requirements. The aforementioned imaging parameters determine scanning time, radiation exposure, and image quality.,
This study focused on the parameter voxel size. A three-dimensional image's smallest distinguishable unit is the voxel, which determines image resolution., Each voxel's size is modified based on the height, width, and its thickness. Smaller voxel size is beneficial because of its high spatial resolution, but dis-benefits are longer scanning and elevated radiation doses. Areas with larger voxel have adequate image resolution and reduced radiation doses.
Although few studies validated voxel size variation's effect on certain dental diagnostic tasks,,,,, scientific substantiation for the standard protocol to be followed for CBCT examination of any specific diagnostic task is scarce, and there is heterogeneous literature evidence for the optimized voxel setting that is ideal for tracking MC in various image sections.
Therefore, this study aimed to evaluate the performance of CBCT images at various voxel sizes to determine a specific voxel size that would enable MC visualization at two different imaging sections.
| Materials and Methods|| |
Clearance for study conduction was accorded by the institutional review board (ethical board, clearance no., Date: SRMDC-IRB, 903, 31/10/2017). Twelve dry human mandibles borrowed from the institution's anatomy department were chosen based on the following:
Inclusion criteria: dry, dentulous human mandibles.
Exclusion criteria:- pathologies, fractured/edentulous mandibles (for this in vitro study, written informed consent and Helsinki declaration were not applicable).
Sample size was calibrated using nMaster Software, version 2.0, with power 80%, alpha error 5%, and sample proportion 86.1%. The minimum sample size per voxel setting derived for this study was 47 teeth regions (per voxel size and each imaging section), which was approximated to 60 teeth regions.
To simulate human cheeks, a soft tissue simulator-base plate wax sheet, 13 mm, was used. Each mandible was then subjected to CBCT scan using Kodak (CareStream)–9500, 3D Dental imaging software, version 6.14 (from a Private Scan centre). The acquisition parameters were as follows: 90 kv voltage, 10 Ma current, 10.8 s exposure time, FOV 18 × 21 cm, 360-degree rotation, 32-bit depth, slice thickness 180 microns, and voxel sizes: 0.5, 0.3, 0.25, and 0.2 mm. Generated images were transferred as DICOM files from the workstation monitor to a Dell laptop (windows 10, Inspiron-3521, 15.8-inch, and resolution 1366 × 768) using the CD.
Two observers/specialists, skilled in oral and maxillofacial radiology, were blinded from the study and independently evaluated all the images on the same laptop under ambient light settings. They assessed MC visibility in two image sections – coronal (buccolingual) and curved (anterior-posterior); each at five mandibular regions (right and left sides): mental foramen, second premolar, first molar, second molar, and third molar. The evaluation was done at 480 mandibular sites with 4-point visibility scoring:
0- canal's corticated border is invisible
1- canal's one corticated border is visible
2- more than one corticated canal border is visible
3- all canal borders are completely visible.
Curved sections (panoramic-like reconstruction):
0- canal's superior and inferior corticated borders are invisible
1- 1/3rd of the canal's corticated borders are visible
2- 2/3rd of the canal's corticated borders are visible
3- superior and inferior corticated canal borders are wholly visible.
CBCT images taken at four voxel settings and in two sections are shown in [Figure 1]a, [Figure 1]b, [Figure 1]c, [Figure 1]d, [Figure 1]e, [Figure 1]f, [Figure 1]g, [Figure 1]h. The procured data were sent to the statistician for analysis.
|Figure 1: CBCT images at various voxel sizes and imaging sections (a) 0.2 mm (curved) (b) 0.25 mm (curved) (c) 0.3 mm (curved) (d) 0.5 mm (curved) (e) 0.2 mm (coronal) (f) 0.25 mm (coronal) (g) 0.3 mm (coronal) (h) 0.5 mm (coronal)|
Click here to view
Statistical tests were conducted using SPSS software (Chicago, USA) version-17 (Microsoft Windows). The data obtained were not normally distributed and were ordinal data. Hence, non-parametric tests were performed. They were represented as mean as well as standard deviation. Furthermore, P < 0.05 was deemed statistically significant.
Tests were the following:
- Mann–Whitney U test: compared the variables among CBCT images – coronal and curved sections; shown in [Figure 2] (specialist 1) and 3 (specialist 2).
- Kruskal–Wallis test: checked whether CBCT's accuracy varied among the voxel settings, depicted in [Figure 4] (coronal) and 5 (curved).
- Post-hoc Tukey HSD multiple-comparison test: examined the variation within every voxel size's potential; at both sections, as shown in [Table 1] (coronal) and 2 (curved).
- Kappa statistics: checked inter-observer variability between two observers. This is illustrated in [Table 3].
|Figure 2: Voxel size Performance in two image sections (Observer 1). Voxel Size [1-0.5 mm, 2-0.3 mm, 3-0.25 mm, 4-0.2 mm]|
Click here to view
|Figure 3: Voxel size Performance in two image sections (Observer 2). Voxel Size [1-0.5 mm, 2-0.3 mm, 3-0.25 mm, 4-0.2 mm]|
Click here to view
|Table 1: Multiple comparison within the voxel sizes and observers - coronal section|
Click here to view
|Table 2: Multiple comparison within the voxel sizes and observers – curved section|
Click here to view
| Results|| |
[Figure 2] and [Figure 3] authenticated that both specialists' interpretations signified statistical evidence of appreciable distinction between the two sections for the designated voxel settings (P < 0.05). Coronal sections/slices grossly showcased finer performance.
[Figure 4] and [Figure 5] suggested a considerable variation statistically, in every voxel size's potentiality, at both sections (P < 0.05 either of the assessors). In comparison, the assessors reported an unequalled MC visibility score for voxel size 0.2 mm.
|Figure 4: Comparison among voxel sizes (Coronal). Voxel Size [1-0.5 mm, 2-0.3 mm, 3-0.25 mm, 4-0.2 mm]|
Click here to view
|Figure 5: Comparison among voxel sizes (Curved). Voxel Size [1-0.5 mm, 2-0.3 mm, 3-0.25 mm, 4-0.2 mm]|
Click here to view
Tables 1 and 2 elucidated that the voxel sizes 0.5 and 0.25 mm, 0.5 and 0.2 mm presented notable statistical variation (P < 0.05 – each specialist; per section). Besides, in curved slices, prominent statistical differentiation was reported the voxel parameters between 0.3 and 0.2 mm (P < 0.05 – first assessor); 0.3 and 0.5 mm (P < 0.05 –second assessor).
Table 3 disclosed an excellent agreement between both observers in coronal sections for voxel 0.25 mm (1.000 – third molar region) and better for voxel 0.2 mm (0.581– second premolar area, 0.571 – first molar region). In curved sections, agreement between them was excellent for voxel 0.2 mm (0.800 – first molar area); better for voxel 0.25 mm (0.625 – third molar region) and for voxel 0.2 mm (0.636 – mental foramen region).
| Discussion|| |
The MC would exhibit increased visibility in dry mandibles because of being devoid of soft tissue structures. In living human mandibles, these soft tissues bring about increased scattered radiation, negatively affecting the diagnostic imaging's accuracy., To account for this discrepancy, a soft tissue simulator was used in the present study. Differences between right and left MC's visibility were not investigated in this study, because literature suggests that there is no variation in bilateral MC structures.,,
A study by Kamburoğlu et al. showed that inter-observer agreement seemed to be much higher in CBCT acquisitions in comparison with other imaging modalities. In our study too, observers' agreement was high (voxel sizes 0.2 and 0.25 mm) for the acquired images.
In routine clinical practice, more than one image section should be assessed, to identify the MC. Therefore, this study included two imaging slices, such as coronal and curved. For both observers, images at coronal sections exhibited superior performance in accomplishing the diagnostic task. An earlier study showed that MC visibility was similar in two image sections: ortho-radial and oblique; this was attributed to the smaller voxel setting (0.2 mm), which resulted in excellent canal visibility in both sections.
In our study, when multiple comparisons were done for specific voxel sizes, there were statistically significant differences among them. Some past studies,,, were not consistent with this study outcome (voxel settings demonstrated insignificant statistical distinction), whereas few others,, showed similar results. Heterogeneity observed in these studies could be owing to: use of different imaging equipment, varying acquisition techniques, differing samples, study designs, or diagnostic tasks.
Diagnostic radiography should incorporate appropriate techniques that would maintain the radiation dose well below conservative limits, as recommended by ICRP (International Commission on Radiological Protection), 2007. The higher the settings, the prettier the image, but sometimes, even at lower settings, it is feasible to provide basic diagnostic information. From the current study, it is obvious that although 0.2 and 0.25 mm voxel settings showed remarkable performance in all the image slices, there were insignificant statistical differences among 0.2, 0.2, 5, and 0.3 mm voxel sizes (coronal slices). Thus, by considering radiation exposure concerns, 0.3 mm voxel size would be more reasonable for this study's objective, and the preferable imaging section to track MC would be the coronal section.
Limitations of this study were as follows:
- Scanning was performed using single equipment; parameters can vary with every CBCT unit.
- Laptop employed in this study might have influenced the observers' performances, although evidence shows that diagnostic accuracy is unaffected by LCD monitors/laptops of different contrast resolutions.
These factors could have possibly influenced this particular study's results. Reproducing similar results considering these factors is questionable and necessitates further experiments. This study contributes as an addition to the existing literature, which has shown inconclusive results. This information could be prospectively used in future research (in vivo quantitative studies assessing CBCT imaging parameters) to arrive at a convincing and definitive conclusion.
| Conclusion|| |
The potential benefits of employing CBCT imaging modality in dentistry should outweigh the risks. Apparently, by taking radiation risks into consideration, 0.3 mm voxel size would be more reasonable for this study's diagnostic task, and the appropriate imaging section would be coronal section.
The authors would be delighted in thanking the institution for their guidance and support; and also the biostatistician for aiding in statistical analyses.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Oliveira-Santos C, Capelozza AL, Dezzoti MS, Fischer CM, Poleti ML, Rubira-Bullen IR. Visibility of the mandibular canal on CBCT cross-sectional images. J Appl Oral Scl 2011;19:240-3.
Kim TS, Caruso JM, Christensen H, Torabinejad M. A comparison of cone-beam computed tomography and direct measurement in the examination of the mandibular canal and adjacent structures. J Endod 2010;36:1191-4.
Miles MS, Parks ET, Eckert GJ, Blanchard SB. Comparative evaluation of mandibular canal visibility on cross-sectional cone-beam CT images: A retrospective study. Dentomaxillofac Radiol 2015;45:20150296. doi: 10.1259/dmfr. 20150296.
Waltrick KB, de Abreu Junior MJ, Corrêa M, Zastrow MD, D'Avila Dutra V. Accuracy of linear measurements and visiblity of the mandibular canal of cone beam computed tomography images with different voxel sizes: An in vitro
study. J Periodontol 2013;84:68-77.
Puri A, Verma P, Mahajan P, Bansal A, Kohli S, Faraz SA. CBCT evaluation of the vital mandibular interforaminal anatomical structures. Ann Maxillofac Surg 2020;10:149-57. [Full text]
Fokas G, Vaughn VM, Scarfe WC, Bornstein MM. Accuracy of linear measurements on CBCT images related to presurgical implant treatment planning: A systematic review. Clin Oral Implants Res 2018;29:393-415.
Pauwels R, Beinsberger J, Stamatakis H, Tsiklakis K, Walker A, Bosmans H, et al
. Comparison of spatial and contrast resolution for cone-beam computed tomography scanners. Oral Surg Oral Med Oral Pathol Oral Radiol 2012;114:127-35.
Ganguly R, Ramesh A, Pagni S. The accuracy of linear measurements of maxillary and mandibular edentulous sites in cone-beam computed tomography images with different fields of view and voxel sizes under simulated clinical conditions. Imaging Sci Dent 2016;46:93-101.
Torres MG, Campos PS, Segundo NP, Ribeiro M, Navarro M, Crusoé-Rebello I. Evaluation of referential dosages obtained by Cone-Beam Computed Tomography examinations acquired with different voxel sizes. Dental Press J Orthod 2010;15:42-3.
Hekmatian E, Jafari-Pozve N, Khorrami L. The effect of voxel size on the measurement of mandibular thickness in cone-beam computed tomography. Dent Res J 2014;11:544-8.
Ibrahim N, Parsa A, Hassan B, van der Stelt P, Aartman IH, Wismeijer D. The effect of scan parameters on cone beam CT trabecular bone microstructural measurements of the human mandible. Dentomaxillofac Radiol 2013;42:20130206. doi: 10.1259/dmfr. 20130206.
Alabdulwahid A, Alfaleh W. Identification of mandibular canal in cone beam computed tomography plane with different voxel sizes. SaudiDent J 2020;32:403-9.
Schropp L, Alyass NS, Wenzel A, Stavropoulos A. Validity of wax and acrylic as soft-tissue simulation materials used in in vitro
radiographic studies. Dentomaxillofac Radiol 2012;41:686-90.
Grauer D, Cevidanes LS, Proffit WR. Working with DICOM craniofacial images. Am J Orthod Dentofacial Orthop 2009;136:460-70.
Shokri A, Shakibaei Z, Langaroodi AJ, Safaei M. Evaluation of the mandibular canal visibility on cone-beam computed tomography images of the mandible. J Craniofac Surg 2014;25:e273-7.
Kamburoğlu K, Yeta EN, Yılmaz F. An ex vivo comparison of diagnostic accuracy of cone-beam computed tomography and periapical radiography in the detection of furcal perforations. J Endod 2015;41:696-702.
Alkhader M, Hudieb M, Jarab F, Shaweesh A. The visibility of mandibular canal on orthoradial and oblique CBCT slices at molar implant sites. Biotechnol Biotechnol Equip 2016;30:770-6.
Yilmaz F, Sonmez G, Kamburoglu K, Koc C, Ocak M, Celik HH. Accuracy of CBCT images in the volumetric assessment of residual root canal filling material: Effect of voxel size. Niger J Clin Pract2019;22:1091-8.
] [Full text]
Afkhami F, Ghoncheh Z, Khadiv F, Kaviani H, Shamshiri AR. How does voxel size of cone-beam computed tomography effect accurate detection of root strip perforations. Iran Endod J 2021;16:43-8.
Pour DG, Sedaghati A, Shamshiri AR. Effect of resolution and bit depth on inferior alveolar canal visualization on exported mandibular cone-beam computed tomography images. J Oral Maxillofac Surg 2020;78:731-7.
Ashmawy MS, Abou-Khalaf AA, Mostafa RA. Effect of voxel size on the accuracy of nerve tracing module of cone beam computed tomography images. Egypt Dent J2017;63:2403-12.
Protection R. ICRP publication 103. Ann ICRP 2007;37:2.
Palomo JM, Rao PS, Hans MG. Influence of CBCT exposure conditions on radiation dose. OralSurg Oral Med Oral Pathol Oral Radiol Endod 2008;105:773-82.
Nemtoi A, Czink C, Haba D, Gahleitner A. Cone beam CT: A current overview of devices. Dentomaxillofac Radiol 2013;42:20120443.
Al-Ekrish AA, Ekram MI, Al Faleh W, Alkhader M, Al-Sadhan RE. The validity of different display monitors in the assessment of dental implant site dimensions in cone beam computed tomography images. Acta Odontol Scand 2013;71:1085-91.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2], [Table 3]