Introduction

In addition to properties, such as resilience, hardness, strength, and fracture resistance, dentin thickness is important for a successful treatment [1]. Dentin thickness affects the pulp vitality and tooth durability. Reduction in residual dentin thickness is associated with a reduction in number of odontoblasts, and can lead to pulpal inflammation, enabling bacteria and their byproducts to easily access the pulp chamber [2].

The residual radicular dentin thickness has a pivotal role in successful root canal preparation process, but over-instrumentation of the root canal system can weaken the root canal walls [3]. Majority of endodontically treated teeth have lost a great portion of their structure, and may require an intracanal post for retention of their coronal restoration [4]. Post space preparation process can also lead to excessive removal of tooth structure and decrease tooth’s fracture resistance [5] and increase the risk of root perforation [6]. For that reason, a minimum of 1 mm of radicular dentin should remain to prevent incidents, such as strip perforation and vertical root fracture [7]. Therefore, knowledge about the root canal anatomy and residual dentin thickness at different parts of the root is imperative.

The highest rate of endodontic treatment failures is found in the maxillary first molars, due to their complex anatomy, number of canals, difficult access for cavity preparation, and canal orifice negotiation (especially the second mesiobuccal canal) [8, 9]. In primary or se­condary endodontic treatment of maxillary molars, attention paid to the mesiobuccal root canal complexity and danger zone located at the distal wall of mesiobuccal root, can maximize the likelihood of a successful endo­dontic treatment [10, 11].

Radiography is the most commonly recommended modality for measurement of residual dentin thickness.

Periapical radiography is among the most commonly used imaging modalities. However, it provides a two-dimensional (2D) image of three-dimensional (3D) structures with their superimposition, and cannot provide accurate information about the buccal and palatal surfaces [12].

Cone beam computed tomography (CBCT) enables anatomical assessment of 3D structures with high accuracy, clarity, and lower distortion than 2D modalities. Also, there are no difficulties of image magnification and superimposition of 2D images in 3D CBCT [13]. CBCT is often requested as an adjunct imaging modality for cases with endodontic treatment complications or incidents, or as primary imaging modality for preoperative assessment of complex cases. CBCT is an accurate and reliable method for most linear measurements; however, its relatively high radiation dose compared to 2D imaging is a drawback. Nonetheless, it may be recommended when conventional imaging modalities fail to provide accurate information, or deliver insufficient or misleading data for treatment planning [14].

Literature is controversial regarding the effects of field of view (FOV) and voxel size on measurement accuracy of CBCT [15, 16]. A previous study reported that FOV and voxel size can affect the accuracy of CBCT linear measurements, such that the accuracy of linear measurements was higher in smaller FOV and voxel size [15]. However, other research reported that reduction of FOV and voxel size was not associated with an increase in the accuracy of CBCT linear measurements [16, 17]. While CBCT provides reliable linear measurements, changes in the exposure setting parameters, including field of view and voxel size, can affect its accuracy. Given the limited number of studies comparing the performance of different CBCT scanners with varying FOVs, especially in relation to dentin thickness measurement, this study aimed to address this gap.

Objectives

Considering the scarcity of studies investigating the simultaneous effects of FOV and type of CBCT scanner on the accuracy of linear measurements, this study aimed to compare the accuracy of two CBCT scanners with two FOVs for the measurement of radicular dentin thickness of maxillary first molars.

Material and methods

This in vitro study included 50 maxillary first molars extracted for purposes not related to this study, i.e., pe­riodontal problems. The study protocol was approved by the Ethics Committee of Hamadan University of Medical Sciences (approval number: IR.UMSHA.REC.1401.875).

Sample size

Sample size was calculated as 50 in each group using a sample size calculation formula, assuming 95% confidence interval, 80% study power, and mean expected difference in dentin thickness as 0.12 mm with a standard deviation of 0.21 mm.

Eligibility criteria

Maxillary first molars with straight roots or minimal curvature (< 10 degrees) were considered. The curvature was measured as the angle formed between two hypotheti­cal lines connecting the apex and the orifice of the canal to the initiation point of the curvature. Teeth initially underwent a digital periapical radiography with phosphor plates, to ensure the presence of straight or minimally curved and fully developed roots. Whereas roots with any evidence of intrapulpal calcification, root fracture, internal or external root resorption, previous endodontic treatment, and root canal obstruction, were excluded.

Data collection

The collected teeth were disinfected and decoronated at the cementoenamel junction by using a #245 diamond fissure bur. Pulp tissue was gently removed with a #15 hand K-file (Micro-Mega, Besancon, France) under irrigation with 5.25% NaOCl and sufficient pressure using a 30-gauge needle. To simulate clinical setting and enhance radiography, the roots were embedded in paraffin blocks. Next, the samples underwent CBCT scanning: once with CS9600 CBCT scanner (Carestream Dental, Paris, France) using 4 × 4 cm and 5 × 8 cm FOVs, 60 kVp tube potential, 6 mA tube current, and 75 µm voxel size, and once with Cranex 3D CBCT scanner (Soredex, Tuusula, Finland) using 4 × 6 cm and 6 × 8 cm FOVs, 60 kVp tube potential, 6 mA tube current, and 136 µm
and 200 µm voxel sizes (Table 1). After scanning, volume­tric data were saved in DICOM format, while OnDemand 3D software (Cybermed Inc., Seoul, South Korea) was employed for volumetric image reconstruction. CBCT scans were anonymized and randomly coded using a computer-generated random number list before analysis.
Two observers were blinded to the scanner type and FOV setting during measurements to minimize bias. Cross-sectional images with 1-mm slice thickness and 0.5-mm slice interval were reconstructed, and the radicular dentin thickness at the desired points was measured on axial sections by two observers, i.e., an oral and maxillofacial radiologist as the first observer and a trained senior dental student as the second observer. The images were viewed on a 20-inch monitor (LG, Seoul, South Korea), with desired brightness, contrast, and magnification; the measurements were made by the two observers. Distance between the root center and root surface of the first mesiobuccal and palatal canals was measured at the level of furcation, 2 mm below the furcation, and 4 mm below the furcation (Figure 1). After 2-week interval, 20% of the samples underwent a second measurement by the two observers to determine intraobserver agreement.

Next, the roots were removed from the paraffin block, and three horizontal lines were drawn on the first mesiobuccal and palatal roots. The first line passed through the root furcation, and the second and third lines were drawn at 2 mm and 4 mm apical to the furcation, respectively. Subsequently, the roots were sectioned hori­zontally at the lines by using a double-blade diamond disc, and the distance between the canal center and root surface was measured by a digital caliper (Mitutoyo, Kawasaki, Japan) with 0.01 mm accuracy to serve as the gold standard. The digital caliper was standardized before measurement sessions according to the manufacturer’s instructions. Each measurement was performed twice at the same pre-marked horizontal sections, and the average was recorded to ensure reliability (Figure 2). Radicular dentin thickness measured on CBCT scans was then compared with the gold standard.

To ensure consistent measurement levels, three horizontal lines were marked on each root prior to sectioning, corresponding exactly to the furcation, and 2 mm and 4 mm below the furcation. These levels were used as fixed landmarks both in CBCT scans and in physical sectioning. The observers were trained to locate these exact levels using cross-sectional views aligned with these markers.

Statistical analysis

Reliability testing: Intraclass correlation coefficient (ICC) was calculated to assess the intra-observer agreement.

Comparison to the gold standard (paired t-tests): Radicular dentin thickness measured on CBCT scans was compared to the gold standard using paired t-test. Comparison of small and large FOVs in Cranex and Carestream was also performed with t-test.

Comparisons between scanner type and between FOV (ANOVA and Tukey’s post-hoc): ANOVA test was used to compare the mean radicular dentin thickness at the furcation and 2 mm and 4 mm below it as well as among the three methods (two CBCT scanners and digital caliper).

Correlation (Pearson’s coefficient): Pearson’s correlation analysis was performed to assess the correlation of two CBCT scanners with the gold standard. Values > 0.8 indicated excellent, and values > 0.6 and < 0.8 denoted moderate agreement.

Agreement visualization (Bland-Altman plot): Finally, Bland-Altman plot was drawn to compare the agreement between the methods.

All statistical analyses were carried out using SPSS version 21.0 (SPSS Inc., IL, USA) at 0.05 level of signi­ficance.

Results

The intraobserver agreement was excellent, with ICCs of 0.995 and 0.975 for the first and second observers, respectively. Interobserver ICCs were also high (> 0.890) across all FOVs and scanners, confirming strong consistency in the measurements. Therefore, the mean values reported by the two observers were used in the final ana­lysis (Table 2).

Table 3 presents the mean radicular dentin thickness measured at the furcation, and 2 mm and 4 mm below it, using the two CBCT scanners with small and large FOVs. Using small FOVs, the mean dentin thickness values ranged from 1.09 to 2.29 mm, while with large FOVs, the values ranged from 1.18 to 2.29 mm. The gold standard measurements obtained by digital caliper ranged from 0.86 to 1.67 mm.

ANOVA test revealed a significant difference in the mean radicular dentin thickness at the three measured areas and among the three methods (p < 0.001).

Pairwise comparisons (Tukey’s test) showed that for small FOVs, both the scanners significantly differed from the gold standard at most levels (p < 0.05) (Table 4).
Cranex and Carestream also differed from each other, except for the furcation and 2 mm below the furcation in the palatal root, and at 2 mm below the furcation in the mesiobuccal root. While for FOVs, both the scanners again differed significantly from the gold standard (p < 0.05), and there was no significant difference between the scanners at any level (Table 5).

Pearson’s correlation analysis (Table 6) demonstrated excellent agreement between CBCT measurements and the gold standard in most samples using small FOVs (> 0.8), with only five measurements in the mesiobuccal root showing moderate correlation (> 0.6 and < 0.8). With large FOVs, correlation was moderate to excellent. All differences were statistically significant (p < 0.05), indicating that the dentin thickness measured by the two scanners with both FOVs was moderately to highly close to the gold standard (positive coefficients).

Comparison of small and large FOVs in Cranex and Carestream scanners using t-test (Table 7) revealed that Carestream had a significantly higher accuracy in large FOV, whereas Cranex had a significantly higher accuracy in small FOV, with all differences being statistically significant (p < 0.001).

Comparison of small and large FOVs irrespective of the scanner type for the measurement of dentin thickness at different areas (Table 8), exhibited a significant difference only at 4 mm below the furcation (p = 0.007), where smaller FOV showed a significantly higher measurement accuracy. The difference at the furcation (p = 0.055) and 2 mm below the furcation (p = 0.088) was not significant between small and large FOVs; the two FOVs had an approximate difference of 0.03 mm. Although statistically significant, the differences observed were within the clinically acceptable margin for clinical purposes.

Considering that the voxel size was constant for both small and large FOVs in Carestream scanner (75 µm), larger FOV demonstrated superior measurement accuracy. In contrast, in Cranex scanner where the voxel size was smaller (136 µm) in small FOV, measurement accuracy was significantly greater than with larger voxel size (200 µm).

Overall, regardless of the scanner type, smaller voxel sizes consistently showed higher agreement with the gold standard compared to scans with larger voxel sizes, indicating that voxel size was a contributing factor to measurement accuracy.

Bland-Altman plots were employed to evaluate the agreement between CBCT measurements obtained from Carestream and Cranex scanners (with both large and small FOVs) and the gold standard. The mean differences (bias) across all conditions were within clinically acceptable ranges, with 95% limits of agreement showing relatively consistent distribution. The plots demonstrated no significant systematic bias, although a slight increase in variability was observed with smaller FOVs and at greater distances from the furcation area. Overall, there was a slight tendency for overestimation in CBCT measurements, particularly in deeper sections and with smaller FOVs, but the agreement remained acceptable for clinical use (Figures 3-6).

Discussion

This study compared the accuracy of two CBCT scanners with two FOVs for the measurement of radicular dentin thickness of maxillary first molars. The results showed that, although the measurement accuracy was higher (closer to the gold standard) in small FOV than large FOV, this difference was slight, and therefore clinically negligible. Elshenawy et al. [15] compared 50 × 50, 61 × 78, 78 × 78, and 78 × 150 mm FOVs of Cranex 3D, and stated that in the same voxel size, smaller FOVs provided a higher accuracy than larger FOVs. Simi­larly in the present study, small FOV of Cranex 3D yielded a higher accuracy than its large FOV. The diffe­rence between the two FOVs was 0.09 mm at the furcation area, 0.07 mm at 2 mm below the furcation, and 0.08 mm at 4 mm below the furcation in Cranex 3D. Thus, the small FOV enabled more accurate measurements than the large FOV. Ganguly et al. [16] reported that decreasing the size of FOV and voxel size was not associated with any increase in the accuracy of linear measurements. Nonetheless, they recommended using a smaller FOV for children to lower the radiation dose. Unlike the present study, they mounted the teeth in a skull covered with soft tissue, which can cause radia­tion scattering. The authors used I-CAT scanner with 16 × 13 cm FOV and Planmeca scanner with 5 × 8 cm FOV, and found no superiority of one scanner over the other in terms of measurement accuracy. Also, Anter et al. [18] compared 80 × 80, 100 × 100, and 100 × 200 mm FOVs, and reported no significant difference among them in measurement accuracy. However, since the risk of distortion due to dental restorations exists in the clinical setting, they recommended using a small FOV to prevent artifacts in the oral environment. Kamburoğlu et al. [17] compared 40 × 40, 60 × 60, and 100 × 100 mm FOVs, and found no significant difference in the measured depth, width, or volume of lesions among the three FOVs; the CBCT measurements were associated with the gold standard. Moshfeghi et al. [19] evaluated the linear measurement accuracy of NewTom VGi CBCT scanner in axial and coronal planes with 0.3- and 0.15-mm voxel sizes (small and large FOVs). Unlike the present study, they mounted the teeth in a dry skull to increase contrast. The authors concluded that CBCT was highly accurate and reproducible for linear measurements of different parts of the maxillofacial region in the coronal and axial planes. They obtained CBCT scans with a larger voxel size for measurement of linear distances. Both voxels showed high accuracy in their study, but the authors recommended a method with lower radiation dose and shorter scanning time.

The present results also showed overestimation of values by CBCT scanners, although it was clinically negligible (the difference with the gold standard was 0.41 mm for the small, and 0.44 mm for the large FOV). Bunn et al. [20] also reported significant overestimation of radicular dentin thickness by CBCT measurements compared to the gold standard, which was greater in the lingual surface than other areas, but not significan­tly. They added that despite the magnification and distortion caused by CBCT, its accuracy was optimal for measurement of dentin thickness. Ganguly et al. [16] stated that CBCT measurements overestimated the values in 40.08%, underestimated in 57.74%, and precisely measured the values in 2.18% of the cases, compared to physical measurements. Van Dessel et al. [21] compared 7 CBCT scanners, one multislice computed tomography, and one microcomputed tomography (gold standard), and found that all CBCT and multislice computed tomography scanners overestimated the values compared to the gold standard. Also, the measurement accuracy of Cranex 3D was higher than that of Carestream 3D in their study. In the present research, Cranex 3D had a higher measurement accuracy than Carestream in the small FOV. However, Cranex 3D and Carestream had a very good correlation (agreement) with each other, which was higher for the small FOV.

The use of CBCT as an adjunctive imaging modality in endodontics plays a crucial role in diagnosis, treatment planning, and follow-up process. It is particularly valuable in evaluating complex root canal anatomy, delineating the extent and boundaries of periapical lesions, detecting root fractures, and identifying root treatment failures. All of these applications require high diagnostic accuracy [22]. As demonstrated in previous studies, as well as in the present investigation, smaller FOVs provide superior linear measurement accuracy and produce fewer image artifacts in CBCT imaging [23, 24]. Therefore, in accordance with the ALARA principle [25], the use of smaller FOVs is clinically more justified in most endodontic applications, where the area of interest is typically limited to one or a few teeth within a single quadrant.

Strict eligibility criteria and difficult sectioning of the teeth that resulted in exclusion and replacement of some of the samples were among the limitations of this study. A systematic review is recommended to reach a definite conclusion regarding the linear measurement accuracy of CBCT scanners with different exposure settings.

Conclusions

Within the limitations of this in vitro study, the results showed that Cranex 3D CBCT scanner was more accurate than Carestream only in small FOV. Although the differences in measurement accuracy between scanners and FOVs were statistically significant, they were minor and likely not clinically meaningful. However, when accurate measurements are needed, a small FOV is preferred due to slightly higher accuracy and the added benefit of reduced radiation exposure.

Acknowledgments

The present study is derived from a thesis with refe­rence number: 14020205696, conducted in the Research and Technology Center of Hamadan University of Medical Sciences.

Disclosures

Author contributions: Conceptualization: A.S.; Metho­dology: F.Z., E.N.; Questionnaire adaptation: Not applicable; Investigation and data collection: F.Z., E.N.; Formal analysis: L.T.; Data curation: F.Z., E.N.; Writing of original draft: F.Z., E.N.; Writing – review and editing: A.S., S.H.; Supervision: M.F.; Project administration: S.H. All authors have read and agreed to the published version of the manuscript.

Funding: This research received no external funding.

Institutional Review Board statement: The study protocol was approved by the Ethics Committee of Hamadan University of Medical Sciences (approval number: IR.UMSHA.REC.1401.875, issued on 25 April 2023).

Informed consent statement: Not applicable.

Data availability statement: The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Acknowledgments: None.

Conflicts of interest: The authors declare no conflicts of interest.

AI use statement: No artificial intelligence tools were used in the preparation of this manuscript.

References