Abstract
The objective of this study was to make a systematic review on the impact of voxel size in cone beam computed tomography (CBCT)-based image acquisition, retrieving evidence regarding the diagnostic outcome of those images. The MEDLINE bibliographic database was searched from 1950 to June 2012 for reports comparing diverse CBCT voxel sizes. The search strategy was limited to English-language publications using the following combined terms in the search strategy: (voxel or FOV or field of view or resolution) and (CBCT or cone beam CT). The results from the review identified 20 publications that qualitatively or quantitatively assessed the influence of voxel size on CBCT-based diagnostic outcome, and in which the methodology/results comprised at least one of the expected parameters (image acquisition, reconstruction protocols, type of diagnostic task, and presence of a gold standard). The diagnostic task assessed in the studies was diverse, including the detection of root fractures, the detection of caries lesions, and accuracy of 3D surface reconstruction and of bony measurements, among others. From the studies assessed, it is clear that no general protocol can be yet defined for CBCT examination of specific diagnostic tasks in dentistry. Rationale in this direction is an important step to define the utility of CBCT imaging.
Keywords: Dentistry, Cone beam CT, Voxel size, Diagnostic outcome, Image quality
Introduction
Cone beam computed tomography (CBCT) is a relatively new imaging technology that provides multi-planar images in submillimeter resolution [1]. In the last years, CBCT has achieved wide acceptance in dentomaxillofacial imaging and has fundamentally replaced conventional tomography for several diagnostic tasks in dentistry [2, 3]. The main advantage of CBCT is its lower acquisition time and patient dose when compared to medical CT scanning [3–6].
In clinical practice, the image quality of CBCT scans and the ability of CBCT to display anatomic features and pathology is influenced by a number of variables such as the scanning unit, the field of view (FOV), examined object, examination time, tube voltage and amperage, and also spatial resolution defined by the voxel size [7]. The size of a voxel is defined by its height, width, and depth, and CBCT voxels are generally isotropic (the three parameters are equal) [8]. The voxel size of a 3D image is equivalent to the pixel resolution in 2D images, and, in this case, a resolution of 300 ppi (pixels per inch) would directly correlate to a voxel size of 0.085 mm. Images acquired in smaller voxel sizes, although “prettier” and sharper from a subjective point of view[9], will increase the radiation dose to the patient but might provide the same diagnostic outcome as lower resolution images [10]. Thus, it is important to ponder that the comparison of CBCT examinations with various voxel settings is relevant to understand the impact of the inherent image quality on the reliability and accuracy of the diagnostic outcome [11].
For medical CT examinations, settings or protocols for any application are commonly discussed in the literature [8, 12, 13]. In opposition to that, rationales for protocols and their impact on CBCT-based diagnosis are not yet available in dentistry. In an attempt to search for a rationale concerning protocols for settings in CBCT imaging, our objective was to perform a systematic review on the impact of voxel size variation on CBCT-based image acquisition, retrieving evidence regarding the diagnostic outcome of those images.
Material and Methods
The MEDLINE (PubMed) bibliographic database was searched from 1950 to May 2012 for reports comparing diverse CBCT voxel sizes. The search strategy was limited to English-language publications using the following combined terms in the search strategy: (voxel or FOV or field of view or resolution) and (CBCT or cone beam CT).
Studies comparing the influence of using two or more voxel sizes on CBCT-based diagnostic outcome qualified for inclusion. Only studies in which information regarding (1) image acquisition, (2) reconstruction protocols, (3) type of diagnostic task, and (4) presence of a gold standard for the true state of disease were selected. For studies based on categorical data, accuracy parameters such as sensitivity, specificity, likelihood ratio, or ROC curves should be present (at least one of them) to qualify for inclusion. For studies using quantitative data, agreement between measurements or accuracy of measurements should be present. All references were screened by one reviewer, and data extraction was verified separately by all authors.
Results
Review Search Results
The search strategy yielded 689 publications in MEDLINE (PubMed). The initial screening of the articles was conducted using the abstracts and keywords, but when these were unclear or unavailable, the full text was used.
Screening yielded 25 citations that potentially met the inclusion criteria, but an additional five papers were excluded. In one of them, although images were generated using diverse voxel sizes, this parameter was not assessed and interpreted in the results, which referred to other parameters (e.g., unit, FOV, and section thickness) [1]. One study tested various voxel sizes used to acquire the images (0.30 and 0.25 mm) but reconstructed the images using the same final resolution (0.25 mm), neglecting a direct comparison of voxel size [14]. The other studies were excluded because no gold standard was employed [15, 16] or because the evaluation was merely subjective of image quality [7].
Twenty publications were identified that assessed the influence of voxel size on CBCT-based diagnostic outcome, and in which the methodology/results fit the inclusion criteria. All included studies were ex vivo. Only one of the included studies cited the contrast settings used to evaluate the images [17]. Various diagnostic tasks were assessed, and according to it, the gold standard varied.
Among the 20 studies, 12 were based on categorical data. Of those, four assessed the detection of root fractures [18–21] (and used the clinical truth (an induced root fracture) as the gold standard), three assessed internal or external root resorption [22–24] also using the clinical truth as the gold standard, three assessed caries lesion detection [25–27] (and used sectioning and histology of the teeth as the gold standard), one assessed the detection of erosions in the temporomandibular joint [28] (using the clinical truth as the gold standard), and one assessed the presence of root canals in molars [29] using sectioning and histology of the teeth as the gold standard. The remaining eight studies were based on numerical outcomes measured in the images. Three studies measured 3D surface reconstructions [17, 30, 31] (and used micro-CT of the teeth as the gold standard), two measured the height and thickness of alveolar bone [11, 32], one assessed linear bone measurements (distance between established anatomical landmarks) [33], one measured the thickness of the soft tissue of the face [34], and one measured tooth and root length [35], and these last five studies used caliper measurements of the skulls as the gold standard. Six studies further compared the CBCT images in different resolutions to images acquired using intraoral radiographic systems [18, 21, 25–27, 35].
Information regarding the protocols of the studies is shown in Table 1. A compact overview of the assessment methods and results of the studies is shown in Table 2. From the studies assessed, it is clear that no protocol is yet defined for any of the evaluated diagnostic tasks.
Table 1.
Summary of the diagnostic tasks, sample definition, and acquisition protocol of the images in the included studies
| Study | Diagnostic task | Sample size | Unit (kV, mA; nd = not described) | Voxel size mm (FOV cm; nd = not described) |
|---|---|---|---|---|
| Wenzel at al. [21] | Detect transverse root fractures | 69 human teeth, 34 with root fractures and 35 without | i-CATa (nd) | 0.125 (nd), 0.25 (nd) |
| Melo et al. [19] | Detect longitudinal root fractures | 180 endodontically prepared human teeth, 90 with root fractures and 90 without. Further, each group of 90 teeth was divided in subgroups of 30 teeth—unfilled, filled with gutta-percha, or filled with a gold-alloy post | i-CATa (120, 3–8) | 0.2 (8), 0.3 (8) |
| Ozer [20] | Detect vertical root fractures | 60 human teeth, 30 with root fractures and 30 without | i-CATa (120, 5) | 0.125 (4), 0.2 (4), 0.3 (4), 0.4 (4) |
| da Silveira et al. [18] | Detect vertical root fractures | 60 single-rooted human teeth, 30 with root fractures and 30 without. Further, each group of 30 teeth was divided in subgroups of 10 teeth—unfilled, filled with gutta-percha, or filled with gutta-percha and a metal post | i-CATa (120, 3–8) | 0.2 (8), 0.3 (8), 0.4 (8) |
| Liedke et al. [24] | Detect external root resorption | 59 human teeth, with surgically simulated root resorption | i-CATa (120, 3–8) | 0.2 (8), 0.3 (8), 0.4 (8) |
| Kamburoglu and Kursun [23] | Detect internal root resorption | 60 human teeth, with surgically simulated root resorption | Iluma Ultrab (120, 3.8) 3D Accuitomoc (65, 2) | 0.1 (14 × 21), 0.2 (14 × 21), 0.3 (14 × 21) 0.125 (6), 0.16 (8) |
| Dalili et al. [22] | Detect external root resorption | 16 human tooth roots, with surgically simulated root resorption | NewTom VGd (110, automated) | 0.125–0.150 (15), 0.2–0.24 (23) |
| Haiter-Neto et al. [26] | Detect approximal and occlusal caries lesions | 200 approximal tooth surfaces (126 sound and 74 diseased) and 100 occlusal tooth surfaces (6 sound and 94 diseased) | NewTom 3Gd (110, automated) 3DX Accuitomoc (60, 3) | 0.16 (15), 0.25 (23), 0.36 (30) 0.125 (4) |
| Kamburoglu et al. [27] | Detect occlusal caries lesions | 130 occlusal tooth surfaces (61 sound and 69 diseased) | Iluma Ultrab (120, 3.8) | 0.1 (nd), 0.2 (nd), 0.3 (nd) |
| Cheng et al. [25] | Detect approximal caries lesions | 90 approximal tooth surfaces (58 sound and 32 diseased) | ProMax 3De (84, 6 and 12) DCT Prof (90, 7.5) | 0.16 (8), 0.32 (8) 0.2 (16 × 7), 0.3 (16 × 7) |
| Librizzi et al. [28] | Detect TMJ erosion | 16 TMJs, with surgically simulated erosion spots | CB MercuRayg (120, nd) | 0.2 (15), 0.3 (23), 0.4 (30) |
| Bauman et al. [29] | Detect mesiobuccal canals in maxillary molars | 24 human maxillary molars, 22 with a second mesiobuccal canal | i-CATa (120, 3–8) | 0.125 (6 × 8), 0.2 (6 × 17), 0.3 (6 × 17), 0.4 (6 × 17) |
| Al-Rawi et al. [30] | Measure 3D reconstructions | 2 fully dentate dry human jaws (maxilla and mandible) | Scanora 3Dh (85, 8) | 0.133 (6), 0.2 (7.5 × 10), 0.25 (7.5 × 14.5) |
| Maret et al. [31] | Measure 3D reconstructions | 70 human teeth | Kodak 9500 3Di (90, 10) Kodak 9000 3Di (85, 2) | 0.2 (9 × 15), 0.3 (18 × 20) 0.076 (5 × 3.7) |
| Damstra et al. [17] | Measure 3D reconstructions | 10 dry human mandibles | KaVo 3D eXamj (120, nd) | 0.25 (10), 0.4 (10) |
| Sun et al. [32] | Measure bone height and thickness | 11 pig maxillae | i-CATa (nd) | 0.25 (nd), 0.4 (nd) |
| Patcas et al. [11] | Measure bone height and thickness | 8 cadaver heads | KaVo 3D eXamj (120, 5) | 0.125 (nd), 0.4 (nd) |
| Torres et al. [33] | Measure bone linearly | 8 dry human mandibles | i-CATa (120, nd) | 0.2 (6), 0.25 (6), 0.3 (6), 0.4 (6) |
| Fourie et al. [34] | Measure soft tissue thickness | 7 cadaver heads | KaVo 3D eXamj (nd) | 0.3 (nd), 0.4 (nd) |
| Sherrard et al. [35] | Measure tooth and root length | 52 human teeth | i-CATa (120, 3–8) | 0.2 (nd), 0.3 (nd), 0.4 (nd) |
aImaging Sciences International Inc., Hatfield, PA, USA
b3M Imtec, Ardmore, OK, USA
cJ Morita Mfg. Corp., Kyoto, Japan
dQuantitative Radiology, Verona, Italy
ePlanmeca Oy, Helsinki, Finland
fVATECH Co. Ltd., Yongin-Si, South Korea
gHitachi Medical, Kyoto, Japan
hSoredex Oy, Tuusula, Finland
iCarestream, Marne-la-Vallee, France
jKaVo Dental AG, Brugg, Switzerland
Table 2.
Summary of the assessment methods and results of the included studies
| Study | Assessment methods | Results |
|---|---|---|
| Wenzel at al. [21] | Six observers scored the images “with” or “without” fracture and fracture location (coronal, middle, apical third of the root). Images from an intraoral PSP system were also evaluated. The clinical truth was the gold standard | Sensitivity was higher in high-resolution CBCT images than in both the lower resolution CBCT and PSP images. The difference between CBCT in high and low resolution was highly significant, and also between CBCT in high resolution and PSP images. There was no difference between CBCT in the low resolution and PSP images |
| Melo et al. [19] | One observer scored the images “with” or “without” fracture. The clinical truth was the gold standard | Sensitivity was significantly higher for the 0.2-mm voxel resolution. Specificity values for CBCT were similar and did not depend on voxel resolution. The presence of gutta-percha or cast-gold posts reduced the overall sensitivity and specificity for both voxel resolutions |
| Ozer [20] | Three observers scored the teeth “with” or “without” fracture, in duplicate, based on images of three planes (axial, frontal, and sagittal). The clinical truth was the gold standard | Sensitivity and specificity were similar for all tested voxel sizes; the accuracy was inversely proportional to the voxel size, but with no statistical significance. The positive likelihood ratio was higher with 0.125- and 0.2-mm voxel size than with 0.3- and 0.4-mm voxel size |
| da Silveira et al. [18] | Three observers scored the images “with” or “without” fracture, in duplicate. Images from an intraoral radiographic method (conventional film), acquired with three different horizontal angles (orthogonal, mesial, and distal-angulated) were also evaluated. The clinical truth was the gold standard | Sensitivity, specificity, and accuracy were similar for 0.2 and 0.3-mm voxel size, for unfilled roots. Conventional radiographs showed similar results compared with 0.2 and 0.3-mm voxel CBCT in roots without endodontic treatment and metallic post. In teeth with root canal treatment or a post, the accuracy was higher for 0.2-mm voxel size |
| Liedke et al. [24] | One observer scored the images regarding the resorption size (small, medium, or large) and its location. The clinical truth was the gold standard | Sensitivity and specificity were similar and independent of voxel size. Likelihood ratio was 16 for 0.3-mm voxel, 12 for 0.2-mm voxel and 6.4 for 0.4-mm voxel size |
| Kamburoglu and Kursun [23] | Two observers scored the images, in duplicate, using a five-step scale to determine if there was internal resorption. The clinical truth was the gold standard | Intra- and inter-observer agreement was higher for the Accuitomo images. ROC curves showed the best results for Accuitomo 0.125-mm voxel size and the worst for Iluma 0.3 mm. No significant differences were found among the values for Accuitomo 0.125 mm, Accuitomo 0.16 mm, Iluma 0.1 mm and Iluma 0.2 mm |
| Dalili et al. [22] | One observer scored the images “with” or “without” resorption and its location (root surface/radicular third). The clinical truth was the gold standard | Diagnostic accuracy was high for the two tested modes of CBCT. The 0.125-mm voxel size was more accurate for small simulated resorption lesions located in the apical third and on lingual surfaces of teeth |
| Haiter-Neto et al. [26] | Six observers scored the images “with” or “without” caries lesions, using a five-step scale. Images from two intraoral radiographic methods (PSP and conventional film) were also evaluated. Histology was the gold standard | Diagnostic accuracy for detection of caries lesions was lower for NewTom 3G (in spite of the voxel size—0.16, 0.25, and 0.36 mm) than intraoral modalities and the 3DX Accuitomo in 0.125-mm voxel size. The Accuitomo 0.125-mm voxel size had a higher sensitivity than the intraoral systems for detection of lesions in dentine, but the overall true score was not higher |
| Kamburoglu et al. [27] | Four observers scored the images “with” or “without” caries lesions, in duplicate, using a five-step scale. Images from an intraoral CCD system were also evaluated. Histology was the gold standard | Voxel size did not affect the detection of occlusal caries. CBCT images based on 0.1-mm voxel size were the most, and intraoral images were the least accurate methods to diagnose superficial enamel, deep enamel, superficial dentine, and deep dentine caries |
| Cheng et al. [25] | Eight observers scored the images “with” or “without” caries lesions, in duplicate, using a five-step scale. Images from an intraoral PSP system were also evaluated. Histology was the gold standard | For enamel caries, no significant difference among CBCT images in 0.16, 0.32, 0.2, and 0.3-mm voxel size or among CBCT in all tested resolutions and PSP images was found. For dentinal caries, no significant difference among CBCT images in 0.16-, 0.32-, 0.2-, and 0.3-mm voxel size were found, but there was significant difference when comparing CBCT in all tested resolutions and PSP images |
| Librizzi et al. [28] | Two observers scored the images “with” or “without” erosion, in duplicate, using a five-step scale. The clinical truth was the gold standard | The agreement tended to be higher for 0.2-mm voxel size. The ROC analysis showed statistically significant difference between the 0.2 and the 0.4-mm voxel size |
| Bauman et al. [29] | Seven observers evaluated videos made with all images from each tooth, scoring the absence, presence of one or two mesiobuccal canals. Histology was the gold standard | Observers were able to detect the correct number of mesiobuccal canals as voxel size decreased. There was a significant difference in accuracy among the tested resolutions except the two highest ones (0.2- and 0.125-mm voxel size) |
| Al-Rawi et al. [30] | One observer superimposed the CBCT-based surface reconstruction of tooth crows to the micro-CT-based 3D reconstruction. Micro-CT (isotropic 25 μm resolution) was the gold standard | CBCT reconstructions were larger than their micro-CT counterparts. For the CBCT-based images, differences were found between the 0.25 and 0.2-mm voxel sizes, but not between the 0.2 and 0.133 mm |
| Maret et al. [31] | One observer measured the images, in duplicate. Micro-CT (isotropic 41 μm resolution) was the gold standard | CBCT voxel sizes of 0.2 and 0.3 mm were not significantly different from 0.076 mm and micro-CT 41 μm, but CBCT slightly underestimated the volumetric measurements |
| Damstra et al. [17] | One observer measured the images, in triplicate. Digital caliper measurements were the gold standard | Statistically, all methods were similar. CBCT underestimated the reference values for approximately 60 % and overestimated for approximately 30 % of the measurements |
| Sun et al. [32] | Two observers measured the images, in duplicate. Digital caliper measurements were the gold standard | Alveolar bone thickness interfered in the accuracy of the bone height measurements. When the thickness was greater than the voxel size, the distance was overestimated. When it was equal or smaller than the voxel size, the distance was underestimated. For bone walls thinner than the voxel size, measures from both tested voxel sizes were different from the gold standard |
| Patcas et al. [11] | One observer measured the images, in duplicate. Digital caliper measurements were the gold standard | Both voxel resolutions were accurate compared to the gold standard, but the 99 % confidence interval for the absolute measurement error was higher in the lower resolution images |
| Torres et al. [33] | One observer measured the images, in duplicate. Digital caliper measurements were the gold standard | The majority of the measurements in CBCT underestimated the true measurements. There was no statistical difference between the tested voxel sizes regarding the overall measurement error |
| Fourie et al. [34] | Two observers measured the images, in triplicate, based on ten anatomic landmarks. Digital caliper measurements were the gold standard | CBCT measurements were reliable when compared to the true measurements. There was a slight but definite difference in the facial soft tissue thickness measurements between the two voxel sizes |
| Sherrard et al. [35] | One observer measured the images in duplicate. Images from two intraoral radiographic methods (CCD and conventional film) were also evaluated. Digital caliper measurements were the gold standard | Intraoral radiography measurements were less accurate than CBCT measurements, underestimating true root length and overestimating tooth length, although without statistical significance. Tooth length measurements for the 0.2-mm voxel size were significantly larger than the other two voxel sizes. The differences between the truth and CBCT measurements were not statistically significant |
Discussion
Under clinical conditions, the subjective image quality of CBCT-based images and the ability of CBCT to display various features are influenced by a number of variables, including the unit itself, the FOV, the tube voltage and current, the voxel size, and other technical factors [7, 36]. Many of these parameters can be varied according to the diagnostic task, but still no protocols have been established for specific diagnostic tasks in dentistry. In this way, the selection of parameters related to image generation and manipulation in CBCT imaging, including the selection of voxel resolution, seems to have been performed almost arbitrarily (“best guess” or availability in the equipment) [37, 38]. In opposition to that, protocols standardizing the selection of these parameters would have a direct impact on the radiation dose that the patient receives during examination (since they interfere with the scanning time), and a better definition would be essential in respecting the “ALARA” [39]. The intention of our review was to seek for a rationale on the impact of voxel size in CBCT-based image acquisition in order to suggest protocols for various diagnostic tasks in dentistry.
It was common in the studies that FOV and voxel size were evaluated together. It is known that larger FOVs may provide less sharp reconstructions because of the greater beam angulation in the superior and inferior volume area and reduced contrast-to-noise ratio [11]. In this review, included studies focused on voxel size, rather than the FOV size. Thus, the influence of FOV size per se was not assessed. Moreover, the influence of the software used to reconstruct the images, which may interfere in the quality of the final reconstructed image, was also not assessed, since the studies were based only in the native software of each unit.
Voxel size can influence the characteristics of the final image in several ways. It may influence noise in the orthogonal sections of an image: the smaller the voxel size, the greater the noise, but of course, the higher the spatial resolution [30]. Depending on the voxel size, radiopaque structures can become invisible. This can be caused by the partial volume averaging effect, which is a common computed tomography artifact and occurs when a voxel lies on the borders of two objects of different densities. This voxel will then reflect the average density of both objects rather than the true value of either object [3]. This “invisibility” of some structures could also be caused by the limitations in contrast resolution related to CBCT units, which determines the ability to distinguish two objects of similar densities and in close proximity [40, 41].
The diagnostic tasks evaluated in the studies included in this review were diverse. Four studies assessed the detection of root fractures [18–21] using the clinical truth as a gold standard. Wenzel et al. [21] and Melo et al. [19] suggested that high-resolution CBCT (voxel size smaller than 0.2 mm) should be used when root fracture is suspected but not visualized in a periapical image. In opposition to that, Ozer [20] and da Silveira et al. [18] did not find significant differences among voxel sizes ranging from 0.125 to 0.4 mm. Three studies evaluated the detection of simulated root resorption, also using the clinical truth as a gold standard. While Liedke et al. [24] tested three voxel sizes (0.4, 0.3, and 0.2 mm) and showed that they produced similar results, Dalili et al. [22] and Kamburoglu and Kursun [23] showed that high-resolution images (voxel size lower than 0.16 mm) were more accurate for the detection of artificially created internal root resorption.
Three studies evaluated caries lesion detection [25–27] and used histology of the tooth sections as a gold standard. Images from intraoral 2D systems were included for comparison. While Kamburoglu and colleagues [27] showed that voxel size did not affect the detection of occlusal caries lesions, Haiter-Neto and coauthors [26] showed that 0.125-mm voxel size (but not 0.16, 0.25, and 0.36 mm) provided more accurate lesion depth estimates for approximal lesions when compared to the histological gold standard than did the intraoral images. Cheng and coauthors [25] found diverse results for intraoral and CBCT-based images for dentinal caries (independent of voxel size) and suggested that CBCT should be a secondary imaging modality to be used when conventional views provide controversial results. For the detection of simulated temporomandibular joint erosion, Librizzi and coauthors [28] showed that images acquired with voxel size of 0.2 mm were significantly more accurate than those acquired using 0.4 mm. Similar results were found regarding the detection of mesiobuccal maxillary canals in molars [29].
Of the three studies which measured the accuracy of 3D reconstructions quantitatively, two used micro-CT of the teeth as a gold standard [30, 31]. Both studies found that CBCT-based images, independent of the voxel size, underestimated, though not statistically significant, the volumes measured with micro-CT. Micro-CT suffers from artifacts similar to those observed in CBCT, since the scanning technology is comparable. Both CBCT and micro-CT suffer from beam hardening artifacts [42]. As an example, specific materials such as gutta-percha and intra-canal metallic posts may contribute to artifact formation, decreasing the diagnostic quality of the image [19, 43]. However, the influence of beam hardening artifacts is lower in micro-CT compared with CBCT because it operates at a higher kVp and offers more uniform filtration as well as a higher spatial resolution [30]. The other study assessing 3D reconstructions used digital caliper measurements as a gold standard [17] and found no difference between the CBCT measurements in images with 0.4- and 0.25-mm voxel resolution compared with the anatomic truth.
Regarding the bone measurements made in CBCT images and compared to digital caliper measurements (as the gold standard), Sun and coauthors [32] (evaluating bone thickness) showed that measurements in images with a voxel size of 0.25 mm were closer to the direct measurements than 0.4 mm images. Patcas and colleagues [11] (evaluating both bone height and width) showed that 0.4-mm voxel images provided results as accurate as 0.125-mm voxel images. This finding is also in agreement with Torres and coauthors [33], who did not find differences between voxel sizes of 0.2, 0.3, and 0.4 mm, when evaluating linear bone measurements.
Regarding the thickness of soft tissue at ten facial landmarks, CBCT measurements were reliable when compared to the digital caliper measurements used as the gold standard. There were differences in the measurements obtained in 0.3- and 0.4-mm voxel images [34]. For tooth and root length measurements, no differences were shown between the gold standard (caliper) and CBCT-based measurements, independent of the selected voxel size (0.2, 0.3, or 0.4 mm) [35].
The relationship between dose and image quality should always be part of the decision making process for the establishment of imaging protocols. Despite the benefits and improvements in CBCT, it is still a source of ionizing radiation to the patient [44]. Voxel size is not only of overriding importance in terms of image quality but is also directly connected to the scanning and reconstruction times of CBCT images, along with other factors such as FOV, amperage, and tube voltage [39].
The benefits of a shorter scanning time (i.e., lower radiation exposure and less patient movement) might outweigh the poorer resolution [17]. It is important to emphasize that the ALARA principle should always be applied; thus, the protocol must be tailored to each case. Without sacrificing image quality and adopting the ALARA principle, the ability to select various voxel settings would be helpful in reducing the radiation dose to the patient [20]. None of the included studies directly evaluated the effective dose associated with the use of different voxel sizes. In one of the studies, this evaluation was made, but it connected the variation of FOV and voxel size at the same time, making it impossible to discuss whether the variation was caused by voxel size or FOV. [28] Another previous study showed similar results that smaller FOVs resulted in lower effective radiation doses and suggested that in general smaller FOVs should be used for dental imaging and larger FOVs restricted to cases where a wide view is required; however, no relationship between FOV and voxel size was commented [45].
One point that should be noticed is that most (with the exception of two [18, 19]) of the studies included in this review argued that the ex vivo ideal conditions for CBCT image generation, excluding metallic restorative materials, soft tissue, and other parameters that could complicate the CBCT-based diagnosis, were drawbacks. These factors, especially metallic materials that produce beam hardening artifacts, may influence the image quality and thus the diagnostic accuracy when a patient is examined and should therefore be considered in future studies.
Conclusion
The number of studies assessing the impact of voxel size variation on the diagnostic outcome in CBCT imaging in dentistry is small. Focusing on the studies, which used a gold standard method as validation for the diagnostic outcome, the lack of systematic information is clear. Although studies dealing with categorical data showed a tendency towards more accurate results connected to higher voxel resolutions, it is not yet possible to suggest general protocols for the different diagnostic tasks, in which CBCT can be applied. Further clinical studies in this area are needed in order to allow the professional radiological society to develop detailed guidelines for the use of CBCT.
References
- 1.Qu X, Li G, Zhang Z, Ma X. Detection accuracy of in vitro approximal caries by cone beam computed tomography images. Eur J Radiol. 2011;79:24–27. doi: 10.1016/j.ejrad.2009.05.063. [DOI] [PubMed] [Google Scholar]
- 2.Scarfe WC, Farman AG, Levin MD, Gane D. Essentials of maxillofacial cone beam computed tomography. Alpha Omegan. 2010;103:62–67. doi: 10.1016/j.aodf.2010.04.001. [DOI] [PubMed] [Google Scholar]
- 3.Scarfe WC, Farman AG. What is cone-beam CT and how does it work. Dent Clin North Am. 2008;52:707–730. doi: 10.1016/j.cden.2008.05.005. [DOI] [PubMed] [Google Scholar]
- 4.Grauer D, Cevidanes LS, Proffit WR. Working with DICOM craniofacial images. Am J Orthod Dentofacial Orthop. 2009;136:460–470. doi: 10.1016/j.ajodo.2009.04.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tsiklakis K, Donta C, Gavala S, Karayianni K, Kamenopoulou V, Hourdakis CJ. Dose reduction in maxillofacial imaging using low dose cone beam CT. Eur J Radiol. 2005;56:413–417. doi: 10.1016/j.ejrad.2005.05.011. [DOI] [PubMed] [Google Scholar]
- 6.White SC, Pharoah MJ. The evolution and application of dental maxillofacial imaging modalities. Dent Clin North Am. 2008;52:689–705. doi: 10.1016/j.cden.2008.05.006. [DOI] [PubMed] [Google Scholar]
- 7.Kamburoglu K, Murat S, Kolsuz E, Kurt H, Yuksel S, Paksoy C. Comparative assessment of subjective image quality of cross-sectional cone-beam computed tomography scans. J Oral Sci. 2011;53:501–508. doi: 10.2334/josnusd.53.501. [DOI] [PubMed] [Google Scholar]
- 8.Hatcher DC. Operational principles for cone-beam computed tomography. J Am Dent Assoc. 2010;141:3–6. doi: 10.14219/jada.archive.2010.0359. [DOI] [PubMed] [Google Scholar]
- 9.Palomo JM, Rao PS, Hans MG. Influence of CBCT exposure conditions on radiation dose. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2008;105:773–782. doi: 10.1016/j.tripleo.2007.12.019. [DOI] [PubMed] [Google Scholar]
- 10.Davies J, Johnson B, Drage N. Effective doses from cone beam CT investigation of the jaws. Dentomaxillofac Radiol. 2012;41:30–36. doi: 10.1259/dmfr/30177908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Patcas R, Muller L, Ullrich O, Peltomaki T. Accuracy of cone-beam computed tomography at different resolutions assessed on the bony covering of the mandibular anterior teeth. Am J Orthod Dentofacial Orthop. 2012;141:41–50. doi: 10.1016/j.ajodo.2011.06.034. [DOI] [PubMed] [Google Scholar]
- 12.Anderson SW, Soto JA. Multi-detector row CT of acute non-traumatic abdominal pain: contrast and protocol considerations. Radiol Clin North Am. 2012;50:137–147. doi: 10.1016/j.rcl.2011.08.009. [DOI] [PubMed] [Google Scholar]
- 13.Miles KA. Perfusion CT for the assessment of tumour vascularity: which protocol. Br J Radiol. 2003;76:36–42. doi: 10.1259/bjr/18486642. [DOI] [PubMed] [Google Scholar]
- 14.Neves FS, Vasconcelos TV, Vaz SL, Freitas DQ, Haiter-Neto F. Evaluation of reconstructed images with different voxel sizes of acquisition in the diagnosis of simulated external root resorption using cone beam computed tomography. Int Endod J. 2012;45:234–239. doi: 10.1111/j.1365-2591.2011.01966.x. [DOI] [PubMed] [Google Scholar]
- 15.Hassan B, Couto Souza P, Jacobs R, de Azambuja Berti S, van der Stelt P. Influence of scanning and reconstruction parameters on quality of three-dimensional surface models of the dental arches from cone beam computed tomography. Clin Oral Investig. 2010;14:303–310. doi: 10.1007/s00784-009-0291-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Waltrick KB, de Abreu Junior MJ, Correa M, Zastrow MD, D'Avila Dutra V: Accuracy of linear measurements and visibility of the mandibular canal on cone-beam computed tomography images with different voxel sizes: an in vitro study. J Periodontol, 2012, doi:10.1902/jop.2012.110524 [DOI] [PubMed]
- 17.Damstra J, Fourie Z, Huddleston Slater JJ, Ren Y. Accuracy of linear measurements from cone-beam computed tomography-derived surface models of different voxel sizes. Am J Orthod Dentofacial Orthop. 2010;137:16–17. doi: 10.1016/j.ajodo.2009.08.023. [DOI] [PubMed] [Google Scholar]
- 18.da Silveira PF, Vizzotto MB, Liedke GS, da Silveira HL, Montagner F, da Silveira HE: Detection of vertical root fractures by conventional radiographic examination and cone beam computed tomography—an in vitro analysis. Dent Traumatol, 2012, doi:10.1111/j.1600-9657.2012.01126.x [DOI] [PubMed]
- 19.Melo SL, Bortoluzzi EA, Abreu M, Jr, Correa LR, Correa M. Diagnostic ability of a cone-beam computed tomography scan to assess longitudinal root fractures in prosthetically treated teeth. J Endod. 2010;36:1879–1882. doi: 10.1016/j.joen.2010.08.025. [DOI] [PubMed] [Google Scholar]
- 20.Ozer SY. Detection of vertical root fractures by using cone beam computed tomography with variable voxel sizes in an in vitro model. J Endod. 2011;37:75–79. doi: 10.1016/j.joen.2010.04.021. [DOI] [PubMed] [Google Scholar]
- 21.Wenzel A, Haiter-Neto F, Frydenberg M, Kirkevang LL. Variable-resolution cone-beam computerized tomography with enhancement filtration compared with intraoral photostimulable phosphor radiography in detection of transverse root fractures in an in vitro model. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2009;108:939–945. doi: 10.1016/j.tripleo.2009.07.041. [DOI] [PubMed] [Google Scholar]
- 22.Dalili Z, Taramsari M, Mousavi Mehr SZ, Salamat F. Diagnostic value of two modes of cone-beam computed tomography in evaluation of simulated external root resorption: an in vitro study. Imaging Sci Dent. 2012;42:19–24. doi: 10.5624/isd.2012.42.1.19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kamburoglu K, Kursun S. A comparison of the diagnostic accuracy of CBCT images of different voxel resolutions used to detect simulated small internal resorption cavities. Int Endod J. 2010;43:798–807. doi: 10.1111/j.1365-2591.2010.01749.x. [DOI] [PubMed] [Google Scholar]
- 24.Liedke GS, da Silveira HE, da Silveira HL, Dutra V, de Figueiredo JA. Influence of voxel size in the diagnostic ability of cone beam tomography to evaluate simulated external root resorption. J Endod. 2009;35:233–235. doi: 10.1016/j.joen.2008.11.005. [DOI] [PubMed] [Google Scholar]
- 25.Cheng JG, Zhang ZL, Wang XY, Zhang ZY, Ma XC, Li G. Detection accuracy of proximal caries by phosphor plate and cone-beam computerized tomography images scanned with different resolutions. Clin Oral Investig. 2011;16:1015–1021. doi: 10.1007/s00784-011-0599-7. [DOI] [PubMed] [Google Scholar]
- 26.Haiter-Neto F, Wenzel A, Gotfredsen E. Diagnostic accuracy of cone beam computed tomography scans compared with intraoral image modalities for detection of caries lesions. Dentomaxillofac Radiol. 2008;37:18–22. doi: 10.1259/dmfr/87103878. [DOI] [PubMed] [Google Scholar]
- 27.Kamburoglu K, Murat S, Yuksel SP, Cebeci AR, Paksoy CS. Occlusal caries detection by using a cone-beam CT with different voxel resolutions and a digital intraoral sensor. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2010;109:63–69. doi: 10.1016/j.tripleo.2009.12.048. [DOI] [PubMed] [Google Scholar]
- 28.Librizzi ZT, Tadinada AS, Valiyaparambil JV, Lurie AG, Mallya SM. Cone-beam computed tomography to detect erosions of the temporomandibular joint: effect of field of view and voxel size on diagnostic efficacy and effective dose. Am J Orthod Dentofacial Orthop. 2011;140:25–30. doi: 10.1016/j.ajodo.2011.03.012. [DOI] [PubMed] [Google Scholar]
- 29.Bauman R, Scarfe W, Clark S, Morelli J, Scheetz J, Farman A. Ex vivo detection of mesiobuccal canals in maxillary molars using CBCT at four different isotropic voxel dimensions. Int Endod J. 2011;44:752–758. doi: 10.1111/j.1365-2591.2011.01882.x. [DOI] [PubMed] [Google Scholar]
- 30.Al-Rawi B, Hassan B, Vandenberge B, Jacobs R. Accuracy assessment of three-dimensional surface reconstructions of teeth from cone beam computed tomography scans. J Oral Rehabil. 2010;37:352–358. doi: 10.1111/j.1365-2842.2010.02065.x. [DOI] [PubMed] [Google Scholar]
- 31.Maret D, et al. Effect of voxel size on accuracy of 3D reconstructions with cone beam CT. Dentomaxillofac Radiol. 2012;41:649–655. doi: 10.1259/dmf/81804525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Sun Z, Smith T, Kortam S, Kim DG, Tee BC, Fields H. Effect of bone thickness on alveolar bone-height measurements from cone-beam computed tomography images. Am J Orthod Dentofacial Orthop. 2011;139:117–127. doi: 10.1016/j.ajodo.2010.08.016. [DOI] [PubMed] [Google Scholar]
- 33.Torres MG, Campos PS, Segundo NP, Navarro M, Crusoe-Rebello I. Accuracy of linear measurements in cone beam computed tomography with different voxel sizes. Implant Dent. 2012;21:150–155. doi: 10.1097/ID.0b013e31824bf93c. [DOI] [PubMed] [Google Scholar]
- 34.Fourie Z, Damstra J, Gerrits PO, Ren Y. Accuracy and reliability of facial soft tissue depth measurements using cone beam computer tomography. Forensic Sci Int. 2010;199:9–14. doi: 10.1016/j.forsciint.2010.02.018. [DOI] [PubMed] [Google Scholar]
- 35.Sherrard JF, Rossouw PE, Benson BW, Carrillo R, Buschang PH. Accuracy and reliability of tooth and root lengths measured on cone-beam computed tomographs. Am J Orthod Dentofacial Orthop. 2010;137:100–108. doi: 10.1016/j.ajodo.2009.03.040. [DOI] [PubMed] [Google Scholar]
- 36.Yeni YN, Christopherson GT, Dong XN, Kim DG, Fyhrie DP. Effect of microcomputed tomography voxel size on the finite element model accuracy for human cancellous bone. J Biomech Eng. 2005;127:1–8. doi: 10.1115/1.1835346. [DOI] [PubMed] [Google Scholar]
- 37.Chadwick JW, Lam EW. The effects of slice thickness and interslice interval on reconstructed cone beam computed tomographic images. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2010;110:e37–e42. doi: 10.1016/j.tripleo.2010.05.008. [DOI] [PubMed] [Google Scholar]
- 38.Spin-Neto R, Marcantonio E, Jr, Gotfredsen E, Wenzel A. Exploring CBCT-based DICOM files. A systematic review on the properties of images used to evaluate maxillofacial bone grafts. J Digit Imaging. 2011;24:959–966. doi: 10.1007/s10278-011-9377-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Ludlow JB, Davies-Ludlow LE, Brooks SL, Howerton WB. Dosimetry of 3 CBCT devices for oral and maxillofacial radiology: Cb Mercuray, Newtom 3G and i-Cat. Dentomaxillofac Radiol. 2006;35:219–226. doi: 10.1259/dmfr/14340323. [DOI] [PubMed] [Google Scholar]
- 40.Ballrick JW, Palomo JM, Ruch E, Amberman BD, Hans MG. Image distortion and spatial resolution of a commercially available cone-beam computed tomography machine. Am J Orthod Dentofacial Orthop. 2008;134:573–582. doi: 10.1016/j.ajodo.2007.11.025. [DOI] [PubMed] [Google Scholar]
- 41.Leung CC, Palomo L, Griffith R, Hans MG. Accuracy and reliability of cone-beam computed tomography for measuring alveolar bone height and detecting bony dehiscences and fenestrations. Am J Orthod Dentofacial Orthop. 2010;137:109–119. doi: 10.1016/j.ajodo.2009.07.013. [DOI] [PubMed] [Google Scholar]
- 42.Thomas RZ, Ruben JL, de Vries J, ten Bosch JJ, Huysmans MC. Transversal wavelength-independent microradiography, a method for monitoring caries lesions over time, validated with transversal microradiography. Caries Res. 2006;40:281–291. doi: 10.1159/000093186. [DOI] [PubMed] [Google Scholar]
- 43.Hassan B, Metska ME, Ozok AR, van der Stelt P, Wesselink PR. Detection of vertical root fractures in endodontically treated teeth by a cone beam computed tomography scan. J Endod. 2009;35:719–722. doi: 10.1016/j.joen.2009.01.022. [DOI] [PubMed] [Google Scholar]
- 44.Farman AG. ALARA still applies. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2005;100:395–397. doi: 10.1016/j.tripleo.2005.05.055. [DOI] [PubMed] [Google Scholar]
- 45.Hirsch E, Wolf U, Heinicke F, Silva MA. Dosimetry of the cone beam computed tomography Veraviewepocs 3D compared with the 3D Accuitomo in different fields of view. Dentomaxillofac Radiol. 2008;37:268–273. doi: 10.1259/dmfr/23424132. [DOI] [PubMed] [Google Scholar]