Introduction

Dental implant treatment has been widely applied for cases with missing teeth. In the pretreatment planning, computed tomography (CT) is widely used to evaluate the bone width and height, and to identify the mandibular canal in the premolar and/or molar regions of the mandible [13].

For CT examinations, either multi-slice CT (MSCT) or cone-beam CT for dental use (CBCT) can be employed. With the latest scanners for both of these approaches, the images are composed of tiny isotropic-voxel-assembled data, so the object can be observed in arbitrary sections with a quality similar to that of the axial images [12]. For evaluating the mandible, curved multi-planar reconstruction (MPR) is usually performed using a dental CT reformatting program [5]. Dental CT was originally developed for implant planning, and can provide two types of sectional image: paraxial and panoramic (cross-sectional) views. Using these images, clinicians can determine the available bone size, and the location of the mandibular canal [4]. The latest version of dental CT has the function of “Outlining the Mandibular Canal in the Panoramic View”. This is the function in which the course of the mandibular canal is automatically drawn on the paraxial view, by identifying it on the panoramic view.

We sometimes encounter cases in which the mandibular canal is poorly visualized on dental CT cross-sectional images. Although there have been several studies evaluating the visibility of the mandibular canal on dental CT images, they mostly evaluated it on paraxial views alone [5, 9]. The aim of this study was to compare the panoramic and paraxial views of dental CT reformatted images in terms of detecting the mandibular canal, and to evaluate the usefulness of the function of “Outlining the Mandibular Canal in the Panoramic View”.

Methods

Patients

The study population included 155 consecutive patients, who underwent MSCT examination for treatment of dental implant to the mandible at the Dental hospital in July and August, 2011. The patients consisted of 102 females and 53 males, with a mean age of 53.9 years. The exclusion criterion from this study was the presence of lesions reaching the mandibular canal, although we did not encounter such cases in this series. This was a retrospective study using the clinical data that had been archived on the image server at our hospital.

MSCT examination

MSCT examination was performed using a Somatom Sensation 64 scanner (Siemens, Forchheim, Germany). Patients were scanned in a supine position, with the scanning plane parallel to their occlusal plane. The scanning parameters were as follows: tube voltage of 120 kV, tube current of 90 effective mAs, collimation of 64 × 0.6 mm and pitch of 0.9. These parameters had been determined based on the low dose technique, and the effective dose per an examination was estimated at approximately 240 μSv [11]. After scanning, 0.6-mm-thick axial images were reconstructed with 0.3 mm increments using a bone kernel of H70h. Using these axial images, curved MPR was performed using a dental CT reformatted program (Siemens), and both panoramic and paraxial cross-sectional views of the mandible were obtained. The reference curve was set to connect the mandibular foramen and the mental foramen, being harmonized to the dental arch. Both paraxial and panoramic views were reconstructed with 1 mm thickness and 1 mm increment. Panoramic and paraxial views were separately printed on film using a dry laser imager DryPix7000 (Fuji Film Medical Co. Ltd., Tokyo Japan) with a magnification of 1.0. The parameters of the imager were fixed to 100 μm pixel size and 14-bit gradation, which had sufficient image quality for observation. In all cases, the window width and the window center were set to 3,700 and 500 HU, respectively.

Evaluation of images

Dental CT images of 155 cases (310 sides) were evaluated for the detectability of the mandibular canal by two raters. On each of the panoramic and paraxial views, the detectability of the canal was evaluated using a 5-point rating scale as shown in Table 1. Because we attempted to evaluate the validity of dental CT images in a clinical setting, the whole mandibular canal was defined as ranging from the mental foramen to the lateral end of the molar region. With this rating scale, cases scored as 4 or 5 were considered to be acceptable for preplanning of implant surgery, whereas those scored as 1, 2 or 3 were unacceptable.

Table 1 The scoring for the mandibular canal detectability
Full size table

One observer (H.W., 15 years of experience as an oral radiologist) evaluated all of the 310 sides twice with a 3-week interval, to determine the intra-observer reliability. Subsequently, another observer (A.T., 15 years of experience as a dental implantologist and 3 years as a radiologist) evaluated the same images to determine the inter-observer reliability. In cases with disagreement between the two observers, a consensus was reached by discussion.

After evaluating the detectability of the mandibular canal on panoramic and paraxial views, we applied the function of “Outlining the Mandibular Canal in the Panoramic View” to the cases considered as unacceptable (score of 1, 2 or 3) on paraxial views. Using this function, the course of the mandibular canal is automatically drawn on the paraxial view by identifying it on the panoramic view; namely, by drawing a line over the canal on the panoramic view, the program immediately displays markers on the paraxial view to indicate the mandibular canal position. We investigated whether this function was useful to identify the location of the mandibular canal on paraxial views.

Statistical analysis

In this study, the right and left mandibular canals of one patient were regarded as two independent cases. The mean scores and standard deviations of the panoramic and paraxial views of the 310 sides were calculated. Intra-class correlation (ICC) was used to evaluate the intra-observer reliability. For evaluating the inter-observer reliability, Spearman’s correlation coefficient was used. The presence of a statistically significant difference between the scores of paraxial and panoramic views was tested by Wilcoxon’s signed rank test, and a p < 0.05 was considered to indicate a significant difference.

Results

In terms of the intra-observer reliability, ICCs for the panoramic and paraxial views of dental CT images were 0.935 and 0.843, respectively, and both values were classified as “almost perfect” [6]. When concerning the inter-observer reliability, Spearman’s correlation coefficients for the panoramic and paraxial views were 0.804 and 0.864, respectively. The results indicated the significant high correlation, with a p < 0.001.

The results for the visibility score for both views are shown in Fig. 1. Cases with a score of 5 or 4 were more frequent in panoramic views than in paraxial views. The number of cases with a score of 1, 2 or 3 (unacceptable) was 59 in panoramic views, whereas it was 128 in paraxial views. The mean scores and standard deviations of panoramic and paraxial views were 4.2 ± 1.1 and 3.5 ± 1.2, respectively, and the former was significantly higher than the latter according to Wilcoxon’s signed rank test (p < 0.001). The relationship of the visibility scores between the two views is shown in Table 2.

Fig. 1
figure 1

Visibility of the mandibular canals on panoramic and paraxial views for 310 sides of 155 patients. A 5-point rating scale was employed for evaluation. The values written inside the bar boxes represent the case numbers with each score

Full size image
Table 2 Cross-linkage table of 5-point rating scores of panoramic and paraxial views
Full size table

Because the detectability of the mandibular canal on panoramic views was significantly higher than that on paraxial views, we attempted to apply the function of “Outlining the Mandibular Canal in the Panoramic View” to the 128 cases with poor scores (1, 2 or 3) on paraxial views. Using this function, the course of the mandibular canal became identifiable on paraxial views in 72 out of the 128 cases. When considering these results, we planned a strategy to improve the identification of the mandibular canal on dental CT paraxial views as shown in Fig. 2; namely, using paraxial views alone, the detectability of the canal was only 59 % (182/310 cases). However, by applying the function of “Outlining the Mandibular Canal in the Panoramic View”, the percentage of identifying the mandibular canal increased to 82 % (254/310 cases).

Fig. 2
figure 2

Strategy for identifying the mandibular canal in dental CT images. The cases evaluated as having a score of 4 or 5 on paraxial views were considered to be acceptable for clinical use, which amounted to 59 % of cases. The remaining 41 % of cases were evaluated as having a score of 1, 2 or 3 (unacceptable), but 23 % of all cases could be salvaged to identify the canal using the function of “Outlining Mandibular Canals on Panoramic Views”. However, 18 % of cases had a score of 1, 2 or 3 on both paraxial and panoramic views, and should be evaluated using another approach

Full size image

Discussion

In this study, we evaluated the detectability of the mandibular canal using dental CT software, which provided two types of sectional image: panoramic and paraxial views. For surgical treatment planning, evaluation of the distance from the alveolar crest to the upper margin of the mandibular canal is necessary, and it should be measured on a paraxial view, but not on a panoramic view. This is because the exact location of the alveolar crest is unclear on the panoramic view. Furthermore, no information about the available bone width, which is also necessary for the treatment planning is provided on panoramic views. To our knowledge, this is the first study to compare the detectability of the mandibular canal between the two types of view of dental CT. This study clearly showed that the detectability of the mandibular canal was significantly higher on panoramic views than on paraxial views. The mandibular canal is a structure surrounded by comparatively thin cortical bone, and the thickness of the upper cortex is approximately 0.9 mm, although there is variability among different regions [2]. To identify the section of the mandibular canal on paraxial views, we have to identify a round lumen surrounded by thin cortical bone, which may sometimes be indistinguishable from the adjacent trabecular bone pattern. On the other hand, for identifying the mandibular canal on panoramic views, it is necessary to find just two parallel continuous lines across the cancellous bone at which the discrimination from other structures is easier. This may explain the results of our study that the panoramic views showed better detectability of the canal than paraxial views.

There have been several studies on the detectability of the mandibular canal, but they were limited to its evaluation on paraxial views alone. Jacobs et al. evaluated the detectability of the mandibular canal on paraxial views of 250 cases using single-slice CT (SSCT), and reported that the canal was detectable in 98 % of cases [5]. However, in 20 %, the course of the mandibular canal was invisible or poorly visible. Paes et al. [10] performed a comparative study between MSCT and SSCT in terms of evaluation of the mandibular canal. Although they revealed that MSCT provided better and more accurate images than SSCT, they did not report the detectability of the mandibular canal. In this study, we used the latest 64-slice MSCT scanner with optimal scanning and reconstruction settings. Nevertheless, when using paraxial views alone, cases with a score of 4 or 5 accounted for only 59 %, whereas those with a score of 1, 2 or 3 accounted for 41 %. These results are considered to indicate the limitation of diagnosis using paraxial views alone.

The dental CT software program has the function of “Outlining of Mandibular Canals on the Panoramic View” as mentioned above. Our study confirmed that this function would be useful for many cases in which the mandibular canal is not clearly visible on paraxial views. Representative cases are shown in Figs. 3 and 4. In Fig. 3, the visibility had a score of 1 on the paraxial view, but a score of 4 on the panoramic view. Similarly, in Fig. 4, the visibility had a score of 2 on the paraxial view because the surrounding trabecular pattern was similar to the mandibular canal, but a score of 5 on the panoramic view. For both cases, using this function, the course of the mandibular canal became clearly identifiable on paraxial views. Among the 128 cases with a poor score (score 1, 2 or 3) on paraxial views, 72 cases had a score of 4 on panoramic views (Fig. 2). As shown in Table 2, the visibility score was quite different between paraxial and panoramic views in some cases: a score of 1 on paraxial and a score of 4 on panoramic views (2 cases), and a score of 2 on paraxial and a score of 5 on panoramic views (one case). For these results, we planned a strategy to identify the mandibular canal on dental CT images (Fig. 2). Cases with a score of 4 or 5 on paraxial views can be utilized solely for implant planning (59 %), but this is not the case for the remaining 41 % of cases (score of 1, 2 or 3). However, 23 % of cases were scored as 4 or 5 on panoramic views; therefore, the function mentioned above could be applicable to them. As a result, the percentage of identifying the mandibular canal on paraxial views of dental CT images increased to 82 % when our strategy was applied. One might think that CBCT would be superior to demonstrate mandibular canals because the spatial resolution of CBCT is higher than that of MSCT [12] and/or the clinicians can observe the three-dimensional images in a dynamic manner. However, Naitoh et al. [8] reported that there was no difference between CBCT and MSCT in terms of the visualization of anatomical bone structures, and another study reported that the CBCT visualized the mandibular canal in 62.9–87.9 % of their cases depending on the regions [9], which was not higher than our results using MSCT. According to the study of Lofthag-Hansen et al. [7] involving the evaluation of CBCT, the mandibular canal was only visible in 10 out of 30 cases (33 %) when using paraxial views alone; however, it could be clearly visible in 26 out of 30 (87 %) when all of paraxial, sagittal and axial images were used for the evaluation. The results should confirm the validity of our strategy as mentioned above, and it would be useful for the identification of the mandibular canal in CBCT using the strategy.

Fig. 3
figure 3

A case in which the mandibular canal was successfully identified on paraxial views by the function of “Outlining the Mandibular Canal in the Panoramic View”. The images were sectioned with 1 mm thickness. The patient was a 70-year-old woman, who underwent CT examination for bilateral missing molar teeth. a The right mandibular canal was thin, but clearly identifiable on panoramic views (score 4). b The molar region of paraxial views. It was too difficult to identify the mandibular canal (score 1). c The outline is indicated at the upper margin of the mandibular canal. d The corresponding points are marked with a plus sign (+) on paraxial views, and one could, thus, determine the location of the upper margin of the mandibular canal

Full size image
Fig. 4
figure 4

A case in which the mandibular canal could not be identified on paraxial views without the function of “Outlining the Mandibular Canals in the Panoramic View”. The images were sectioned with 1 mm thickness. The patient was a 35-year-old woman, who underwent CT examination for missing #16 tooth. a The left mandibular canal was fairly clearly depicted from the mental foramen to the molars (score 5). b On paraxial views, there were several round bone structures similar to a mandibular canal; therefore, it was hard to detect (score 2). c A tracing line was drawn to the upper border of the mandibular canal. d The mandibular canal was clearly identifiable because its location was marked with a plus sign (+) on paraxial views

Full size image

Our strategy was considered valid to evaluate the distance between the alveolar crest and the mandibular canal on paraxial views. However, the course of the mandibular canal was still unclear in 18 % of the cases, even when this strategy was applied. In addition, it would be a limitation of this study that we could not detect the mandibular nerve itself because the objects were living bodies. One potential candidate overcoming these issues is magnetic resonance imaging examination (MRI), which uses no ionizing radiation and clearly shows the cortical bone, cancellous bone, nerve and blood flow with different signal intensities. It enables us to observe the inferior mandibular nerve and artery rather than the mandibular canal wall, and there have been some trial studies that evaluated its detectability of the mandibular canal structures [1, 3]. However, MRI has not become a routine examination for patients undergoing dental implant surgery because of its low availability, high cost, and the existence of some contraindications and magnetic susceptibility of artifacts.

Conclusion

In this study, we investigated the detectability of the mandibular canal on dental CT reformatted images of patients before implant operation. The detectability was significantly higher in panoramic views than in paraxial views. Using the function of “Outlining the Mandibular Canal in the Panoramic View”, the precision of identifying the canal on paraxial views was considered to be improved.