Critical anatomic region of nasopalatine canal based on tridimensional analysis: cone beam computed tomography

The study aim of this was to define the critical anatomic region of the premaxilla by evaluating dimensions of nasopalatine canal, buccal bone plate (BBP) and palatal bone plate (PBP). 230 CBCTs were selected with both, one or no upper central incisors present (+/+, −/+, −/−) and periodontal condition was evaluated. T-student test, ANOVA, Pearson´s correlation and a multivariant-linear regression model (MLRM) were used. Regarding gender, significant differences at level 1 (lower NC) were found for: buccal-palatal, transversal and sagittal NC diameters, and NC length (NCL). Regarding dental status, significant differences were found for: total BBP length (tBL) and PBP width (PW2) at level 2 (NCL midpoint). NCL was correlated with PW2, tBL, and PBP length at level 3 (foramina of Stenson level). An MLRM had a high prediction value for NCL (69.3%). Gender is related to NC dimensions. Dental status has an influence on BBP dimensions, but does not influence on NC and PBP. Periodontal condition should be evaluated for precise premaxillae analysis NC diameters at the three anatomical planes are related to each other, while NCL is related to BBP and PBP lengths. A third of premaxilla is taken up by NC, thus, establishing the critical anatomic region.


Level 1.
It is located based on the axial slice when the incisive foramen of NC is completely closed.
In axial slice the following measurements were taken: (1) buccal-palatal diameter (bpD1) and transversal diameter (tD1) of NC. The bpD1 corresponded to maximum diameter between buccal and palatal bone cortical of NC; the tD1 is the maximum diameter perpendicular to bpD1; and, (2) BBP widths. The distance from BBP to buccal wall of NC at three points of NC: left (LW1), central (CW1) and right (RW1) (Fig. 1).
In coronal slice the following measurement was taken on the horizontal line: (1) NC diameter at level 1 (CD1) (Fig. 1).
In sagittal slice the following measurements are taken on the horizontal line: (1) BBP width at level 1 (BW1), (2) NC sagittal diameter at level 1(SD1), and (3) PBP width at level 1 (PW1). In addition, the following measurements are taken from the horizontal line to the inferior edges of BBP and PBP: (1) BBP length at level 1(BL1), and (2) PBP length at level 1 (PL1) (Fig. 1). Level 2. It is located by moving the horizontal line to the middle point of NC length (NCL) on the sagittal plane. NCL is defined as the distance from the incisive foramina to foramina of Stenson.
In axial slice the following measurements were taken: (1) buccal-palatal diameter (bpD2) and transversal diameter (tD2) of NC. The bpD2 corresponded to maximum diameter between buccal and palatal bone cortical of NC; the tD2 is the maximum diameter perpendicular to bpD2; and, (2) BBP widths. The distance from BBP to buccal wall of NC at three points of NC: left (LW2), central (CW2) and right (RW2) (Fig. 2).
In coronal the following measurement was taken on the horizontal line: (1) NC diameter at level 2 (CD2) (Fig. 2).
Scientific RepoRts | 5:12568 | DOi: 10.1038/srep12568 In sagittal slice the following measurements were taken on the horizontal line: (1) BBP width at level 2 (BW2), (2) NC sagittal diameter at level 2(SD2), and (3) PBP width at level 2 (PW2). In addition, the following measurements were taken from the horizontal line to the inferior edges of BBP and PBP: (1) BBP length at level 2 (BL2), and (2) PBP length at level 2 (PL2) (Fig. 2). Level 3. It is located by moving the horizontal line to the foramina of Stenson on sagittal plane. In coronal slice the following measurement was taken on the horizontal line: (1) NC diameter at level 3 (CD3) (Fig. 3).
In sagittal slice the following measurements are taken on the horizontal line: (1) PBP width at level 3 (PW3), (2) the width from the most posterior point of the transverse palatine suture to the palatal wall of NC, (3) NC sagittal diameter at level 3 (SD3), and (4) PBP width at level 3 (PW3). In addition, the following measurements are taken from the horizontal line to the inferior edges of BBP and PBP: (1) BBP length at level 3 (BL3). (2) PBP length at level 3 (PL3), and, (3) the total BBP length (tBL) was also calculated by joining midpoints of coronal and apical BBP width (Fig. 3).
Periodontal condition of the remaining teeth. Periodontal bone loss was evaluated on sagittal plane by measuring distance from the cemento-enamel junction to BBP in buccal surface and from the cemento-enamel junction to PBP in palate surface of teeth 11 and 21. Periodontal condition was defined as the mean value of these measurements for both teeth 11 and 21. Subjects were classified into three groups. The first group was normal periodontal condition ≤ 3 mm, and was based on the research by de Faria Vasconcelos et al. 12 . The other groups were added by our team to reflect different levels of periodontitis: moderate periodontitis (> 3-≤ 6 mm) and severe periodontitis (> 6 mm). Intraobserver agreement. One month later, the same observer assessed the NC measurements in 20 CBCTs to check the intraobserver variability, using the intraclass correlation coefficient for the NC measurements.
Statistical analysis. All data were first analyzed using descriptive statistics. Differences in CBCT measurements were compared among subjects with and without teeth in the premaxila according to their dental status (− /+ , + /+ , − /− ) by ANOVA with post hoc Bonferroni, Scheffe and Tukey-b tests; and according to gender by Student's-t test. The proportion of NC diameter with respect to BBP in axial and sagittal slices was also evaluated. Pearson's correlation was used to evaluate the association between measurements, periodontal condition and patient age. A multivariant linear regression model (MLRM) was used to predict the relation between measurements, dental status, and gender. The level of intraobserver agreement was assessed for anatomical measurements using the intraclass correlation coefficient. Statistical significance was set at p ≤ 0.05. Analyses were performed using SPSS 20.0 software for Windows (IBM SPSS Statistics, Chicago, IL, USA).

Results
A total of 224 CBCT were included in the study. Six CBCTs (out of 230) were not included due to the following reasons: implant placed or bone grafting in the anterior maxilla, poor image quality, presence of maxillary expander, or a nasopalatine duct cyst. The study group comprised 108 males (48.2%) and 116 females (51.8%) with a mean age of 47.28 years (Table 1). Dental status was 81.7% for dentulous (+ /+ ) patients, 11.6% for edentulous (− /− ) and 6.7% for partially edentulous (+ /− ) patients.

Measurements for three anatomical planes of NC.
A summary of NC, BBP, and PBP characteristics can be seen in Table 1. With respect to gender, the following variables present a significantly greater mean for males (p ≤ 0.05): bpD1, tD1, BBP width at axial levels 1 and 2 (LW1/2, CW1/2, RW1/2), CD1, SD1, BW1/2, NCL, PW2/3, and PL2/3 ( Table 2). With respect to dental status, the following variables present significant differences between each group: age, and a lower mean value for edentulous subjects: age, LW1, CW1, RW1, BW1, BL1, BL2, BL3, tBL (Table 3). In Table 4 we can see the correlations between NC dimensions in the three anatomical planes, total BBP length and BBP width at sagittal level 2. Age had a direct correlation with CD1, that is, canal diameter is larger in older subjects.
Multiple linear regression models (MLRM) were obtained for variables as follows: (1)   Intraobserver variability. The intraobserver variability was an intraclass correlation coefficient value ranging from 0.80 to 0.86 (95% confidence interval ranging from 0.57 to 0.94).

Discussion
The use of three-dimensional images does not necessarily imply three-dimensional analysis. Unlike previous research based on CT or CBCT reconstructions 3,13-16 , the present study analyses NC and adjacent bone at the three anatomical planes simultaneously. 3D analysis of premaxilla dimensions is fundamental for determining adequate preoperative treatment. To the best of our knowledge, this is the first study to do so. Our findings regarding axial NC diameters are in line with Song et al. 17 , who found transversal diameter to be greater than buccal-palatal diameter. Insofar as gender, significant differences were found for axial diameter only at level 1, but Thakur et al. 18 found no such differences at any level in a study based on a hindu population. We found no differences regarding dental status, but this is not comparable with Mardinger et al. 13 because these authors use the Lekholm and Zarb classification 19 . Regarding the correlation of transversal diameters, these were found to be directly proportional to each other at every level, with a moderate-strong correlation. In addition, they were proportional to sagittal and coronal diameters. There was an inverse correlation between BBP dimensions and axial diameters. Notably, the MLRM we obtained for predicting axial transversal diameter at level 2 had a predictive value of over 40%. Furthermore, it was the first time this analysis had been carried out. With respect to sagittal NC diameters, we found significant differences according to gender at level 1, as did Tözüm et al. 20 . Other authors, such as Bornstein et al. 14 and Thakur et al. 18 , found no such differences. However, we should point out that Bornstein et al. 14 used the oblique measurement at this level instead of the horizontal measurement, resulting in greater values. Like other authors 14,20 , we found no differences with respect to dental status. This is in line with the centripetal resorption of maxillary bone theory [21][22][23][24][25] , indicating that anterior maxillary resorption occurs mainly in the buccal plate instead of the palatine plate.
Regarding coronal NC diameters, we obtained a similar mean value to Liang et al. 3 (3.4 ± 0.9 mm) at level 1. The fact that these authors 3 carried out their analysis on skulls is another confirmation of CBCT accuracy [26][27][28][29][30] . With respect to gender, we found male diameter to be significantly higher, but this is not supported by previous research. Insofar as dental status, our results were in line with Liang et al. 3 , who found no significant differences. Also like these authors 3 , we found a direct correlation with age; that is, diameter increased along with age.  With respect to NCL, the mean value we obtained is in line with other authors 14,20 . However, Liang et al. 3 and Mraiwa et al. 16 found a lower mean value. We should note that the sample size in the present study was considerably higher than the samples in those studies 3,16 . Regarding the influence of gender, like previous studies 14,17,18,20,31 , males presented significantly greater values. With respect to dental status, our results were nearly significant (p = 0.076), while previous studies 3,20 reported significantly greater NCL in the dentulous group. Like other studies 13,17 , we found NCL in the dentulous group to be approximately 2-mm longer than the edentulous group, but these studies make no mention of significance. The present study found that NCL had a moderate to strong direct correlation with BBP and PBP length at levels 2 and 3 (BL2-3, PL2-3) as well as with BtL, while previous research did not analyze these correlations. Like Tözüm et al. 20 , we found no correlation between NCL and diameter. While other authors 14,20 found an indirect correlation with age, we found no such correlation. Nevertheless, we did find that edentulous status increased with age (p < 0.001), thus suggesting a decrease in BBP length. In fact, when we applied an MLRM, we found that BBP length explained NCL with a predictive value of almost 70%.
To the best of our knowledge, this is the first study to analyze BBP in axial slice. This plane is the basis of CBCT, because coronal and sagittal planes are both determined in relation to axial slice 32 . We found that BBP width was significantly higher in males, and that it was higher in the coronal ridge region in the dentulous group, which is in line with others authors [23][24][25] , who explain that remodeling only occurs in coronal bone while basal bone is genetically determined. We found that axial BBP width was directly correlated with tBL and indirectly correlated with NC diameter in the three anatomic planes. We believe that this indirect correlation can be explained by a biological dimension adjustment, in other words, the proportion of bpD1/2 to CW1/2 remains constant. Thus, we have a + b = c, where a = bpD1/2 and b = CW1/2 and c is constant. Regardless of dental status, wider canals entail straighter ridges, and straighter canals entail wider ridges.
Like other authors 14,16,20 we observed that sagittal BBP width increased to apical ridge and was greater in males 14,20,33 . As with the axial plane, only the coronal ridge region was significantly influenced by dental status, coinciding with other studies 13,14,20 . A previous study by our team 5 did not find this difference because it only evaluated ridge width at one-third of the coronal BBP length. The present study demonstrates the importance of analyzing BBP ridge at different levels, because of its peculiar bone remodeling. We found sagittal width and length to be proportional, and Van der Weijden et al. 34 reported that width presented more bone loss than length.
Regarding the BBP length, we found differences in dental status at all levels. Like Mardinger et al. 13 and Tözüm et al. 20 , BBP length was significantly greater in the dentulous group. Moreover, the presence of only one central incisor (+ − ) also entailed a significantly greater value of BL2 and tBL. This is not supported by previous studies, nor is the direct correlation between BBP and PBP lengths.
In addition to dental status, the periodontal condition of remaining central incisors should be considered when the bone level and NC dimensions are evaluated. Several studies 12,35,36 have demonstrated the accuracy of CBCT in the measurement of periodontal bone defects.
For higher quality images of periodontal structures, the voxel size should be inferior to 0.3 mm 29 as in the present study. The mean age in the present study was similar to de Faria Vasconcelos et al. 12   et al. 29 and Mol et al. 37 , therefore, like these authors, we considered a measurement of higher than 3 mm between the cemento-enamel junction and the alveolar crest to indicate periodontitis. Regarding PBP, width landmarks in the present study were not exactly the same as Tözüm et al. 20 , though both found a progressive increase and greater value for males in PBP width. Unlike these authors 20 , we found no significant differences regarding dental status. As occurred with BBP, PBP width and length were directly proportional. Insofar as PBP length, we obtained similar mean values for PL3 and differences for gender and dental status.
The critical anatomic region of premaxilla is defined by NC and BBP dimensions. We found that on average the NC diameter takes up approximately 1/3 of premaxilla. Although males presented greater NC and BBP dimensions in this area, both genders had similar biological dimension adjustment. Furthermore, we found that edentulous patients presented a higher proportion of NC diameter, entailing lower BBP width. The fact that bone resorption is related with tooth loss and mainly occurs at the coronal ridge helps to explain why we found a bone decrease of 6% at coronal ridge, 2% at middle ridge, and 1% at apical ridge.
With respect to limitations, the present study is based on the analysis of CBCT images, and patients gave their consent for this. Therefore, the present study had no other way of determining whether the subjects had had prior surgery.

Conclusions
Gender is related to NC dimensions, BBP width and apical dimensions of PBP. Dental status has an influence on BBP dimensions, but does not influence on NC and PBP dimensions. Periodontal condition should be evaluated for a precise premaxillae analysis. NC diameters at the three anatomical planes are related to each other, while NCL is related to BBP and PBP lengths. BBP and PBP dimensions are proportional. A third of the premaxilla is taken up by the NC, thus, establishing the critical anatomic region.