Morphological types of sella turcica bridging and sella turcica dimensions in relation to palatal canine impaction: a retrospective study

2.1 Subjects

Treatment records of 120 subjects (43 males, 77 females; mean age 18.8 ± 7.8 years) with unilateral or bilateral PCI were collected retrospectively from patients who had visited the Orthodontic Department of Clinic of Stomatology, Faculty of Medicine, Masaryk University and St. Anne’s University Hospital Brno, Czech Republic. The palatal position of canine was identified by an experienced clinician (AO) using dental panoramic radiographs, anterior occlusal radiographs, and computed tomography/cone-beam computed tomography (CBCT) and later confirmed during surgery. The data of subjects were collected blindly, without any analysis of the lateral cephalometric radiograph. The exclusion criteria included a low-quality or missing lateral cephalometric radiograph, the presence of craniofacial anomalies, a cleft lip or palate, a trauma or previous orthodontic treatment.

The controls consisted of the same number of healthy individuals (44 males, 76 females; mean age 18.1 ± 5.6 years) with erupted canines in the dental arch and complete permanent dentition. The exclusion criteria were the same as for the subjects with PCI.

2.2 Data collection

Lateral cephalometric radiographs were taken in a standardized manner using the same cephalostat and X-ray device (PM 2002 CC Proline, Planmeca OY, Helsinki, Finland) in the Department of Dental Radiology. The cephalometric radiographs were scanned using a scanner (Epson Perfection V700, Epson Portland Inc., Portland, OR, USA) and then measured with imaging software (PicPick, Version 4.0.0, NTeWORKS, Daejeon, Republic of Korea). The actual size of each image was calibrated based on the known distance on the cephalometric radiograph.

2.3 Definition of measurement

Fig. 1 illustrates the linear dimensions of the sella turcica in the sagittal plane. The length, depth and anteroposterior diameter were measured according to the study of Silverman [27] and Kisling [28], and the interclinoid distance was measured by the method used by Sundareswaran and Nipun [29]. The definition of the parameters is as follows:

Fig. 1.The contour of a sella turcica with landmarks and linear dimensions. TS: tuberculum sellae; DS: dorsum sellae; PClin: posterior clinoid; AClin: anterior clinoid; SP: sella posterior; L: length; ID: interclinoid distance; APD: anteroposterior diameter; D: depth.

∙ Length (L): the distance from the tip of the dorsum sellae to that of the tuberculum sellae.

∙ Depth (D): the line’s distance dropped perpendicularly from the line above to the deepest point of the sella floor.

∙ Anteroposterior diameter (APD): the distance from the tip of the tuberculum sellae to the farthest point on the inner wall of the sella turcica.

∙ Interclinoid distance (ID): the shortest distance between the anterior clinoid and posterior clinoid processes of the sella turcica.

The length-to-diameter ratio according to Leonardi et al. [19], the ratio between the interclinoid distance and length according to Sundareswaran and Nipun [29] and the area size (the product of length and depth) were also calculated.

2.4 Measurement

The subjects with PCI and the controls were divided into two groups: sella turcica with no fusion (Fig. 2a) and sella turcica with a bridge. These groups were classified using three methods. In the group with no fusion, the level of calcification was assessed based on the length and anteroposterior (AP) diameter of the sella turcica as suggested by Leonardi et al. [19], method 1:

∙ Class I (no bridging): length >3/4 of the AP diameter.

∙ Class II (partial sella turcica bridging): length ≤3/4 of the AP diameter.

However, method 1 does not consider the distance between the clinoid processes, therefore, this study used an additional method according to Sundareswaran and Nipun [29], described as method 2:

∙ Class I (no bridging): interclinoid distance >1/3 of the length of the sella turcica.

∙ Class II (partial sella turcica bridging): interclinoid distance <1/3 of the length of the sella turcica.

In the group with complete sella turcica bridging, the morphology of the bridge according to Becktor et al. [26] was evaluated and classified into two subgroups:

∙ Type A: ribbon-like fusion (Fig. 2b).

∙ Type B: extension of the anterior and/or posterior clinoid process, where these two meet anteriorly, posteriorly or in the middle, with a thinner fusion (Fig. 2c).

Fig. 2.Radiographic image obtained from the lateral cephalogram and a corresponding contour of a sella turcica. (a) no bridging; (b) Type A sella bridging; (c) Type B sella bridging.

The cephalograms were randomly numbered and then assigned to the subjects with PCI or controls after the sella turcica bridging was evaluated and measured. The jaws and teeth were not visible due to the use of an opaque card in the computer screen. Classification of sella turcica bridging was determined independently by two experienced clinicians (AO, PV) who were blinded to the treatment assignment of individual images. There was 100% agreement between the clinicians on the incidence and type of complete sella bridging. Measurements of linear dimensions were performed by one of them (AO). The same investigator re-evaluated 30 randomly selected lateral cephalometric radiographs two weeks after the first analysis to establish the error of linear measurements. The method error of linear measurements was calculated according to Dahlberg’s formula [30]. The errors of duplicate measurements were generally small with a standard deviation of 0.26 mm for the length, 0.23 mm for the diameter and 0.26 mm for the depth.

2.5 Statistical analysis

Power of the study was calculated as a component of a-priori designed study protocol. The sample size was optimized to account for the relatively wide range of comparisons performed, where the outcome depends on the stratified subsamples and the occurrence of the analyzed endpoints. In the case of binary and ordinary endpoints (Fisher’s exact test, goodness-of-fit test) the achieved sample size is sufficient to detect differences ranging from 10% to approximately 21% in all comparisons performed with a power of 80%, overall tests working with the whole sample (2 × N = 120) are powerful to detect a 10% difference in these endpoints. For the quantitative parameters, the sample size was optimized to detect as statistically significant difference of 0.4–0.5 times the standard deviation, also with 80% power. These effect size ranges appear to be clinically relevant and consistent with the results and designs of other already published studies (cited in discussion). The analysis was computed using IBM SPSS Statistics 29 Software (IBM Corp, Armonk, NY). The two-tailed Fisher’s exact test with a significance level of 0.05 was used to compare the presence of sella turcica bridging between the subjects with PCI and the control group. The cephalometric data were analysed using Student’s t-test (two-tailed, unequal variance) to determine the differences between the two groups, also with a significance level of 0.05. Statistical analysis was conducted using Microsoft Office Excel 2016 (Microsoft Corporation, Redmond, Washington, United States).

4.1 Main observations

This study confirmed previous findings of a significantly higher occurrence of complete sella turcica bridging in subjects with PCI. In addition, the effect of the type of complete bridging according to Becktor et al. [26] for PCI was assessed for the first time and found to be insignificant. Partial bridging was classified according to Leonardi et al. [19] (method 1), which is a relatively common method; however, it often exhibits confusion between the interclinoid distance and the sella turcica length. Therefore, the method of Sundareswaran and Nipun [29] (method 2), which works with the interclinoid distance as the shortest distance between clinoid processes and which has not been previously reported for PCI subjects, was also used. Both methods for the identification of partial bridging showed a non-significant association with PCI. Finally, several parameters characterizing the sella turcica dimensions were tested for the association with PCI. Controversially, the length, AP diameter and area were found to be higher for subjects with PCI and in the male population.

4.2 Sella turcica bridging

In most relevant studies, a higher prevalence of sella bridging was observed in the subjects with PCI. The highest prevalence in subjects (25.8%) was found by Ali et al. [23]. Moderate prevalence, consistent with this work, was reported by Leonardi et al. [19] (16.7%), Leonardi et al. [22] (18.4%), Haji et al. [24] (14.3%) and Divya et al. [16] (17.9%). Al-Nakib and Najim [25], Amelinda et al. [11], Dadgar et al. [12], and Kaya et al. [17] showed a relatively low presence of complete sella turcica bridging in subjects with PCI (5.0%, 5.9%, 4.4% and 6.3%, respectively), but still higher than controls. Although these results differ, all findings were statistically significant, except those by Kaya et al. [17]. The differences may be due to the variance in age and genetic background between the samples. It is not clear from the literature whether these studies were blinded or open. Typically, dental structures and the sella turcica are visible together on a lateral cephalometric radiograph, leading to a potential bias. Only four studies [17, 19, 22, 23] described the diagnostic method of PCI. All studies reported PCI except for that of Al-Nakib and Najim [25]. They included maxillary palatal and buccal impacted canines, which could affect their results.

The prevalence of Type A and Type B complete sella bridging according to Becktor et al. [26] has not yet been reported in subjects with PCI. Becktor et al. [26] found complete sella turcica bridging in 18.6% of subjects, with 5.6% being Type A and 13% Type B. All subjects had severe craniofacial deviation and underwent combined orthodontic and surgical treatment. Although Becktor et al. [26] used a sample population with different pathology, these values correlate with those of our results.

More pronounced differences between published studies were in the classification of partial bridging. All the authors who studied canine impaction used the method proposed by Leonardi et al. [19], which involves the sella length-to-diameter ratio (method 1). Except for the study by Majeed et al. [14], the prevalence of partial bridging was higher in subjects with impacted canines. The difference was statistically significant in seven studies [12, 13, 16, 19, 22, 23, 24] and insignificant in two studies [14, 17]. Nevertheless, this study showed contradictory results.

The clinical significance of the classification according to method 1 is questionable for several reasons. Firstly, the exact definition of the “tip of the dorsum sellae”, as one of the points defining the length, differs across the above mentioned studies. Some studies define the tip as the most superior tip of the dorsum sellae, as in our work, while other studies consider the sella length as the closest distance between the tuberculum sellae and dorsum sellae. The latter method underestimates the length, resulting in the overestimation of the occurrence of partial bridging. This may lead to a relatively high prevalence of partial bridging in both groups (>50%), making the classification less valuable from a clinical point of view. Secondly, the distance from the tip of the dorsum sellae to the tuberculum sellae does not reflect in all respects the degree of calcification that occurs between the anterior and posterior clinoid processes.

Hence, a method according to Sundareswaran and Nipun [29] (method 2) was used in the current study. This method is based on the interclinoid distance, which reflects the variable extent of calcification of the clinoid processes. Nevertheless, no significant differences in groups with partial and without sella turcica bridging between the subjects with PCI and the controls were found.

4.3 Sella turcica dimensions

The length of the sella turcica was significantly higher in subjects with PCI. The difference was higher in males. A higher length in subjects with PCI is inconsistent with data reported in previous studies [14, 15, 23, 25]. An in-depth review of the studies revealed that their results differed from each other in terms of length, even in the controls.

There is an apparent discrepancy in the graphic interpretation of the “length”, although the definition is the same, i.e., the distance from the tuberculum sellae to the tip of the dorsum sellae. As already mentioned, some authors consider the tip to be the most superior tip of the dorsum sellae [25, 29], while others consider the tip as the point closest to the tuberculum sellae [13, 14, 15, 17, 23, 24]. In our opinion, the first definition is considerably more appropriate because it corresponds better to measurements from CBCT or directly on the skull, where the length is evaluated in the midsagittal plane and is not affected by the lateral clinoid processes.

Moreover, according to the second definition, the authors confuse the length with the sagittal interclinoid distance [12, 14, 15, 17, 23]. This is inaccurate since the interclinoid distance connects to the anterior clinoid process rather than the tuberculum sellae. Therefore, the definition of “interclinoid distance” varies across studies.

The AP diameter is used to describe the size of the sella turcica, and thus, its definition is more straightforward and should not lead to misinterpretation. The other studies have shown a relatively large variance in mean values. The discrepancies are usually attributed to the differences in the sample population. However, significant differences can also be found between studies with the same population [13, 14, 15, 23]. Our results correspond well with those of Ali et al. [23], Al-Nakib and Najim [25] and Axelsson et al. [8]. The differences between the subjects and controls are mostly inconclusive. In this study, the mean AP diameter increased significantly in subjects with PCI, but the difference is on the border of statistical significance.

Considering the length-to-AP diameter ratio in Table 3, we can observe that the difference between the groups is practically negligible, although the parameters were calculated with the inclusion of patients with sella turcica bridging. Thus, the ratio used by Leonardi et al. [19] shows an insignificant difference in the occurrence of partial bridging. Interclinoid distance seems to be a much more appropriate parameter. In our study, this is the only parameter that is smaller in subjects with PCI. However, it shows a very high standard deviation which makes the difference statistically insignificant. For the same reason, the difference in the interclinoid distance-to-length ratio is not significant.

The mean of the area increased significantly in subjects with PCI. However, in contrast to recent studies by Sato and Endo [31] and Antonarakis et al. [32], no detailed evaluation of the sella surface area, was performed in this study. No other studies investigated sexual dimorphism in subjects with PCI.

When comparing linear dimensions of sella turcica between sexes, we found a significantly higher length in males compared to females. This difference was reflected in the length-to-diameter ratio, which was also significant. Greater length in males compared to females was found only in the 18-year-old age group in the study by Axelsson et al. [8], which corresponds to the mean age in this study. Nevertheless, no significant sex-related differences were reported more frequently [14, 23, 25, 37]. Previous studies have suggested that the change in size with age and sex reflects the adolescent growth spurt in females 2–3 years earlier [27, 35, 36]. Silverman [27] and Axelsson et al. [8] also agreed that both depth and diameter increase during puberty, while the length remains the same.

4.4 Clinical relevance

Both classifications of partial bridging appear to be inappropriate for everyday orthodontic practice. Although there are significant differences in dimensions, these differences are clinically insignificant and not useful for the classification of partial bridging. Correct evaluation of the linear dimensions of the sella turcica in the midsagittal plane requires some training, and one may encounter an unclear definition of cephalometric landmarks, which significantly affects the resulting prevalence. Moreover, the cephalometric radiograph should be of good quality, and irregularities in the sella shape can make the classification process difficult. On the other hand, complete sella turcica bridging was observed twice as often in subjects with PCI and we therefore recommend careful monitoring of the eruption of permanent canines in cases of complete sella turcica bridging, as mentioned by Jankowski et al. [33].

4.5 Limitation

Limitations must be considered when evaluating the results from cephalometric studies. Measurements of sagittal dimensions on the lateral cephalometric radiograph may be misleading due to the superimposition of structures. In particular, in the case of the interclinoid distance, the measurement may be performed between the opposite left and right structures. The most accurate assessment of sella turcica bridging and dimensions has been reported in CBCT studies [34, 35, 36, 37, 38, 39]. The cephalometric evaluation was found to be a suitable diagnostic method for diagnosing complete sella turcica bridging but it overestimates partial bridging [37]. Acevedo et al. [37] found that lateral cephalogram could reveal all cases of complete bridging, but in addition, the same number of pseudo bridges was classified. CBCT did not show any association between sella turcica bridging and canine impaction [38], and there was no difference between the two groups in terms of sella dimensions. Although these conclusions enhance the diagnostic accuracy of CBCT imaging, it has limited relevance in orthodontic practice. CBCT scans usually display dentition and the surrounding jaws, and the cranial base with sella turcica is usually not visible. CBCT scans showing the entire craniofacial region require a high radiation dose compared to conventional radiography, and they are not necessary for a regular orthodontic visit. For this reason, the CBCT examination of the sellar area is not usually indicated, and a lateral cephalometric radiograph is still preferred.

Future studies should focus on the appropriate definition of measurement points and dimensions in terms of their clinical relevance. CBCT evaluation of the relationship between the presence of sella turcica bridging and other genetically determined dental anomalies should be also considered. In addition, a more detailed comparison of the cephalometric and CBCT results could contribute to the development of an appropriate classification.