The aim of the present comparative study was to analyse whether the visibility of the MSS is comparable using PR versus CBCT and also to investigate whether the visibility on PR and CBCT images depends on cortical bone thickness in the maxilla or the actual dimensions of the septa. The hypothesis stated that the MSS would be more commonly identified on CBCT versus PR scans and in addition the visibility of the septa is influenced by the thickness of the buccal cortical bone wall of the maxilla and the dimension of the septa.

For the present study, 549 patients were initially selected from the database of a dental practice in Stuttgart, Germany after they underwent PR and CBCT imaging between February 2010 and January 2017. Both PR and CBCT images were already available prior to the start of the study. Three-hundred thirty-seven patients fulfilled the inclusion criteria of whom 173 patients were female and 165 were male patients. The average age of female patients was 62.1 years, and the average age of male patients was 58.5 years.

In the preliminary stages, measurement integrity was verified by an expert. The measurements were repeated at 2-week intervals. For both intra- and inter-observer reliability, Cohen’s kappa coefficient was computed as 1.0 with a 95% confidence interval of 0.92–1.00. Thus, a high degree of agreement was obtained.

The findings of our study indicate that on the CBCT images, the MSS could be detected in 266 cases in the maxillary sinus. However, only about half of these (135) were also visible on PR images (Fig. 5). Thirteen of the detected MSS on PR images were not visible on the CBCT images. A direct comparison of the corresponding images revealed that this difference may be caused by translucencies, such as those due to the hard palate or by imaging the dorsum of the tongue. Both the palate and dorsum are superimposed on the maxillary sinus on PR images and could result in misdiagnoses.

Sinus lifting has become a common surgical intervention aimed at producing an increase in alveolar bone height prior to dental implant placement in the posterior maxilla [17]. However, the occurrence of specific intraoperative complications, such as Schneiderian membrane perforation, bleeding from the antral alveolar artery, and/or the presence of septa must be considered [7, 18, 19]. Septa are common anatomical structures and are most often found in the first or second molar region on the floor of the maxillary sinus [15]. By contrast, Ulm et al. [19] and Pommer et al. [20] report that septa can occur in all maxillary sinus regions. Systematic reviews reported a frequency of the occurrence of maxillary sinus septa between approximately 13%–35% [20, 21]. The group around Irinakis et al. even report a 48.1% occurrence of septa [22].

Wen et al. performed a literature search and found that sinus augmentations are subject to an increased complication rate in the presence of septa [6]. The orientation of the septa can cause sinus augmentations to be even more difficult [6, 22].

Based on the results of their studies, some authors recommend a classification system of septa and treatment strategies to reduce complications during sinus augmentation [6, 21, 22].

A thorough knowledge of the anatomy of the maxillary sinus and its possible variations is essential to prevent complications during sinus augmentation [23].

Both PR and CBCT are available for diagnosis and planning before dental procedures [10, 11]. Due to the size and distribution of anatomical structures, not every area of interest can be clearly visualized on PR images and can therefore be negatively influenced by various structures [24]. In their experimental and comparative diagnostic study, Dau et al. found that PR alone remained insufficient for evaluating pathologies in the maxillary sinus. Furthermore, the presence of pathologies can affect the visibility of anatomical structures [25]. Pommer et al. concluded. That septa diagnosis using panoramic radiographs yielded incorrect results in 29% of cases [20]. Anamali et al. report in their study that CBCT images show highly beneficial information regardless of the presence of intra-sinus pathology [26]. CBCT images appear to enable an examiner-independent assessment of anatomical structures as they leave little room for interpretation of the findings [27]. Therefore, some authors suggest a thorough three-dimensional radiographic examination of the sinus prior to surgery [15, 20, 28, 29].

The results of our study show that septa generally occur in all regions of the maxillary sinus, both in CBCT and in PR. In their systematic reviews, Pommer et al. [20] and Maestre-Ferrín et al. [21] provide similar results. In our study, septa were found in 13.15% of the cases on CBCT images based on the regions (n = 2022). On CBCT images, however, the septa were most frequently visible in regions M2 and M1 followed by P2 and M3 at a similar level. The septa were observed much less frequently in region P1 and only to a negligible degree in region C1. The study thus shows similar observations regarding the visibility of the MSS and is confirmed by the results of the above-mentioned studies [19,20,21]. The visibility of the septa in PR images was highest in the P2 region. Otherwise, the septa behave similarly on CBCT images (Fig. 5). Overall, the number of septa observed on CBCT images was approximately twice as high as on PR images (Fig. 5). When directly comparing the visibility of septa between the corresponding regions, septa were detected in 34.5% of cases on CBCT and in 24.5% of cases on PR images (Tables 1 and 2, respectively). In our study, CBCT images show a significantly higher detectability of MSS in all regions (p < 0.05) compared to PR images with the exception of the canine region (p > 0.05).

For daily practice, the results of our study can be interpreted as follows: the differences in the detectability of MSS between PR and CBCT in the anterior region are smaller than in the posterior region. While in regions M3 (p < 0.000) and M2 (p < 0.017) the detectability of MSS in CBCT is significantly higher than in PR, this difference is not significant in region C (p > 0.625). In regions M1 (p < 0.041) and P1 (p < 0.039) the significance is marginal. Accordingly, the importance of the PR for detecting the septa in the anterior region is considered to be high. In the posterior area, CBCT would be more useful to identify the septa.

Regarding the influence of the thickness of the buccal cortical bone on the visibility of the MSS in a direct comparison between CBCT and corresponding PR images, there is a lack of specialist literature. In our study, the buccal cortical bone thickness showed in region M3 (p = 0.043) and M1 (p = 0.047) a significant influence on the consistency of MSS visibility in CBCT and PR measurements, if they were only visible in CBCT. The visibility of the MSS was significantly lower in PR the thicker the buccal cortical bone was. In all other regions no significant correlation was found between the visibility of the MSS and the buccal cortical bone thickness (p > 0.05).

The results here are a bit confusing. It could actually be the case that the thickness of the buccal compact influences the visibility of the septa only in regions M3 and M1. The study by Simsek et al. provides an indication of this [30]. The authors noted that the highest thickness of the lateral wall of the sinus was found in the first molar region. However, this does not explain the high significance in region M3.

On the other hand, it is also possible that the visibility of MSS in PR is influenced by anatomical variations as described by Shiki et al., which are stronger in the posterior region of the maxillary sinus [24]. The same rationale applies to the presence of pathologies that affect the visibility of anatomical structures [25,26,27]. These aspects were not examined in our study and thus cannot be sufficiently proven by studies. Further studies are necessary to investigate these aspects.

Another aspect of our study was to investigate the influence of the size of the septa on their visibility between the PR and CBCT images. The results of our study show that the measured values for the height of the septa on the CBCT (range 7.44–26.48 mm) images are higher than on PR (range 6.17–20.36 mm) images in all regions. However, significance could only be shown for the measurements in region M1 (Fig. 7; Wilcoxon test: p = 0.013). For the width of the septa, the values between PR (range 1.70–2.33 mm) and CBCT (range 1.57–1.74 mm) images differed only slightly. Significance was only found for region P2 (Wilcoxon test: p = 0.034) as shown in Fig. 8. Furthermore, the depth of the septa was considered a possible influencing factor for their visibility on PR images. The lowest values for the depth of the septa were measured in region P1, and the values were similar in all other regions.

The measured values for the dimension of the septa as an influencing factor for their visibility on PR images do not reveal a clear trend. Significance was found in all three measurements for the visibility of MSS in PR only in P2 (Mann–Whitney U test, width p = 0.041, height p = 0.001, and depth p = 0.007). Isolated cases of further significance in region M3 for width (Mann–Whitney U test p = 0.043), in region M2 for height (Mann–Whitney U test p = 0.024), and in region P1 for depth (Mann–Whitney U test p = 0.034) were found.

In the specialist literature, septa height showed great variability with a mean value of 8.72 mm (SD ± 4.26; range 3.7–18.4 mm) [23]. In a systematic review, Pommer et al. report that the height of the septa is on average 7.5 mm [20]. In their literature review, Maestre-Ferrín et al. provide values between 2 and 13 mm for the height of the septa [21]. The values measured in our study correspond to the previously reported studies. Nevertheless, no direct comparison was made between PR and CBCT in these studies.

The results of our study are consistent with the results of other studies in terms of the frequency and distribution of MSS [19,20,21]. Our study shows that septa generally occur in all regions of the maxillary sinus, basically being visible both on CBCT and PR images. However, septa can be detected almost twice as often on CBCT compared to PR. For the influence of the dimension of the septa on their visibility, no clear trend could be found in our study. Only in isolated cases was there a significant influence due to the septa dimension. This aspect should be investigated in more detail in further studies. The influence of the thickness of the buccal cortical bone on the visibility of MSS seems to be more promising. A significant influence due to the thickness was observed in regions M1 and M3. This aspect has not yet been investigated in the literature.

Operations on the maxillary sinus are complicated by the presence of antral septa, the visibility and dimensions of which determine the degree of surgical difficulty. For the purposes of reducing such complications, it is recommended that a thorough radiological assessment of the maxillary sinus be carried out in the sinus region prior to surgical intervention [8, 9]. The location, dimension and course of the septa can have an influence on the planning of maxillary sinus surgery. Wen et al. [6] were able to show in their literature review that knowledge of these parameters is essential when developing a strategy for planning maxillary sinus surgery.

In principle, it must be noted that CBCT is clearly superior to PR for the detection of MSS. Similar results were reported in the systematic review by Pommer et al. The authors found that the diagnosis of MSS by PR was incorrect in 29% of cases, which is a lot [20]. One reason for this is that, due to the technique, only structures that lie within the tomographic plane are clearly visible on PR images. Objects that are behind the tomographic plane appear wider and those in front of them appear narrower [12]. Horizontal and vertical distortions can be observed outside the tomographic plane. These biases are the reasons why the measurements in PR are unreliable. Incorrect positioning of the head can also lead to distortion of the anatomical structures. Incorrect positioning of the head relative to the midline causes horizontal errors, so that anatomical structures in the posterior region appear wider or narrower. Incorrect positioning of the head relative to the horizontal plane causes vertical errors, making structures appear longer or shorter [12]. It also plays a role that CBCT allows for fewer interpretations and the findings are less dependent on the examiner [27].

However, CBCT also has some disadvantages, e.g. CBCTs show poor soft tissue contrast [31] and artefacts, especially when metal or root canal fillings are in the area of interest [32]. Another typical problem with CBCT are artefacts that arise from movements of the head. The cause of this is insufficient fixation of the head. But the time required to produce a CBCT can also be a possible cause. Since larger FOVs take longer to produce, CBCTs with small FOVs focusing on the region of interest are more advantageous [33].

Finally, the question arises as to whether CBCT can have an influence on the planning of a surgical procedure on the maxillary sinus. The study by Kaeppler et al. [34] showed that creating a CBCT can have an influence on treatment planning. However, this only affected 9.5% of cases with suspected mandibular fractures. However, this aspect was not examined in our study. It remains to be seen whether further studies will provide clarity for other surgical indications.

Further studies are necessary here.

Therefore, our hypothesis that MSS would be more commonly identified on CBCT versus PR images can be verified. The assumption that the dimension in addition to the thickness of the buccal cortical bone influences the visibility of the septa can only be verified to a limited extent. Here, not all regions show significance.