Introduction

Anatomical characteristics of the nose and paranasal sinuses exhibit considerable variations. Many of those perturbations vary depending on age, gender, geography, race, and ethnicity, though some differences aren’t significant statistically [1,2,3,4,5,6,7,8]. Nevertheless, quite a few of these variations have been linked to aetiopathogenesis of inflammatory sinonasal diseases [9,10,11,12]. Knowledge about these variations is not only relevant diagnostically, but also play a significant role in reducing the intraoperative difficulties and post-operative complications of endoscopic or open sinus surgery/skull base surgery [5, 13,14,15,16,17]. Computed tomography (CT) of the nose and paranasal sinus is of paramount importance in detecting the type and extent of anatomical perturbations in this region. Though the prevalence rates and patterns of the variations have been studied radiologically in the past, studies looking into almost all of the anatomical differences comprehensively in the same cohort are very sparse. Most of the existing studies, including those from this region are of lesser sample size and have evaluated limited parameters [7, 18,19,20,21]. The present study with larger sample size describes almost all the anatomical variations of paranasal sinuses in the study cohort. Furthermore, the study uniquely explores some of the clinico-radiological parameters like tooth root location in the maxillary sinus, the maxillary septation, the unsafe nature of the sphenoid septation, most of which have not been defined in the previous studies.

The primary objective of this study was to evaluate the anatomical variations in CT images of the nose and paranasal sinuses in a cohort of patients having inflammatory sinonasal diseases, in the form of chronic rhinosinusitis. Subset analysis in the study was meant to investigate associations between the various anatomical perturbations if any. And lastly, the study also tries to establish the clinically meaningful correlations between the prominent anatomical variations and corresponding clinical manifestations of sinonasal diseases.

Materials and Methods

It is a retrospective study done at a tertiary care medical center in the southern part of India. After obtaining the institute’s ethical committee approval, the CT images of the nose and paranasal sinuses of chronic rhinosinusitis patients managed by the Department of Otorhinolaryngology between June 2016 and November 2018 were retrieved for the analysis in this study. The CT on these patients was done using 64 slice Philips Brilliance, Philips Healthcare Suzhou Co Ltd, China. The 3 mm axial images were acquired cranio-caudally from the highest point of frontal sinus to the lowest part of the maxillary sinus and then reconstructed to 0.6 mm slices multiplanar images. The anatomical variations and their characteristics noted in the CT images included the deviated nasal septum (DNS), the concha bullosa (CB), the Haller cell (HC), the Onodi cell (OC), pneumatization of the anterior clinoid process (ACP), the pterygoid base (PB), the superior turbinate (ST), the inferior turbinate (IT), the crista galli (CG) and the nasal septum. Apart from the prevalence rates of these variations, the classifications, and the laterality were studied, and the relevant associations were investigated. DNS was classified radiologically into obstructive or simple, depending on whether or not the septum is touching the lateral wall or the turbinate on the ipsilateral side. The CB was classified as bullous, lamellar or extensive based on which part of the middle turbinate [MT] is pneumatized, whether bulla, lamella, or both. The variations in the depth of olfactory fossa, the relationship of the optic nerve to sphenoid sinus wall and the sphenoid pneumatization were analyzed separately.

The olfactory fossa depth was represented by the height of the lateral lamella of the cribriform plate and was classified into three types as per the Keros’ classification system [22]. Similarly, the relationship of the optic nerve to the sphenoid sinus wall was categorized as per Delano et al. [21, 23]. The sphenoid pneumatization itself was classified into either of conchal, pre-sellar, sellar and post-sellar varieties depending on the extent of exposure of sella turcica into the sphenoid sinus [3, 7, 24,25,26,27]. The study also explores the prevalence of the unsafe type of sphenoid septation that is having an attachment to the intra-cavernous carotid artery or optic nerve, and the implications of protrusion of tooth root into the maxillary sinus in its opacification. Accordingly, to verify the association of maxillary sinus opacification with tooth root, DNS, CB or HC, the maxillary sinuses and these characteristics were considered separately on each side, except for DNS, for which the presence of maxillary opacification on either of side was deemed to be positive.

The first two authors together evaluated the coronal, the sagittal, and the axial images of all the patients and assigned each anatomical characteristic an agreed type or grade or both, wherever applicable. In case of any disagreements or ambiguity in any parameter, the report given by the radiologist was taken into consideration, and the said features were assigned accordingly. CT images of those patients who were of age less than 18 years and those who had undergone surgery in the past were excluded from the study. The demographic data, the clinical manifestations, and the radiological characteristics were tabulated separately into a spreadsheet. Using the STATA 13.1 (StataCorp. 2013, College Station, TX: StataCorp LP), the appropriate statistical tests were performed to find out clinically relevant associations between these variables.

Results

The included patients were either partial responders or non-responders to maximal medical therapy for chronic rhinosinusitis and were managed by surgical intervention encompassing balloon sinoplasty, middle meatal antrostomy, frontoethmoidectomy or functional endoscopic sinus surgery, depending on the extent of the disease. These patients had been subjected to plain CT of the nose and paranasal sinuses before their operation, which was retrieved from the digital repository of the institute and was analyzed in this study. A total of 151 patients’ CT images have been analyzed in this study. All the patients included were adults aged between 18 and 71 years, with a mean age of 42 years. The majority were male patients [n = 101] with a male: female ratio of 2:1. The anatomical variations and their types/grades noted in the CT images of the study cohort have been summarized in Table 1. Representative CT images of the anatomical differences identified in this study are shown in Figs. 1, 2, 3 and 4. The most common anatomical variation identified was deviated nasal septum seen in 126 patients. Pneumatization of the MT (Fig. 1a–c), ST and IT were respectively seen in 74, 6, and one patient. The presence of the HC (Fig. 1d, e) and the OC (Fig. 1f) were noted in 59 and 35 patients, respectively. Sixty-five of our patient had pneumatization of the CG (Fig. 2a), 41 each of the nasal septum (Fig. 2b) and the ACP (Fig. 2c), and 71 had that of PB (Fig. 2d–f). In our study, the most common type of sphenoid pneumatization was post-sellar type (Fig. 3d) followed by sellar type (Fig. 3c). Majority of the patients had Kero’s type II olfactory fossa (Fig. 3j), and 5% of patients had low lying cribriform plate (Fig. 3k). The optic nerve’s relationship to the sphenoid sinus wall was commonly representative of Delano type II (Fig. 3f).

Table 1 Summary of anatomical variations seen in the present study
Full size table
Fig. 1
figure 1

Coronal images of plain computer tomography of paranasal sinus showing, a bulbous type, b lamellar type and c extensive type of concha bullosa [yellow arrows], d, e haller cell [blue arrows], f onodi cell [white arrow]

Full size image
Fig. 2
figure 2

Coronal images of plain computer tomography of paranasal sinus showing, a pneumatized crista galli [yellow arrow] and extensive type of concha bullosa [asterisk], b pneumatized nasal septum [white arrow], c pneumatized anterior clinoid process [green arrow], d–f pneumatized pterygoid base [green circle], exposed maxillary nerve [red arrows] and exposed vidian nerve [red arrowheads]

Full size image
Fig. 3
figure 3

Plain tomography of paranasal sinuses, first row of sagittal images showing sphenoid pneumatization pattern, a conchal type, b pre-sellar, c sellar and d post-sellar; second row of coronal images showing optic nerve relationship with sphenoid sinus wall as per Delano et al. classification, e Type I, f Type II, g Type III [triangle pointing the exposed optic nerve], h Type IV [star inside Onodi cell]; third row consisting of coronal images representing the depth of olfactory fossa as per Keros classification, i Keros I, j Keros II and k Keros III [blue line representing the depth], l Type III Delano with exposed optic nerve [yellow arrow]

Full size image
Fig. 4
figure 4

Coronal images of plain computer tomography of paranasal sinus; first row showing protrusion of tooth root into the maxillary sinus, a with maxillary sinus opacification, b mucosal thickening and C no opacification; second row showing unsafe sphenoid septa attached, d to both optic nerve and carotid artery, e to carotid artery and f to optic nerve

Full size image

Interestingly, the maxillary sinus had septation in 9 of our patients, all of them extending from the lateral wall to the floor. In 37 of study patients the sphenoid septation had an attachment to either of the optic nerve (Fig. 4f) or carotid artery (Fig. 4e) or both (Fig. 4d), constituting an unsafe pattern of septation. The association between the anatomical characteristics with each other were calculated by using a contingency table and Pearson Chi Square test, as shown in Table 2. However, most of the associations could not attain statistical significance as except a few. In our study, only the presence of OC was found to be significantly related to septal pneumatization (p = 0.03) and pneumatization of CG (p = 0.05). In our set of patients with chronic rhinosinusitis, radiological opacification of maxillary sinus showed significant association separately with the DNS (p = 0.03) and with the CB (p = 0.04). Statistically, the location of tooth root protruding through the maxillary floor into the sinus (Fig. 4a–c) did not have any relationship with the maxillary sinus opacification.

Table 2 Association between various anatomical variations, and between the clinically relevant anatomical variations with corresponding radiological abnormalities
Full size table

Discussion

This study has evaluated almost all of the clinically relevant anatomical variations that could be appreciated in the CT images of paranasal sinuses belonging to more than 150 patients. The results of our study are in line with the rates reported by different authors over the last two decades, concerning most of the variations except a few as summarized in Table 3. In fact, for many anatomical variations, the prevalence rates in the reported literature itself exhibit non-uniformity and a wide range of distribution. The primary reason for this variability across the studies could be attributed not only to the geographical, racial and ethnic differences between the study participants in different studies, but also to the differences in methodological considerations among these studies. While most studies have reported the prevalence rates for the number of included patients, some studies’ reports are based on the number of sides or sinuses evaluated and not the patients involved, as shown in Table 3. To make it comparable with either of these type of reports, we have reported prevalence rates of all anatomical factors separately for the number of patients and the number of sides as the denominator, except for the variations in the unpaired midline structures like nasal septum, and CG.

Table 3 Comparison of anatomical variations reported by different authors over the last two decades
Full size table

The anatomical variations like DNS, CB, HC, OC, pneumatization of PB, and ACP had been studied in detail by various authors with the inclusion of participants from across the globe. Our results were within the reported range in the literature for all of these factors. On the other hand, only a few studies have discussed the pneumatization status of nasal septum and CG, and the prevalence rates of these two in our study were considerably more than the several previous reports. However, studies by Mladina et al. respectively showed pneumatization in the perpendicular plate of ethmoid in 34.4% and pneumatization of CG in 66.6%, by doing cone-beam CT of human skull bones [40, 41]. Contrary to pneumatized nasal septum seen in 2% of the cases reported by Dua et al. [20] that were involving vomer bone of septum, we identified 27% our cases to have nasal septal pneumatization, and all were seen involving the perpendicular plate of the ethmoid, similar to findings of Mladina and group [40]. 43% of pneumatized CG seen in our study is between the 22.9% reported by cobzeanu et al. and the 66.6 by Mladina et al [32, 41]. The literature review of the other existing studies rates the CG pneumatization to vary between 3 and 37.5% [32, 41]. Further studies in this regard can validate the appropriate prevalence rates of these.

Presence of DNS was seen in 83% of our patients, most of whom were symptomatic in terms of nasal obstruction, independent of the type of deviation. As reported previously [46], most of the other anatomical variations were found to be higher in patients with DNS, however, we could not find any statistically significant association between them. With respect to CB, though the frequent type in our study was ‘extensive type’ like in other studies, we had more of unilateral CB than what is reported in the literature [42, 46, 47]. Statistically, the presence of CB had no bearing on the pneumatization of any other ethmoidal or sphenoidal bony landmark in paranasal sinuses of our study cohort. Some authors have shown a significant correlation between the surface area of CB and the number of aerated ethmoid structures [48]; however, such an analysis was beyond the scope of our objective and methodology. Nevertheless, in our study the presence of OC showed significant association with the aeration of septal part of the perpendicular plate of the ethmoid, but not with that of other ethmoidal or sphenoidal bony parts like ST and CG or PB and ACP, respectively. In practical, the pneumatization status of PB and ACP are very relevant both prognostically as well as therapeutically. The PB is related to and can expose the maxillary nerve and the vidian nerve into the sphenoid or posterior ethmoid sinus. Similarly, the ACP pneumatization can play a role in protrusion or dehiscence of the optic nerve and less commonly the internal carotid artery into the sphenoid sinus [17, 35, 37, 49, 50], making them vulnerable to the disease process or the surgical trauma. Therapeutically, the aeration pattern of PB can be utilized as a sinonasal corridor to control lesions of middle and posterior skull base through endoscopic endonasal transpterygoid [16], intrasphenoidal or transsphenoidal approaches [17]. As discussed earlier, the pneumatization rates of PB and ACP in our study were very much comparable to those in the literature. But what needs to be critically appreciated here is the rate of asymmetry of these anatomical structures between the two sides in the same patient. The asymmetry across the various anatomical characteristics can vary from 10 to 50% [26, 36, 42, 47], as was the case in our study also.

Another important anatomical consideration that is crucial for pre-operative planning of surgical approaches to structures or lesions surrounding the sphenoid like the pituitary gland, cavernous sinus, vidian nerve and maxillary nerve, is the pneumatization pattern of sphenoid sinus itself [3, 16, 17, 27, 35]. Post-sellar type of pneumatization being the most prevalent type like in our study as well as in other studies [3, 7, 26, 31] or second common type [15, 24, 25, 27] following sellar, provides a wide surgical window through sphenoid or pterygoid base during skull base surgery. However, attention has to be given to some of the other anatomical perturbations in the sphenoid sinus that could also be detrimental in preventing surgical morbidity. In our study, septation of sphenoid was found to be unsafe in 24% of the patients, most of which were having an attachment to the carotid canal, some to the optic canal and a few to both the optic and carotid canal. Attachment of sphenoid septation to the carotid canal has been reported by others in up to 37.5% of the cases [13,14,15, 31]. In 2 of our cases, the carotid artery was dehiscent inside the sphenoid sinus, and optic nerve was coursing through the sphenoid sinus (Type III Delano) in 22.5% of our patients. The reported rate of ‘protrusion or bony dehiscence of internal carotid artery’ into the sphenoid sinus is as high as 37. 5% and that of ‘protruding or dehiscent optic nerve’ is up to 30% [13, 21, 23, 24, 28, 30, 35, 37, 43, 45, 49]. Knowledge about these variations in sphenoid sinus anatomy including the pneumatization pattern, the attachment of inter-sinus septa and the vulnerable neurovascular anatomy is of utmost importance to ensure surgical safety. When it comes to prevention of surgical morbidity, the depth of olfactory fossa is another critical area which needs to be assessed in the pre-operative setting to prevent any inadvertent damage to the anterior skull base and resultant cerebrospinal fluid leak. As seen in our cases and in the literature, though majority would have a favorable ‘Kero’s type II olfactory fossa’ [28], the low lying ‘Kero’s type III olfactory fossa’ is not uncommon and needs a careful planning as well as meticulous handing of skull base while performing sinus or skull base surgeries.

Finally, we hypothesized the possible involvement of tooth root protrusion into the maxillary sinus in the causation of maxillary sinus opacification, independent of other anatomical variations like DNS, CB, and HC. However, on analysing our data, only the DNS and the CB showed significant association with radiologically opacified maxillary sinuses; but the tooth root protrusion and HC did not. Though the location of tooth root especially had no bearing on maxillary sinus opacification in this study, we could not incorporate the status of the tooth as healthy or as caries, in the analysis. It is still possible that the chronically diseased tooth having its root protruding into the sinus may have some implication in causing maxillary sinusitis. On the other hand, though we identified maxillary septa in 6% of cases, the rate that is variable in literature [20, 51], and the implication of maxillary sinus septa in pathophysiology of maxillary sinusitis is yet to be determined. Further studies are needed to understand these associations. Our study, though unique in many ways has some limitations. Being retrospective in nature, no comparison or association of anatomical variations with the endoscopic findings could be incorporated in our study, and similarly, we could not analyse the surgical results, nor could we correlate them with the anatomical variations.

Conclusions

Almost all of the anatomical variations noted in this study had the prevalence rates that were comparable with the literature. DNS, CB are the two significant anatomical variations that were encountered, and both were found to have an association with maxillary sinus opacification in chronic rhinosinusitis patients. The associations between most of the anatomical changes were not significant statistically. Similar to previous reports in the literature, the asymmetry between the sides in the same patient is quite commonly seen in our study. Overall, considering the gross variability in the critical neurovascular and bony anatomy of paranasal sinuses, it is imperative that these variations be studied thoroughly before any interventions involving the sinuses.