There were 74.7% sellar type, 24.9% presellar type and 0.4% conchal type (Figure 1).

Protrusion of carotid canal (Figure 2) and optic canal were found in 48.3% and 13.1% of patients, respectively. Dehiscence was observed in 10.2% of patients for carotid canal and 5.8% for optic canal (Figure 3).

Table 1 showed the detail of protrusion and dehiscence of carotid canal and optic canal.

The Protrusion of maxillary and vidian (Figure 4) nerves was found in 18% and 9.5% of the patients, respectively. Dehiscence of maxillary nerves was observed in 6% of patients. Table 2 presented the detail of protrusion and dehiscence of maxillary and vidian nerves.

Pneumatization of lesser and greater wings of the sphenoid (Figure 5); of anterior clinoid processes (Figure 6) and of pterygoid processes (Figure 7) was found in 4.6%, 3.3%, 7.1% and 7.3% of patients, respectively. The distribution of these data according to their unilateral or bilateral location is presented in Table 3.

In our series, sellar type pneumatization was the most common, 74.7%. In the literature, variable proportions of different types of pneumatization have been reported. Sellar type was the most frequent. Wang et al. [6] reported in 2010, 98% of sellar type pneumatization and 2% for the presellar type. Other authors have found prevalence closer to ours. Notably Lu et al. [8] in China and El Kammash et al. [9] in Egypt, which reported 63.5% and 85.7% of sellar type pneumatization; 28.5% and 12.6% of presellar pneumatization respectively.

In the literature, there is great variability in the prevalence of protrusion and dehiscence of carotid canal and optic canal. This variability could be due to the criteria considered to define them. In our study, we considered as a protrusion the presence of at least one third of the canal circumference in the sinus cavity and dehiscence was defined as the absence of visible bone density separating the sinus from the course of the canal. In their work, Fasunla et al. [14] , Sirikci et al. [11] , Unal et al. [5] recognized protrusion as the presence of more than half of the canal circumference in the sinus. For Hewaidi et al. [12] , the presence of the canal in the sinus cavity to any degree was sufficient to define protrusion. Fasunla, Sirikci, Unal found respectively 27.3%, 26.1% and 30.3% for the prevalence of carotid canal protrusion. Our rate of 48.3% was close to that of Hewaidi which was 41%. This great variability is also observed at the optic canal. The prevalence of the protrusion of optic canal varies from 8% to 70% according to the authors [13] . In our series the prevalence of the optical canal protrusion was 13.1%. This wide range of prevalence may also indicate that the relationship of the carotid canal or the optic canal to the sphenoidal sinus is different between races and ethnic groups, but the variability of the definition criteria does not help to confirm this hypothesis.

If the surgeon operating the sphenoid sinus is not aware of the dehiscence or protrusion of the carotid canal, a wound of this vessel, even a small one, can be fatal, as it is barely possible to control the bleeding of an internal carotid artery injured in the sphenoidal sinus. Also, neurological sequelae may be unavoidable. Similarly, sphenoiditis can also make internal carotid artery vulnerable [11] .

In protrusion or dehiscence of optic canal, optic nerves injury may occur in the event of surgical trauma or as a complication of sphenoid sinusitis. The risk of blindness is high if the surgeon injures the nerve in the sinus [17] . Also, visual deficits can result from an infection of the sphenoid or a mucocele from compressing the optic canal. Compression of the optic nerve can cause ischemia and congestion of the ophthalmic vein. In addition, the optic canal is where the optic nerve is least vascularized, making it very susceptible to injury [18] .

Protrusion prevalence of maxillary and vidian nerves observed in our series was significantly lower than those found in the literature. We observed protrusion of the maxillary nerve in 18% of patients. Hewaidi et al. [12] , Unal et al. [5] , reported a prevalence of 24.3% and 30.3%, respectively.

We found protrusion of vidian nerve in 9.5% of our patients. Hewaidi et al. [12] found this protrusion in 27% of cases.

The divergence between these prevalences may reflect ethnic differences between the populations studied.

In endoscopic surgery, knowledge of these variants will contribute to improved outcomes and decrease the complications of sinus surgeries [9] .

For Hewaidi et al. [12] , it is likely that thin CT scan sections are more precise for the detection of pneumatization of the adjacent structures to the sphenoid sinuses. Thick slice may underestimate these anatomic variants especially when they are not joined. In his series, on 2 mm slice thickness, 20% had pneumatization of greater wings, 15.3% had pneumatization of anterior clinoid processes, and 29% had pneumatization of pterygoid processes. In our study, on 1 mm slice thickness, we found a much lower prevalence of pneumatization of adjacent structures to the sphenoid sinuses. 4.6% of our patients had pneumatization of greater wings, 7.1% had pneumatization of anterior clinoid processes, and 7.3% had pneumatization of pterygoid processes. On 3 mm slice thickness, Lane and Bolger [19] and Unal et al. [5] found a prevalence of 13% and 24.1% for pneumatization of anterior clinoid processes, respectively. These results suggest that beyond the thickness of the sections, the wide variability in the prevalence of pneumatization of adjacent structures can also be attributed to ethnic differences in the populations studied. The pneumatization of the pterygoid process, when present, is an important route for access to the skull base. It allows, for example, extensive transnasal endoscopic approaches to reach the pterygoid process through the medial part of the posterior maxillary wall [17] . These techniques may provide a pathway for endoscopic repair of cerebrospinal fluid leaks and for endoscopic biopsy of skull base lesions. Such information may be important for the surgeon in preoperative planning for skull base surgery. Also, the pneumatization of the pterygoid process thins the bone floor of the scaphoid fossa, resulting in an intimate relationship between the sinus and the auditory tube [20] .

The number of cases and the absence of additional readers to compare the results of the first reading are limits to this study.