Morphometric Alterations of the Sphenoid Ostium and other Landmarks in Acromegaly: Anatomical Considerations and Implications in Endoscopic Pituitary Surgery
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.288996
Source of Support: None, Conflict of Interest: None
Keywords: Acromegaly, endoscopic, landmarks, morphometry, sphenoid ostium
The trans-sphenoidal route is the standard approach to pituitary adenomas and other sellar lesions. Thorough knowledge of the anatomical structure of the sphenoid and its variations is necessary to ensure a safe and effective surgical strategy.,, The sphenoid ostium (SO) is one of the most important landmarks available to the operating surgeon, and its intraoperative visualisation is the first key step in trans-sphenoidal surgery., Changes in size and position of the ostium in relation to other skull base structures in patients with acromegaly have not been adequately evaluated. The aim of this study was to assess the morphology of various skull base landmarks in acromegaly with particular reference to the SO and analyze these values with respect to the preoperative growth hormone (GH) levels. The skull base anatomy was also compared between acromegalics and nonacromegalics.
Study design and patient population
A retrospective case–control study of 36 adult patients who underwent endoscopic trans-sphenoidal surgery for pituitary adenomas between January 2014 and January 2016 was performed. Preoperative images of these patients were retrieved from the hospital database, and their medical records were reviewed for data related to their demographic profile, clinical parameters, and hormonal analysis. The study group consisted of 18 consecutive patients with GH adenomas who had undergone endoscopic trans-sphenoidal surgery. Their preoperative radiological and hormonal values were compared with those of 18 age- and sex-matched controls who had undergone endoscopic surgery for nonfunctioning pituitary adenomas (NFPAs).
All patients had undergone preoperative computed tomography (CT) of their nasal cavities and paranasal sinuses and magnetic resonance imaging (MRI) brain with contrast. Standardized imaging protocols for both CT and MRI were used for all patients during the study period. The CTs were performed on a GE Discovery 750HD helical CT scanner (GE Medical Systems, Milwaukee, WI, USA) with 1 mm thickness cuts and two-dimensional reconstruction. The CT parameters used were as follows: 512 matrix size, 120 kVp, 220 mAs, and field of view 180 mm × 180 mm. MRI brain was performed on a HDi 1.5 Tesla magnet (GE Signa, Milwaukee, WI, USA) using a standard NV coil. After retrieving the images from the hospital radiographic system (Synapse; Fujifilm Health Systems, USA), ImageJ 1.x software  wasused for evaluation of the nasal, sphenoid, and sellar anatomy.
The radiological evaluation was done by two independent blinded observers (VH, a radiology consultant, and NR, a neurosurgery resident) who were blinded to the group to which the patient belonged (GH adenoma vs. NFPA groups). The means of their readings for all the measurements were used for the analysis. The size of the SO on either side was measured in the vertical and horizontal dimensions, and their respective distances from the midline were measured. The mean values obtained from either side were calculated for the analysis.
The location of the SO relative to the vertical height of the sphenoid face (in relation to the anterior skull base) was determined using the “SO–skull base” and “skull base–choana” distances [Figure 1]. The “SO–skull base” distance was measured from the upper margin of the SO to the lowest point of the olfactory fossa. The “skull base–choana” distance was measured from the lowest point of the olfactory fossa to a line drawn tangential to the highest point on the floor of the choana (hard palate). As it is difficult to measure absolute distances in view of the concave and down-sloping sinus floor, we used the ratio of the SO–skull base and skull base–choana distances to determine the relative position of the SO.
The sphenopalatine foraminae (SPFs) were identified on the best-visualized axial CT sections [Figure 2]. The distance between the SO [Figure 3] and the SPF was calculated using the formula for calculation of the distance between two points in a three-dimensional plane, that is d = (x2 − x1) 2 + (y2 − y 1) 2 + (z2 − z1) 2 where x2, y2, and z2 and x1, y1, and z1 were the three-dimensional coordinates of the SPF and SO, respectively.
The dimensions of the sphenoid rostrum (anteroposterior and width) were recorded. Distances between various landmarks were calculated [Figure 4]. The distance between the pyriform aperture and the SO, SO and the anterior wall of the sella, total depth of the surgical field or working distance (distance from the pyriform aperture to the dorsum sella), and the anteroposterior sellar dimensions were recorded.
Sphenoid sinus pneumatization was recorded as being conchal, presellar, sellar, or postsellar. The number of intrasphenoid septae and their deflection toward or attachment to the bony wall covering the internal carotid artery (ICA) were noted. ICA protrusion, defined as the protrusion of more than half of the diameter of the vessel into the sphenoid sinus, was looked for.
T2-weighted MRI coronal sequences were used for measuring the diameter of the intracavernous carotid and the intercarotid distance. Intracavernous carotid diameters were measured on the best-visualized cross-sectional view of the cavernous ICA. The diameters were measured perpendicular to the long axis for both the right and left internal carotid arteries. Diameters for each side were measured separately, and the mean artery diameter was then calculated. Intercarotid distance was measured between the intracavernous ICAs in mid-sellar coronal sequences.
The sample size was calculated using the formula for paired means. The minimum difference value and standard deviation values were obtained from a previous study that had compared the nasal morphometry of acromegalics and nonacromegalics. Using an alpha value of 0.05, a sample size of 12 in each group was calculated to be adequate to obtain a power of 0.8. Data were entered in an Excel spreadsheet (Microsoft Office 2007) and analyzed using SPSS version 20. Means and standard deviations were computed for continuous variables. The paired t-test and z-test for proportions were used to compare differences between the measurements in the GH adenomas and the NFPA groups. Correlations of the preoperative GH levels with various anatomical measurements were analyzed using Pearson's correlation test. A P value of <0.05 for any of the results was taken to be significant. Interobserver variability was measured using the intraclass correlation coefficient and standardized ratings of agreement.
Demographics and tumor characteristics
The age- and sex-matched groups had 8 males and 10 females each, with a mean age at presentation being 35.5 ± 8.15 years. The maximum tumor diameter was similar in both the groups (mean value of 3.19 ± 1.32 cm for the GH-adenoma group and 2.92 ± 1.21 cm for the NFPA group). None of the patients had radiological evidence of apoplexy.
Morphology and location of the SO
The mean vertical and horizontal measurements of the SO in acromegalics were 2.66 ± 1.74 and 2.34 ± 1.71 mm, whereas in the NFPA group they were significantly smaller at 1.21 ± 0.66 and 1.03 ± 0.60 mm, respectively (P = 0.005). The left and right ostia were symmetric in location, within and between the groups. The SO was located further from the midline and closer to the SPF in the acromegalics [Table 1]. The SO–skull base distance and the skull base–choana distance were similar in the two groups (5.63 ± 2.83 and 6.38 ± 3.56 mm; 41.18 ± 3.25 and 40.09 ± 2.55 mm for the GH and NFPA groups, respectively). The location of the SO in the sphenoid face in relation to the skull base was thus similar in both the groups.
Other anatomical landmarks
Various distances between anatomical landmarks [Table 2], including the total depth of the surgical corridor from the pyriform aperture to the dorsum sella, were not significantly different in the two groups. The dimensions of the sphenoid rostrum and the pneumatization pattern of the sphenoid sinus were similar in both the groups [Table 2]. The number of patients with accessory septae and carotid-directed septae was higher in the GH-adenoma group, although this was not statistically different [Table 2]. Carotid artery protrusion was similar between the two groups. Acromegalics demonstrated significantly larger carotid diameters and smaller intercarotid distances [Table 2].
Correlations of GH levels with various anatomical landmarks
Various parameters in the acromegalics were analyzed for correlations with preoperative GH titers. Increasing GH levels were found to have a strong negative correlation with the SO–skull base distance [Figure 5] and a positive correlation with the distance of the SO from the midline [Figure 6]. Higher GH levels also demonstrated a strong negative correlation with the SO–SPF distance [Figure 7]. A moderate negative correlation was obtained between increasing GH levels and intercarotid distance (r = −0.46, P = 0.03). None of the other measurements demonstrated any significant correlations with the GH levels (P < 0.05).
The agreement between the two observers was substantial for the measurement of the various anatomical landmarks on CT (weighted kappa coefficient ranging from 0.81 to 0.90) and almost perfect for measurement of tumor size, carotid diameter, and intercarotid distance (weighted kappa coefficients of 0.97, 0.92, and 0.96, respectively).
Morphological changes in acromegaly
GH-secreting pituitary adenomas cause the clinical syndrome of acromegaly. Along with systemic changes, acromegalics have extensive craniofacial and skull base dysmorphic changes including hyperpneumatization of the sinuses, nasal bone hypertrophy, mandibular overgrowth, maxillary widening, and frontal bossing. Endoscopic trans-sphenoidal surgery is accepted as being more difficult in acromegalics because of the hypertrophic bony changes and mucosal thickening that compromises the working space and endoscopic view. Most of the existing morphometric studies in acromegalics have assessed mucosal and turbinate thickening, sphenoid sinus pneumatization and septations, surgical working distance,, carotid artery caliber, and intercarotid distance. However, alterations in the morphology and location of the SO and the related endoscopic corridor has not been evaluated. Knowledge of these changes would contribute to a safer and more effective surgical strategy in these patients.
Morphology and location of the SO
The sphenoid sinus is the common passage that is used to access the anterior, middle, and posterior cranial fossae in endoscopic skull base surgery and the SO forms the most consistent and reliable route of entry into it. It is also a good anatomical landmark for the surgeon to orient the approach.,, Even though many studies have been conducted on the location and morphology of the SO,,,,,,,,,, there is a dearth of information on the differences in their anatomy between acromegalics and nonacromegalics. This is surprising, considering that localization of the ostia is known to be more elusive in acromegalics due to their characteristic mucosal hypertrophy.
There seem to be considerable variations in the size of the SO across populations. For example, the average (vertical and horizontal) dimensions of the SO have been reported to be 2.78 ± 1.35 × 1.48 ± 1.01 mm in the Europeans, 5.62 × 3.12 mm in the South Americans, and 2–6 mm in an Indian population. In our study, the SO was found to be much smaller with mean dimensions of 1.21 ± 0.66 mm (vertical) and 1.03 ± 0.60 mm (horizontal) in the NFPA group. The ostia in the GH-adenoma group, however, were almost double this size.
Previous studies, including those done in the Indian population,,, have found the SO to be at the midpoint of the sphenoid face. In our study, the SO was found to be located in the upper half of the sphenoid face. The distance of the SO from the midline in our study was around 4–5 mm, with the distance being significantly more in the acromegalics. This lateral displacement of the SO was also found to correlate with increasing GH levels. This is of importance during the nasal phase of endoscopic surgery already rendered difficult by the mucosal hypertrophy and relatively less intranasal working space seen in these patients.
Relation of the ostium to the SPF
The sphenopalatine artery within the SPF is situated slightly superior to the posterior end of the middle turbinate and is the limiting factor for lateral and inferior enlargement of the SO during surgery. Previous studies have reported the SO– SPF distance to be approximately 7–13 mm., In our study, the SO–SPF distance was significantly less in the GH-adenoma group. Also, increasing preoperative GH titers had strong negative correlations with the SO–SPF distance. This means that the more the preoperative GH value, the more the SO is displaced laterally on the sphenoid face, closer to the SPF. This decreased SO–SPF distance translates to a potentially higher risk of injury to the sphenopalatine artery, with the risk being proportional to the preoperative GH titres. In situ ations where a nasoseptal flap is planned for skull base reconstruction, any damage to the artery will preclude the usage of the flap.
Other anatomical variations in acromegaly
The total depth of the surgical field or working distance in endoscopic pituitary surgery is the distance from the pyriform aperture to the dorsum sella. Saeki et al. and Carrabaet al. in their CT-based studies found that the sella was not deeper in acromegalic patients; this finding is borne out in our study as well. This implies that access is not more difficult or unfavorable to the surgeon in acromegaly. The type of sphenoid sinus and intrasinus septae in our study were similar across the two groups, except for a slightly higher number of accessory and carotid-directed septations in the GH-adenoma group. The carotid artery may bulge into the sphenoid sinus in 65%–72% of patients, and in 4%–8% of cases, the thin bone separating the two may be absent.,, Sasagawa et al. reported that carotid artery protrusion and dehiscence were more frequent in acromegalics. In our study as well, the incidence of carotid artery protrusion was found to be higher in acromegalics.
Changes related to the carotid artery
Besides causing obvious macroscopic anatomical changes, chronic GH elevation also causes microvascular changes in the endothelium. This is known to result in an increased caliber and tortuosity of the carotid arteries and a higher tendency for aneurysm formation.,, In our study, the acromegalics were found to have larger carotid diameters and smaller intercarotid distances in line with findings described previously.,, These vascular changes can be potentially catastrophic if overlooked during the generous sellar floor exposures that endoscopic surgery entails. The negative correlation of the intercarotid distance that we found with preoperative GH titers translates to a need for added surgical caution in patients with higher GH values.
The study has the inherent disadvantages of a retrospective study with regard to recruitment of cases and retrieval of data. Although the study was well-powered, a larger sample size may have yielded more significant results.
The dimensions of the SO are larger in patients with acromegaly. With increasing GH levels, the ostium tends to be displaced upward and laterally on the sphenoid face and closer to the SPF, thus predisposing to a potentially higher risk of injury to the sphenopalatine artery. Patients with GH adenomas demonstrate larger carotid diameters and smaller intercarotid distances than patients with NFPAs. Awareness of such morphological alterations is important for the endoscopic pituitary surgeon.
Presentation at a meeting
Best-paper Award presentation at SKULLBASE CON 2016 in Jaipur.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]
[Table 1], [Table 2]