Introduction
Tumours in the infratemporal fossa are notoriously difficult to operate upon. Most tumours occurring in this region do not originate there, but arise from nearby structures such as the paranasal sinuses, parotid gland, middle cranial fossa and nasopharynx.Reference Tiwari, Quak, Egeler, Smeele, Waal and Valk1 Typically, pathologies occurring in these structures range from benign nasopharyngeal juvenile angiofibromas, gliomas, schwannomas and meningiomas to more aggressive malignancies including adenoid cystic carcinoma, adenocarcinoma, squamous cell carcinoma and fibrosarcomatous lesions with extensive infiltration of surrounding tissues.Reference Tiwari, Quak, Egeler, Smeele, Waal and Valk1–Reference Eloy, Murray, Friedel, Tessema and Liu3 Indeed, it is well established that adenoid cystic cancer in particular shows a predilection for perineural spread along structures such as the trigeminal nerve branches.
Traditionally, tumours within the infratemporal fossa have been operated upon using a transfacial and/or transcranial approach to the skull base. This involved extensive facial incisions and osteotomies, with numerous complications arising from the surgery, including unilateral hearing loss, masticatory difficulties, cerebrospinal fluid (CSF) leakage and hypoaesthesia.Reference Eloy, Murray, Friedel, Tessema and Liu3, Reference Theodosopoulos, Guthikonda, Brescia, Keller and Zimmer4
Surgical techniques in otorhinolaryngology have changed dramatically over the years, especially with the introduction of minimal access endoscopic procedures. Indeed, for anterior skull base surgery, endoscopic approaches are now regarded as standard procedure.Reference Ong, Gore, Donnellan, Kertesz and Teo5, Reference El-Sayed, Pletcher, Russell, McDermott and Parsa6 However, pathologies occurring in the infratemporal fossa and around the foramen ovale remain among the more difficult to approach endoscopically.
More recently however, there has been a great deal of interest in access to this anatomical territory using the endonasal, transmaxillary route to access the infratemporal fossa.Reference Hosseini, Razfar, Carrau, Prevedello, Fernandez-Miranda and Zanation2, Reference Theodosopoulos, Guthikonda, Brescia, Keller and Zimmer4, Reference El-Sayed, Pletcher, Russell, McDermott and Parsa6–Reference de Almeida, Snyderman, Gardner, Carrau and Vescan14 However, to our knowledge there has been no previous, comprehensive data compilation and statistical analysis of the major anatomical, radiological and surgical landmarks which can be used to access the foramen ovale via the infratemporal fossa.
Therefore, in the present study we aimed to examine several key landmarks that could be used radiologically and surgically to guide an endoscopic, endonasal, transmaxillary, transpterygoid approach to the infratemporal fossa, in order to access pathologies arising in and around the foramen ovale. We then tested this approach on a cadaver to assess the ease and accuracy of accessing the foramen ovale.
Materials and methods
Radiological measurements
All radiological measurements were made using axial computed tomography (CT) brain scans analysed with the BrainLAB iPlan 1.1 Cranial software program (BrainLAB, Feldkirchen, Germany). The study initially included 183 patients, selected from a cohort that had undergone endoscopic or skull base surgery at the ear, nose and throat department of Glasgow Royal Infirmary between February 2005 and October 2012. All CT scans were pre-operative; if multiple scans were available for one patient, the earliest scan of acceptable quality was selected.
Four sets of measurements were assessed using the software: (1) the distance from the posterior maxillary sinus wall to the anterior border of the foramen ovale; (2) the distance from the anterior nasal spine to the anterior border of the foramen ovale; (3) the distance from the posterior base of the lateral pterygoid plate to the antero-inferior border of the foramen ovale; and (4) the angle between the anterior border of the foramen ovale, the anterior nasal spine and the sphenoid rostrum (at the level of the foramen ovale in the coronal plane). The key landmarks and measurement sites are illustrated in Figures 1 to 4. All the measured distances were taken as the shortest distance between the two points. The position of the foramen ovale was taken to be its location as identified in the first axial CT slice in which the full extent of the foramen could be identified (moving from inferior to superior). When measuring the distance between the posterior base of the lateral pterygoid plate to the foramen ovale, the foramen ovale was measured from the point at which its most anterior and inferior point could be identified on the axial and coronal scans. The base of the lateral pterygoid plate was defined as its most superior, lateral and posterior point. Measurements were assessed on the left and right sides of every scan twice by one of the authors, with at least one week between measurements.
Fig. 1 Pre-operative computed tomography screen capture images showing BrainLAB iPlan 1.1 displays, including (a) three-dimensional reconstruction, and (b) axial, (c) sagittal and (d) coronal views showing the foramen ovale marked in pink. The position of the maxillary sinus and the posterior maxillary sinus wall are shown. Measurement of the distance from the posterior maxillary sinus wall to the foramen ovale is shown in part (b). A = anterior; P = posterior; R = right, L = left; H = head
Fig. 2 Pre-operative computed tomography screen capture images showing BrainLAB iPlan 1.1 displays, including (a) three-dimensional reconstruction, and (b) axial, (c) sagittal and (d) coronal views showing the foramen ovale marked in pink. The position of the anterior nasal spine is shown. Measurement of the distance from the anterior nasal spine to the foramen ovale is shown in part (b). A = anterior; P = posterior; R = right, L = left; H = head
Fig. 3 Pre-operative computed tomography screen capture images showing BrainLAB iPlan 1.1 displays, including (a) three-dimensional reconstruction, and (b) axial, (c) sagittal and (d) coronal views showing the foramen ovale and the medial and lateral pterygoid plates in pink. The yellow line indicates measurement of the distance from the base of the lateral pterygoid plate to the foramen ovale, best recorded in the sagittal plane, as shown in part (c). A = anterior; P = posterior; R = right, L = left; H = head
Fig. 4 Pre-operative computed tomography screen capture images showing BrainLAB iPlan 1.1 displays, including (a) three-dimensional reconstruction, and (b) axial, (c) sagittal and (d) coronal views showing the foramen ovale in pink. The position of the anterior nasal spine and sphenoid rostrum is shown. Yellow lines indicate measurement of the angle between the foramen ovale, the anterior nasal spine and the sphenoid rostrum, best recorded in the axial plane, as shown in part (b). A = anterior; P = posterior; R = right, L = left; H = head
Statistical analysis
The PASW (Predictive Analytics Software) Statistics 18 software program (SPSS, Quarry Bay, Hong Kong, China) was used to examine and process the statistical data. A two-tailed, paired sample t-test was used to compare the right and left sides for all measurements. To determine whether there was a difference in measurements between male and female patients, a two-tailed, independent sample t-test was used for all measurements. For both types of t-test, a p value of less than 0.05 was considered statistically significant.
Dissection
In order to assess the feasibility of an endoscopic, transmaxillary, transpterygoid approach to the foramen ovale, and to test the radiological data obtained, surgery was performed on a fresh frozen (non-embalmed) cadaver selected from the regular stock within the Laboratory of Human Anatomy, University of Glasgow. All work was undertaken within the auspices of the Anatomy Act 1984 and the Human Tissue (Scotland) Act 2006, Part 5. No craniofacial or obvious otorhinolaryngological abnormalities were present in the cadaver. Dissection was performed by one of the authors (GWM), an experienced endoscopic anterior skull base surgeon working at a major university teaching hospital. The approach was undertaken on the left side of a 67-year-old male. Dissection was performed using a Karl Storz 4 mm, 0° and 30° rod-lens endoscope, and a three-chip endocamera. High-resolution images were obtained from the monitor used for visualisation of the surgical approach, and imported into the Adobe Photoshop CS4 application (Adobe Systems, San Jose, California, USA).
An endonasal, transmaxillary, transpterygoid approach was utilised. Initially, a 0° endoscope was used. Firstly, the middle turbinate was identified and displaced medially. Next, the ethmoid bulla and uncinate process were identified. To improve access to the natural maxillary os, the uncinate process was removed. For the next stages, we used a 30° endoscope. The natural maxillary sinus ostium was clearly identified. Posteriorly, the sphenopalatine artery was identified and the nasolacrimal duct opening was located anteroinferiorly from the opening into the maxillary sinus. To improve access, an endoscopic medial maxillectomy was performed.
On entering the maxillary sinus, the mucosa was elevated from the bone and removed. The bone of the posterior maxillary wall was removed in order to access the pterygopalatine fossa. After dissecting through Bechet's fat pad, the internal maxillary artery was identified. Posterior to the artery, the pterygopalatine ganglion and maxillary nerve were encountered.
The lateral pterygoid muscle was identified and elevated to provide access to the lateral pterygoid plate of the sphenoid. The lateral pterygoid plate was used as the guiding anatomical landmark to allow identification of the mandibular nerve exiting the foramen ovale.
Results and analysis
Radiological measurements
Initially, the BrainLAB iPlan 1.1 Cranial software program was used to assess the CT brain scans of 183 patients who had undergone endoscopic or skull base surgery between February 2005 and October 2012. However, 74 patients were excluded from the study: 50 for poor scan quality, 21 for distorted pathology of the region where measurements were to be analysed, and 3 for being paediatric cases. Of the remaining 109 patients, 62 (56.9 per cent) were male and 47 (43.1 per cent) female. At the time of surgery, the patients' ages ranged from 19 to 84 years, with a mean age of 50.8 years for men and 55.6 years for women.
Table I gives the mean values, 95 per cent confidence intervals and standard deviations for the following measurements: (1) the distance between the posterior maxillary sinus wall and the anterior border of the foramen ovale; (2) the distance between the anterior nasal spine and the foramen ovale; (3) the distance between the posterior base of the lateral pterygoid plate and the foramen ovale; and (4) the angle between the foramen ovale, the anterior nasal spine and the sphenoid rostrum (at the level of the foramen ovale in the coronal plane).
Table I Measurements
PMW = posterior maxillary wall; FO = foramen ovale; ANS = anterior nasal spine; PBLPP = posterior base of the lateral pterygoid plate; CI = confidence interval; SD = standard deviation
Distances from the posterior maxillary wall to the foramen ovale and from the anterior nasal spine to the foramen ovale were significantly greater in the men (with a mean ± standard deviation (SD) of 19.8 ± 2.2 mm on the left and 20.2 ± 2.4 mm on the right, and 80.2 ± 3.4 mm on the left and 80.4 ± 3.5 mm on the right, respectively) compared with the women (being 18.3 ± 2.2 mm on the left and 18.5 ± 1.9 mm on the right, and 74.0 ± 3.1 mm on the left and 74.1 ± 2.9 mm on the right, respectively) (p < 0.05 for all comparisons). The other two measurements (the distance from the posterior base of the lateral pterygoid plate to the foramen ovale, and the angle between the foramen ovale, the anterior nasal spine and the sphenoid rostrum) showed no statistically significant differences between the sexes. In addition, there were no statistically significant differences between the left and right sides, for any of the four measurements.
Dissection
Using the 0° endoscope, the middle turbinate was clearly identified (see Figure 5a) and displaced medially. After identifying the ethmoidal bulla and uncinate process, and removing this process, the maxillary antrum was entered and artificially widened (Figure 5b). Following this, the posterior wall of the maxillary antrum was opened up to gain access to the pterygopalatine fossa (Figure 5c), and the maxillary nerve was easily identified (Figure 5d). Deeper dissection along the lateral pterygoid plate, on the undersurface of the lateral pterygoid muscle, led to the foramen ovale and the emerging mandibular nerve, as shown in Figure 5(e). The foramen ovale was located approximately 6 mm posterior to the posterior border of the lateral pterygoid plate.
Fig. 5 Endoscopic images taken during performance of an endonasal, transmaxillary, transpterygoid approach to the foramen ovale on a fresh frozen (non-embalmed) cadaver. (a) The nasal cavity, showing the maxillary os, uncinate process, ethmoidal bulla and middle turbinate. (b) The maxillary sinus opened out from its medial aspect, with the middle turbinate shown towards the left edge of the image. (c) Performance of a posterior maxillectomy. Note the eggshell thinness of the posterior maxillary sinus wall, shown by the surgical dissection equipment. (d) The pterygopalatine fossa, with the maxillary nerve clearly seen, together with the closely related maxillary artery. (e) The mandibular nerve emerging from the foramen ovale, with the edge of the lateral pterygoid plate also seen. M = medial; L = lateral
This dissection indicated that endoscopic access to the foramen ovale is feasible, using an endonasal, transmaxillary, transpterygoid approach.
Discussion
The aim of our study was to establish an accurate, comprehensive dataset of measurements for anatomical, surgical and radiological landmarks relevant to endoscopic access to the foramen ovale (and thus the mandibular nerve) via the nose and maxillary sinus, in order to supplement current knowledge.
Our study findings demonstrated that, in the horizontal plane, there is a significant difference between men and women as regards the distance between the posterior maxillary sinus wall and the anterior border of the foramen ovale (20 mm in men and 18 mm in women), as well as the distance between the anterior nasal spine and the foramen ovale (80 mm in men and 74 mm in women) (p < 0.05 for both comparisons).
We also found that the angle between the foramen ovale, the anterior nasal spine and the sphenoid rostrum was consistent in both men and women, and on both the left and right sides (being approximately 19°). In addition, we found no statistically significant difference in the distance from the posterior base of the lateral pterygoid plate to the foramen ovale, comparing men versus women and the left versus right sides of the same patient.
Historically, the surgical treatment of pathologies in the infratemporal fossa was fraught with difficulty due to the deep, posterior location of this area. Indeed, pathologies generally do not originate in this site but, rather, infiltrate from surrounding structures.Reference Tiwari, Quak, Egeler, Smeele, Waal and Valk1 The traditional approach to this region was an open transfacial and/or transcranial skull base approach, radical surgery which by its very nature resulted in high morbidity rates.Reference Eloy, Murray, Friedel, Tessema and Liu3, Reference Theodosopoulos, Guthikonda, Brescia, Keller and Zimmer4
Nowadays, minimally invasive, minimal access endoscopic surgical approaches have replaced many extensive open surgical techniques. Endoscopic access is now a well established anterior skull base surgical technique, and has been shown to have equivalent or better rates of gross tumour resection, and less CSF leakage, compared with traditional open techniques.Reference Ong, Gore, Donnellan, Kertesz and Teo5, Reference El-Sayed, Pletcher, Russell, McDermott and Parsa6, Reference Komotar, Starke, Raper, Anand and Schwartz8 Compared with open tumour resection techniques, endoscopic methods have lower morbidity rates, better tumour clearance, shorter hospitalisation time and (in several studies) improved post-operative quality of life.Reference Eloy, Vivero, Hoang, Civantos, Weed and Morcos9–Reference de Almeida, Snyderman, Gardner, Carrau and Vescan14
Although the anterior skull base is easier to access endoscopically through the nose, the endoscopic approach to the lateral skull base, infratemporal fossa and foramen ovale is much more complex. Access to the infratemporal fossa has been studied by a number of authors, assessing both clinical and anatomical scenarios, with a variety of modifications proposed for each approach.Reference Hofstetter, Singh, Anand, Kacker and Schwartz15–Reference McCoul, Schwartz and Anand21
Two major routes of access to the infratemporal fossa have been described: the lateral and the anterior approach.Reference Eloy, Murray, Friedel, Tessema and Liu3, Reference Vilela and Rostomily22–24
The lateral approach adds great complexity to the procedure as it generally involves resection of the parotid gland (and related facial nerve dissection), or dissection or resection of the temporalis muscle and its related attachment (the mandibular condyle). In addition, the post-operative complications of this approach are very serious, and include facial paralysis, hearing loss, and temporomandibular joint pain and/or dysfunction.
The anterior approach can be divided into traditional transmaxillary and endoscopic transmaxillary approaches. A variety of options have been proposed to facilitate access to the infratemporal fossa via the anterior route. These include the use of 30°, 45° and 70° endoscopes,Reference Kassam, Gardner, Snyderman, Mintz and Carrau19, Reference Herzallah, Germani and Casiano20 the performance of a nasal septectomy to aid the angle of approach,Reference El-Sayed, Pletcher, Russell, McDermott and Parsa6, Reference Harvey, Sheehan, Debnath and Schlosser7 and the combination of the endonasal, transmaxillary, transpterygoid route with other approaches.
Sun and colleagues have reported using sublabial and buccolabial incisions to assist the removal of a large juvenile nasopharyngeal angiofibroma with extensive infiltration into the infratemporal fossa.Reference Sun, Li, Liu, Hu, Yu and Wang25 Anatomical studies have also assessed the use of a Caldwell–Luc approach with an anterior antrostomy,Reference Theodosopoulos, Guthikonda, Brescia, Keller and Zimmer4 and an endoscopic, sublabial, transmaxillary approach with the maxillary sinus accessed by an anterior (rather than medial) approach, as methods of reaching the infratemporal fossa.Reference Ong, Gore, Donnellan, Kertesz and Teo5
Use of the Caldwell–Luc approach, especially for access to the infratemporal fossa, can result in a number of complications, including damage to the anterosuperior alveolar nerve leading to dental denervation, facial swelling, infection, and haematoma.Reference El-Sayed, Pletcher, Russell, McDermott and Parsa6, Reference Defreitas and Lucente26, Reference Low27 In order to access the infratemporal fossa, Abuzayed and colleagues proposed a lateral sublabial approach, drawing on their anatomical study.Reference Abuzayed, Tanriover, Canbaz, Akar and Gazioglu28 However, Eloy and co-workers stated that this approach would have ‘minimal freedom’ and would be unsuitable for the resection of larger tumours within the infratemporal fossa.Reference Eloy, Murray, Friedel, Tessema and Liu3 They recommended using a combination of approaches (i.e. endoscopic, contralateral, nasal septectomy and endoscopic, ipsilateral, Caldwell–Luc approach) to improve visualisation and potential for manipulation in this poorly accessible area.Reference Theodosopoulos, Guthikonda, Brescia, Keller and Zimmer4
The present study adds considerable anatomical, radiological and surgical detail to the body of evidence in this field. We present a highly accurate dataset of measurements of key landmarks, which can assist surgical navigation to the foramen ovale. We examined a significantly larger number of cases, during data collection, than previous, similar studies. The anatomical studies undertaken by Theodosopoulos et al. Reference Theodosopoulos, Guthikonda, Brescia, Keller and Zimmer4 Herzhallah et al. Reference Herzallah, Germani and Casiano20 and Cai et al. Reference Cai, Zhang, Yang, Wang, Liu and Li29 only examined 4, 5 and 11 cadaveric heads, respectively, regarding the endoscopic approach to the infratemporal fossa. Eloy and colleagues' assessment of the anatomy of the infratemporal fossa examined CT data for only 15 patients.Reference Eloy, Reyes, Germani, Liu and Casiano30 In contrast, we examined CT data for 109 patients, including both left and right sides. In addition, almost all these previous authors stated that there were no statistically significant differences for the landmarks they assessed. The only results similar to those of the present study were published by Eloy et al. whose reported distances from the anterior nasal spine to the foramen ovale (i.e. mean ± SD of 80.3 ± 4.3 mm in males and 77.0 ± 6.5 mm in females) were comparable to our own findings (i.e. 80.2 ± 3.4 mm on the left and 80.4 ± 3.5 mm on the right in men; and 74.0 ± 3.1 mm on the left and 74.1 ± 2.9 mm on the right in women).Reference Eloy, Reyes, Germani, Liu and Casiano30 However, Eloy et al. did not find a statistically significant difference between the sexes regarding these measurements, whereas we found a clear sex difference for two of our four foramen ovale landmark measurements (i.e. distances from the posterior maxillary wall and from the anterior nasal spine).Reference Eloy, Reyes, Germani, Liu and Casiano30
This difference is unlikely to be explained by any sex difference in the shape of the foramen ovale itself. Ray et al. examined the foramen ovale morphology and found no statistically significant differences between men and women, regarding size or shape.Reference Ray, Gupta and Ghose31
The more probable reason for our statistically significantly greater values in men (regarding distances from the posterior maxillary wall, and from the anterior nasal spine, to the foramen ovale) is the fact that these measurements were taken in the horizontal, anteroposterior plane. The male skull is generally larger and more robust than the female skull.Reference Nafte, Saunders and Denning32, Reference White, Folkens and Folkens33 It could be postulated that the sex differences observed in our study were due to the male skulls simply being larger in the anteroposterior plane. Skull size could also vary with patients' height and/or race, with resultant alteration of relevant landmark measurements, although this was not analysed in the current study. In addition, it may be that the maxillary sinus and general skeletal structure of males are larger compared with those of females, accounting for the observed, significant differences in landmark measurements.Reference Ilayperuma34
Despite such sex-based variations, the relevance of the present study findings is supported by the fact that they derive from a patient cohort undergoing surgery of the infratemporal fossa and foramen ovale, and are drawn from a major hospital within a large UK city.
In addition to recording measurements for key foramen ovale landmarks, we also successfully performed an endonasal, transmaxillary, transpterygoid approach to the foramen ovale on a non-embalmed cadaver. This exercise adds to the relevance of our findings, as it was carried out in a laboratory setting and we were thus able to spend as long as necessary accurately identifying all structures. The regional anatomy was able to be clearly identified, and the cadaver was a fresh frozen (non-embalmed) specimen and therefore mimicked the real-life operative situation. With our large sample size, coupled with the success of our experimental endoscopic approach, we believe that our findings have direct clinical relevance.
• Traditional open surgical approaches to the infratemporal fossa are invasive, with significant morbidity
• A proposed endoscopic, endonasal approach has substantial advantages
• This study measured dimensions for key anatomical and radiological landmarks for such surgery
• Some dimensions were greater in males than females
• Study findings will aid an endonasal, transmaxillary, transpterygoid approach to the foramen ovale
• This approach is feasible without additional incisions
With growing interest in pre-operative planning and image guidance, and the necessity for accurate localisation of structures during surgical navigation, the measurements recorded in the present study have direct relevance for image-guided surgery. Rather than being prescriptive, they should serve as a guide to the relevant anatomical structures. We do not foresee these measurements replacing surgical experience or detailed surgical dissection of the infratemporal fossa and foramen ovale, but rather as providing a reference point. The presented distances between the posterior maxillary wall and the foramen ovale will be extremely useful to surgeons after leaving the sinus, and will give an indication of one's proximity to the mandibular nerve. The angle between the foramen ovale, the anterior nasal spine and the sphenoid rostrum gives an accurate indication of where, in the horizontal plane, to exit the maxillary sinus at its posterior aspect. On exiting the sinus, our data on the distance between the posterior portion of the lateral pterygoid plate and the foramen ovale give a clear indication of the distance to the mandibular nerve, in the final approach to that structure. The distance from the anterior nasal spine to the foramen ovale gives an overall idea of the endoscope length required to approach the deeper structures. Indeed, our accurate and comprehensive measurements could also be used in the future to assist the design of endoscopic surgical instrumentation for operations in this field.
Conclusion
This study aimed to create a comprehensive dataset, with statistical analysis, of the major anatomical, radiological and surgical landmarks which could be used to perform endoscopic, endonasal, transmaxillary, transpterygoid surgery to access the foramen ovale via the infratemporal fossa.
The distances from the posterior maxillary sinus wall and from the anterior nasal spine to the foramen ovale were statistically significantly greater in men than women. The distance from the base of the lateral pterygoid plate to the foramen ovale, and the angle between the foramen ovale, the anterior nasal spine and the sphenoid rostrum, were similar in men and women. There were no statistically significant differences between the left and right sides, for any measurement.
These measurements could serve as an anatomico-radiological guide to the endoscopic surgical approach to the foramen ovale and infratemporal fossa, obviating the need for nasal septectomy, anterior maxillectomy and additional incisions.
In order to identify whether an endonasal, transmaxillary, transpterygoid approach to the foramen ovale is feasible using these landmarks, we also performed this procedure on a fresh cadaver (mimicking the operative setting). This experiment indicated that an endoscopic, endonasal, transmaxillary, transpterygoid approach to the foramen ovale via the infratemporal fossa is easily viable, without additional incisions.
Image guidance is becoming increasingly popular in this, and related, fields. Whilst measurements are not a replacement for surgical training, anatomical knowledge or good intra-operative decision-making, they will prove extremely useful when navigating to the infratemporal fossa and foramen ovale, and could also assist the development of future endoscopic equipment for this territory.
