Introduction
The anterior ethmoidal artery is an important structure in functional endoscopic sinus surgery (FESS). Iatrogenic injury may result in retraction of the artery into the orbit, with consequent intra-orbital bleeding and possible blindness.
The anterior ethmoidal artery is a branch of the ophthalmic artery, which itself arises from the internal carotid artery. The anterior ethmoidal artery pierces the medial orbital wall at the frontoethmoidal suture line to form the anterior ethmoidal foramen on the medial wall of the orbit. It then enters the ethmoidal sinuses and follows a variable course to supply the mucosa and to send branches to the upper part of the nasal cavity. It may be closely apposed to the skull base, or traverse the ethmoids, either freely or within a bony canal.Reference Ducasse, Delattre, Segal, Desphieux and Flamant1 Such anatomical differences are vital to the endoscopic sinus surgeon wishing to minimise the risk of damage to the artery and the possible consequences.
Many centres routinely organise pre-operative coronal computed tomography (CT) scans of the sinuses in order to improve the safety of FESS. These are typically displayed in theatre during the operation, and provide a reference for key anatomical landmarks. Despite the wealth of published information about the anatomy of the anterior ethmoidal artery,Reference Ducasse, Delattre, Segal, Desphieux and Flamant1–Reference Badia, Lund, Wei and Ho10 no previous authors have reported their findings in a way that relates to this practice.
We present a reliable method of radiologically detecting the position of the anterior ethmoidal artery as it enters the ethmoid sinuses, which we believe will be useful to otolaryngologists undertaking FESS.
Aims
This study aimed to answer a number of questions concerning the identification of the anterior ethmoidal artery in coronal CT scans. Firstly, how reliably can the anterior ethmoidal foramen be identified on routine sinus CT scans? Secondly, what proportion of CT scans enable identification of the anterior ethmoidal artery within the ethmoids? Thirdly, what is the vertical height relationship between the anterior ethmoidal foramen and the skull base? Fourthly, what is the distance between an identifiable anterior landmark and the position of the anterior ethmoidal foramen?
Methods
Fifty consecutive CT scans of the paranasal sinuses, dating from 1 January 2003, were acquired retrospectively from the radiology department at Southmead Hospital, Bristol. The spiral CT scanner (Philips MX8000, Philips, Reigate, UK) used employed a 2.5 mm contiguous scan technique to generate coronal images of the sinuses. In order to minimise radiation exposure to the eyes, the most anterior images were 5 mm apart.
Two observers (SM and PR) independently assessed each scan, according to a pre-arranged protocol. Firstly, the image depicting the anterior lacrimal crest, anterior wall of maxillary sinus or lacrimal fossa was identified. Sequential coronal images were then viewed from anterior to posterior, and the presence and position of other anatomical features were noted. The anterior ethmoidal artery within the ethmoid sinuses was defined for the purposes of this study as a linear soft tissue structure demarcated by a clear bony outline, extending medially from and continuous with the anterior ethmoidal foramen.
Where anatomical information could be confidently identified, the two observers independently recorded the position of the image in the scan sequence (present on each image as a distance in millimetres from an arbitrary starting point). If an anatomical structure was present on more than one image on the film, the observer selected the image which best captured the anatomy. The anatomy of the anterior ethmoidal foramen and anterior ethmoidal artery was sought separately for the patient's left and right sides. For the anterior landmarks, the single frame which best represented the anatomy was chosen, taking both sides into account. When both observers were able to confidently identify the landmarks, the results are presented below. When both observers identified a landmark but chose a different image, they reviewed the scans together to reach a consensus. Inter-observer concordance is presented in the results.
Results
Of the 50 scans assessed, eight were found to be low definition coronal images reconstructed from another plane. The results presented below relate to the remaining 42 scans. Percentages are rounded to the nearest 0.5 per cent.
Landmark identification and inter-observer concordance
Results for inter-observer concordance are shown in Table I. Both observers were able to identify the lacrimal fossa, the maxillary sinus (most anterior image) and at least one anterior ethmoidal foramen in all 42 scans (100 per cent). The lacrimal crest was only identified on 29 scans (69 per cent) and the vertical attachment of the middle turbinate on 23 scans (57 per cent). The observers agreed more frequently on which image best depicted the anterior landmarks than the posterior landmarks.
Landmark identification and inter-observer concordance
* Percentage data relate to n = 42. †Percentage data relate to n = structure identified by both observers. AEF = anterior ethmoid foramen; AEA = anterior ethmoid artery
Identification of anterior ethmoidal foramen and artery
The anterior ethmoidal foramen was identified bilaterally on 40 of the 42 scans assessed. In 30 (71 per cent) cases, they were seen on the same coronal image. In eight cases (19 per cent), the anterior ethmoidal foramina were identified on adjacent images (2.5 mm apart), and in two cases (5 per cent) they were two images (5 mm) apart. Table II shows the number of scans in which both observers identified the anterior ethmoidal foramen and anterior ethmoidal artery.
Identification of anterior ethmoidal artery and foramen*
* n = 42. AEF = anterior ethmoid foramen; AEA = anterior ethmoid artery
Relationship between anterior ethmoidal foramen height and skull base
Table III shows the relationship between the level of the anterior ethmoidal foramen and the skull base. In the majority of cases (30 scans; 72 per cent), the anterior ethmoidal foramen was at the same level as the skull base. In the remainder, either opacified or pneumatised air cells were identified above the anterior ethmoidal foramen; examples are show in Figure 1.
Coronal computed tomography scans showing: (a) anterior ethmoidal arteries with opacified air cells above; (b) anterior ethmoidal arteries emerging immediately below the level of the skull base; and (c) anterior ethmoidal arteries with air cell above on right.
Vertical relation of anterior ethmoid foramen to skull base*
* n = 42.
Distance between anterior landmarks and anterior ethmoidal foramen
The distance between the anterior landmarks and the anterior ethmoidal foramen on each side is presented in Table IV.
Distances between landmarks
Data shown are means, with standard deviations in parentheses. *n = 29; †n = 42. AEF = anterior ethmoid foramen
For 36 of the 42 scans, the most anterior view of the lacrimal fossa and the maxillary sinus was present in the same image. For the remaining were scans, the two landmarks were seen on adjacent images. In three cases, the lacrimal fossa was anterior to the maxillary sinus, and in three the maxillary sinus was the more anterior structure. For these six cases, distances were measured from the more anterior of the two structures.
Discussion
A total of 10 studies were identified by a thorough Medline search of the English language literature relating specifically to the anatomy of the anterior ethmoidal artery. Four papers related to cadaveric dissection only.Reference Ducasse, Delattre, Segal, Desphieux and Flamant1–Reference Ohnishi and Yanagisawa4 One compared cadaveric dissection to cadaveric CT images.Reference Moon, Kim, Lee, Chung and Yoon5 Five were studies of CT images,Reference Basak, Karaman, Akdilli, Mutlu, Odabasi and Erpek6–Reference Badia, Lund, Wei and Ho10 while three specifically assessed coronal scans.Reference Chung, Dhong and Kim7, Reference Zacharek, Han, Allen, Weissman and Hwang9, Reference Badia, Lund, Wei and Ho10 These studies reported great variability in the anatomy of the ethmoid sinuses.
However, none of the above studies approached the anatomy from the perspective of the FESS surgeon, whose priority is identification of surgical danger areas.
Our results show high inter-observer concordance for the identification of anatomical landmarks, particularly the lacrimal fossa and maxillary sinus anteriorly and the anterior ethmoidal foramen posteriorly. Difficulty in identifying the lacrimal crest is likely to be due to it being an anterior, vertically-orientated landmark, and also to our CT scan protocol, which aimed to reduce orbital radiation exposure. The vertical anterior attachment of the middle turbinate was the least likely of the anterior landmarks to be reliably identified. This is probably due both to natural inter-patient anatomical variation and to disease within the ethmoid sinuses obscuring it from view.
When both observers were able to identify an anatomical structure, they selected the same image as best representing this structure on 76 to 92.5 per cent of occasions. The lower figure of 76 per cent refers to the identification of the anterior ethmoidal foramen, and is likely to represent the phenomenon of partial volume averaging, which causes an anatomical structure to appear on consecutive images despite being smaller in width than the distance between them.
From our results, the anterior ethmoidal foramen was reliably identified bilaterally on 95 per cent of scans included in the study, and on at least one side in the remaining 5 per cent. The anterior ethmoidal artery was less easily identified, and was seen in only 33 per cent of cases.
• The anatomy of the anterior ethmoidal artery and foramen is known to be variable
• No previous studies have approached the anatomy of the anterior ethmoidal artery and anterior ethmoidal foramen from the perspective of functional endoscopic sinus surgery (FESS), in which the priority is identification of surgical danger areas
• In this study, two observers independently attempted to identify the anterior ethmoidal artery and foramen on coronal computed tomography scans of the paranasal sinuses. There was good inter-observer concordance, and the anterior ethmoidal foramen was reliably identified bilaterally in 95% of cases
• In approximately one-third of cases, there was an air cell above the level of the anterior ethmoidal foramen. Identification of this forewarns the surgeon performing FESS
Using the anterior ethmoidal foramen to infer the site of anterior ethmoidal artery entry to the medial aspect of the ethmoid sinuses provides important information about the position of the artery relative to the level of the skull base. The anterior ethmoidal foramen was at the level of the skull base in 72 per cent of sides studied. In the remainder, either opacified cells or air-filled cells were observed above the level at which the artery enters the sinuses.
These results show a variable distance between the anterior landmarks and the anterior ethmoidal foramen. The authors feel that this can be accounted for partly by previously documented inter-patient variation, and partly by deviation from the true coronal plane in the CT images assessed. The mean distance from the lacrimal crest to the anterior ethmoidal foramen compares well with cadaveric data.Reference Ducasse, Delattre, Segal, Desphieux and Flamant1
Conclusion
Previous work on the anatomy of the anterior ethmoidal artery has not approached the subject from the perspective of the FESS surgeon studying pre-operative coronal CT images. This study shows that paying particular attention to identifying the anterior ethmoidal foramen on coronal CT images could provide useful information about the probable position of the artery within the ethmoid sinuses. In 28 per cent of hemi-ethmoids studied, the anterior ethmoidal artery was below the level of the skull base and therefore at greater risk of damage during surgery. The method described identifies this anomaly prior to the start of the operation.
