Keywords
anterior ethmoidal artery - CT - volume rendering
Introduction
The anterior ethmoidal artery (AEA), a branch of the ophthalmic artery, crosses three
cavities along its course: it arises in the orbit, reaches the ethmoidal labyrinth
passing through the bony anterior ethmoidal canal (AEC) together with its homonymous
vein and nerve, and finally enters the olfactory fossa, through the lateral lamella
of the cribriform plate, along the so-called anterior ethmoidal sulcus,[1] where it becomes the anterior falx artery[2].
Several studies demonstrated that the AEC is not always continuous in its bony structure,
but it may show a partial or complete bone dehiscence in its most caudal portion.[1]
[3]
[4] Significant intrasubject side-to-side variability may exist so that the bony canal
can be complete on one side and partially or completely open on the other one.
Moreover, useful information about the distance of the AEA from the columella,[5] the middle turbinate axilla and the nasal valve[1]
[6] has been provided.
The AEA is one of the most important landmarks in endoscopic surgery, representing
a fundamental orientation point for ethmoidal fovea and anterior cranial fossa.[7]
[8]
In the surgical approach to the frontal recess, the AEC marks its posterior border.[9]
[10] Likewise, in the external approaches, AEA identification in the fronto-ethmoidal
suture marks the ventral limit of the anterior cranial fossa.[11]
In addition to the marking function, the AEA, together with the arteries coming out
from the sphenopalatine foramen, plays a key role for vascular ligation in nasal surgery
to control epistaxis during endoscopic sinus surgery (ESS).[12]
[13]
Considering the huge variety of its anatomical localization, course, and length, the
preoperative assessment of AEA is of clinical relevance. Indeed, severe iatrogenic
injuries may lead to liquor fistulas, intracranial bleeding and blindness, if not
decompressed within an hour.[14] As shown by Cassano et al,[12] if a good hemostatic control is performed under endoscopic view during the nasal
surgery, the nasal precautionary intraoperative packing is not justifiable, considering
the low percentage of postoperative epistaxis. Furthermore, ligation of the sphenopalatine
complex and of the AEA is a good way to gain control of epistaxis.[15]
The application of preoperative imaging techniques allows to obtain information about
the development of the paranasal sinuses, variants of pneumatization, as well as anatomical
variants of the vessels in relation to the adjacent rhino-ethmoidal structures, thus
allowing accurate surgical planning and increasing safety in rhino-sinusal endoscopic
surgical procedures.[16]
[17]
[18]
During the last decades, the development of many postprocessing techniques and the
modulation of the image presets have allowed for subject- and condition-based interpretations
of magnetic resonance imaging (MRI),[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41] ultrasonography[42] and computed tomography (CT) data,[43]
[44]
[45] thus providing the real 3D morphology of several structures and better highlighting
anatomical details. In this regard, the direct volume rendering (DVR) is a direct
technique to visualize primitive volumes without any intermediate conversion of the
volume data to surface presentation.[46]
However, despite its clinical usefulness, to the best of our knowledge, only a few
3D-reconstruction studies have demonstrated the course of the AEA and its relationship
with the skull base and adjacent structures through non-invasive imaging in living
human subjects.[47]
[48]
Based on these simple premises, the present work is aimed at i) visualizing topographic
anatomical details, such as the AEA course and its relations with the skull base;
ii) understanding whether preoperative detailed information on the course of the AEA
in its AEC or its possible dehiscence can be achieved non-invasively and on living
human subjects by direct volume rendering (DVR) on multidetector computed tomography
(MDCT) with contrast medium of supra aortic trunks (SAT).
Materials and Methods
Participants
We retrospectively evaluated 18 patients, 10 of whom were male and 8 female, aged
between 28 and 82 years old (63.4 ± 17.9), who underwent MDCT of the SAT because of
suspected cerebral aneurism between the 2015 and 2016.
Data Acquisition
In all patients, MDCT of the cranium was performed by a 64-banks multislice CT scanner
Philips MX-8000 EXP v25 model (Philips Medical Systems, New Hartford, CT, USA) with
axial volumetric acquisition and with transfemoral infusion of contrast medium (iopromide)
at a dose of 120 mg/ml.
We evaluated the lengths and distances between the AEA at its entry point in the nasal
cavity and the skull base, the middle portion of the frontal beak and the head of
middle turbinate on all patients ([Fig. 1]).
Fig. 1 Distance between the anterior ethmoidal artery and topographical reference points. Paramedian sagittal section. Black arrows show the entry point of the anterior ethmoidal
artery in the nasal cavity; its distance from the frontal beak (yellow line) and from
the anterior limit of the middle turbinate (black line) are also shown.
Direct Volume Rendering
A 3D reconstruction was performed by using a freeware graphic software package Osirix
3.3.1, 64-bit (Pixmeo SARL, Geneva, Switzerland). The reconstruction was performed
with fields of view of 21 × 21cm, with the result that the voxel size of the in-plane
became 0.41 × 0.41 mm with a 512 × 512 matrix. Once the CT scan was acquired, the
data were sent to the workstation Apple Mac Pro 2.7 GHz 12-Core Intel Xeon E5 equipped
with 64 GB 1866 MHz DDR3 RAM and AMD FirePro D500 3072 graphics, where the images
were viewed as consecutive axial sections and analyzed by using the Osirix 3.3.1 software
(Pixmeo SARL).
A voxel-intensity histogram was created, and then a transfer function, in the form
of a window level, mapped each voxel in the histogram for color and opacity. The voxel
positioned under the ramp of function was displayed with a color scale to reflect
the volume averaging of different tissue types.[49]
[50] The brightness and opacity of the voxels have been regulated by increasing or decreasing
them with up-regulation of the transfer function height.
Decreasing and increasing the window level permitted to display higher or lower attenuation
structures so that the final image comprises the pixels of interest. Indeed, we were
able to modulate the final image working on the window's width and level, to create
the 3D image; the final image was adjusted at the workstation to choose, selectively,
the structure to highlight (i.e., vessels, muscle, or bones) and to increase the conspicuity
of the selected tissue, maintaining, at the same time, the vascular 3D relationships.
Through a process of digitization and interpolation, it was possible to enlarge and/or
shrink, translate, and rotate the 3D image.
The Mann-Whitney U test was used to assess differences in measurements between the
genders and the sides. A 95% of confidence level was considered, and the statistical
significance was set at p < 0.05.
Results
The topographical location of 36 AEAs was assessed in vivo in a total of 18 patients.
We evaluated the lengths and distances of the AEA from the skull base, from the middle
portion of the frontal beak and from the head of the middle turbinate ([Fig. 1]). Statistical analysis reported no significant differences between the left and
right sides ([Table 1]) nor gender-related variations ([Table 2]).
Table 1
Quantitative data on the length of the anterior ethmoidal artery and its distance
from important anatomical landmarks (comparison between the left and right sides).
No significant differences emerged. Mean values are expressed in millimeters (mm)
|
Left
mean ± SD
|
Right
mean ± SD
|
p
Mann-Whitney U test
|
|
Anterior ethmoidal artery length
|
4.75 ± 2.26
|
5.13 ± 2.29
|
0.73
|
|
Distance from the head of the middle concha
|
24.35 ± 4.80
|
25.55 ± 3.95
|
0.60
|
|
Distance from the skull base
|
2.64 ± 2.54
|
2.51 ± 2.54
|
0.76
|
|
Distance from the frontal beak
|
13.06 ± 2.33
|
12.94 ± 2.73
|
0.95
|
Abbreviation: SD, standard deviation.
Table 2
Quantitative data on the length of the anterior ethmoidal artery and its distance
from important anatomical landmarks (comparison between male and female subjects).
No significant differences were found. Mean values are expressed in millimeters (mm)
|
Female
mean ± SD
|
Male
mean ± SD
|
p
Mann-Whitney U test
|
|
Anterior ethmoidal artery length
|
4.30 ± 2.42
|
5.75 ± 1.75
|
0.10
|
|
Distance from the head of the middle concha
|
24 ± 2.99
|
26.15 ± 5.54
|
0.45
|
|
Distance from the skull base
|
2.35 ± 2.60
|
2.86 ± 2.44
|
0.62
|
|
Distance from the frontal beak
|
13.54 ± 2.46
|
12.33 ± 2.45
|
0.37
|
Abbreviation: SD, standard deviation.
Relations between the AEA and its bony canal were assessed: 10 AEAs (27.8%) were found
to be dehiscent, 20 (55.5%) were intracanal, and, in 6 cases (16.7%), the AEC was
incomplete and open antero-inferiorly and the arteries were strictly adherent to the
skull base.
Applying the algorithm of DVR and extracting the isosurface on a transverse plane,
we obtained a three-dimensional image ([Fig. 2]) in which it was possible to distinguish the bone, the vascular, muscular and connective
structures, as well as cavity spaces.
Fig. 2 Axial plane of the skull base as obtained by means of direct volume rendering. The section is conducted at the level of the frontal sinus. The large wing of the
sphenoid, the ethmoid, the crista galli, and a section of the frontal bone can be
observed. The arrows show different anatomical structures as they are clearly represented
and distinguished by means of the volume rendering: vascular structures: arteries
of the circle of Willis (white arrow); muscular structures: ocular recti muscles,
lateral rectus (green arrow), medial rectus (gray arrow); optic nerves (cyan arrow),
eyeballs (yellow arrow), brain parenchyma (blue arrow) and skull (magenta).
Zooming on the transverse plane at the ethmoidal level, it was possible to obtain
a better visibility of the AEA and AEC.
The two AECs, from the medial wall of the orbit (lamina papiracea), reached the lateral
portion of the lamina cribrosa; they moved to the neurovascular peduncle, including
the artery, the vein, and the anterior ethmoidal nerve. In [Fig. 3], the CT and 3D DVR images are compared.
Fig. 3 Comparison between the same image as obtained by means of 3D direct volume rendering (A) and computed tomography (B). In the axial plane, the anterior ethmoidal artery (white arrows), the frontal sinuses
(green arrows), and the supraorbital recesses (yellow arrows) are visible.
By navigating through the skull base, modulating the parameters of intensity, opacity
and transparency, mitigating the bone and amplifying the vascular signal, it was possible
to better highlight the vascular structures, as depicted in [Fig. 4]. The magnification shows the origin of the AEA from the ophthalmic artery to its
end at the level of the lamina cribrosa, where it becomes the anterior meningeal artery.
([Fig. 5]).
Fig. 4 Axial section of the skull after modulation of intensity and transparency parameters
and amplification of the vascular signal to better visualize the vascular structures. The image shows: the internal carotid arteries (blue arrows), the ophthalmic arteries
(green arrow), left anterior ethmoidal artery (AEA) (white arrow), right AEA (yellow
arrow). As clearly demonstrated by the image, in this patient, the left AEA was extracanal,
while the right AEA is intracanal, within the AEC.
Fig. 5 A more detailed axial section allows to show the complete course of the left anterior
ethmoidal artery. The anterior ethmoidal artery (white arrow) originates from the ophthalmic artery
(cyan arrow) in an advanced point of its path, between the exit from the carotid siphon
(green arrow) and the terminal branches in the anterior wall of the frontal bone (white
asterisk); then, it goes along its intraorbital tract (black arrow), passes the lamina
papyracea, and then becomes intranasal (white arrow). At the level of the lamina cribrosa
(green asterisk), it turns back and becomes the anterior falx artery (yellow arrow).
Discussion
The AEA originates from the ophthalmic artery within the orbital cavity, moves to
the nasal cavity through the anterior ethmoidal foramen, and crosses antero-medially
the ethmoidal sinus to reach the cribriform plate.
The identification of the AEA during an ethmoidectomy is of major importance to better
detect the fronto-ethmoidal recess, which is always separated from the first ethmoidal
foveola. The AEA is closely related to the ethmoidal fovea, and where the ethmoidal
fovea inserts at the skull base, the lateral lamella of the cribriform plate (the
thinnest bone in the skull base) can be found.[51] This is the point of highest surgical risk.[52] Several methods have been proposed to identify the AEC: Kirchner et al, in 1961,[53] used the maxillo-lacrimal suture, while Lang[54] suggested the cribriform plate as a landmark. However, these attempts to search
for points of landmarks are difficult to accomplish during ESS (i.e., Araujo Filho
studied the distance between the AEA and the anterior nasal spine).[1]
In 1988, Kainz and Stammberger described the AEA with an upturned half-moon shape
resting on the skull base and with an average width of ∼ 5 mm.[55] Recent CT and dissection studies assessed the localization of the AEA between the
second and third lamella in 87.1% of the cases, while in the remaining 12.9% it was
inside the 2nd or the 3rd lamella; in no cases the AEA emerged anterior to the bulla ethmoidalis.[3] Moreover, in some cases, the AEC may be incomplete or even absent, thus exposing
the artery to direct trauma.[3]
[56]
Several variations of the AEA have been previously described. Simmen et al evaluated
34 cadaver heads, showing that the AEAs were located in the retrobullar recess or
roof of the bulla ethmoidalis (14.7%) and in the recess between the 2nd and 3rd lamella (85.3%).[4] Yang et al studied 15 cadavers and found that 85% of the AEAs were posterior to
the bulla ethmoidalis, 10.7% were in the roof of the frontal recess, and 3.6% were
in the roof of the posterior ethmoid sinus.[57]
During the last decade, consistent effort has been made to allow for a reliable pre-
and/or intraoperatory identification of the AEA, either by using anatomy and topographic
landmarks[1]
[4]
[14]
[58]
[59]
[60]
[61] or imaging techniques. To date, three-dimensional spin digital subtraction angiography
(DSA) is the gold standard for preoperative assessment of anatomy and relations of
the AEA and the AEC.[48] Many CT scans, both on cadavers and on living patients, have been performed to identify
the AEA and assess its course.[4]
[59]
[60]
[62]
[63]
[64]
[65]
[66]
[67]
[Table 3] summarizes the most important findings of the last decade.
Table 3
Summary of the studies of the last decade on the assessment of the anterior ethmoidal
artery
|
Authors
|
Year
|
Patients
|
Controls
|
Techniques
|
Results
|
|
Yenigun et al
|
2016
|
184 CT scans, excluding patients with nasal polyposis, sinus anomalies, previous sinus
surgery.
|
|
CT scans
|
Positive correlation between the AEC, the superior–inferior depth of the lateral lamella
of the cribriform plate LLCP, and the anterior–posterior length.
|
|
Wong et al
|
2014
|
1 Waardenburg syndrome and chronic rhinosinusitis
|
–
|
CT scan
|
The AEA was not identified on preoperative CT. Intraoperatively, the AEA was found
to be anterior to the bulla ethmoidalis, within the frontal recess, and was inadvertently
injured.
|
|
Ding et al
|
2012
|
32 (30 intracranial aneurisms, 2 meningiomas)
|
–
|
3D spin DSA; 320-slice CTA
|
The entire course of the AEA was seen in 100% (32) of the cases with 3D spine DSA
and in 29 of 32 cases (90.1%) with 320-slice CTA, with no significant difference.
|
|
Monjas-Cánovas et al
|
2011
|
–
|
20 cadavers
|
CT scan; MIP
|
AEA identified in 95% (38/40) of the scans. Mean length of the anterior ethmoidal
canal: 8.43 ± 0.74 mm. The angle performed into the skull base was 37.3 ± 5.48 degrees.
Topographical location: between the second and third lamella in 90% of the cases (36/40).
|
|
Joshi et al
|
2010
|
–
|
50
|
CT scan
|
AEA identified in 97% of the scans. Mean distance between artery and skull base: 1.5 mm
in absence of the supraorbital cell; 4.86 mm in the presence of the supraorbital cell.
|
|
Yang et al
|
2009
|
–
|
15 randomly selected adult
cadavers
|
Dissection
|
Average diameter of the AEA: 0.80 ± 0.24 mm. Topographical location: between the second
and third lamella in 87.5% of the cases; over the roof of the frontal recess cells
in 10.7%; over the roof of the posterior ethmoid sinus in 3.6%. The angle formed by
the long axis of the artery and the lamina papyracea was 60.5 ± 16.4 degrees.
|
|
Souza et al
|
2009
|
198 CT scans of patients excluding: patients aged below 12 years, a history of surgery
or trauma in the paranasal sinuses or the skull base, congenital anomalies of the
face, paranasal sinus malignancies, osteofibrous lesions, sinus diseases.
|
–
|
CT scan
|
AEC identified in 41% of the scans. Anterior ethmoidal sulcus identified in 98% of
the scans; anterior ethmoidal foramen identified in 100% of scans.
|
|
Han et al
|
2008
|
–
|
24 cadaver heads
|
Dissection
|
Mean distance from the axilla: 17.5 mm. Mean distance from the posterior ethmoid artery:
14.9 mm Topographical location: anterior to 31%, at 36%, or immediately posterior
to 33% the superior attachment of the basal lamella.
|
|
McDonald et al
|
2008
|
|
50
|
CT scan
|
Anterior ethmoidal foramen identified in 95% of the cases bilaterally and in the remaining
5% unilaterally. AEA identified in 33% of the scans.
|
|
Riehm et al
|
2008
|
–
|
9 cadavers
|
Dissection, CT scans
|
Correlation between CT and surgical findings allowed identification of 2 criteria
predictive of AEA origin: 1 presence of an ethmoid bulla above the AE C2 AEC located
below the ethmoid roof, anterior to the bulla.
|
|
Pandolfo et al
|
2007
|
20 18 subarachnoid hemorrhage, 2 meningiomas
|
78 (25 chronic sinusitis, 40 massive poliposis, 13 cases of single polyp
|
Dual volume angiography and MDCT patients; MDCT controls
|
On patients, the entire course of the AEA was identified by means of DVA angiography:
AEA behind the frontal recess (18/20); AEA near to the posterior wall of frontal recess
2/20. On controls, MDCT only (110/156) AEAs were identified only indirectly; 46 out
of 156 were shown in their entirety.
|
|
Lannoy-Penisson et al
|
2007
|
|
9 heads
|
Dissection, CT scan
|
The arteries were identifiable on CT scans. Topographical location: arteries were
included in the roof of the ethmoid in 8 cases; prominent under the roof in 3 cases;
distant from the roof in 7 cases, in association with pneumatization of the floor
of the orbit.
|
|
Erdogmus et al
|
2006
|
–
|
19 adult cadavers
|
Dissection
|
The AEA was observed in all specimens except for one case. Average diameter: 92 +/−
0.2 mm on the right and 0.88 +/− 0.15 mm on the left. Four variations of its course
were differentiated and described.
|
|
Floreani et al
|
2006
|
–
|
22 cadaver heads
|
CT scan
|
AEA unable to be clipped in 66% of the cases; clipped effectively, all in a mesentery
in 20% of cases; not effectively clipped in 14%.
|
|
Araujo Filho et al
|
2006
|
–
|
25 cadavers with no past nasosinusal injury/surgery and altered anatomy
|
Dissection
|
Average intranasal length of the AEA: 5.2 mm. AEC dehiscence in 66.7%. Average distance
between the artery middle point to the anterior nasal spine: 61.72 mm. Average distance
from the nasal axilla: 64.04 mm. Average distance from the anterior axilla to the
middle turbinate:21.14 mm.
|
Abbreviations: AE, ; AEA, anterior ethmoidal artery; AEC; anterior ethmoidal canal;
CTA, computed-tomography angiography; DSA, digital subtraction angiography; DVA, dynamic
volumetric angiography; LLCP, lateral lamella of the cribriform plate; MDCT, multidetector
computed tomography; MIP, maximum intensity projection.
However, the AEA may be difficult to identify, especially in the presence of pathologies,
on CT scan because of the morphological and structural complexity of this anatomical
region that may lead to several issues. In patients with chronic sinusitis or polyposis,
Pandolfo et al[47] were able to identify the AEA in its entirety only in 23 of 78 patients (46 AEA);
McDonald and colleagues successfully identified the AEA only in 33% of the CT scans[62]; Souza et al[63] were able to find the AEC only in 41% of the exams; finally, an anatomical variation
of the AEA course, described by Wong et al[67] in a case study of a patient with Waanderburg Syndrome and chronic rhinosinusitis,
was not detected in preoperative CT and then injured during the surgery.
Taking into account the aforementioned limitations, we applied the 3D-DVR technique
to evaluate the quantitative and qualitative features of the AEA and its relations
with the skull base and neighboring structures. Taken together, our results show that
3D-DVR is a useful technique for a detailed characterization of the localization and
anatomical relationships of the AEA.
Conclusion
In conclusion, we showed the 3D reconstruction of the AEA in its entirety and topographic
complexity. However, further studies are needed to assess the usefulness of in vivo
CT-based 3D volume rendering as a tool for preoperative assessment. Nevertheless,
we would like to stress the great potential of this technique for visualization purposes,
ranging from clinical evaluations to anatomical studies or even educational images.
What has been pursued and achieved is summarized in the perception of anatomical,
morphological and topographic detail in a new tridimensional framework, which gives
added value to any physician, no matter his/her area of specialization, from surgery
to radiology, from anatomy to clinic.
