Keywords
middle fossa approach - Kawase's rhomboid - cochlea - greater superficial petrosal
nerve
Introduction
In the extended middle fossa approach, a portion of the anteromedial petrous bone
known as Kawase's rhomboid can be resected to provide exposure of the posterior fossa
through a middle fossa corridor. Kawase's rhomboid is bordered by the petrous ridge
medially, the greater superficial petrosal nerve (GSPN) and internal carotid artery
(ICA) laterally, the arcuate eminence posteriorly, and the V3 division of the trigeminal
nerve anteriorly.[1] This area, in turn, can be bisected by the internal acoustic canal (IAC) into pre-
and post-meatal triangles. During bony resection of Kawase's rhomboid, the cochlea—in
addition to various other structures—is placed at significant risk. Damage to the
cochlea results in ipsilateral hearing loss and vertigo, and it may be associated
with facial weakness if the nearby facial nerve is injured.
Minimization of risk during bony resection of the rhomboid construct is contingent
on extensive knowledge of the microsurgical anatomy of the cochlea and its association
with several external landmarks. This study was conducted to evaluate the relationship
of the cochlea to Kawase's rhomboid and provide useful information for surgeons planning
to perform an anterior petrosectomy to augment posterior fossa exposure.
Methods
Cadaveric Dissection
A total of 11 cadaveric specimens were dissected, including wet bench dissection of
6 cadaveric heads and 5 dry temporal bones. Each of the 6 cadaveric heads was positioned
in pins for a standard middle fossa craniotomy. A curvilinear “question mark” incision
was fashioned about 1 cm in front of the tragus, initially curving posteriorly then
anteriorly. Burr holes were placed superior to the zygomatic arch and at the superior
margin of the craniotomy. A craniotome was used to turn a 5-cm bone flap located approximately
one third behind and two thirds in front of the external auditory canal. A Leksell
rongeur was then used to extend the craniotomy flush with the middle fossa floor.
Subtemporal dissection was accomplished in a posterior to anterior dissection until
the medial extent of the petrous ridge could be visualized. Following completion of
the extradural, subtemporal dissection, the porus trigeminus could be visualized anteromedially
and the arcuate eminence posteriorly.
Following subtemporal dissection, the petrous apex was drilled in the manner previously
described by Miller et al[2] until the dura of the seventh/eighth nerve complex could be visualized ([Fig. 1]). Specifically, the arcuate eminence was first “blue lined” (e.g., drilled until
the membrane could be visualized) until the membranous superior semicircular canal
(SSCC) was visible. The GSPN was identified and tracked posteriorly to its exit from
the facial hiatus. The angle between the superior semicircular canal and GSPN was
bisected to reveal the approximate location of the IAC.[3] Drilling over the expected location of the IAC was continued until the dura of the
seventh/eighth nerve complex was visualized. Drilling was then extended anterolaterally
until the basal turn of the cochlea had been “blue lined”; blue lining of the cochlea
was performed to improve measurement accuracy—as differentiation of the cortical bone
of the cochlea from surrounding cortical bone of the petrous temporal bone is not
always straightforward (blue lining the cochlea would not otherwise have been performed
in a standard anterior petrosectomy). A diamond drill was then used to expose the
carotid artery under the GSPN. Once the seventh/eighth nerve complex, petrous ICA,
cochlea, and SSCC had been exposed, the remainder of the anterior petrosal bone was
then resected.
Fig. 1 Anatomy encountered in extradural extended middle fossa approach (photo obtained
following dissection of cadaveric head as described in text). Left of picture is anterior,
bottom of picture is medial. GSPN, greater superficial petrosal nerve; ICA, internal
carotid artery; SSC, superior semicircular canal; V3, third division of trigeminal
nerve.
Following completion of the steps described above, a postoperative computed tomography
(CT) scan was obtained to verify skeletonization. After review and approval of the
CT images, three morphologic measurements were made on the cadaveric specimen. Quantification
of the following distances, as depicted in [Fig. 2], was then made with an industrial grade Starrett No. 274 3-inch caliper: (1) The
extrapolated junction of the GSPN and facial nerve to the membranous anterior border
of cochlea (measured along the GSPN). The anteriormost point corresponded to the anterior
limit of the basal turn of the cochlea. (2) The extrapolated junction of the GSPN
and facial nerve to the medial membranous border of cochlea (measured along cranial
nerve [CN] 7). The medialmost point corresponded to the medial margin of the basal
turn of the cochlea. (3) The average maximum distance from the extrapolated junction
of the GSPN and facial nerve to the basal turn of the membranous cochlea.
Fig. 2 Illustration of anatomic measurements that were obtained. A drilled, dry temporal
bone following skeletonization (A). The green arrow depicts the distance from the
extrapolated junction of the greater superficial petrosal nerve (GSPN) and facial
nerve (base of the arrow) to anterior border of the basal turn of the cochlea (tip
of the arrow). This distance is measured along the GSPN. The white arrow depicts the
distance from the extrapolated junction of the GSPN and facial nerve (base of arrow)
to the medial border of the basal turn of the cochlea (tip of the arrow). This distance
is measured along the meatal trajectory of cranial nerve 7. The red arrow depicts
the maximum distance from the extrapolated junction of the GSPN and facial nerve to
the basal turn of the cochlea (tip of arrow). Extrapolated junction of the GSPN and
facial nerve represent the approximate location of the geniculate ganglion. Depiction
of the aforementioned measurements in an illustration, courtesy of Jackler et al (B).[12]
The procedure used in analysis of the dry temporal bones (visualized in [Fig. 2A]) was similar. The GSPN was immediately visible in selected dry temporal bone specimens.
Bone underneath the GSPN was drilled until the petrous ICA was visible. The IAC in
these specimens could be directly visualized and was drilled until the seventh and
eighth nerve complex had been completely unroofed. The drilling was extended anterolaterally
until the cochlea could be identified. The above measurements were repeated following
completion of the anterior petrosectomy.
It is important to mention that, although portions of the technique described on cadaveric
specimens in this study are routinely used to gain access to the petrous apex, other
maneuvers described (e.g., “blue-lining” the cochlea) do not represent standard surgical
procedure and would introduce significant risk to eloquent structures if utilized
in an intraoperative setting. Thus, the methods of cadaveric dissection described
in this study should not be interpreted as standard surgical technique.
Radiologic Assessment
A cohort of 25 patients who had received thin-cut CT imaging though the temporal bones
was obtained from the neuro-otology database. These patients underwent imaging for
a variety of reasons, but all were deemed to have bilateral morphologically normal
inner ears. The above measurements were repeated using standard PACS software ([Fig. 3]). It is important to note that the position of the GSPN was estimated, in these
scans, using the bony groove of the GSPN that is visible deep to this structure.
Fig. 3 Computed tomography image with greater superficial petrosal nerve (GSPN) canal, cochlea,
and meatus of 7/8 complex visible (A). The green arrow depicts the distance from the
extrapolated junction of the GSPN and facial nerve (base of the arrow) to the anterior
border of the basal turn of the cochlea (tip of the arrow). This distance is measured
along the GSPN. The white arrow depicts the distance from the extrapolated junction
of the GSPN and the facial nerve (base of arrow) to the medial border of the cochlea
(tip of the arrow). This distance is measured along the meatal trajectory of cranial
nerve 7. The red arrow depicts the maximum distance from the extrapolated junction
of the GSPN and facial nerve to the anterior edge of the basal turn of the cochlea
(tip of arrow). Depiction of the aforementioned measurements in an illustration, courtesy
of Jackler et al (B).[12]
Results
[Table 1] details the anatomic measurements defining the relationship of the cochlea and adjacent
structures. Special attention is paid to the maximum distances encountered from the
extrapolated junction of the GSPN and facial nerve to the membranous basal turn of
the cochlea, as these figures represent the distance, outside which drilling is associated
with a low degree of risk to the membranous cochlea (highlighted by red box). Review
of this data indicates that drilling outside of a radius of 11 mm from the extrapolated
junction of the GSPN and facial nerve is associated with a low degree of risk to the
membranous cochlear apparatus. [Table 2] additionally details measured dimensions of cochlea obtained following review of
thin-cut CT imaging in 25 patients.
Table 1
Measurements of extrapolated junction of greater superficial petrosal nerve and facial
nerve to cochlear borders
|
Measurement
|
Cadaveric specimen
|
Mean
|
Median
|
Minimum
|
Maximum
|
SD
|
|
EJ-medial CO (white arrow)
|
Cadaveric head
|
8.4
|
8.3
|
6.9
|
9.0
|
.766
|
|
Temporal bone
|
8.0
|
8.0
|
6.9
|
8.9
|
.760
|
|
EJ-anterior CO (green arrow)
|
Cadaveric head
|
7.8
|
8.0
|
6.4
|
8.1
|
.880
|
|
Temporal bone
|
7.5
|
7.5
|
6.6
|
8.7
|
.862
|
|
EJ-basal turn CO (red arrow)
|
Cadaveric head
|
9.4
|
9.4
|
8.2
|
10.3
|
.776
|
|
Temporal bone
|
9.3
|
9.3
|
8.3
|
10.0
|
.709
|
Abbreviations: CO, cochlea; EJ, extrapolated junction of greater superficial petrosal
nerve and facial nerve (approximate location of geniculate ganglion); SD, standard
deviation.
Table 2
Cochlear measurements from computed tomography (CT) imaging
|
Measurement
|
Mean
|
Median
|
Minimum
|
Maximum
|
SD
|
|
EJ-medial CO (white arrow)
|
7.9
|
7.8
|
6.3
|
9.3
|
.064
|
|
EJ-anterior CO (green arrow)
|
5.7
|
5.6
|
4.7
|
7.3
|
.062
|
|
EJ-basal turn CO (red arrow)
|
8.7
|
8.7
|
7.4
|
10.5
|
.061
|
Abbreviations: CO, cochlea; EJ, extrapolated junction of greater superficial petrosal
nerve and facial nerve (approximate location of geniculate ganglion); SD, standard
deviation. Note: Measured dimensions of cochlea obtained following review of thin-cut
CT imaging in 25 patients. Values include measurement of 25 right and 25 left temporal
bones.
Discussion
The complexity of the extended middle fossa approach relates to the close anatomic
proximity of the internal acoustic canal, cochlea, semicircular canal, GSPN, and ICA.
In this procedure, the anteromedial petrous bone—also known as Kawase's rhomboid—is
removed to improve visualization of the posterior fossa. During removal of the Kawase's
rhomboid, skull base surgeons must be aware of the location of the cochlea at all
times. Previous authors have attempted to delineate the location of the cochlea in
relation to surgical landmarks. Diaz Day et al described a pre-meatal triangle bordered
by the GSPN laterally, the IAC posteriorly, and the medial petrous ridge and porus
trigeminus medially. In addition to its referencing its borders, this triangle was
described by three points: the internal carotid artery genu, the geniculate ganglion,
and the medial lip of the IAC. The authors reported that the cochlea could reliably
be found in the lateral half of this triangle—and stressed the close proximity of
the cochlea to the genu of the petrous carotid.[1] In a manuscript with similar conclusions, Isolan et al reported that the cochlea
could be exposed by drilling the lateral aspect of the posteromedial triangle—whose
borders are defined by the porus trigeminus, cochlea, and mandibular division of the
trigeminal nerve.[4] Sameshima et al noted that the cochlea resides in a plane deeper than the IAC and
GSPN, lateral to the genu of the ICA.[5] Mastronardi et al described two theoretical “fans” with a 90-degree relationship
to one other—useful in identifying the vascular, nervous, and osseous structures,
including the cochlea, of Kawase's rhomboid.[6] In a study describing exposure of the horizontal petrous carotid artery in preparation
for intrapetrous carotid bypass, Dew et al reported the average distance between the
cochlea and carotid to be 4.3 mm.[7] In a separate manuscript, Rhoton reported this distance to be an average of 2.1
mm.[8]
Two previous studies have reported data particularly relevant to this discussion.
Jung et al dissected 32 cadaveric specimens and found the average distance between
the cochlea and geniculate ganglion to be 3.0 ± 0.8 mm.[9] Though useful, this information is limited in the sense that it does not help the
skull base surgeon define the anteromedial border of the cochlea—the border placed
at greatest risk during anterior petrosectomy. A second study by Sennaroglu et al
provided an extensive review that included 39 anatomic measurements of the petrous
temporal bone obtained from 10 cadaveric specimens.[10] One of these measurements was from the IAC apex to the anteromedial cochlear boundary—roughly
corresponding to the measurement (depicted by the red arrow above) from the extrapolated
junction of the GSPN and facial nerve to the basal turn of the cochlea. In this study,
the mean value of this distance was 10.7 mm and the maximum distance encountered was
12 mm. These values correlate well with the values obtained from cadaveric specimens
in our study, which averaged a composite mean of 9.4 mm and maximum of 10.2 mm (of
note, it is possible the average 1.3 mm of difference relates to our practice of “blue-lining”
the cochlea in this study and measuring to the border of the cochlear membrane). The
other two measurements reported in this manuscript were not assessed in the Sennaroglu
study or, to the authors' knowledge, in any other studies.
Although the aforementioned studies are useful in regards to approximate localization
of the cochlea, they fail to provide the skull base surgeon with a precise method
of localization using microsurgical landmarks. The current study describes a method
used to predict cochlear location following identification of the extrapolated junction
of the GSPN and the facial nerve ([Fig. 4A]). Once this junction has been identified, the anterior and medial boundaries of
the cochlea can be estimated using the measurements obtained above ([Fig. 4B]). Using morphologic and radiologic data obtained in this study, the membranous cochlea
was not encountered more than 11 mm from this extrapolated junction. In the event
the skull base surgeon wishes to avoid transgression of the cortical bone of the cochlea,
as is almost always the case, this figure should be increased to 12.5 mm, based on
previous data estimating the thickness of the cochlear bone at 1.1 to 1.7 mm.[11]
Fig. 4 Method of cochlear localization. (A) Following identification of the greater superficial
petrosal nerve (GSPN) and 7/8 complex (arrows), the extrapolated junction of the two
structures is estimated (A). The membranous cochlea was encountered not more than
11 mm from this approximate junction (junction depicted via intersection of two line
segments; B).
Data from this study obtained from wet bench cadaveric dissection of six heads correlated
extremely well with data obtained following dissection of five dry temporal bones
([Table 1]). Cadaveric data, in turn, correlated well with radiologic data ([Table 2]), with one exception. The average distance of the extrapolated junction of the GSPN
and facial nerve to the anterior border of the membranous cochlea was, on average,
2.0 mm less in the radiologic set. This discrepancy likely relates to difficulty with
obtaining this specific measurement in a prefixed axial plane but may also relate
to the chosen method of estimating the position of the GPSN based on the underlying
bony groove—which has not undergone previous, rigorous evaluation. Interestingly,
the other two parameters that were measured demonstrated a high level of agreement
between radiologic and morphologic data.
Review of the data obtained indicates that a “danger zone” exists within 11 mm of
the extrapolated junction of the GSPN and facial nerve. Drilling of the anteromedial
petrous bone within of a radius of 11 mm from the extrapolated junction of GSPN and
facial nerve appears to be associated with a high degree of risk to the membranous
cochlea. Drilling outside of a radius of 12.5 mm from the extrapolated junction of
GSPN and facial nerve appears to be associated with a low degree of risk to the bony
cochlear apparatus.
In conclusion, our study demonstrates that, when drilling Kawase's rhomboid, it is
useful to locate the extrapolated junction of the GSPN and the facial nerve. Drilling
of the anteromedial petrous bone outside of a radius of 12.5 mm from the extrapolated
junction of GSPN and facial nerve appears to be associated with a low degree of risk
to the cochlear apparatus.
