The Laryngoscope

V
C 2012 The American Laryngological,
Rhinological and Otological Society, Inc.

Petrous Apex Cholesterol Granulomas: Endonasal Versus

Infracochlear Approach

Tiago Fernando Scopel, MD; Juan C. Fernandez-Miranda, MD; Carlos D. Pinheiro-Neto, MD;
Maria Peris-Celda, MD; Alessandro Paluzzi, MD; Paul A. Gardner, MD; Barry E. Hirsch, MD;
Carl H. Snyderman, MD, MBA

Objectives/Hypothesis: The aim of this study was to investigate and compare the surgical anatomy of two different
routes to access and drain petrous apex (PA) cholesterol granulomas: the expanded endonasal approach (EEA) and the trans-
canal infracochlear approach (TICA).
Study Design: Anatomic and radiologic study.
Methods: The EEA and TICA to the PA were performed in 11 anatomic specimens with the assistance of imaging guid-
ance. The PA was categorized into three zones: superior PA, anterior-inferior PA, and posterior-inferior PA. The maximum
drainage window achieved by each approach was calculated using the imaging studies of each anatomic specimen.
Results: The EEA was able to reach superior PA and anterior-inferior PA in all specimens and posterior-inferior PA in
90%. The TICA did not provide access to superior PA in any case. The TICA was suitable to reach anterior-inferior PA in 80%
of specimens and posterior-inferior PA in 60%. Based on the radiologic study, the EEA provided a drainage window three
times larger than the TICA.
Conclusions: The transnasal approach provides reliable access to the PA when combined with internal carotid artery
exposure and allows for large drainage window. The transcanal approach is less versatile and more limited than the trans-
nasal approach but provides access to the most posterior and inferior portion of the PA without Eustachian tube transection.
Here we propose a new surgical classification that may help to decide the most suitable approach to the PA according to the
location and extension of the lesion.
Key Words: Petrous apex, cholesterol granuloma, skull base, ear surgery, endoscopic endonasal surgery.
Level of Evidence: 5
Laryngoscope, 122:751761, 2012

INTRODUCTION lobe retraction is necessary, which could result in brain

The petrous apex (PA) is one of the least accessible
areas of the skull base. Cholesterol granuloma (CG) is
the most common lesion of the PA.1 A variety of surgical
approaches have been used to drain CG when clinically
indicated. If the patient presents with preserved hear-
ing, four routes are possible: transcanal infracochlear,
transsphenoidal, infralabyrinthine, and middle cranial
fossa. If hearing is not preserved, there are two more
options: translabyrinthine and transcochlear.2,3 The
infralabyrinthine approach may provide limited access,
especially in patients with high jugular bulb (JB).3,4 The
middle cranial fossa approach does not allow a perma-
nent drainage pathway, and some degree of temporal
From the Department of Otolaryngology (T.F.S., C.D.P.-N., B.E.H.,
C.H.S.); and Department of Neurological Surgery (J.C.F.-M., M.P.-C., A.P.,
P. A .G ., B . E . H ., C . H . S .), University of Pittsburgh School of Medicine,
Pittsburgh, Pennsylvania, U.S.A.
Editors Note: This Manuscript was accepted for publication
October 13, 2011.
The authors have no funding, financial relationships, or conflicts
of interest to disclose.
Send correspondence to Juan C. Fernandez-Miranda, MD, Depart-
ment of Neurological Surgery, University of Pittsburgh School of Medicine,
200 Lothrop Street, PUH B-400, Pittsburgh PA 15213.
E-mail: fernandezmirandajc@upmc.edu
DOI: 10.1002/lary.22448
injury.2 Despite the high rate of recurrence, the transca-
nal infracochlear approach (TICA) is the traditional
procedure for treatment of PA CG with hearing preser-
vation.5 The transsphenoidal approach is appropriate for
CGs that protrude medially into the sphenoid sinus;
however, it is hindered when the lesion is located more
laterally.
Advances in minimally invasive skull base surgery,
through careful study of the endoscopic endonasal surgi-
cal anatomy allied with accurate intraoperative image
guidance and devoted instrumentation, have provided
safe access to regions distant from the midline.2,6 Some
authors have completed anatomic dissection studies and
calculated the distances between relevant structures
involved with the infracochlear and endonasal accesses
to the PA.7,8 Nevertheless, there are no previous
anatomic and/or surgical studies that have compared
these two routes or performed a systematic description of
which approach would be more suitable for the treatment
of CG according to the surgical anatomy of the PA.
The aim of this study was to compare two different
approaches to the PA: the endoscopic endonasal approach
(EEA) and TICA. The degree of exposure achieved in
each approach was evaluated. A new clinical-surgical
classification of the PA region is proposed.

Laryngoscope 122: April 2012 Scopel et al.: Endonasal Versus Infracochlear Approach
751
MATERIALS AND METHODS
Eleven human anatomic specimens (11 heads, 22 petrosal
apices) were prepared for dissection at the Surgical Neuroanat-
omy Laboratory of the University of Pittsburgh. This research
was approved by the Committee for Oversight to Research
Involving the Dead (CORID). The common carotid arteries, ver-
tebral arteries, and internal jugular veins were dissected,
isolated, and canulated with flexible tubing. Each vessel was
flushed to remove blood clots with approximately 500 mL of
warm water in two separate sessions. The vessels were injected
with 50 mL of a polymethyl siloxane/silicone conglomerate, in a
2:1 preparation for arteries and 1:1 for veins, dyed with red or
blue water-soluble pigments, respectively. Dilaurate calcium
carbonate (5 mL) was added as a catalyst immediately before
infusion. Specimens were refrigerated overnight and then pre-
served in 200 mL proof ethyl alcohol diluted in water as a 70%
solution. Five screws were placed, distributed in the cranial
area, and high-resolution computed tomography (CT) with
1 mm of separation acquisition was performed for neuronaviga-
tion protocol (Stryker, Kalamazoo, MI). Two dry skull bases
were prepared to study the osseous relationships.
Each head was stabilized with a Mayfield head holder
simulating the surgical position obtained in the operating room.
Endonasal approaches were performed in five heads (10 sides)
under endoscopic visualization (Karl Storz, Tuttlingen,
Germany; 4 mm, 18 cm, Hopkins II, 0 and 45 degrees). Transca-
nal infracochlear approaches were performed in another five
heads (10 sides) under optic magnification (3
to 40
) with a
surgical microscope (OPMI; Zeiss, Oberkochen, Germany); one
extra head was used to perform both approaches in the same
specimen. The measurements to estimate the maximal drainage
window for both approaches were evaluated using the high-
definition CT scans in all heads.
Dissection Technique
To compare the accessibility and clinical applicability of
both approaches (EEA and TICA), the PA was divided into three
zones in accordance with three constant landmarks, foramen
lacerum, carotid foramen, and jugular foramen, as follows:
superior PA (sup PA), between the level of Dorellos canal
(petrosal segment of sixth nerve) and foramen lacerum; ante-
rior-inferior PA (ant-inf PA), between the foramen lacerum and
the carotid foramen; posterior-inferior PA (post-inf PA), between
the carotid foramen and the jugular foramen (Figs. 1 and 2).
The zones were considered accessible when optimum visualiza-
tion and instrumentation were achieved in the PA. Also, to
increase the consistency of our method, each approached zone
was verified using neuronavigation system in all specimens.
Expanded endonasal approach to the PA. There are
three ways to access the PA through the EEA: 1) transsphenoi-
dal transclival approach, 2) transsphenoidal transclival
approach with paraclival internal carotid artery (ICA) laterali-
zation (medial PA approach), and 3) transpterygoid infrapetrous
or sublacerum approach (inferior to the horizontal petrous
ICA).2 We performed the three different approaches on each
side in all specimens aiming to obtain the largest possible win-
dow into the PA.
TRANSSPHENOIDAL TRANSCLIVAL APPROACH. Initially, a
wide bilateral sphenoethmoidectomy was done. Approximately 1
cm of the posterior edge of the nasal septum was resected using
back-biting rongeurs to expose the sphenoid rostrum. A wide
sphenoidotomy was performed with removal of the rostrum.
Septations within the sphenoid sinus were removed, and imag-
ing guidance was used to define the course of the ICA. Other
anatomic landmarks were identified: planum sphenoidale, sella,
clival recess, optic canal, and medial and lateral optic-carotid
Laryngoscope 122: April 2012
752
recesses. The sphenoid floor was drilled back to the clival
recess. Subsequently a transclival approach was performed, and
the bone between the paraclival ICAs was drilled until the
clival dura was exposed. The paraclival ICAs were not uncov-
ered (Fig. 3).
TRANSSPHENOIDAL TRANSCLIVAL APPROACH WITH ICA
LATERALIZATION. This approach consisted of an additional step
after the previous one was completed. The bone covering the
paraclival ICA from its medial to lateral surface was removed
to allow the vessel to be gently mobilized in a lateral direction;
the bone behind the paraclival ICA (medial-posterior wall of the
paraclival carotid canal) was removed to gain access to the
petroclival fissure and most medial aspect of the PA. Then, the
petrous bone between the paraclival ICA and the inferior petro-
sal sinus (IPS) was removed from the level of Dorellos canal to
the foramen lacerum to maximize the surgical access to sup PA.
Special attention was taken in this stage to avoid injury to the
sixth nerve superiorly along its petrous segment (Fig. 3).9
TRANSPTERYGOID INFRAPETROUS OR SUBLACERUM
APPROACH. After the two previous modules were completed,
bilateral maxillary antrostomies were performed. The posterior
wall of the maxillary sinus was removed exposing the pterygo-
palatine fossa. Medially, the palatosphenoidal (palatovaginal)
artery, a terminal branch of the maxillary artery, was trans-
ected, and soft tissues of the pterygopalatine fossa were
mobilized laterally to facilitate identification of the vidian
nerve.10 The pterygoid base was drilled out, preserving the
vidian nerve. The drilling proceeded circumferentially along the
vidian canal (keeping vidian nerve and artery intact) until the
foramen lacerum soft tissue and anterior genu of the ICA were
reached.10 These are the most important anatomic landmarks of
this module.
After completing these steps, we transected the foramen
lacerum soft tissue and drilled the bone between the horizontal
petrous ICA and Eustachian tube (ET). Hence, we completed a
sublacerum route into the ant-inf PA. To further expose post-inf
PA, we transected the ET and drilled the petrous carotid canal
inferiorly to provide complete exposure of the horizontal petrous
ICA. We followed the horizontal petrous carotid segment in the
direction of the carotid and jugular foramina and identified the
angle formed between the horizontal petrous ICA and posterior
vertical petrous ICA. This provided full access to the posterior-
inferior zone of PA (Fig. 4).
Transcanal infracochlear approach. A standard post-
auricular incision was performed, the auricle was reflected
anteriorly, and the external auditory canal was transected at
the bony-cartilaginous junction. A tympanomeatal flap was
then elevated from the 2 oclock to the 10 oclock position, leav-
ing the tympanic membrane attached at the umbo. Using a
high-speed drill with continuous suction and irrigation, bone
was removed over the anterior, inferior, and posterior portions
of the bony canal wall to expose the hypotympanum. A small di-
amond burr was used to remove air cells below the cochlea to
expose the course of the carotid artery and JB. Bony dissection
was then carried out in an anterior-medial direction, using the
carotid artery as the anterior limit, the round window as the
superior limit, and the JB as the posterior limit. This provided
exposure of ant-inf PA and post-inf PA. We also extended the
dissection anteriorly toward the temporomandibular joint, with-
out its violation; and posteriorly, following the chorda tympani
nerve, to achieve the largest window possible (Fig. 5).
Radiologic Study
Cranial CT scans of 11 human heads (22 petrosal bones)
were studied. We generated sagittal oblique cuts to simulate the
maximal drainage window obtained into the PA by EEA and
Scopel et al.: Endonasal Versus Infracochlear Approach
Fig. 1. Anatomy and surgical classification of the petrous apex (PA) in human skull bases. (A) Superior view of skull base. (B) Medial view
of the PA showing its zones. (C) Front view simulating endoscopic endonasal view of PA zones in the operation room. Note the relationship
of PA zones with internal carotid artery (ICA) divisions. Also, note the illustrative extension reached by expanded endonasal approach (EEA)
and transcanal infracochlear approach (TICA) on PA. Superior PA: relationship with paraclival ICA. Antero-inferior PA: relationship with hori-
zontal petrous ICA. Posterior-inferior PA: relationship with posterior vertical petrous ICA. Inf inferior; Fiss fissure; Sup superior; Ant-
inf anterior-inferior; Post-inf posterior-inferior; VI six nerve; VII facial nerve; VIII vestibulocochlear nerve; V2 maxillary nerve;
V3 mandibular nerve; Paraph parapharyngeal; IX glossopharyngeal nerve; X vagus nerve; XI accessory nerve.

TICA (Fig. 6A and 6B). The measurements to predict the maxi- permits a dynamic analysis of the images, thus allowing a three-
mal drainage window from both approaches were performed on dimensional localization of the structures. The data were ana-
petrous bones using the software OsiriX for Macintosh, version lyzed using software SPSS 17.0 for Macintosh (SPSS, Inc.,
3.9.1 (The OsiriX Foundation, Geneva, Switzerland), that Chicago, IL), and P  .05 was considered significant.

Laryngoscope 122: April 2012 Scopel et al.: Endonasal Versus Infracochlear Approach
753

Fig. 2. Computed tomography (CT) scans before dissection of
specimens. (A) CT scan predissection of the superior petrous apex
(PA) (sup PA); note the straight relation of paraclival internal carotid
artery (ICA) with sup PA. (B) CT scan predissection of the anterior-
inferior PA (ant-inf PA); note the relationship of horizontal petrous
ICA with ant-inf PA. (C) CT scan predissection of the posterior-
inferior PA (post-inf PA); note the correlation between posterior
vertical petrous ICA and post-inf PA. Inf inferior; Fiss fissure.
RESULTS
Using the transnasal route it was possible to access
all three zones of the PA in nine out of 10 (90%) speci-
mens. In one specimen, the post-inf PA could not be
Laryngoscope 122: April 2012
754
approached endonasally because the posterior vertical
petrous ICA was located very close to the IPS and the
brainstem.
The infracochlear route did not provide access to
sup PA because the cochlea blocked the superior exten-
sion of the dissection in all specimens. This approach
offered access to ant-inf PA in eight of 10 (80%) speci-
mens and post-inf PA in six of 10 (60%) specimens. In
two specimens we could not access any area of the PA by
TICA because the posterior genu of ICA was located ad-
jacent to the JB. In two other specimens, the access to
ant-inf PA was achieved, but the inferior extension to
post-inf PA was prevented by the posterior vertical
petrous ICA, which was situated close to the JB.
In one extra head we conducted a demonstrative
EEA and TICA to the PA in the same specimen (right
side) to provide a simultaneous comparison between
both approaches. By EEA we were able to find the route
previously created by TICA into the PA and enlarge it
inferiorly and superiorly (Figs. 6C, 6D, and 7).
Comparison of Surgical Approaches to the PA
To access the sup PA by EEA, we performed a
transsphenoidal transclival approach with ICA laterali-
zation; this approach also allowed for preservation of the
foramen lacerum soft tissue and ET. The approach to
ant-inf PA was accomplished by a transpterygoid infra-
petrous approach with transection of the foramen
lacerum soft tissue and preservation of the ET. In one
specimen, in addition to the transection of the lacerum
soft tissue, we required ET sacrifice to access the ant-inf
PA. To access post-inf PA, we always required transec-
tion of the ET.
By applying the TICA, we first reached the ant-inf
PA at the level of posterior genu of ICA. Then we fol-
lowed the dissection in an anteromedial direction,
drilling parallel to the horizontal petrous portion of the
ICA. After completing this step, we followed the poste-
rior genu of ICA inferiorly to enter post-inf PA, until we
reached the corner between posterior vertical petrous
ICA and JB. Most of the time, the maximal drainage
window created by TICA was a small triangle between
the ICA, JB, and basal turn of the cochlea (Fig. 5D).
Using the EEA, the maximal drainage window had a
large oval shape between the ICA, IPS, and brainstem
(Fig. 4D).
On the radiologic study of specimens, the average of
the maximal drainage window provided by the EEA was
1.19 cm2 and for TICA was 0.38 cm2 (P < .001) (Table I).
Illustrative Case
Presentation and examination. A 66-year-old
female presented with headache on the right side, dizzi-
ness, tinnitus, and previous diagnosis of PA CG. She had
two operations (TICA) to drain the CG at another insti-
tution, with recurrence of the granuloma. Preoperative
CT findings were compatible with a CG of the PA (sup
PA and ant-inf PA), without protrusion into the sphenoid
sinus, and with intracystic septations (Fig. 8A).
Scopel et al.: Endonasal Versus Infracochlear Approach
Fig. 3. Cadaveric stepwise dissection by expanded endonasal approach. Transsphenoidal transclival approach and transsphenoidal trans-
clival approach with internal carotid artery (ICA) lateralization to petrous apex (PA) on right side. Photographs obtained using 0-degree
endoscope. (A) Nasal cavity inspection with sphenoid ostium identification. (B) Wide sphenoethmoidectomy exposing anterior ethmoidal ar-
tery (ant eth art), posterior ethmoidal artery (post eth art), sphenoid sinus, and resection of posterior nasal septum. (C) Wide bilateral sphe-
noethmoidectomy. (D) Close-up view of transclivus approach; the clivus was drilled until a thin layer of bone; the right paraclival ICA was
completely exposed. (E) Exposure of superior PA. View with 45-degree endoscope: the dura is exposed, and the bone between paraclival
ICA and inferior petrous sinus was drilled, opening the superior PA (sup PA); note that inferior petrous sinus was removed to show its rela-
tionship with petrous segment of the sixth nerve. (F) Endoscopic endonasal neuronavigation. The PA was approached medially, and the
neuronavigation confirmed that we achieved the sup PA; note the straight relationship between paraclival ICA and sup PA. Op carot
optic-carotid; VI (petr seg) abducens nerve (petrosal segment); Inf inferior.

Laryngoscope 122: April 2012 Scopel et al.: Endonasal Versus Infracochlear Approach
755
Fig. 4. Cadaveric stepwise expanded endonasal approach. Transpterygoid infrapetrous or sublacerum approach on the right side. Photo-
graphs obtained using 0-degree endoscope. (A) The components of the pterygopalatine fossa (Pterygop fossa) were lateralized to show
the palatosphenoidal artery. (B) Transclival and transpterygoid approach. The pterygoid base was drilled out, preserving the vidian nerve.
(C) The Eustachian tube (ET) was partially opened, and a lapel pin was placed to show its lumen. The bone between horizontal petrous
(Horiz petr) internal carotid artery (ICA) and ET was drilled, the lacerum soft tissue was partially transected, and the jugular tubercle was
exposed. (D) View with 45-degree endoscope. The carotid canal was drilled out and horizontal petrous ICA was complete exposed. A small
window was opened in the parapharyngeal ICA (Paraph ICA) to evidence its relationship with ET. The petrous bone between the segments
of ICA and inferior petrosal sinus were removed to achieve all zones of petrous apex (PA). (E) Endoscopic endonasal neuronavigation show-
ing the correlation between horizontal petrous ICA and anterior-inferior PA (ant-inf PA). (F) Endoscopic endonasal neuronavigation exhibiting
the relation between posterior vertical petrous (Post vert petr) ICA and posterior-inferior PA (post-inf PA). Proc pal process of palatine;
Inf inferior; VI sixth nerve; XII hypoglossal nerve; Fiss fissure; Sup superior.

Laryngoscope 122: April 2012 Scopel et al.: Endonasal Versus Infracochlear Approach
756
Fig. 5. Transcanal infracochlear approach (TICA) on the right side. Microscopic view. (A) External auditory canal and tympanic membrane
view after standard postauricular incision and transection of the bony-cartilaginous junction. (B) The inferior, anterior, and posterior canal-
oplasty was performed, the tympanic membrane was elevated, and the chorda tympani, promontory, and hypotympanum air cells were
identified. (C) Demonstrative view obtained using 0-degree endoscope. Exposition of middle ear and hypotympanum is shown; note the
course of facial nerve (tympanic segment) and Jacobson nerve. (D) The window obtained by TICA (triangle). The air cells below the
cochlea were removed to expose the posterior genu petrous internal carotid artery (ICA), posterior vertical petrous ICA (post vert petr
ICA), and jugular bulb; note the lateral projection of anterior-inferior petrous apex (PA) (ant-inf PA) and posterior-inferior PA (post-inf PA)
under the cochlea. (E) Neuronavigation obtained by TICA. Note the relationship between horizontal petrous ICA (petr ICA) and ant-inf
PA. (F) Neuronavigation obtained by TICA. Note the relation between post vert petr ICA and post-inf PA. Seg segment; Sup
superior.

Laryngoscope 122: April 2012 Scopel et al.: Endonasal Versus Infracochlear Approach
757
Fig. 6. Demonstrative expanded endonasal approach (EEA) and transcanal infracochlear approach (TICA) to petrous apex (PA) in the same
specimen (right side) and radiologic study of petrous bone. (A) Left sagittal oblique view obtained through the cut at the level of petroclival
fissure in axial computed tomography (CT) scan. The maximal drainage window provided by petrous bone for EEA in one specimen. (B)
Left sagittal oblique view obtained through the cut at the level of posterior genu internal carotid artery (ICA) and jugular bulb in axial CT
scan. The maximal drainage window provided by petrous bone for TICA in one specimen. (C) Endoscopic endonasal view with 45-degree
endoscope. The EEA found the route previously built by TICA into PA and enlarged it inferiorly and superiorly as well as exposed a portion
of membranous cochlea; note the structures surrounding PA that can be involved in a real case of cholesterol granuloma. (D) Endoscopic
endonasal view with 45-degree endoscope. Close-up view in the corridor created by TICA into PA; note the tympanic membrane and mem-
branous cochlea through endoscopic endonasal view. Petr seg VI petrous segment of abducens nerve; Cis seg VI cisternal segment
of abducens nerve; Inf inferior; IX glossopharyngeal nerve; X vagus nerve; XI accessory nerve; Jug tub jugular tubercle; Hip
hypoglossal; P vert pe posterior vertical petrous; Horiz pe horizontal petrous; Lac lacerum; Inf petr inferior petrosal; Sup supe-
rior; Ant-inf anterior-inferior; Post-inf Posterior-inferior; * carotid foramen; ** posterior genu petrous ICA.

Operation and postoperative course. A fully en-
doscopic endonasal transsphenoidal transclival approach
combined with a transpterygoid infrapetrous approach
was performed. After wide bilateral sphenoidotomies,
the clivus was drilled until a thin layer of cortical bone
was left covering the clival dura. The vidian canal was
drilled out until the nerve reached to the lateral portion
of the lacerum segment of the ICA (Fig. 8B). The right
paraclival ICA was completely skeletonized up to the lac-
erum foramen to allow gentle lateralization of the ICA
and access to the PA (Fig. 8C). Neuronavigation was
used to identify the course of ICA and PA areas
(Fig. 8D); the lacerum soft tissue was identified and par-
tially transected, and the IPS was identified on the most
lateral aspect of the clivus, running at the petroclival fis-
sure (Fig. 8C). Finally, the CG was readily drained by

Laryngoscope 122: April 2012 Scopel et al.: Endonasal Versus Infracochlear Approach
758

Fig. 7. Schematic three-dimensional figure to provide subjective
comparison between both approaches, expanded endonasal
approach (EEA) and transcanal infracochlear approach (TICA), to
petrous apex (PA). Frontal view simulating the endoscopic endo-
nasal approach in the operating room. Note the limits of each
zone of PA and the structures surrounding it, as well as the main
direction provided by each approach. Sp superior PA; Ai an-
terior-inferior PA; Pi posterior-inferior PA; VI abducens nerve;
AE arcuate eminence; TI trigeminal impression; Co coch-
lea; JV jugular vein, ET Eustachian tube; CF carotid fora-
men; L lacerum; VN vidian nerve; VA vidian artery; VC
vidian canal; Pt pituitary gland; St stalk; P genu posterior
genu petrous ICA; Horiz Horizontal petrous ICA; pICA para-
clival ICA; A genu Anterior genu petrous ICA; ICA Internal ca-
rotid artery.
removing the cystic petrous bone behind the paraclival
ICA in sup PA and under horizontal petrous ICA in ant-
inf PA; a Silastic tube (Dow Corning, Corp., Midland,
MI) was placed in the new cavity created into the PA
(Fig. 8E). Immediate postsurgical CT showed the Silastic
in the PA (Fig. 8F). The patient was discharged 2 days
postoperatively, and at 3-month follow-up the Silastic
tube was clearly visualized in the sphenoid sinus and
removed under endoscopic visualization in the outpa-
tient clinic. The patient showed complete resolution of
preoperative symptoms.
DISCUSSION
In this study, we show the surgical anatomy of two
distinct approaches to the PA: transnasal and infraco-
chlear. Both approaches are currently used for drainage
of PA CGs, aiming to aerate the cyst and prevent recur-
rence. Size and location of the lesion, level of functional
hearing, and neurovascular anatomy are critical factors
needed to determine which approach to use for drain-
age11; however, recent advances in endoscopic endonasal
skull base surgery mandate reevaluation of current cri-
teria for surgical approach selection.
Here we present a surgical classification of the PA
that allows correlation between the topography of the
lesion and the segments of the ICA. Importantly, the
ICA is the most critical structure at risk when approach-
ing the PA. A recent study categorizes the petrous ICA
Laryngoscope 122: April 2012
into five different segments, from proximal to distal: pos-
terior vertical segment, posterior genu, horizontal
segment, anterior genu (lacerum segment), and anterior
vertical segment (paraclival ICA) (Fig. 7).12 The para-
clival ICA, horizontal petrous ICA, and posterior vertical
petrous ICA sit at the level of sup PA, ant-inf PA, and
post-inf PA, respectively (Fig. 2). This proposed classifi-
cation is merely a schematic way of analyzing the
location and extension of CG within the PA. Most CGs
will not be limited exclusively to one zone. This is partic-
ularly true when dealing with large granulomas that
may require surgical drainage. The simplified classifica-
tion proposed here may help the skull base surgeon in
choosing the most suitable approach according to the to-
pography of the lesion on preoperative imaging.
Our results showed the endonasal route was more
consistent because it reached sup PA and ant-inf PA in
all specimens and post-inf PA in most specimens. The
transcanal infracochlear route was never suitable for
accessing sup PA but provided direct access to ant-inf
PA and post-inf PA without transection of the ET when-
ever the anatomy was favorable (large window between
cochlea, ICA, and JB). Importantly, based on our radio-
logic study, the EEA can provide a large drainage
window, allowing placement of a Silastic stent three
times larger than the infracochlear approach; this statis-
tically significant difference in size of the drainage
window may have a direct impact on postoperative re-
currence rates.
In our experience, the EEA provides a comfortable
route with wide operative field, direct access to the mid-
line, and potential to extend the approach to the
superior and inferior PA (Figs. 1C, 6A). When CGs pro-
trude into or are directly behind the posterior wall of the
sphenoid sinus, the endonasal approach is technically
straightforward and is the preferred surgical option.2
Furthermore, by EEA, the surgeon is able to drain an
uncommon bilateral CG13,14 using the same corridor.
Others advantages of the endonasal route include wide
drainage window into the PA connected with the natural
nasal corridor and sinonasal mucociliary system and
postoperative endoscopic surveillance in the office.8 It
also can be performed in patients with high JB,15 carries
no risk for the facial nerve or the auditory and vestibu-
lar systems, and allows for fast recovery with minimal
postoperative symptoms, no external scars, and short
hospitalization.2
On the other hand, the advantages of the TICA
include preservation of normal external and middle ear
conduction mechanism. It also allows drainage of the
granuloma in a well aerated portion of the middle ear,
TABLE I.
Maximal Areas Obtained in the Petrous Bones for Each Approach
on Radiologic Study of 11 Cranial Computed Tomography Scans.
Areas Minimum, cm2 Maximum, cm2 Mean 6 SD, cm2
EEA 0.71 1.77 1.19 6 0.3
TICA 0.1 0.64 0.38 6 0.16
SD standard deviation; EEA expanded endonasal approach;
TICA transcanal infracochlear approach.
Scopel et al.: Endonasal Versus Infracochlear Approach
759
Fig. 8. Illustrative case. (A) Computed tomography (CT) preoperative scan showing the petrous apex (PA) cholesterol granuloma (Cholest
Granul) without protrusion to sphenoid sinus. (B) Intraoperative view of expanded endonasal approach (EEA) with 0-degree endoscope. The
transsphenoidal transclival approach and transpterygoid infrapetrous approaches to PA were performed to access superior PA (sup PA) and
anterior-inferior PA. The pterygoid base was drilled out preserving the vidian nerve. The bone of clivus was drilled until a thin layer (transclival
approach). The paraclival internal carotid artery (ICA) was skeletonized towards lacerum segment of ICA to allow lateralization of ICA and infe-
rior extension of dissection, at the level of superior PA to anterior-inferior PA. (C) Endoscopic view with 45-degree endoscope on the right side.
Window created into PA by EEA; note the paraclival ICA as anterior limit and inferior (Inf) petrosal sinus as posterior limit of dissection. (D) Intra-
operative neuronavigation showing the point achieved into PA through paraclival ICA lateralization. (E) Endoscopic view with 0-degree endo-
scope. The Silastic was placed into PA, creating permanent pathway drainage connected with natural nasal corridor. (F) CT of the first
postoperative day evidencing the Silastic in PA. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

near the ET,5 fast postoperative recovery without Nevertheless, the access to PA by EEA can be diffi-
requirement several office visits for debridements, and cult when the brainstem and anterior vertical segment
re-exploration through a simple inferior myringotomy. of the ICA are nearby or when there is a poor

Laryngoscope 122: April 2012 Scopel et al.: Endonasal Versus Infracochlear Approach
760
pneumatization of the sphenoid sinus. However, intrao-
perative image guidance and meticulous surgical
technique allows the precise removal of bone from the
clival region with minimal risk to the ICA to reach PA
medially. Similarly, the access to PA by TICA can be dif-
ficult when there is a high or anteriorly placed JB,16,17 if
the lesion is located anterior and superior to the cochlea,
and if there is poor pneumatization of hypotympanum.18
In addition to some risk of ICA injury, the potential
risks of EEA include diplopia from abducens nerve palsy
and dry eye from vidian nerve damage. All of these risks
can be minimized with the use of imaging guidance and
exquisite surgical anatomy knowledge.2 Another theo-
retic risk would be otitis media with effusion due to
section of the ET, but this is not fully elucidated. The
potential risks of TICA include injury of the ICA and
JB, damage to the cochlea with complete hearing loss14
due to narrow field of dissection7; facial nerve and
chorda tympani nerve injury; and conductive hearing
loss secondary to postoperative chronic inflammation of
the middle ear and intraoperative damage of the Jacob-
sons nerve.19
Contemporary skull base surgery requires multidis-
ciplinary evaluation, selection, and execution of surgical
approaches. Rather than reducing surgical options to a
single approach, different alternatives should be avail-
able to offer patients the best choice for each individual
case. Here we propose the endonasal approach as the
preferred surgical option for CGs that require drainage
and extend into the sphenoid sinus and those in the sup
and ant-inf PA. The infracochlear approach would be
ideal for those located in the post-inf PA, because the
endonasal access may require ET transection. Finally,
these challenging endonasal approaches should only be
attempted after significant and successful experience
has been acquired with less demanding surgical cases.
CONCLUSION
The ICA is of critical importance for approaches to
the PA; hence it is useful to categorize the PA into zones
based on the relationship with ICA. The EEA provided a
wide surgical field to PA with full preservation of impor-
tant structures such as cochlea, vestibular system, and
facial nerve. The endonasal route afforded exposure to
all zones of PA in most specimens (90%); moreover, on
radiologic study of specimens, the petrous bones pro-
vided a large drainage window for EEA, three times
larger than TICA. TICA did not achieve access to sup
Laryngoscope 122: April 2012
PA in any specimen but allowed access to ant-inf PA in
most specimens (80%) and to post-inf PA in 60% of speci-
mens; it did not entail transection of the ET. The choice
of surgical approach to PA CGs depends on preoperative
status of the patients hearing, the surgeons skills, and
exhaustive study of PA anatomy and its relationship
with the pathology on preoperative image.
BIBLIOGRAPHY
1. Lo WW, Solti-Bohman LG, Brackmann DE, Gruskin P. Cholesterol granu-
loma of the petrous apex: CT diagnosis. Radiology 1984;153:705711.
2. Zanation AM, Snyderman CH, Carrau RL, Gardner PA, Prevedello DM,
Kassam AB. Endoscopic endonasal surgery for petrous apex lesions. La-
ryngoscope 2009;119:1925.
3. Gore MR, Zanation AM, Ebert CS, Senior BA. Cholesterol granuloma of
the petrous apex. Otolaryngol Clin North Am 2011;44:10431058.
4. Griffith AJ, Terrell JE. Transsphenoid endoscopic management of petrous
apex cholesterol granuloma. Otolaryngol Head Neck Surg 1996;114:
9194.
5. Brackmann DE, Toh EH. Surgical management of petrous apex cholesterol
granulomas. Otol Neurotol 2002;23:529533.
6. Kassam AB, Gardner P, Snyderman C, Mintz A, Carrau R. Expanded
endonasal approach: fully endoscopic, completely transnasal approach to
the middle third of the clivus, petrous bone, middle cranial fossa, and
infratemporal fossa. Neurosurg Focus 2005;19:E6.
7. Giddings NA, Brackmann DE, Kwartler JA. Transcanal infracochlear
approach to the petrous apex. Otolaryngol Head Neck Surg 1991;104:
2936.
8. Chatrath P, Nouraei SA, De Cordova J, Patel M, Saleh HA. Endonasal en-
doscopic approach to the petrous apex: an image-guided quantitative an-
atomical study. Clin Otolaryngol 2007;32:255260.
9. Barges-Coll J, Fernandez-Miranda JC, Prevedello DM, et al. Avoiding
injury to the abducens nerve during expanded endonasal endoscopic
surgery: anatomic and clinical case studies. Neurosurgery 2010;67:
144154; discussion 154.
10. Prevedello DM, Pinheiro-Neto CD, Fernandez-Miranda JC, et al. Vidian
nerve transposition for endoscopic endonasal middle fossa approaches.
Neurosurgery 2010;67:478484.
11. Mosnier I, Cyna-Gorse F, Grayeli AB, et al. Management of cholesterol
granulomas of the petrous apex based on clinical and radiologic evalua-
tion. Otol Neurotol 2002;23:522528.
12. Osawa S, Rhoton AL Jr, Tanriover N, Shimizu S, Fujii K. Microsurgical
anatomy and surgical exposure of the petrous segment of the internal
carotid artery. Neurosurgery 2008;63:210238; discussion 239.
13. Jaberoo MC, Hassan A, Pulido MA, Saleh HA. Endoscopic endonasal
approaches to management of cholesterol granuloma of the petrous
apex. Skull Base 2010;20:375379.
14. Castillo MP, Samy RN, Isaacson B, Roland PS. Petrous apex cholesterol
granuloma aeration: does it matter? Otolaryngol Head Neck Surg 2008;
138:518522.
15. Haberkamp TJ. Surgical anatomy of the transtemporal approaches to the
petrous apex. Am J Otol 1997;18:501506.
16. Dhanasekar G, Jones NS. Endoscopic trans-sphenoidal removal of choles-
terol granuloma of the petrous apex: case report and literature review.
J Laryngol Otol 2011;125:169172.
17. Sanna M, Dispenza F, Mathur N, De Stefano A, De Donato G. Otoneuro-
logical management of petrous apex cholesterol granuloma. Am J Oto-
laryngol 2009;30:407414.
18. Shoman NM, Samy RN, Pensak ML. Double-barrel stenting in infraco-
chlear approach for drainage of petroux apex cholesterol granulomas.
Laryngoscope 2011;121:574576.
19. Ceylan A, Goksu N, Kemaloglu YK, Ugur B, Akyurek N, Bayazit YA.
Impact of Jacobsons (tympanic) nerve sectioning on middle ear func-
tions. Otol Neurotol 2007;28:341344.
Scopel et al.: Endonasal Versus Infracochlear Approach
761