DOI: 10.3171/2012.5.

JNS112334

Interactive virtual simulation using a 3D computer graphics

model for microvascular decompression surgery

Clinical article
MAKOTO OISHI, M.D., MASAFUMI FUKUDA, M.D., TETSUYA HIRAISHI, M.D.,
NAOKI YAJIMA, M.D., YOSUKE SATO, M.D., AND YUKIHIKO FUJII, M.D.
Department of Neurosurgery, Brain Research Institute, Niigata University, Niigata, Japan

Object. The purpose of this paper is to report on the authors’ advanced presurgical interactive virtual simulation
technique using a 3D computer graphics model for microvascular decompression (MVD) surgery.
Methods. The authors performed interactive virtual simulation prior to surgery in 26 patients with trigeminal
neuralgia or hemifacial spasm. The 3D computer graphics models for interactive virtual simulation were composed
of the brainstem, cerebellum, cranial nerves, vessels, and skull individually created by the image analysis, includ-
ing segmentation, surface rendering, and data fusion for data collected by 3-T MRI and 64-row multidetector CT
systems. Interactive virtual simulation was performed by employing novel computer-aided design software with
manipulation of a haptic device to imitate the surgical procedures of bone drilling and retraction of the cerebellum.

Results. In all patients, interactive virtual simulation provided detailed and realistic surgical perspectives, of

patients) obtained by review of 2D images only (p < 0.05). Surgeons evaluated interactive virtual simulation as hav-
ing “prominent” utility for carrying out the entire surgical procedure in 50% of cases. It was evaluated as moderately
useful or “supportive” in the other 50% of cases. There were no cases in which it was evaluated as having no utility.
The utilities of interactive virtual simulation were associated with atypical or complex forms of neurovascular com-
pression and structural restrictions in the surgical window. Finally, MVD procedures were performed as simulated in
23 (88%) of the 26 patients .
Conclusions. Our interactive virtual simulation using a 3D computer graphics model provided a realistic envi-
ronment for performing virtual simulations prior to MVD surgery and enabled us to ascertain complex microsurgical
anatomy.
(http://thejns.org/doi/abs/10.3171/2012.5.JNS112334)

KEY WORDS

T
HREE-DIMENSIONAL analysis techniques for radio-
logical imaging data have become more conve-
nient and are providing more realistic images than
ever before. Furthermore, high-quality computer graphics
have also recently become possible using commercially
available software on a personal computer. For perform-
ing neurosurgical procedures, a precise understanding of
Abbreviations used in this paper: AICA = anterior inferior cer-
ebellar artery; CAD = computer-aided design; CISS = constructive
interference in steady state; CPA = cerebellopontine angle; CTA =
CT angiography; MRA = MR angiography; MVD = microvascular
decompression; PICA = posterior inferior cerebellar artery; SCA =
superior cerebellar artery; STL = standard triangulated language;
VA = vertebral artery.
-
sential. Introducing 3D imaging tools into neurosurgical
planning is expected to greatly enhance both the certainty
and the safety of surgical procedures.13,15,20,21,26 There are
recent reports demonstrating a 3D interactive computer
simulation system applicable not only to surveying 3D
data visually but also to manipulating data by employing
13,15,26
Neurovascular compression syndromes, such as tri-
hyperactive dysfunctions involving the cranial nerves
This article contains some figures that are displayed in color
online but in black-and-white in the print edition.

J Neurosurg / June 29, 2012 1


and have been explained as being caused by anatomi-
cal factors including vascular compression of the cranial
nerve.6 The success of MVD surgery, the only estab-
lished curative treatment for neurovascular compression
syndromes,10,11 largely depends on proper knowledge of
individual neurovascular anatomical structures in the
CPA cistern.8 The advent of MRI, especially heavily T2-
weighted imaging such as CISS imaging and MRA, has
contributed to improved diagnosis of neurovascular com-
pression syndromes.2,3,5,28,29 Recent application of the 3D
image-analysis technique to MRI data of patients with
neurovascular compression syndromes has provided di-
rect visualization of detailed anatomical structures that
alone.1,4,7,9,12,17–19,22–25,27
We developed a novel surgical simulation method.
This interactive virtual simulation is characterized by
creating a high-quality 3D computer graphics model
from MRI and CT data and then simulating surgical pro-
cedures on the unique CAD software with manipulation
of a haptic device. Herein, we demonstrate the utility of
this system for MVD surgery, especially for understand-
ing detailed and variable microsurgical anatomy and pre-
dicting microscopic perspectives during surgery.
Patients
Twenty-six patients with neurovascular compres-
sion syndromes (11 with trigeminal neuralgia and 15 with
hemifacial spasm; 11 men and 15 women; age range 26–82
years) underwent presurgical evaluation and simulation
for MVD by our interactive virtual simulation method.
All patients were determined to be candidates for MVD
surgery after assessment of clinical symptoms and MRI
studies, and all underwent surgery at the Niigata University
Hospital between February 2010 and July 2011.
Data Collection
Magnetic resonance images were acquired using a
3-T system (Verio, Siemens AG). Heavily T2-weighted
images were obtained by 3D CISS sequence with the fol-
angle 70°, fov 200 mm, matrix 512 384, slice thickness
0.80 mm, 1 acquisition. Magnetic resonance angiography
angle 20°, fov 200 mm, matrix 320 320, slice thickness
0.90 mm, 1 acquisition. Computed tomography images
were acquired using a 64-row multidetector CT scanner
(Aquilon, Toshiba) with a 0.4-mm slice thickness. For
scanning CTA, immediate and delayed scanning sessions
after venous injection of iodized contrast media were per-
formed to obtain arterial and venous information.14,16,22
Data Processing
As preparation for creation of the entire 3D computer
graphics model consisting of all structures required for
simulation of MVD surgery, all DICOM-formatted im-
age data were processed on image-analysis software
2 J Neurosurg / June 29, 2012
M. Oishi et al.
(ZedView, LEXI Co., Inc.). Individual structures were ex-
tracted from the appropriate data set: the brainstem, cer-
ebellum, tentorium, and cranial nerves were from MR-
CISS data; the skull from CT data; and arteries and veins
from MRA and CTA data. First, we created a mask of
the images to extract the targeted structure by determin-
ing the extent of interest and adjusting the window level.
Continuous voxels presenting a similar image-intensity
or density in the mask were automatically segmented.
Additionally, we aligned the structure to deal with unclear
margins by manual outlining. The extracted data were
converted to 3D models formed as triangular polygons by
the surface image-rendering technique. The coordinates
the individual structural margins on the multidirection-
ally reconstructed planes. Finally, we completed creation
of each structural 3D model of the brainstem, cerebellum,
tentorium, cranial nerves, vessels (arteries and veins), and
the skull, and output these data individually in the stan-
dard triangulated language (STL) format.
Interactive Virtual Simulation on a High-Quality 3D
Computer Graphics Model
To complete creation of a 3D computer graphics
model and to perform interactive virtual simulation us-
ing this model, we employed a commercially available
CAD system on a personal computer (Fig. 1). This system
was characterized by unique design software (FreeForm
Modeling v.10.2, Sensable Technologies, Inc.), which al-
lowed us to freely modify the created 3D model as re-
quired by employing intelligible 3D tools for cutting, re-
moving, and deforming, and a haptic device (PHANTOM,
SensAble Technologies, Inc.), which provided a true 3D
interface with force feedback that enabled operators to
use their sense of touch to manipulate the virtual mod-
els just like real models. The computer we used (HP Z
400 Workstation, Hewlett-Packard) had an Intel Xeon
3-GHz processor with 8 GB of main memory, equipped
with an NVIDIA Quadro 4000 graphics card, and the
operating system was Windows XP (64-bit edition). All
STL-formatted data were individually imported onto the
same working space and displayed in different colors to
indicate the different structures. First we observed the
entire 3D computer graphics model from various angles
by rotating it freely and sequentially making structures
translucent or invisible to understand anatomical relation-
ships among structures in the targeted region. Then we
performed interactive virtual simulation on the model to
predict the surgical perspective by imitating craniotomy
and retraction of the cerebellum (Video 1).
VIDEO 1. Case 3. Video clip showing interactive virtual
simulation on the 3D computer graphics model in a patient with
right trigeminal neuralgia. The 3D computer graphics model
was created from MRI, MRA, CT, and CTA data and consists
of a brainstem, cranial nerves, and arteries. It identified the
neurovascular compression by the SCA with indentation on the
trigeminal nerve (CN V). On the complete 3D computer graphics
model, which includes veins, tentorium, and cranium, presurgi-
cal interactive virtual simulation is performed with surgical
positioning (the right side is the top). The virtual craniotomy is
performed with a drilling tool, and then, to allow visualization of
the CPA region, the cerebellum is retracted by a virtual retractor
Interactive virtual simulation for microvascular decompression
Fig. 1. A: FreeForm modeling system (Sensable Technologies,
Inc.) characterized by unique 3D design software and a haptic device
(PHANTOM, Sensable Technologies, Inc.). B: In combination with a
3D display, interactive virtual simulation is performed under 3D visual-
ization through the 3D glasses.
(translucent orange ball). The operative perspective predicted
by interactive virtual simulation shows detailed microsurgical
anatomy including the form of the trigeminal nerve compression
by the SCA and also petrosal veins passing through the surgical
field. PVs = petrosal veins. Click here to view with Windows
Media Player. Click here to view with Quicktime.
Comparison Between the Interactive Virtual Simulation
and Operative Findings
To determine the diagnostic value of 3D computer
graphics models for neurovascular compression, we com-
pared the diagnoses determined by 3D model observa-
neurosurgeons reviewing 2D MRI. Both diagnoses were
- pression nor deformation, on axial 2D images, whereas
traoperatively.
All of the participating surgeons knew the simula-
- (Fig. 4). In 3 hemifacial spasm patients (Cases 12, 15, and
tive virtual simulation for performing entire procedures
during MVD surgery was evaluated in individual patients
J Neurosurg / June 29, 2012
based on surgeon consensus, as follows: The utility was
considered “prominent” if the surgeon found that it pro-
vided information that was important for accomplishing
-
able using only 2D images. It was considered “support-
ive” if the surgeon found that it provided information
of some minor structures, and thereby could enhance sur-
applied if the surgeon found that it might be misleading
or confusing with respect to the actual MVD procedure.
Patient characteristics, simulation results, and sur-
computer graphics model with excellent quality was cre-
ated to evaluate surgical anatomy, and interactive virtual
simulation was successfully performed, allowing predic-
tion of the surgical perspective via the lateral suboccipital
route. The time required for a whole process of creating a
3D computer graphics model was 1–2 hours, and interac-
tive virtual simulation was usually performed for approx-
imately 1 hour. In patients with trigeminal neuralgia (Fig.
2), simulation imitating the infratentorial supracerebellar
route predicted the surgical perspective surrounded by
the tentorium, pyramidal wall and retracted cerebellum,
and exposed the form of the trigeminal nerve that was
distorted, indented, or constricted by vascular compres-
sion or sometimes adhesion. The coursing patterns of
bridging veins such as the petrosal vein or transverse pon-
tine vein and bony protrusion above the internal acoustic
canal (the suprameatal process) in front of the targeted
site were useful information for understanding the work-
ing space and the appropriate trajectory for surgical ma-
nipulation. In patients with hemifacial spasm (Fig. 3),
route predicted the surgical perspective surrounded by
the pyramidal wall and retracted cerebellum and allowed
survey of the root exit zone of the facial nerve from below
The position of the VA, which is occasionally involved
in neurovascular compression, and the bony shape of the
jugular tubercle were useful information for considering
the space and the optimal means of accomplishing MVD.
Diagnostic Value of a 3D Computer Graphic Model for
Neurovascular Compression
-
lar compression by observing 3D computer graphics mod-
in 19 of 26 cases) achieved by reviewing only 2D images
(p < 0.05). In 3 trigeminal neuralgia patients (Cases 5–7),
the trigeminal nerve looked normal, with neither com-
distortion of the nerve due to adhesion of surrounding tis-
sues, especially veins, was suspected based on 3D models
25), one of the multiple vessels contributing to neurovas-
cular compression could not be determined on 2D images
3
4
Determination of Neurovascular Compression Utility of Interactive Virtual Simulation
Case Age (yrs),
No. Sex Dx Side On 2D Images On 3D Models During Surgery Operative Procedures Overall Utility Outcome
1 53, M TN lt SCA SCA SCA as planned prominent protruding SMP, abundant PVs, & running excellent
course of SCA
5 68, F TN rt SCA, AICA SCA, AICA SCA, AICA as planned prominent extra-large PVs & running course of SCA excellent
6 33, M TN rt SCA, PV SCA, PV SCA, PV as planned prominent protruding SMP excellent
7 60, F TN lt SCA, PV SCA, PV SCA, PV as planned prominent complex PVs excellent
2 74, M TN rt not determined PV, adhesion PV, adhesion as planned prominent nerve deformity excellent
3 54, F TN rt not determined SCA, AICA, SCA, AICA, adhe- as planned prominent nerve deformity excellent
(adhesion?) sion
4 59, M TN rt not determined SCA, TPV, PV, SCA, TPV, PV, ad- as planned prominent complex NVC, nerve deformity, & complex excellent
(adhesion?) hesion PVs
8 73, M TN rt SCA SCA SCA as planned supportive excellent
9 45, M TN rt SCA SCA SCA as planned supportive good
10 51, M TN lt SCA, PV SCA, PV SCA, PV as planned supportive excellent
11 58, F TN lt SCA AICA AICA restricted by perforators supportive excellent
12 45, F HFS rt AICA, VA assist AICA, PICA, VA AICA, PICA, VA as planned prominent complex NVC & relationship w/ VA excellent
assist assist
13 26, F HFS rt AICA AICA AICA as planned prominent unusual compression pattern excellent
14 68, F HFS rt AICA AICA AICA as planned prominent unusual compression pattern excellent
15 60, F HFS lt AICA, VA assist AICA, PICA, VA AICA, PICA, VA as planned prominent complex NVC & relationship w/ VA good
assist assist
16 51, F HFS rt AICA AICA AICA as planned prominent overhanging JT excellent
17 45, F HFS lt AICA, VA assist AICA, VA assist AICA, VA assist as planned prominent overhanging JT excellent
18 29, F HFS lt AICA AICA AICA as planned supportive excellent
19 42, F HFS lt PICA PICA PICA as planned supportive excellent
20 51, M HFS rt AICA, VA assist AICA, VA assist AICA, VA assist as planned supportive excellent
21 43, M HFS lt AICA, VA assist AICA, VA assist AICA, VA assist as planned supportive excellent
22 82, F HFS lt AICA, VA assist AICA, VA assist AICA, VA assist as planned supportive excellent
23 68, M HFS rt PICA PICA PICA as planned supportive excellent
24 62, F HFS rt PICA PICA PICA as planned supportive excellent
25 61, M HFS lt AICA br 1, VA AICA br 1, VA AICA br 1, br 2, VA additional procedure for 1 supportive excellent
assist assist assist more branch
26 34, F HFS lt AICA,VA assist AICA,VA assist AICA, VA assist restricted by rich perforators supportive excellent

* Br = branch; HFS = hemifacial spasm; JT = jugular tubercle; NVC = neurovascular compression; PV = petrosal vein; SMP = suprameatal process; TN = trigeminal neuralgia; TPV = transverse pontine
vein.

J Neurosurg / June 29, 2012

M. Oishi et al.
Interactive virtual simulation for microvascular decompression

Fig. 2. Case 8: right trigeminal neuralgia. A–F: (yellow arrows)

to be the offending vessel compressing the trigeminal nerve and the 3D computer graphics model (C) directly visualizing the
neurovascular compression with indentation by the SCA on the trigeminal nerve (black arrow). Skull (D) and veins (E), especially
petrosal veins (white arrow)—extracted from CT data—are integrated, and then the 3D computer graphics model for the simula-
tion is completed (F). G–I: Performing interactive virtual simulation with surgical positioning (the right side is the top in each
photo). The craniotomy location is determined on the translucent skull image (G) and the virtual craniotomy is performed on the
3D computer graphics model with a drilling tool (H). To view the CPA region, the cerebellum is retracted by a virtual retractor
(translucent orange ball) (I). J and K: The operative perspective predicted by interactive virtual simulation (J) shows detailed
microsurgical anatomy, including the form of the trigeminal nerve compression by the SCA and also petrosal veins passing
through the surgical field and is completely concordant with the intraoperative microscopic view (K). cbll = cerebellum; PV =
petrosal vein; pyr = pyramis; tent = tentorium; V = CN V (trigeminal nerve); VIII = CN VIII (acoustic nerve).

J Neurosurg / June 29, 2012 5

 M. Oishi et al.

Fig. 3. Case 26: left hemifacial spasm. A: Sequential CISS (upper) and TOF (lower -
ship of the elongated VA and AICA at the root exit zone of the facial nerve (VII) (yellow arrows). B: The 3D computer graphics
model directly visualizes the AICA as running along the VA, which is shown as partially translucent, and compressing the root
exit zone of the facial nerve. C and D: The predicted perspective on interactive virtual simulation (C), in which the cerebellum
is retracted to imitate the infrafloccular approach, shows that the AICA compresses the root exit zone of the facial nerve (black
arrow) with a contribution from the VA in the microsurgical window, and is confirmed to be concordant with the intraoperative
microscopic view (D). Flo = cerebellar flocculus; VI = CN VI (abducens nerve); VII = CN VII (facial nerve); IX-X = CN IX and CN
X (glossopharyngeal and vagus nerves).

because of the complicated form of the neurovascular
complex entangled with multiple vessels. The 3D model
and 15) but only partially in the other (Case 25).
Overall Utility of Interactive Virtual Simulation Compared
With Operative Findings
The overall utility of interactive virtual simulation was
in 10 (38%), and “supportive” in 3 (12%). There were no
cases in which it was rated as worthless. It was evaluated
as prominent and supportive in 7 (64%) and 4 (36%) of 11
trigeminal neuralgia cases and in 6 (40%) and 9 (60%) of
15 hemifacial spasm cases, respectively, but the difference
-
ation of prominent were related to the atypical or complex
form of the neurovascular compression pattern in 7 (27%)
of 26 patients and structural restriction of the surgical win-
dow with abundant bridging veins or unexpectedly over-
hanging bony structures in 7 (27%) of 26 patients (Fig.
5). Furthermore, the courses of offending vessels requir-
ing transposition without excessive tension, which would
cause kinking, were visualized in 4 (15%) of 26 patients.
Most notably, atypical forms of neurovascular compression
involving veins in 3 trigeminal neuralgia patients (Cases
5–7) and vascular compression at the unusual site of the
root exit zone in 2 hemifacial spasm patients (Cases 13 and
models. Complex neurovascular entanglements involving
multiple vessels in 2 hemifacial spasm patients (Cases 12
and 15) were also correctly ascertained on the 3D models.
The MVD procedures were performed in essentially the
same manner as planned by interactive virtual simulation
in 23 (88%) of 26 patients. The main factors preventing the
planned procedure were restriction due to unpredictably
abundant or short perforators from the targeted vessels.
Surgical Outcomes
All but one of the patients were symptom free with-
out neurological complications. The one exception was a
pain reduction but occasionally requires carbamazepine
treatment.
Herein, we have described our novel surgical simu-

6 J Neurosurg / June 29, 2012

Interactive virtual simulation for microvascular decompression

Fig. 4. Case 4: trigeminal neuralgia with atypical pattern of neurovascular compression. A:

showing multiple vessels (arrows) around the trigeminal nerve with no marked findings of compression. B: The 3D computer

adhering to the nerve. C and D: The predicted perspective on interactive virtual simulation (C) shows that the trigeminal nerve
may adhere to the petrosal vein (PV2) and the transverse pontine vein (TPV) and is confirmed to be concordant with the intra-
operative microscopic view (D). E:
coagulation of the transverse pontine vein are performed, and postoperative pain relief is thereby obtained.

lation method, termed interactive virtual simulation, us-
ing a 3D computer graphics model for MVD surgery in
patients with neurovascular compression syndromes. The
3D computer graphics model constructed with triangular
details of the complex anatomical structures in CPA re-
gions. The unique CAD software for modifying these 3D
computer graphics models provides an extremely realistic
environment for simulating MVD surgery by using vari-
ous 3D design functions to imitate surgical manipulations
such as drilling into bone or retracting structures, as well
as haptic feedback providing tactile and kinesthetic sen-
system can run on a personal computer without any large-
scale equipment and thus can be easily implemented in
a neurosurgical department. The software is easy to use
and requires no advanced technical knowledge. We pre-
viously reported the utility of virtual endoscopy analy-
sis of CISS data obtained by MRI, which allowed us to
review 3D anatomical images from a “virtual camera”
set in the cistern.22,23,27 The 3D views obtained by virtual
endoscopy were almost identical to those seen under the
operative microscope, but this analysis permitted only
observation from a viewpoint like that of “sightseeing”—
that is, without interaction. To predict the microsurgical
views in MVD surgery more realistically, positional al-
terations after craniotomy, CSF leakage, and retraction
of the cerebellum during actual surgery should be taken
into consideration.8 In this respect, the present simula-
tion enabled us to imitate surgical procedures and create
realistic surgical perspectives, allowing for the presurgi-
cal understanding of microsurgical anatomy, essential for
perspectives on interactive virtual simulation showed ex-
cellent consistency with the intraoperative microscopic
views in individual patients.

J Neurosurg / June 29, 2012 7

 M. Oishi et al.

Fig. 5. Case 7 (A–F) and Case 17 (G–K): representative interactive virtual simulations demonstrating restricted surgical
windows. A–F: Case 7—left trigeminal neuralgia. The CISS images (A and B) show arterial compression (yellow arrow) of
the trigeminal nerve and multiple large bridging veins in the surgical field (white arrow). On the 3D computer graphics model
(C), anatomical relationships are directly determined. Interactive virtual simulation predicts the restricted operative window sur-
rounded by the large transverse pontine vein and petrosal vein, suprameatal process (SMP), tentorium, and cerebellum (D), and
during the actual operation it was necessary to manipulate structures in this predicted restricted surgical window (E and F). The
SCA behind the nerve in the window (D) is identified based on interactive virtual simulation knowledge (F). G–K: Case 17—left
hemifacial spasm. The CISS image (G) identifies a vessel passing by the root exit zone (yellow arrow) and the CT image (H)
shows protrusion of the jugular tubercle (JT, white arrow). The 3D computer graphics model shows neurovascular compression at
the root exit zone by the AICA from the common trunk (I), and interactive virtual simulation predicts a restricted operative window
due to marked protrusion of the jugular tubercle (J). Manipulation within this restricted window was necessary during the actual

8 J Neurosurg / June 29, 2012

Interactive virtual simulation for microvascular decompression
Magnetic resonance imaging studies combining heav-
ily T2-weighted imaging and MRA sequences have shown
good sensitivity in diagnosing neurovascular compression
syndromes and identifying the offending vessel.2,3,5,28,29 This
is because heavily T2-weighted imaging can clearly iden-
tify delicate nerves and vessels in the cistern with excellent
the arteries. Since the advent of 3D image-analyzing tech-
niques, various authors have reported the diagnostic value
of reviewing 3D rather than only 2D images for the pur-
pose of identifying the offending vessels.1,4,7,9,12,17–19,22–25,27
identifying the cause of neurovascular compression be-
tween reviewing 2D images and viewing the 3D computer
graphics model. Especially in trigeminal neuralgia patients
in whom the offending artery was unclear on 2D images,
venous compression or distortion of the trigeminal nerve
was directly visualized on our 3D computer graphics mod-
el and was critical for determining surgical indications. In
patients with hemifacial spasm, the offending artery was
occasionally in contact with the root exit zone of the facial
nerve as part of a complex combination of several vessels
such as the AICA, PICA, VA, and their branches. Direct
and careful inspection of the root exit zone on the 3D com-
- endoscopy analysis27 but was important for creating a suf-
plex vascular structures.
The utility of our system in simulating MVD surgery
is based not on identifying the offending vessel, and there-
by diagnosing neurovascular compression syndrome, but
rather on precisely evaluating the 3D anatomy surrounding
study, surgeons assessed the simulation system as showing
“prominent” utility for carrying out entire surgical proce-
dures in 50% of cases based on obtaining 3D anatomical
information that could not be clearly ascertained by re-
viewing only 2D images. Of course, in cases in which it
it was assessed as “supportive,” the surgeons appeared to
after simulation. Thus, we believe that interactive virtual
simulation has high utility for helping surgeons to perform
emphasized the role of a “déjà-vu effect” produced by elab-
orate presurgical simulation on 3D imaging in contribut-
13,15,20,21,26
In this respect, MVD surgery is optimal for 3D simula-
tion because it requires extremely precise knowledge of the
structural relationships in the CPA cistern, which are char-
acterized by marked individual variation. The space for
MVD surgery via retrosigmoid craniotomy is surrounded
by the retracted cerebellum, tentorium, and pyramidal wall
and is usually too narrow and too deep for optimal access.8
The perspective allowing observation of the neurovascular
compression site is sometimes restricted by various obsta-
cles, including large vessels like the petrosal vein or VA,
other nerves, and abnormal bony protrusions, depending
on individual anatomical variations. In our series, surgeons
gave the assessment of “prominent” in trigeminal neural-
gia patients more often than in hemifacial spasm patients,
because the course and thickness of petrosal veins in front
J Neurosurg / June 29, 2012
to deal with the offending vessel, not only the form of the
neurovascular compression site but also the entire path of
the relevant vessel from proximal to distal to the neuro-
vascular compression site is important. When the vessel
has a meandering or looping course, knowledge of the en-
tire path, even blinded portions, of the vessel facilitates the
avoidance of vessel kinking when performing MVD. Thus,
interactive virtual simulation contributed to selecting the
best surgical trajectory to the neurovascular compression
site among various structures and planning the appropriate
MVD procedure in individual patients.
The quality of the 3D computer graphics model for
visualizing nerves and vessels in detail depends on the
resolution of the original images and the thickness of the
structures themselves. We used advanced 3-T MRI and
multidetector CT systems to collect appropriate data for
3D image analysis. Then, the individual structures of the
brainstem, cerebellum, cranial nerves, tentorium, skull, ar-
-
propriate original images and reconstructed by a surface
image-rendering technique. This process was performed
image by image with careful attention; it was a labor-inten-
sive task in comparison with volume rendering and virtual
With the present method, it was possible to combine clear-
ly depicted and delicate venous structures obtained from
CTA data after an additional analyzing process for raw
data obtained by multidetector CT.14,16 Previous reports
3D imaging simulation.4,22 The quality of our 3D computer
graphics model consisting of structures based on various
-
spectives more realistically than in previous studies.
-
active virtual simulation using a 3D computer graphics
model in educating medical students and training young
neurosurgeons. The no-risk environment of interactive vir-
tual simulation using 3D computer graphics models can
enhance microsurgical senses and skills through harmless
repetition. Other authors have also discussed the role of
interactive simulation using advanced 3D computer tech-
nologies in neurosurgical training as a highly effective
means of bridging the substantial gap between textbooks
and actual operative practices.13,15,26 The high level of real-
ity of our simulation system provided by the high quality of
the 3D computer graphics model and the haptic feedback
-
ing intuitive comprehension of 3D anatomy and a realistic
experience of microscopic neurosurgery—and with a sys-
tem that can be run on a personal computer. Furthermore,
the original data are obtained from a living person, not a
cadaver. We also believe that interactive simulation using
a 3D computer graphics model has potential as an alterna-
tive or even advanced tool for education and training in
surgical skills. For the advanced use of interactive virtual
simulation, a collaborative virtual reality environment,13
which can display the real-time data on a large screen by
the projection system in front of a large audience while the
surgeon or trainee is actually performing interactive vir-
9

tual simulation, should be useful for discussing or learning
neurosurgical approaches and strategies.
Our novel presurgical method of interactive virtual
simulation, which is characterized by creating a high-
quality 3D computer graphics model from MRI and CT
data and simulating surgical procedures on the unique
CAD software with manipulation of a haptic device, pro-
vided a realistic environment for performing virtual simu-
lations prior to microvascular decompression surgery and
enabled us to ascertain detailed and variable microsurgi-
cal anatomy and predict microscopic perspectives during
surgery. The no-risk environment and the extreme reality
of our system can enhance microsurgical senses and skills
through harmless repetition. We also believe that our sys-
- 26:159–162, 1967
dents and training young neurosurgeons.
The authors report no conflict of interest concerning the mate-
rials or methods used in this study or the findings specified in this
paper.
Author contributions to the study and manuscript preparation
include the following. Conception and design: Oishi. Acquisition
of data: Oishi, Fukuda. Analysis and interpretation of data: Oishi.
Drafting the article: Oishi. Critically revising the article: all authors.
Reviewed submitted version of manuscript: all authors. Approved
the final version of the manuscript on behalf of all authors: Oishi.
Administrative/technical/material support: Hiraishi, Yajima, Sato.
Study supervision: Fukuda, Fujii.
The authors thank Go Ishida of the Department of Clinical
Radiology, Shinjinkai Kita-Nihon No-Shinkei Geka Hospital;
Yoshiyuki Noto of the Department of Radiology, Niigata University
Hospital; and Kenzo Yoshida and Machiko Takahashi of 3D
Solution at Toyotsu Machinery Corp. for their technical assistance.
1. Akimoto H, Nagaoka T, Nariai T, Takada Y, Ohno K, Yoshino
N: Preoperative evaluation of neurovascular compression in
patients with trigeminal neuralgia by use of three-dimension-
al reconstruction from two types of high-resolution magnetic
resonance imaging. 51:956–962, 2002
2. Anderson VC, Berryhill PC, Sandquist MA, Ciaverella DP,
Nesbit GM, Burchiel KJ: High-resolution three-dimensional
magnetic resonance angiography and three-dimensional
spoiled gradient-recalled imaging in the evaluation of neuro-
vascular compression in patients with trigeminal neuralgia:
a double-blind pilot study. 58:666–673, 2006
3. Boecher-Schwarz HG, Bruehl K, Kessel G, Guenthner M,
the diagnosis of neurovascular compression in patients with
- 68 (1
ings. 40:88–95, 1998
4. Du ZY, Gao X, Zhang XL, Wang ZQ, Tang WJ: Preoperative
evaluation of neurovascular relationships for microvascular de-
compression in the cerebellopontine angle in a virtual reality
environment. Technical note. 479–485, 2010
5. Fukuda H, Ishikawa M, Okumura R: Demonstration of neu-
10
M. Oishi et al.
rovascular compression in trigeminal neuralgia and hemifa-
cial spasm with magnetic resonance imaging: comparison
59:93–100, 2003
6. Gardner WJ: Concerning the mechanism of trigeminal neu-
ralgia and hemifacial spasm. 19:947–958, 1967
7. Hastreiter P, Naraghi R, Tomandl B, Bonk A, Fahlbusch R:
Analysis and 3-dimensional visualization of neurovascular
compression syndromes. 10:1369–1379, 2003
8. Hitotsumatsu T, Matsushima T, Inoue T: Microvascular de-
compression for treatment of trigeminal neuralgia, hemifacial
spasm, and glossopharyngeal neuralgia: three surgical ap-
proach variations: technical note. 1436–
1443, 2003
9. Ishimori T, Nakano S, Kagawa M, Yokoe K, Togami T,
Asakura H, et al: Virtual endoscopic images by 3D FASE
cisternography for neurovascular compression.
2:145–149, 2003
10. Jannetta PJ: Arterial compression of the trigeminal nerve at
the pons in patients with trigeminal neuralgia.
11. Jannetta PJ, Abbasy M, Maroon JC, Ramos FM, Albin MS:
-
cial spasm. Operative techniques and results in 47 patients. J
47:321–328, 1977
12. Kin T, Oyama H, Kamada K, Aoki S, Ohtomo K, Saito N:
Prediction of surgical view of neurovascular decompression
using interactive computer graphics. 65:121–
129, 2009
13. Kockro RA, Stadie A, Schwandt E, Reisch R, Charalampaki
C, Ng I, et al: A collaborative virtual reality environment for
neurosurgical planning and training. 61 (5
379–391, 2007
14. Lell MM, Anders K, Uder M, Klotz E, Ditt H, Vega-Higuera
F, et al: New techniques in CT angiography.
26 S45–S62, 2006
15. Malone HR, Syed ON, Downes MS, D’Ambrosio AL, Quest
DO, Kaiser MG: Simulation in neurosurgery: a review of
computer-based simulation environments and their surgical
applications. 67:1105–1116, 2010
16. Matsumoto M, Kodama N, Sakuma J, Sato S, Oinuma M,
Konno Y, et al: 3D-CT arteriography and 3D-CT venogra-
phy: the separate demonstration of arterial-phase and venous-
phase on 3D-CT angiography in a single procedure. AJNR
26:635–641, 2005
17. Miller J, Acar F, Hamilton B, Burchiel K: Preoperative visu-
alization of neurovascular anatomy in trigeminal neuralgia. J
108:477–482, 2008
18. Naraghi R, Hastreiter P, Tomandl B, Bonk A, Huk W,
Fahlbusch R: Three-dimensional visualization of neurovascu-
lar relationships in the posterior fossa: technique and clinical
application. 100:1025–1035, 2004
19. Ogiwara M, Shimizu T: Surface rendered three-dimensional
MR imaging for the evaluation of trigeminal neuralgia and
hemifacial spasm. 11:840–844, 2004
20. Oishi M, Fukuda M, Ishida G, Saito A, Hiraishi T, Fujii Y:
Prediction of the microsurgical window for skull-base tumors
by advanced three-dimensional multi-fusion volumetric im-
aging. 51:201–207, 2011
21. Oishi M, Fukuda M, Ishida G, Saito A, Hiraishi T, Fujii Y:
Presurgical simulation with advanced 3-dimensional multifu-
sion volumetric imaging in patients with skull base tumors.
188–199, 2011
22. Oishi M, Fukuda M, Noto Y, Kawaguchi T, Hiraishi T, Fujii
pattern of transverse pontine vein: diagnostic value of three-
dimensional multifusion volumetric imaging.
89:226–233, 2011
23. Oishi M, Fukuda M, Takao T, Ishida G, Sato M, Fujii Y: [The
J Neurosurg / June 29, 2012
Interactive virtual simulation for microvascular decompression

utility of presurgical simulation of microvascular decompres- with 3D CISS MR imaging and MR angiography.
sion by MR virtual endoscopy.] 1087– 228:539–545, 2003
1095, 2007 (Jpn)
24. Satoh T, Onoda K, Date I: Fusion imaging of three-dimen- M, Brückmann H, et al: Visualization of cranial nerves I-XII:
sional magnetic resonance cisternograms and angiograms for value of 3D CISS and T2-weighted FSE sequences.
the assessment of microvascular decompression in patients 10:1061–1067, 2000
with hemifacial spasms. 106:82–89, 2007
25. Satoh T, Onoda K, Date I: Preoperative simulation for micro-
vascular decompression in patients with idiopathic trigemi- Manuscript submitted January 2, 2012.
nal neuralgia: visualization with three-dimensional mag- Accepted May 24, 2012.
netic resonance cisternogram and angiogram fusion imaging. Please include this information when citing this paper: pub-
60:104–114, 2007 lished online June 29, 2012; DOI: 10.3171/2012.5.JNS112334.
26. Stadie AT, Kockro RA, Reisch R, Tropine A, Boor S, Stoeter Supplemental online information:
P, et al: Virtual reality system for planning minimally invasive
neurosurgery. Technical note. 108:382–394, 2008 akamai.com/21492/wm.digitalsource-na-regional/jns11-2334_
27. Takao T, Oishi M, Fukuda M, Ishida G, Sato M, Fujii Y: Three- video.asx (Media Player).
dimensional visualization of neurovascular compression: http://mfile.akamai.com/21488/mov/digitalwbc.download.akamai.
presurgical use of virtual endoscopy created from magnetic com/21492/qt.digitalsource-global/jns11-2334_video.mov
resonance imaging. (1 ONS139– (Quicktime).
ONS146, 2008 Address correspondence to: Makoto Oishi, M.D., Department of
28. Yoshino N, Akimoto H, Yamada I, Nagaoka T, Tetsumura A, Neurosurgery, Brain Research Institute, Niigata University, 1-757
Kurabayashi T, et al: Trigeminal neuralgia: evaluation of neu- Asahimachidori, Chuo-Ku, Niigata 951-8585, Japan. email: mac.
ralgic manifestation and site of neurovascular compression oishi@mac.com.

J Neurosurg / June 29, 2012 11