04.01 - Clinical Neuroendos PDF

CLINICAL NEUROENDOSCOPY
CONTENTS
Preface
Rick Abbott
ix
History of Neuroendoscopy
Rick Abbott
Since the beginnings of medicine, physicians have sought minimally invasive ways to
peer into body cavities. It is only in the last several decades that the promises of endoscopy have begun to be answered. What follows is a brief outline of the development of
endoscopic technology and its application to the nervous system both for diagnostic and
therapeutic procedures.
The Endoscopic Management of Arachnoidal Cysts

Rick Abbott
There is little doubt that most arachnoidal cysts will be managed endoscopically in the
future given the advances we have seen over the last decade in our instrumentation.
Excitement to employ this new technology should be governed by the reality that we
are still learning and that our current success rate is not quite as good as what can be
expected when using microneurosurgery.
Basic Principles and Equipment in Neuroendoscopy

Vit Siomin and Shlomi Constantini
19
Understanding some of the basic principles of endoscopy and awareness of available resources can potentially be of considerable help to experienced neurosurgeons as well as
beginners in selection of the most appropriate tools for different procedures and making
cost-effective choices when browsing through multiple commercial advertisements and
purchasing new equipment. Although numerous advantages in science and industry
have made it possible to offer a wide variety of neuroendoscopes and tools, we believe
the major achievements in this field are yet to occur. This particularly refers to the development of smaller fiberoptic scopes with better image quality and three-dimensional endoscopes and to the invention of more efficient tools for endoscopic tumor removal with
the same degree of safety as in open surgery.
The Anatomy of the Ventricular System

David G. McLone
33
The embryology of the ventricular development of the brain assists in understanding the
final relations between structures forming these cavities. An accurate concept of this
anatomy allows the endoscopist to maneuver within the ventricular system.
VOLUME 15
NUMBER 1 JANUARY 2004
Selecting Patients for Endoscopic Third Ventriculostomy

Harold L. Rekate
39
Endoscopic third ventriculostomy has been used for about 70 years in the treatment of
hydrocephalus but was generally abandoned with the development of valve-regulated
shunts. With improvements in the understanding of the pathophysiology of hydrocephalus and technical equipment improvements for endoscopy, there has been a resurgence of interest in the procedure. Late-onset aqueductal stenosis is the ideal pathologic
condition responding to this treatment, but there are multiple other conditions that are
potentially responsive to internal bypass. All patients in whom the ventricles expand at
the time of shunt failure should be considered as candidates.
Techniques of Endoscopic Third Ventriculostomy

Douglas Brockmeyer
51
Modern techniques of endoscopic third ventriculostomy (ETV) are based on the concept
of establishing a natural conduit for cerebral spinal fluid (CSF) flow through the floor of
the third ventricle. Through the years, a wide variety of techniques have been used as a
means to this end and have included both open and closed approaches. However, the
relatively recent application of endoscopic technology to intraventricular surgery has
allowed neurosurgeons to perform third ventriculostomies in a minimally invasive
fashion. Advances in third ventriculostomy technique have been based on a detailed
understanding of third ventricular anatomy, surgical trajectories, and improved instrumentation. The goal of this article is to discuss these issues in detail and to point out
the relevant risks and known complications associated with them.
Complications of Third Ventriculostomy

Marion L. Walker
61
As experience with ETV grows, the procedure will be performed by an increasing number of neurosurgeons. Although the technique has been greatly refined since its advent
almost a century ago, todays neurosurgeon must never forget that this seemingly simple
procedure holds the potential for a number of devastating complications. Appropriate
training and experience are important to the success of ETV and for avoiding complications. It is imperative that surgeons continue to report their experience with the complications of ETV so that the procedure can continue to be made as safe as possible.
Results of Endoscopic Third Ventriculostomy

Mark R. Iantosca, Walter J. Hader, and James M. Drake
67
Endoscopic third ventriculostomy is emerging as the treatment of choice for aqueductal

stenosis caused by anatomic, inflammatory, and selected neoplastic etiologies. The technique has also proven useful in the pathologic diagnosis and treatment of these conditions. Long-term results of this procedure and comparison to standard shunting
procedures are necessary to define indications for patients with pathologic findings in
the intermediate response groups. Development of new studies for preoperative assessment of cerebrospinal fluid absorptive capacity and quantitative postoperative measures
of ventriculostomy function would be invaluable additions to our ability to assess candidates for this procedure and their eventual outcome. Further study and technical refinements will, no doubt, lead to many more potential uses for these procedures in
the treatment of hydrocephalus and its associated etiologies. The challenge for neurosurgeons will be to define the operative indications and outcomes, while refining techniques
for safely performing these useful procedures.
vi
CONTENTS
Loculated Ventricles and Isolated Compartments in Hydrocephalus: Their

Pathophysiology and the Efficacy of Neuroendoscopic Surgery
Shizuo Oi and Rick Abbott
77
Trapped cerebrospinal fluid spaces can, on occasion, complicate the management of

hydrocephalus and present the surgeon with a treatment dilemma. This condition can
be categorized into one of two types: those arising as a complication of shunting and
those that arise as a complication of an inflammatory process within the ventricles.
Whatever the cause, the result is a significant escalation in the complexity of the management of the patient. Neuroendoscopy is typically viewed as an attractive treatment alternative in such a setting because of its minimalistic and thus seemingly simplistic nature.
We have learned that nothing could be further from the truth. This article reviews the
various entities that can arise in the hydrocephalic patient, how they can be managed
endoscopically, and what sort of result can be expected.
Neuro-oncologic Applications of Endoscopy

Charles Teo and Peter Nakaji
89
Neuro-oncology, in all its aspects, provides an ideal venue for the application of endoscopy. The main obstacle to its use has been neurosurgeons lack of familiarity with the
techniques and their advantages. As the neuro-oncologic surgeon uses the endoscope
more, endoscopy will take its rightful place in the surgeons armamentarium. The advantages of improved visualization of intraventricular pathology, better management
of tumor-related hydrocephalus, less morbid biopsies, and minimally invasive removal
of intraventricular tumors are invaluable adjuncts to traditional tumor management.
Furthermore, endoscopy is the logical next step for surpassing the limitations of traditional microsurgery. Endoscopy is still in its infancy. Rigorous application of the technology is increasingly allowing us to provide our patients the most maximally effective and
minimally invasive surgery possible.
Index
CONTENTS
105
vii
FORTHCOMING ISSUES
April 2004
Traumatic Neurovascular Surgery
J. Paul Elliot, MD, Guest Editor
July 2004
Pain Treatment
Gary Heit, MD, Guest Editor
October 2004
Metastatic Spine Tumors
Meic Schmidt, MD, Guest Editor
RECENT ISSUES
October 2003
Intraventricular Tumors
Andrew T. Parsa, MD, PhD, and
Mitchel S. Berger, MD, Guest Editors
July 2003
Neuroaugmentation for
Chronic Pain
Jaimie M. Henderson, MD, Guest Editor
April 2003
Surgery for Psychiatric Disorders
Ali R. Rezai, MD, Steven A. Rasmussen, MD,
and Benjamin D. Greenberg, MD, PhD
Guest Editors
THE CLINICS ARE NOW AVAILABLE ONLINE!

Access your subscription at
http://www.TheClinics.com
Neurosurg Clin N Am 15 (2004) ix
Preface
Clinical neuroendoscopy
Rick Abbott, MD
Guest Editor
Neuroendoscopy has come of age over the last

decade and is now an invaluable tool to the practice of neurosurgery. This is particularly the case
in practices with large volumes of patients with
hydrocephalus. This issue of the Neurosurgery
Clinics of North America is dedicated to neuroendoscopy with an emphasis on its use in the treatment of hydrocephalus. Readers will nd helpful
articles on instrumentation and anatomy as well
as several articles on various aspects of use for
hydrocephalus and related conditions. With a
look toward the future there is also an article on

the use of the endoscopy in the removal of brain
tumors. By reading this issue, readers will gain
an appreciation of the state-of-the-art of intracranial neuroendoscopy.
Rick Abbott, MD
Inn, Beth Israel Medical Center
170 East End Avenue
New York, NY 10128, USA
E-mail address: rabbott@bethisraelny.org
1042-3680/04/$ - see front matter 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.nec.2003.12.001
Neurosurg Clin N Am 15 (2004) 17
History of neuroendoscopy
Rick Abbott, MD
Clinical Neuroendoscopy, INN, Beth Israel Medical Center, 170 East End Avenue, New York, NY 10128, USA
The insertion of tubes into the body for either

the deliverance of a therapy or diagnostic
purposes has taken place throughout recorded
medicine. The rst documented instance of the use
of a tube to inspect the rectum was in Hippocrates time. Natural light was used to illuminate
the eld [1]. The Babylonians described using
vaginal speculums in 500 AD, with illumination
being provided by ambient light [2]. Abulkaism
(9801037) and Giulio Cesare Aranzi (15301589)
reported on the use of mirrors to reect ambient
light down their endoscopic tubes to allow for
the inspection of deeper body cavities [3]. In 1805,
the Alert Faculty in Vienna heard Philippe
Bozzinis report on an instrument he had developed, which used candle light reected by
a concave mirror for the inspection of the bladder
and rectum [4]. They were not impressed, censoring Bozzini for his inappropriate curiosity and
rejecting the magic lantern.
In 1867, Antonin Desormeaux published a description of an endoscope whose illumination had
been improved by using an alcohol and kerosene
burning candle with a chimney [5]. The light this
candle generated was collected and focused down
the shaft of the scope using a lens. This was to be
the rst successful design for a cystoscope, and it
was presented to the Academy of Medicine in
Paris. Shortly thereafter, there were several reports of therapeutic uses for endoscopes. First,
Bevan reported on successfully using an endoscope to extract foreign material from the
esophagus, and Pantaleoni then reported on using
a scope to inspect the uterus of a women bothered
by postmenopausal bleeding, discovering an intrauterine polyp and cauterizing it with silver
E-mail address: Rabbott@bethisraelny.org
nitrate [4]. In 1870, Kussmaul demonstrated using

a rigid scope to inspect the stomach [5]. The
assistance of a professional sword swallower was
required for this, however.
All these described systems suered from a lack
of magnication. They were simply tubes that
directed illumination down to their distal tip. In
1879, Max Nitze described the rst system that
contained a series of lens [6]. Working with several
opticians, he described a scope with an illumination source to the distal tip, a platinum wire that
glowed when current was conducted through it.
The light produced was then projected through
a prism. This wire produced heat, however,
necessitating a water coolant system. Shortly
thereafter, Edison invented the light bulb, and
Newman then described modifying Nitzes scope
in 1883 by substituting a small light bulb at the
distal tip for the platinum wire [7]. Boisseau du
Rocher was the next to modify the system by
fabricating an outer sheath to contain the
telescope, reporting on this in 1889 [4]. This
modication allowed the surgeon to interchange
dierent scopes during a procedure without
having to renavigate through the body to the
working site.
By the turn of the century, the promise of
endoscopy had been demonstrated, but its acceptance was slowed because of the poor illumination. Indeed, Pantaleonis students remarked on
not being able to appreciate his work, claiming an
inability to see anything with his hysteroscope [4].
The light source evolved during the rst half of
the twentieth century, and by 1950, it consisted of
a tungsten bulb at the distal tip of the scope [8].
Even this was inadequate, providing poor illumination and signicant color distortion. In large
part, these diculties were the result of design
aws inherent in the Nitze endoscope design (ie,
doi:10.1016/S1042-3680(03)00065-2
R. Abbott / Neurosurg Clin N Am 15 (2004) 17
a telescope containing a series of lens housed in an

air-lled tube) [9]. The conduction of light in an
endoscope is a function of the refractory index of
the conducting medium, in this case, air. The refractory index is dened as the dierence in the
speed of light conduction in a vacuum and the
medium in question. As it turns out, the refractory
index for glass is 1.5 times greater than that for
air. Harold Hopkins, a British optical physicist,
used this observation to construct a new endoscope designed to improve on light conduction.
His scope interchanged glass for air and vice
versa, resulting in a series of air lens housed in
a glass tube, or a series of glass rods. Not only did
this result in an increase in the systems refractory
index but also in an increase in the scopes eld of
view because of a decrease in spherical aberration
at the perimeter of the lens (thus increasing the
lenss light gathering capacity). The net gain in
light transmission of Hopkins scope design over
Nitzes design was ninefold. Hopkins also used
coating on the lens to minimize refraction of light
at the lens interface with the glass rods and
calculated the layout of the lens to eliminate
ghosting and chromatic distortion in addition to
further increasing light transmission. With these
improvements, serious attention could be given to
using scopes within the human body, such as the
brains ventricles, where ambient light could not
be directed.
An additional problem with the Nitze design
was the heat generated by its illumination source
housed at the distal tip of the scope. The answer
seemed to be using a light generator outside of
the body and conducting the light down to the
distal tip. The rst report of use of an external
light source was by Fourestier, Bladu, and
Valmier in 1952, when a scope design was
described where light was transmitted down
a quartz rod from the scopes external light
source [8]. This transmitted enough light to make
endoscopic photography possible. An improvement on this shortly followed based on work
done by Heinrich Lamm in 1932 [4]. At that
time, Lamm had demonstrated that light could
be conducted through a bundle of glass bers. In
1954, Harold Hopkins picked up on this idea
and with a colleague, N.S. Kapany, discussed
applying this nding to design a more eective
light conducting system [10]. They described two
types of ber bundles, so-called incoherent and
coherent bundles. Incoherent bundles referred
to bundles of glass bers whose orientation with
neighbors was chaotic or without order. This
type of bundle could be used to transmit light

for illumination. Conversely, a coherent bundle
maintained the orientation of bers through its
length so that any given ber would maintain its
exact orientation to it is neighbors throughout
the length of the bundle. This type of bundle
could be used to transmit an image from one end
of the bundle to the other. Hopkins and Kapany
observed that this type of bundle could be used
in the design of a exible endoscope whose
shaft could be bent to a degree and still allow for
the conduction of an image from one end of the
scope to the other. In the same issue of Nature,
there was a paper describing a technique for
coating the bers to minimize loss of light and
degradation of image quality [5].
Hirschowitz visited Hopkins laboratory shortly after the publication of his paper in 1954 and
was impressed with the promise of the observations made by Hopkins and Kapany [5]. He
returned to the University of Michigan, where he
collaborated with Wilbur Peters and Lawrence
Curtiss to develop a beroptic gastroscope. By the
end of 1956, Curtiss had developed a glass coating
for the optical bers that was permanently
adherent and capable of conducting an image
for over a meter. In January 1957, they completed
construction of their rst beroptic gastroscope.
The next month, it was used to inspect a gastric
ulcer in the wife of a dental student. Within 3
years, a commercial beroptic endoscope, American Cystoscope Makers No. 4990 (American
Cystoscope Makers, New York, NY), became
available. By 1962, Hirschowitz was able to publish a series of 500 gastroesophageal endoscopies.
In 1963, Guiot described an endoscope developed for intracranial work that had a powerful
external light source whose illumination was
conducted via a quartz rod to the scopes distal
tip [9]. This allowed for color photography of the
ventricle. That same year Scar described improving his ventriculoscope by substituting a ber lighting system with an external light source
for the previously used incandescent bulb at the
scopes distal tip [9]. In 1983, the Welch Allyn
Company (Skaneateles Falls, NY) released the
rst endoscope with the charged couple device,
allowing for conduction of a high-quality image
from the scope to a television screen [5]. Other
companies quickly followed with miniaturization
of the scopes, resulting in an improved ability to
document and teach. The ability to introduce rigid
and exible endoscopes into the body and to
obtain good-quality images was thus established.
As the use of endoscopes for diagnostic

purposes increased, not surprisingly, so too did
the desire to render therapy. In 1887, Felix M.
Oberlander described the rst scope designed to
treat postgonorrheal strictures [6]. The scope
allowed for direct visualization of the urethra
and contained an instrument channel allowing for
the introduction of various knives for cutting the
strictures under direct vision. In 1910, LEspinasse
reported to a local medical society in Chicago on
the use a cystoscope to fulgurate the choroid
plexus in two infants with hydrocephalus [11].
One child died immediately after surgery, but the
other lived for 5 years; thus was reported the
rst therapeutic neuroendoscopic case. Erich
Wossidlo, a German urologist, reported on using
a galvanocaustic hook to treat urethral strictures
[6]. In 1937, Ruddock introduced a scope that
contained an electrocautery unit and biopsy
forceps as well as ancillary biopsy instruments
for use in the peritoneal cavity [12]. In 1939,
Crafoord and Frenckner described using sclerotherapy to treat esophageal varices [4]. In the
1970s, lasers were introduced into the armamentarium of medicine. In 1973, Nath and his
colleagues [13] experimentally used an Nd:YAG
laser ber with an endoscope to demonstrate its
feasibility. Two years later, Fruhmorgan et al [14]
reported on its use on a patient.
Although not as aggressive as some specialists,
neurosurgeons have been actively engaged in the
development of endoscopic applications for the
central nervous system. As mentioned earlier,
LEspinasse rst described use of the endoscope
in the central nervous system in 1910. In 1922,
Dandy [15] reported performing an endoscopic
choroid plexectomy after a previously reported
experience in performing open choroid plexectomy on four patients. The attempt to perform the
endoscopic choroid plexectomy was unsuccessful,
and he did not attempt any further such cases.
The next year, Fay and Grant [16] reported on
successfully photographing the interior of the
ventricles of a hydrocephalic child. The fact that
a 40-second exposure was required speaks to the
rather poor illumination available at that time. In
the same year, Mixter [17] performed the rst
successful endoscopic third ventriculostomy using
a cystoscope. This did not gain acceptance, presumably because of the poor visualization. In
1932, Dandy [18] reported on using a cystoscope
to remove the choroid plexus. His results were
similar to those experienced when he did the
surgery via a formal craniotomy. Shortly there-
after, Putnam [19] reported on the eectiveness of

simply cauterizing the choroid plexus using a scope
of his design. Scar [20] then wrote a paper on
his experience in developing a ventriculoscope
that allowed cauterization of the choroid plexus.
He went on to develop a scope that contained
a movable electrode for cauterization of the
plexus. In 1943, Putnam [21] reported on performing endoscopic choroid plexectomy on 42 patients. There were 10 (25%) perioperative deaths.
Fifteen of the patients failed to respond to the
treatment, but 17 experienced success in relief
of their intracranial hypertension. In 1970, Scar
[22] reviewed all available series of endoscopic
choroid plexus cauterization for the treatment
of hydrocephalus. Of 95 patients so treated, 14
(15%) had died, whereas 52 (60%) had initial
successful results. Scar also looked at his
patients (39 of the 95 patients) at more than 5
years after their surgery and found that 7 had died
of causes unrelated to their hydrocephalus and the
rest required no further treatment. More recently,
Bucholz and Pittman [23] reported on using the
Nd:YAG laser to cauterize the choroid plexus of
a shunted infant with ascites, eectively decreasing cerebrospinal uid (CSF) production by
50%. Pople and Grith [24] commented on their
20-year experience treating 156 individuals with
choroid plexus cauterization, nding a 35% longterm success rate.
As stated earlier, Guiot reported on an
endoscope with a powerful external light source
sucient to allow for color photography of the
ventricle. Guiot reported on using this system
to perform third ventriculostomy. The diculty
with his system was its diameter, 9.1 mm. This
prevented a wider use of his scope, and Guiot
ultimately ceased performing third ventriculostomy with his instrument. Vries [25] described his
experience with ve hydrocephalic patients on
whom he performed endoscopic third ventriculostomies in 1978. Although he showed that the
procedure was technically feasible, none of his
patients remained without a shunt over the long
term. Jones et al [26] reported a dierent experience in 1990, describing a 50% success rate in
long-term management of hydrocephalus in 24
patients who underwent a third ventriculostomy.
Modern series have improved on this gure, with
most now reporting 60% to 90% long-term
success with the technique [2733].
Neurosurgeons have used endoscopes for other
indications in the central nervous system. In 1938,
Pool [34] reported on using an endoscope, termed
a myeloscope, to visualize dorsal nerve roots of the

cauda equina. Several years later, he described
using the instrument to view the spinal cord and its
conus. In 1990, the group at Hopital Necker
in Paris described their evolution in the treatment
of suprasellar arachnoidal cysts [35]. From rst
performing stereotactically guided fenestration of
the cyst, they moved to performing the procedure
endoscopically. Others have reported on endoscopically fenestrating other arachnoidal cysts [3641].
Although technically more challenging, the
endoscope can be used to fenestrate other intraventricular cysts such as arise after severe
ventriculitis. In 1986, Powers [42] described two
infants with postinfectious multicystic ventricles
who he managed using a exible endoscope and
argon laser. He was able to fenestrate the cysts
into the ventricles successfully and cure the
lingering infections. In 1992, Zamorano et al
described using a stereotactically guided endoscope to treat cystic ventricles and other entities
[64]. The endoscope has also been used to assist in
the evacuation of intracranial hematomas [4345].
The greatest success has been with chronic
hematomas, but surgeons have described attacking even acute clots [11,46,47].
There are now numerous reports in the
literature of the successful biopsying or removal
of intracranial mass lesions endoscopically. In
1983, Powell et al [48] rst reported on the
removal of a colloid cyst using the endoscope.
There have been many reports since [40,4951,53],
and in 1994, Lewis et al [52] showed that this type
of resection required less surgical time and
postoperative convalescence for their patients.
Fukushima [54] reported on using a exible
endoscope to biopsy tumors in 1978. Many have
reported similar success over the past several
decades [5560]. Endoscopic surgery has also been
used to manage cystic tumors, such as craniopharyngiomas, fenestrating the cysts into ventricles or cisternal spaces [6165]. One particular
successful application for the neuroendoscope
has been to biopsy pineal region tumors when
performing third ventriculostomies to treat the
associated hydrocephalus [66,67]. There has been
one report, however, of secondary seeding of the
endoscopes tract after such a biopsy [68]. As early
as 1979, surgeons reported on using the endoscope
to resect pituitary lesions, and this has become
extremely popular over the last decade [65,6981].
Finally, there are several surgeons investigating
the utility of using an endoscope as an assisting set
of eyes when performing microneurosurgery
[60,61]. This has allowed for a much narrower

surgical corridor and for the surgeon to look
around corners or on the back side of structures,
such as arteries. Fries and Perneczky [11] have
also spoken of improved appreciation of microanatomy not apparent with the microscope.
Other applications are nding their way into
the literature as neurosurgeons gain comfort in
using endoscopic equipment. Jimenez and Barone
[82,83] have reported on using the endoscope to
perform sagittal strip craniectomies in conjunction with molding helmets for the treatment of
scaphocephaly. As early as 1993, neurosurgeons
reported on using endoscopes to perform thoracic
sympathectomies in a minimally invasive fashion
[8489]. In 1994, Schaer [90] reported on resection of an intervertebral disk using an endoscope, and several surgeons have described using
the endoscope to perform various spinal surgeries
since [9195].
It seems clear that the endoscope is a unique
tool and that it has a place in the armamentarium
of the modern neurosurgeon. Its applications will
only broaden as we gain instrumentation and
experience in using this system.
References
[1] Rosin D. History. In: Rosin D, editor. Minimal
access medicine and surgery. Oxford: Radclie
Medical Press; 1993. p. 19.
[2] Gorden A. The history and development of
endoscopic surgery. In: Sutton C, Diamond M,
editors. Endoscopic surgery for gynaecologists.
London: WB Saunders; 1993. p. 37.
[3] Gotz F, Pier A. The history of laparoscopy. In:
Gotz F, et al, editors. Color atlas of laparoscopic
surgery. New York: Thieme; 1993. p. 35.
[4] Lau W, Leow C, Li A. History of endoscopic and
laparoscopic surgery. World J Surg 1997;21:44453.
[5] Hirschowitz B. Development and application of
endoscopy. Gastroenterology 1993;104:33742.
[6] Schultheiss D, Truss M, Jonas U. History of direct
vision internal urethrotomy. Urology 1987;52(4):
72934.
[7] Dameword M. History of the development of
gynecologic endoscopic surgery. In: Azziz R,
Murphy A, editors. Practical manual of operative
laparoscopic and hysteroscopy. New York: SpringerVerlag; 1992. p. 714.
[8] Bordelon B, Hunter J. Endoscopic technology. In:
Green F, Ponsky J, editors. Endoscopic surgery.
Philadelphia: WB Saunders; 1994. p. 617.
[9] Grith H. Endoneurosurgery: endoscopic intracranial surgery. In: Symon L, editor. Advances and
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
technical standards in neurosurgery. New York:

Springer-Verlag; 1986. p. 224.
Hopkins H, Kapany N. A exible light transmitting
tube. Nature 1954;173:3941.
Fries G, Perneczky A. Intracranial endoscopy. In:
Symon L, editor. Advances and technical standards
in neurosurgery. New York: Springer-Verlag; 1999.
p. 2160.
Haubrich W. History of endoscopy. In: Sivak M,
editor. Gastroenterologic endoscopy. Philadelphia:
WB Saunders; 1987. p. 218.
Nath G, Gorish W, Kiefhaber P. The rst
endoscopy via a ber-optic transmission sytem.
Endoscopy 1973;5:20812.
Fruhmorgan P, et al. The rst endoscopic laser
coagulation in the human GI tract. Endoscopy
1975;7:15660.
Dandy W. Cerebral ventriculostomy. Johns Hopkins Hospital Bull 1922;33:189.
Fay T, Grant F. Ventriculostomy and intraventricular photography in internal hydrocephalus. JAMA
1923;80:4613.
Mixter W. Ventriculoscopy and puncture of oor of
third ventricle. Boston Med Surg J 1923;188:2778.
Dandy W. The brain. In: Lewis D, editor. Practice
of surgery. Hagerstown: WF Prior Company; 1932.
p. 24752.
Putnam T. Treatment of hydrocephalus by endoscopic coagulation of the choroid plexus. Description of a new instrument and preliminary report of
results. N Engl J Med 1934;210:13736.
Scar J. Endoscopic treatment of hydrocephalus.
Description of ventriculoscope and preliminary
report of cases. Arch Neurol Psychiatry 1936;35:
85360.
Putnam T. The surgical treatment of infantile hydrocephalus. Surg Gynecol Obstet 1943;76:17182.
Scar J. The treatment of nonobstructive (communicating) hydrocephalus by endoscopic cauterization of the choroid plexus. J Neurosurg 1970;
33:118.
Bucholz R, Pittman T. Endoscopic coagulation of
the choroid plexus using the Nd: YAG laser: initial
experience and proposal for management. Neurosurgery 1991;28:4216.
Pople I, Grith H. Control of hydrocephalus by
endoscopic choroid plexus coagulationlong-term
results and complications. Eur J Pediatr Surg 1993;
3(Suppl 1):178.
Vries J. An endoscopic technique for third ventriculostomy. Surg Neurol 1978;9:1658.
Jones RF, Stening WA, Brydon M. Endoscopic
third ventriculostomy. Neurosurgery 1990;26(1):
8692.
Choi JU, Kim DS, Kim SH. Endoscopic surgery
for obstructive hydrocephalus. Yonsei Med J 1999;
40(6):6007.
Elbabaa SK, Steinmetz, Ross J, Moon D, Luciano
MG. Endoscopic third ventriculostomy for obstruc-
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
tive hydrocephalus in the pediatric population:

evaluation of outcome. Eur J Pediatr Surg 2001;
11(Suppl 1):S5254.
Jones RF, Kwok BC, Stening WA, Vonau M.
Neuroendoscopic third ventriculostomy. A practical alternative to extracranial shunts in noncommunicating hydrocephalus. Acta Neurochir
Suppl (Wien) 1994;61:7983.
Murshid WR. Endoscopic third ventriculostomy:
towards more indications for the treatment of noncommunicating hydrocephalus. Minim Invasive
Neurosurg 2000;43(2):7582.
Panigrahi M, Vinas FC, Guthikonda M. Endoscopic third ventriculostomy: outcome analysis of
100 consecutive procedures. Neurosurgery 2000;
46(6):15345.
Rieger A, Rainov NG, Brucke M, Marx T, Holz C.
Endoscopic third ventriculostomy is the treatment
of choice for obstructive hydrocephalus due to
pediatric pineal tumors. Minim Invasive Neurosurg
2000;43(2):836.
Tuli S, Alshail E, Drake J. Third ventriculostomy
versus cerebrospinal uid shunt as a rst procedure in
pediatric hydrocephalus. Pediatr Neurosurg 1999;
30(1):115.
Pool J. Direct visualization of dorsal nerve roots of
the cauda equina by means of a myeloscope. Arch
Neurol Psychiatry 1938;39:130812.
Pierre-Kahn A, Capelle L, Brauner R, Sainte-Rose
C, Renier D, Rappaport R, et al. Presentation and
management of suprasellar arachnoid cysts. Review
of 20 cases. J Neurosurg 1990;73:3559.
Santamarta D, Morales F, Sierra JM, de Campos
JM. Arachnoid cysts: entrapped collections of
cerebrospinal uid variably communicating with
the subarachnoid space. Minim Invasive Neurosurg
2001;44(3):12834.
Hopf NJ, Perneczky A. Endoscopic neurosurgery
and endoscope-assisted microneurosurgery for the
treatment of intracranial cysts. Neurosurgery 1998;
43(6):13307.
Paladino J, Rotim K, Heinrich Z. Neuroendoscopic
fenestration of arachnoid cysts. Minim Invasive
Neurosurg 1998;41(3):13740.
Decq P, Brugieres P, Le Guerinel C, Djindjian M,
Keravel Y, Nguyen JP. Percutaneous endoscopic
treatment of suprasellar arachnoid cysts: ventriculocystostomy or ventriculocystocisternostomy?
Technical note. J Neurosurg 1996;84(4):696701.
Hellwig D, Bauer BL, List-Hellwig E. Stereotactic
endoscopic interventions in cystic brain lesions.
Acta Neurochir Suppl (Wien) 1995;64:5963.
Caemaert J, Abdullah J, Calliauw L, Carton D,
Dhooge C, van Coster R. Endoscopic treatment of
suprasellar arachnoid cysts. Acta Neurochir (Wien)
1992;119(14):6873.
Powers SK. Fenestration of intraventricular cysts
using a exible, steerable endoscope and the argon
laser. Neurosurgery 1986;18(5):63741.
[43] Auer LM. Ultrasound stereotaxic endoscopy in

neurosurgery. Acta Neurochir Suppl (Wien) 1992;
54:3441.
[44] Bauer BL, Hellwig D. Minimally invasive endoscopic neurosurgerya survey. Acta Neurochir
Suppl (Wien) 1994;61:112.
[45] Karakhan VB, Khodnevich AA. Endoscopic surgery of traumatic intracranial haemorrhages. Acta
Neurochir Suppl (Wien) 1994;61:8491.
[46] Auer LM, Holzer P, Ascher PW, Heppner F. Endoscopic neurosurgery. Acta Neurochir (Wien) 1988;
90(12):114.
[47] Hellwig D, Bauer BL. Minimally invasive neurosurgery by means of ultrathin endoscopes. Acta
[48] Powell MP, Torrens MJ, Thomson JL, Horgan JG.
Isodense colloid cysts of the third ventricle: a diagnostic and therapeutic problem resolved by
ventriculoscopy. Neurosurgery 1983;13(3):2347.
[49] Rodziewicz GS, Smith MV, Hodge CJ Jr. Endoscopic colloid cyst surgery. Neurosurgery 2000;46(3):
65562.
[50] King WA, Ullman JS, Frazee JG, Post KD,
Bergsneider M. Endoscopic resection of colloid
cysts: surgical considerations using the rigid endoscope. Neurosurgery 1999;44(5):110311.
[51] Deinsberger W, Boker DK, Samii M. Flexible
endoscopes in treatment of colloid cysts of the
third ventricle. Minim Invasive Neurosurg 1994;
37(1):126.
[52] Lewis AI, Crone KR, Taha J, van Loveren HR,
Yeh HS, Tew JM Jr. Surgical resection of third
ventricle colloid cysts. Preliminary results comparing transcallosal microsurgery with endoscopy.
J Neurosurg 1994;81(2):1748.
[53] Heikkinen ER, Heikkinen MI. New diagnostic and
therapeutic tools in stereotaxy. Appl Neurophysiol
1987;50(16):13642.
[54] Fukushima T. Endoscopic biopsy of intraventricular tumors with the use of a ventriculoberscope.
Neurosurgery 1978;2(2):1103.
[55] Greenberg IM. Intraoperative ultrasonography with
a cystoscope for the biopsy of a deep-seated brain
lesion: case study. Neurosurgery 1986;19(1):4958.
[56] Otsuki T, Yoshimoto T, Jokura H, Katakura R.
Stereotactic laser surgery for deep-seated brain
tumors by open-system endoscopy. Stereotact
Funct Neurosurg 1990;5455:4048.
[57] Zamorano L, Chavantes C, Moure F. Endoscopic
stereotactic interventions in the treatment of brain
lesions. Acta Neurochir Suppl (Wien) 1994;61:927.
[58] Cohen AR. Ventriculoscopic surgery. Clin Neurosurg 1994;41:54662.
[59] Jallo GI, Morota N, Abbott R. Introduction of a
second working portal for neuroendoscopy. A technical note. Pediatr Neurosurg 1996;24(2):5660.
[60] Fries G, Perneczky A. Endoscope-assisted brain
surgery: part 2analysis of 380 procedures. Neurosurgery 1998;42(2):22632.
[61] Cheng WY, Chang CS, Shen CC, Wang YC, Sun
MH, Hsieh PP. Endoscope-assisted microsurgery
for treatment of a suprasellar craniopharyngioma
presenting precocious puberty. Pediatr Neurosurg
2001;34(5):24751.
[62] Abdullah J, Caemaert J. Endoscopic management
of craniopharyngiomas: a review of 3 cases. Minim
Invasive Neurosurg 1995;38(2):7984.
[63] Caemaert J, Abdullah J, Calliauw L. Endoscopic
diagnosis and treatment of para- and intra-ventricular cystic lesions. Acta Neurochir Suppl (Wien)
1994;61:6975.
[64] Zamorano L, Chavantes C, Dujovny M, Malik G,
Ausman J. Stereotactic endoscopic interventions in
cystic and intraventricular brain lesions. Acta
[65] Halves E, Bushe KA. Transsphenoidal operation
on craniopharyngiomas with extrasellar extensions.
The advantage of the operating endoscope [proceedings]. Acta Neurochir Suppl (Wien) 1979;28(2):362.
[66] Ferrer E, Santamarta D, Garcia-Fructuoso G,
Caral L, Rumia J. Neuroendoscopic management
of pineal region tumours. Acta Neurochir (Wien)
1997;139(1):1221.
[67] Ellenbogen RG, Moores LE. Endoscopic management of a pineal and suprasellar germinoma with
associated hydrocephalus: technical case report.
Minim Invasive Neurosurg 1997;40(1):136.
[68] Haw C, Steinbok P. Ventriculoscope tract recurrence after endoscopic biopsy of pineal germinoma.
Pediatr Neurosurg 2001;34(4):2157.
[69] Liston SL, Siegel LG, Thienprasit P, Gregory R.
Nasal endoscopes in hypophysectomy. J Neurosurg
1987;66(1):155.
[70] Gamea A, Fathi M, el-Guindy A. The use of the
rigid endoscope in trans-sphenoidal pituitary surgery. J Laryngol Otol 1994;108(1):1922.
[71] Rodziewicz GS, Kelley RT, Kellman RM, Smith
MV. Transnasal endoscopic surgery of the pituitary
gland: technical note. Neurosurgery 1996;39(1):
18999.
[72] Jho HD, Carrau RL. Endoscopic endonasal transsphenoidal surgery: experience with 50 patients.
J Neurosurg 1997;87(1):4451.
[73] Heilman CB, Shucart WA, Rebeiz EE. Endoscopic
sphenoidotomy approach to the sella. Neurosurgery
1997;41(3):6027.
[74] Aust MR, McCarey TV, Atkinson J. Transnasal
endoscopic approach to the sella turcica. Am J
Rhinol 1998;12(4):2837.
[75] Moreland DB, Diaz-Ordaz E, Czajka GA, Zugger
CM. Endoscopic resection of pituitary lesions
through the nostril. Semin Perioper Nurs 1998;
7(3):1939.
[76] Moses RL, Keane WM, Andrews DW, Goel R,
Simeone F. Endoscopic transseptal transsphenoidal
hypophysectomy with three-dimensional intraoperative localization technology. Laryngoscope
1999;109(3):50912.

[77] Shen CC, Wang YC, Hua WS, Chang CS, Sun MH.
Endoscopic endonasal transsphenoidal surgery for
pituitary tumors. Zhonghua Yi Xue Za Zhi (Taipei)
2000;63(4):30110.
[78] Moreland DB, Diaz-Ordaz E, Czajka GA. Endoscopic endonasal hemisphenoidotomy for resection
of pituitary lesions conned to the sella: report of 3
cases and technical note. Minim Invasive Neurosurg
2000;43(2):5761.
[79] Jho HD. Endoscopic pituitary surgery. Pituitary
1999;2(2):13954.
[80] Jarrahy R, Berci G, Shahinian HK. Assessment of
the ecacy of endoscopy in pituitary adenoma
resection. Arch Otolaryngol Head Neck Surg
2000;126(12):148790.
[81] Ogawa T, Matsumoto K, Nakashima T, Okano M,
Ono Y, Fukushima K, et al. Hypophysis surgery
with or without endoscopy. Auris Nasus Larynx
2001;28(2):1439.
[82] Barone CM, Jimenez DF. Endoscopic craniectomy
for early correction of craniosynostosis. Plast
Reconstr Surg 1999;104(7):196575.
[83] Jimenez DF, Barone CM. Endoscopic craniectomy
for early surgical correction of sagittal craniosynostosis. J Neurosurg 1998;88(1):7781.
[84] Wang YC, Sun MH, Lin CW, Chen YJ. Anatomical location of T23 sympathetic trunk and Kuntz
nerve determined by transthoracic endoscopy.
J Neurosurg 2002;96(1 Suppl):6872.
[85] Kim BY, Oh BS, Park YK, Jang WC, Suh HJ,
Im YH. Microinvasive video-assisted thoracoscopic
sympathicotomy for primary palmar hyperhidrosis.
Am J Surg 2001;181(6):5402.
[86] Wahlig JB Jr, Welch WC, Weigel TL, Luketich JD.

Microinvasive transaxillary thoracoscopic sympathectomy: technical note. Neurosurgery 2000;
46(5):12548.
[87] Vanaclocha V, Saiz-Sapena N, Panta F. Uniportal
endoscopic superior thoracic sympathectomy. Neurosurgery 2000;46(4):9248.
[88] Chung HY, Seo CG, Lee SG. Video endoscopic
sympathectomy using a beroptic CO2 laser to
treat palmar hyperhidrosis. Neurosurgery 1993;
32(2):3279.
[89] Kao MC. Percutaneous radiofrequency upper
thoracic sympathectomy. Neurosurgery 1997;40(1):
2167.
[90] Schaer JL. Endoscopic discectomy. J Neurosurg
1994;81(6):9656.
[91] Adamson TE. Microendoscopic posterior cervical
laminoforaminotomy for unilateral radiculopathy:
results of a new technique in 100 cases. J Neurosurg
2001;95(1 Suppl):517.
[92] Jho HD. Endoscopic transpedicular thoracic discectomy. J Neurosurg 1999;91(2 Suppl):1516.
[93] Karahalios DG, Apostolides PJ, Vishteh AG,
Dickman CA. Thoracoscopic spinal surgery. Treatment of thoracic instability. Neurosurg Clin North
Am 1997;8(4):55573.
[94] Frank E. Removal of a lateral disc herniation with
malleable endoscopic forceps: technical note. Neurosurgery 1997;41(1):3113.
[95] Onik G, Richardson D, Amaral J, Jennings W,
Sholes A. Percutaneous anterior discectomy under
ultrasound guidance. Minim Invasive Neurosurg
1995;38(2):905.
The endoscopic management of arachnoidal cysts

Rick Abbott, MDa,b,*
a
Department of Neurosurgery, INN, Beth Israel Medical Center, 170 East End Avenue, New York, NY 10128, USA
b
Department of Neurosurgery, Albert Einstein School of Medicine, New York, NY, USA
Arachnoidal cysts are a not uncommon lesion

for a neurosurgeon, particularly a pediatric neurosurgeon, to be called on to treat. In one autopsy
series, the incidence of these cysts was 0.1% [1]. Of
lesions arising intracranially, arachnoidal cysts
comprise approximately 1% [2]. There has been
active debate as to how to manage these cysts best,
with the controversy driven by their benign
behavior, subtle clinical sequelae, and the potential
for treatment failure or complications. Not surprisingly, the introduction of the neuroendoscope
into a neurosurgeons armamentarium typically
leads to the consideration of its use to treat these
entities in the hope of accomplishing what a craniotomy can while avoiding the associated morbidity. In the following article, the wisdom of such an
approach is explored.
Pathology and pathogenesis
In 1831, Bright [3] rst described an arachnoid
cyst as . . .a serous cyst forming in connection with
the arachnoid, and apparently lying between its
layers. . . In his book, he discussed two cases,
stating that these cysts seemed to be chronic, to
have a low potential for growth, and to be of sizes
varying from that of a pea to as large as or larger
than an orange. During the past century, these
observations were conrmed with the introduction
of microscopic neuropathology, and in 1978,
Rengachary et al [4] published a photomicrograph
illustrating Brights observation. This photomicro-
* Department of Neurosurgery, INN, Beth Israel

Medical Center, 170 East End Avenue, New York, NY
10128, USA.
E-mail address: Rabbot@bethisraelny.org
graph demonstrated splitting of arachnoidal membrane at the margin of the cyst and a lack of
trabecula within the cyst, showing the cyst to arise
within the arachnoidal membrane and not within
the subarachnoid space. The cyst membranes
contained hyperplasic arachnoidal cells and a thick
layer of collagen.
The pathogenesis, however, has been more
controversial. Until the 1970s, authors argued
over arachnoid cysts being either secondary
phenomena occurring in regions of agenesis of
the brain or primary events of dysgenesis of the
arachnoid investing the brain. In 1955, Robinson
[5] published a series of 15 patients with middle
fossa arachnoidal cysts, hypothesizing that they
were cerebrospinal uid (CSF) collections passively lling a space left by an agenesis of a portion
of the temporal lobe. Starkman et al [6] put forth
a countertheory in 1958 in a report of three
autopsies done on individuals with middle fossa
cysts, nding these cysts to be surrounded by
arachnoidal membrane, which led him to conclude that the cysts had arisen as a result of
splitting or duplication of the arachnoid during
development. With the introduction of CT scanning and the ability to image the brain before and
after treatment of these cysts, it became apparent
that there was a capacity for expansion of the
temporal lobe into space provided by decompression of the cyst. In 1971, Robinson [2] withdrew
his hypothesis of agenesis, stating that the cysts
were caused by maldevelopment of the arachnoid.
Robinson [2], the primary proponent of these
cysts being secondary phenomena, abandoned his
position after observing that neurologic sequelae
to these cysts were not in proportion to the cysts
size and that brain re-expansion could be seen on
CT after the cysts were treated.
doi:10.1016/S1042-3680(03)00071-8
10
Arachnoidal cysts typically arise within normal

arachnoidal cisterns. They are believed to arise
around week 15 of gestation after the rupture of
the roof of the closed fourth ventricle and creation
of the foramina of Luschka and Magendie. This
allows for the escape of CSF into the subarachnoid
space, where it replaces intracellular ground substance lling the subarachnoid space. Duplication
or splitting of the arachnoid at the time of cistern
formation provides the anlage for the arachnoidal
cyst. Similarly, splitting or duplication of the
ependymal lining of the lateral ventricles results
in the formation of intraventricular or ependymal
cysts. Arachnoidal cysts most commonly arise in
the middle fossa (30%50%), but 10% arise on the
hemisphere convexity, 10% in the suprasellar
cistern, 10% in the quadrigeminal cistern, 10%
in the cerebellopontine (CP) angle, and 10% in the
midline of the posterior fossa [7].
Intraventricular cysts typically arise in or near
the atria of the lateral ventricles. Their walls can be
formed by arachnoidal cells (intraventricular
arachnoidal cysts), ependymal cells (ependymal
cysts more correctly viewed as being periventricular), neuroepithelial cells, or choroidal cells. Most
of these cysts, excluding the choroidal cysts, probably represent a dysgenic process and are therefore not uncommonly associated with dysfunction
of adjacent brain parenchyma as manifested by
focal seizures and cognitive disabilities [8].
Presentation
Intracranial CSF cysts typically become symptomatic in patients before the age of 20 years, with
most doing so within the rst decade [9]. There is
some variation in age at presentation according to
the location of the cyst, however. Arachnoidal
cysts of the middle fossa present in patients before
the age of 16 years [10]. Cysts of the quadrigeminal cistern present earlier, usually by the age of 12
months because of their compression of the
aqueduct of Sylvius [8], whereas arachnoidal cysts
of the suprasellar cistern present later in childhood, with more than 60% presenting in patients
between 1 and 20 years of age [8]. Only 14% of
suprasellar cysts present after the age of 20 years.
The signs and symptoms at presentation are
referable to the location and size of the cyst. With
the exception of cysts of the CP angle, the most
common symptoms present at presentation are
those of a slowly growing mass or of hydrocephalus (ie, headache). The headache can be generalized, or it may localize to the site of the cyst.
Other symptoms at presentation point to the

location of the cyst. Cysts of the middle fossa can
be associated with partial or secondarily generalized seizures. Arai et al [11] found a 39% incidence
in 77 patients having cysts of the middle fossa,
whereas Ciricillo et al [12] found the same incidence in 39 patients. Boop and Teo [13] have
reported that more than 80% of their patients with
such cysts manifest behavioral abnormalities,
attention decit disorder, or other problems in
school. Cysts of the quadrigeminal cistern can be
associated with nystagmus, Parinauds syndrome,
hearing disturbances, and, rarely, motor decits.
Such symptoms are rare because of the relatively
early onset of hydrocephalus associated with such
lesions. Suprasellar cysts can be associated with
visual disturbances (papilledema, optic atrophy, or
bitemporal eld cuts), endocrinopathies, and the
bobble head doll sign that is pathognomic for
lesions in this location. Cysts of the CP angle
typically have a long history of slowly evolving
symptoms referable to stretching of cranial nerves
and distortion of the cerebellum. Vertigo, hearing
disorders (eg, tinnitus, hearing loss), hemifacial
spasm, facial paresis, trigeminal neuralgia, decreased corneal sensation, nystagmus, intention
tremor, ataxia, and dysmetria can be features of
these cysts. If ignored too long, obstruction of the
outlets of the fourth ventricle can result in an
obstructive hydrocephalus. Clival arachnoidal
cysts distort the brain stem, causing compression
of the corticospinal tracts (paresis in the extremities, hyperreexia, and Babinski sign) and stretching of cranial nerves (diplopia). When these extend
into the supratentorial space, endocrinopathies
can evolve because of compression of the pituitary
apparatus. Other than signs of increased intracranial pressure, intraventricular cysts can be associated with focal seizures, ataxia and other gait
disturbances, blurred vision, or frank diplopia.
Diagnosis
The imaging tool of choice with these cystic
CSF lesions is MRI (Fig. 1). Cysts of the middle
fossa, clivus, and CP angle must be dierentiated
from epidermoid cysts. Diusion-weighted MRI
is useful in such cases, because the two types of
cysts have dierent signal characteristics in the
imaging sequence. Suprasellar arachnoid cysts
need to be dierentiated from craniopharyngiomas and Rathke cleft cysts. The suprasellar
arachnoid cysts are typically larger; grow symmetrically and upward, resulting in the Mickey
11
Fig. 1. (A) Coronal T1-weighted MRI scan of middle fossa arachnoid cyst. (B) Axial T2-weighted scan of suprasellar
arachnoid cysts base with ow void seen at point of fenestration (arrow). (C) T2-weighted sagittal MRI scan of
quadrigeminal arachnoidal cyst. (D) T1-weighted coronal MRI scan of cerebellopontine angle arachnoidal cyst.
Mouse sign; and do not contain calcication

commonly seen in craniopharyngiomas. In this
situation, CT may be warranted to exclude
a craniopharyngioma given this modalitys sensitivity for calcium.
Intraventricular CSF cysts can be dierentiated from epidermoids, dermoids, and parasitic
cysts using MRI. The dierential diagnosis should
be discussed with your neuroradiologist when
ordering imaging of the lesion.
12
Treatment
The rst question to be answered is that of
whether any treatment should be oered. Implied
in this question is whether further growth of the cyst
is expected and whether symptomatic hydrocephalus is present. The latter question is easily
addressed by careful history taking and examination of the patient. The former can be more dicult.
There are few reports available that oer rules for
treatment based on the observed natural history of
these cysts. One based on a retrospective study of
adults with arachnoidal cysts showed that cysts
tended to grow over time if there was distortion of
neural structures adjacent to the lobe containing
the cyst or to bony structures [14]. This does not
oer assistance in decision making for children,
however, given its retrospective nature and the fact
that because the study was conducted in a group of
adults, it represented a more benign subgroup of
patients with this condition. Most reports dealing
with the pediatric population recommend treatment at the time of discovery of the cyst unless it is
of a small size with minimal distortion of surrounding tissues and has been discovered incidentally
[13,15,16]. Cysts that distort surrounding neural
tissues have been shown to alter cerebral blood ow
[10,17]. Presumably, this explains the atrophy that
can occur over time, as shown when there is a failure
of parenchymal re-expansion when cysts are
treated in older individuals.
Next to be answered when treatment has been
elected is how the cyst should be treated. In the
1980s, most articles discussing treatment of these
cysts favored the use of shunts. Presumably, this
was the result of a relatively high incidence of
postoperative complications after craniotomies in
the 1960s and 1970s. As operative techniques and
the use of microneurosurgery have evolved, a reduction in the incidence of postoperative complications has occurred. With this and a greater
appreciation of the life history of a shunted patient,
there has been a natural shift in preference from
treating these cysts with shunts to surgical fenestration [18]. Theoretically, to fenestrate a cyst is to
cure it; thus, a lifelong dependence on a shunt is
avoided. This shift in preference toward fenestration is only being accelerated with the introduction
of the endoscope.
Endoscopic treatment of intracranial CSF cysts
is technically challenging and should only be
considered by experienced neuroendoscopists.
The anatomy can be obscure, and the number
and complexity of the instruments used are greater
than those used in the more standard neuroendoscopic cases, such as third ventriculostomy.
To start, careful planning is required. Thought
should be given to the trajectory taken to the
target(s) for fenestration. This is especially critical
when multiple fenestrations are being considered,
as is the case with a suprasellar cyst, or when a third
ventriculostomy is needed in addition to cyst
fenestration, as is case with a quadrigeminal cyst.
An incorrectly placed burr hole because of a poorly
planned trajectory results in an injury to the
cortical tissues surrounding the scope as it is
rocked back and forth to reach fenestration
targets. When grossly o, reinsertion of the guide
cannula to reach a second target may not even be
possible and a new burr hole might be required.
Ideally, when the target abuts or is in a ventricle, it
is best to approach it via the ventricle so that
normal anatomic structures can be used for
guidance. It can be dicult to visualize structures
on the other side of a cysts wall, and it is a common
occurrence to punch holes into brain tissue when
attempting to fenestrate a cyst to an adjacent CSF
space. Thought should also be given to structures
that are to be avoided during the introduction and
advancement of the scope. Ideally, the trajectory
should be established to avoid these structures. If
this is impossible, thought should then be given as
to how to avoid their injury, such as establishing
early visualization with the scope before encountering the structure so as to avoid injury (eg,
visualization of the fornix at the foramen of Monro
so as to avoid it as one traverses the foramen).
Next to be considered is whether a selfretaining holder for the scope is to be used. The
use of a self-retaining system adds time and
complexity to the case. In many cases, an assistant
can navigate the scope while the surgeon works
with instrumentation through the scope. In this
setting, the assistant should be experienced in the
use of the scope and familiar with the working
habits of the surgeon so that the two work as
a uid team. The surgeon cannot do an adequate
job if he or she is worrying about or ghting with
the assistant. In such a setting, the surgeon might
correctly elect to use a self-retaining system. The
other setting where the use of such a system
should seriously be considered is when one
anticipates being at the target site for more than
a few minutes because of the need for a delicate
or extensive surgical dissection. Also, if one is
considering the use of a second instrument
channel, it is best to have the scope held by
a self-retaining system.
With regard to the use of a second instrument

channel, this refers to the introduction of a second
catheter that approaches the target via a dierent
trajectory than that of the endoscope [19]. It can be
useful when one anticipates the need for instruments that are larger than the working channel of
the scope or when there is a need to view the
working of the instruments from an angle. This can
be useful when one anticipates pulling material out
of the surgical eld, because such a maneuver not
uncommonly blinds the endoscopist as the material reaches the head of the scope. The surgeon also
gains a greater appreciation of the amount of
distortion that occurs during the maneuvering of
the instrumentation. Use of the second channel
does add another level of complexity to the case,
however. First, there is the issue of capturing and
maintaining visualization of the channel. This
becomes increasingly dicult as the distance
between the burr holes of the scope and instrument
channel increases and as the depth of the target
increases. Second, there can be diculty in coordinating the manipulation of the two channels.
This takes some practice, and my advice is to make
small movements that do not result in the complete
loss of visualization of the instrument channel. The
instrument channel is rst moved partially out of
the eld of view in a desired direction, and the
scope is then moved to recenter the instrument
channel in its eld of view. Computer-assisted
guidance systems can be of great use in overcoming
these problems, with trackers being placed on the
scope as well as on the instrument channel.
At every step in the surgery, absolute attention
must be paid toward homeostasis. It has been
established that less than a teaspoon of blood
within the intraventricular uid space blinds the
endoscopist. Consequently, after the burr hole has
been drilled and before the dura is opened,
homeostasis must be complete. This is also the
case after the dura is opened before penetrating the
arachnoid. Midline cysts, such as suprasellar and
quadrigeminal cysts, are typically approached via
the ventricles. A guide sleeve is typically introduced
into the lateral ventricle, its stylet is removed, and
the endoscope is then advanced into the lateral
ventricle. During this approach, it is wise to remain
oriented to the trajectory of the scope so as to
appreciate the general location of the scopes tip.
When the anatomy is disorienting, the rst step
should be to check that ones conceptualization of
the scopes trajectory is correct. More than once, I
have mistakenly thought that I was in the region of
the foramen of Monro only to discover that the
13
scope was pointing posterior with the tip in the

ventricles atrium. This is especially common with
a burr hole placed too anteriorly. Normal anatomic structures are searched for to further
establish orientation. Once the choroid plexus is
found, it can be followed anteriorly to nd the
foramen of Monro or posteriorly to point to where
a quadrigeminal cyst is herniating through the
oor of the lateral ventricle. If the scope feels
awkward in its movement, the most common
reason is that the camera is not oriented to the
scope (ie, the cameras 12-oclock position is not
the scopes 12-oclock position). The scope should
be withdrawn from the head and its cameras
orientation checked.
Laterally projecting cysts, such as those of the
middle fossa and CP angle, are approached
through the overlying parenchyma laterally. This
parenchyma can be present to a varying degree.
Draped over the surface of the cysts are cortical
veins, especially in the case of cysts of the middle
fossa. Schroeder et al [20] have suggested not
disturbing the lateral wall of such cysts so as to
avoid traction of cortical veins stretched over their
surface. Once within these cysts, arterial vessels of
the Sylvian ssure can be seen coursing the medial
wall. They can be followed down to the basilar
cisterns, where the fenestrations can be made. The
membrane of these cysts has been shown to be rich
in collagen; consequently, it can be dicult to
penetrate. Scissors or a knife can be used to cut an
opening. It can then be enlarged with sharp
dissection or by repeated ination of a balloon. It
has not been established what constitutes an
adequate opening. Until such time as this has been
established, one should attempt to duplicate what
one would accomplish using microneurosurgery
via a small craniotomy. If diculty is encountered
in doing this, judgment is used, but there should
not be any hesitancy to convert to a craniotomy to
accomplish ones surgical goals. On more than one
occasion, we have done this on our service; the
families of our patients are always told that this is
a possibility, and consent is obtained for just such
a possibility.
Earlier mention was made of the use of
computer-assisted surgical guidance systems or
frameless stereotaxis. These systems can be of great
use when working within cysts containing no
anatomic landmarks for guidance. By using such
systems, one moves from navigating by dead
reckoning based on surface landmarks to imagebased navigation, where the tip of scope and or its
channel or a second instrument channel can be seen
14
on MRI or CT. Before performing the surgery, such

systems can be used for surgical planning. Structures to be avoided can be incorporated in planning
the trajectory, and this information is, in turn, used
to establish the optimum location for the burr hole.
Depending on the guidance system being used,
there are several options that can be employed for
image guidance. Some systems require the use of
dedicated pointing devices. In such settings, a special pointer can be manufactured to replace either
the endoscopes guidance sleeve or its stylet.
Another alternative is to have the company modify
a rigid endoscope so that its system can track it
directly. I have preferred tracking the stylet of
a peel-away catheter used to establish a channel for
the endoscope because it aords the greatest
exibility for scope selection and is the simplest
and least expensive to manufacture. At least one
system has removable tracking devices that can be
attached to an instrument or catheter and then
registered to the system to allow for tracking by the
system. Some accuracy is sacriced when using
such a device, and it can take several minutes to
establish registration. Also, more than once, I have

been unable to accomplish registration when attempting to use these attachable trackers.
After registering the patient and pointing
devices to the system, the burr hole is located
using the system and the adequacy of the trajectory
is conrmed before making a skin incision. It is
wise to take several moments to conrm the
appropriateness of instrument setup at this point
so as to prevent discovering that movement of the
scope is limited by something in the surgical eld or
tracking is lost when the scope is in a certain
position within the head. Once this is done, the
burr hole is made and the dura opened using the
same care as mentioned previously. The appropriateness of the planned trajectory to the target is
conrmed one nal time using dead reckoning; the
scopes channel is then advanced toward the target.
Advancement toward the target is done slowly
while retaining the appropriate trajectory by maintaining it on at least two image planes (Fig. 2).
Some systems give a birds eye view or so-called
target view showing the tip being tracked as well
Fig. 2. Image showing approach to target (marked with +) by outer sleeve of endoscopy (its tip is the point of the cross
hairs). For this cyst, we were targeting the point where the cyst abutted the ambient cistern.
as the target; this is quite helpful, because the goal

is simply to overlay the two cross hairs. Side to side
and upward or downward corrections in route to
the target cause distortion in the brains parenchyma, because the catheter cannot slice through brain
tissue. This is critical to appreciate when a exible
catheter for the scope is being used, because once
the stylet is removed to allow for the introduction
of the scope, the catheter deects as the parenchyma restores itself to its resting conguration. The
result is that the catheter tip will no longer be on
target. Consequently, the surgeon must judge when
there has been too great a need for correction of
trajectory. When there is concern about this, the
catheter should be removed and a second attempt
made to advance it to the target. Much has been
made of the potential for brain shift imparting
inaccuracy to the guidance data set as the surgery
proceeds. Obviously, if one is to be working at only
one target, this should not be of great concern if
things move along at a pace such that little CSF
escapes before the catheter is on target. At the
point that the scope is looking at the target, the
accuracy of the guidance data set is no longer of
concern. Alternatively, if there is more than one
target, thought should be given to the problem. In
my experience, structures that are close to the
center of the head and vertically inferior to my
pointer seem to maintain accurate registration to
the data set, whereas those near the surface or
lateral to a deating structure or lesion tend to lose
registration as the case proceeds. This should be
kept in mind when determining the order of
targeting. It is also important to remember that
there is always a tomorrow and to discuss with
a family member the possibility of needing multiple
sessions to accomplish the surgical goal.
There will be cases in which shunting of a cyst is
required. We had one case of a quadrigeminal cyst
that underwent three endoscopic fenestrations into
the ventricles (twice into the lateral ventricle and
once into the third) and two open fenestrations,
only to fail all procedures and become symptomatic because of an increasing mass eect from the
cyst (Fig. 3). It was then decided to insert
a cystoperitoneal shunt. The placement of a catheter into an arachnoid cyst can be extremely
dicult, because the cyst wall is extremely prone
to deecting the catheter to the side when insertion
is attempted. Use of the endoscope to conrm
entry into the cyst can avoid postoperative
frustration. One can use a larger scope to cut
a fenestration in the cyst wall. The scope and its
guide sleeve are then advanced into the cyst, the
15
Fig. 3. Second recurrence of quadrigeminal cyst after

two prior fenestrations into lateral and third ventricles.
At the third operation, the cyst was fenestrated and
a catheter placed for a cystoperitoneal shunt.
scope is removed, and a proximal catheter of

a cystoperitoneal system is then fed down the guide
sleeve into the cyst. It is wise to determine the depth
of scope insertion to establish the length of catheter
needed before removal of the scope in such a setting
and to use a peel-away guide sleeve to simplify
catheter placement. Alternatively, a small intraluminal scope can be used as the catheters stint
during attempted placement of the proximal
catheter. Once within the cyst, the scope can be
used to conrm cyst penetration. If the wall has not
been penetrated, unipolar cautery can be applied to
the scope to complete penetration unless there is
concern about adjacent vessels, as might be the
case for a quadrigeminal cyst. Before doing this, it
is wise to back the scope out partially to visualize
the wall better and conrm that it has indeed not
been penetrated. On occasion, I have been fooled
into thinking that I had not penetrated the cyst
when, in fact, I had, and I was simply up against the
far wall of the cyst. In such a case, one can
immediately see the interior of the cyst as opposed
to a trailing cyst wall.
Results
There are now a few papers describing outcomes in small series of patients who have
undergone endoscopic management of their intracranial CSF cysts. Paladino et al [21] described
6 of their patients treated endoscopically. One had
16
a parasagittal cyst, and the other 5 had middle

fossa cysts. Four of 6 cysts were successfully
managed endoscopically, with 2 others requiring
insertion of a cystoperitoneal shunt. Kim [22]
reported on 7 patients with arachnoid cysts. Three
had cysts in the posterior fossa, 2 had cysts in the
suprasellar cistern, 1 had a cyst in the middle
fossa, and 1 had a cyst on the convexity, with
all experiencing symptomatic resolution. Radiographically, 4 of 7 cysts diminished in size,
whereas 3 of 7 disappeared. Ruge et al [23]
reported on successfully managing 2 children with
quadrigeminal plate cysts endoscopically, and
Furuta et al [24] reported a similar case managed
successfully using an endoscope to place a cystoperitoneal shunt after a failed attempt to place the
catheter stereotactically. Brunori et al [25] and
Hayashi et al [26] have reported experiences
similar to those of Ruge et al [23] in successfully
communicating quadrigeminal cysts to the ventricles. Hopf and Perneckzy [27] reported on 36
patients with various forms of arachnoid cysts
treated endoscopically, with 26 improving symptomatically. Wagner et al [28] reported on 2
children with arachnoid cysts managed successfully endoscopically, and Decq et al [29] reported
on 2 children with suprasellar arachnoid cysts
successfully managed. Walker et al [30] reported
a 64% success rate in managing 14 children with
arachnoidal cysts endoscopically. In summary, as
series numbers increased, so did failures, with
larger series reporting around a 33% failure rate
in endoscopically managing arachnoid cysts. This
compares with a 30% failure rate for shunted
cysts and a 19% failure rate in cyst fenestration in
the European Cooperative Study [9]. The more
aggressive cyst resection surgery had a 7% failure
rate in the same study. Fewel et al [18] reported
a 27% failure rate in 102 arachnoidal cysts
managed with either surgical fenestration or
resection. Undoubtedly, there is a learning curve
associated with this technique, and the failure rate
currently being seen in the endoscopically managed cases will lessen over time. Until such time as
the failure rate duplicates that seen with open
surgery, thoughtful consideration should be given
to the adequacy of the fenestration accomplished
endoscopically and to whether or not the procedure should be converted to an open operation
if there is concern about its adequacy. Additionally, a frank discussion should be undertaken with
the family as to the pros and cons of the various
approaches so that they can make an informed
decision.
Complications
As with any surgery, there are potential
complications when endoscopically managing
arachnoid cysts. Minimally invasive should not
be construed to mean minimal risk by the surgeon,
and the family or patient should in no way be sold
a bill of goods that this technique has fewer risks
than open surgery. Kim [22] reported that 1 of his 7
patients treated endoscopically experienced significant bleeding during her surgery, requiring abandonment of the endoscopic procedure with
conversion to an open procedure with successful
control of the hemorrhage and completion of the
surgical goal of fenestration. Hopf and Perneczky
[27] reported a 14% complication rate in imaging 36
patients with arachnoidal cysts. Four patients
experienced subdural hematomas or hygromas
after their surgery, with 2 of them also developing
meningitis. This may be an important observation,
because we have noted infections in some of our
patients who have experienced intraoperative
hemorrhaging, which required extraventricular
drainage after surgery. This is a discussion we have
with all our families before surgery. Another of the
patients reported on by Hopf and Perneczky [27]
experienced hemorrhaging after treatment of a posterior fossa cyst, resulting in hydrocephalus that
required a subsequent third ventriculostomy. Robinson and Cohen warn of the risks of injuring blood
vessels and other structures by the guide sleeve
when advancing the scope because of the sleeve not
being visualized by the scope [8]. Finally, structures
can be injured during the actual fenestration
process when visualization is poor or excessive
force is used to penetrate the tough membrane.
Summary
I have little doubt that most arachnoidal cysts
will be managed endoscopically in the future given
the advances we have seen over the last decade in
our instrumentation. Our excitement to employ
this new technology should be governed by the
reality that we are still learning and that our
current success rate is not quite as good as what
can be expected when using microneurosurgery.
References
[1] Shaw C, Alvord EJ. Congenital arachnoid cysts and
their dierential diagnosis. In: Vinken P, Bruyn G,
editors. Handbook of clinical neurology, vol. 31.
Congenital malformations of the brain. Amsterdam: North Holland; 1977. p. 75136.

[2] Robinson R. Congenital cysts of the brain: arachnoidal malformations. Prog Neurol Surg 1971;4:
13374.
[3] Bright R. Reports of medical cases, selected with
a view of illustrating the symptoms and cure of
diseases by a reference to morbid anatomy, vol. 2.
Diseases of the brain and nervous system. London:
Lonham, Rees, Orme, Brown, Green, and Highley;
1831.
[4] Rengachary S, Watanabe I, Brackett C. Pathogenesis of intracranial arachnoid cysts. Surg Neurol
1978;9:13944.
[5] Robinson R. The temporal lobe agenesis syndrome.
Brain 1964;87:87106.
[6] Starkman S, Brown T, Linell E. Cerebral arachnoid
cysts. J Neuropathol Exp Neurol 1958;17:484500.
[7] Hogg J, Peterson A, El-Kadi H. Imaging of cranial
and spinal cerebrospinal uid collections. In:
Kaufmann H, editor. Cerebrospinal uid collections. Park Ridge, IL: American Association of
Neurological Surgeons; 1998. p. 1957.
[8] Kaufmann H, editor. Cerebrospinal uid collections. Park Ridge, IL: American Association of
Neurological Surgeons; 1998.
[9] Oberbauer R, Haase J, Pucher R. Arachnoid cysts
in children: a European cooperative study. Childs
Nerv Syst 1992;8:2816.
[10] Boop F, Young R, Scott R. Arachnoid cysts of the
middle cranial fossa and convexity. In: Kaufmann
H, editor. Cerebrospinal uid collections. Park
Ridge, IL: American Association of Neurological
Surgeons; 1998. p. 6796.
[11] Arai H, Sato K, Wachi A, Okuda O, Takeda N.
Arachnoid cysts of the middle cranial fossa:
experience with 77 patients who were treated with
cystoperitoneal shunting. Neurosurgery 1996;39(6):
110812.
[12] Ciricillo S, Cogen PH, Harsh GR, Edwards MS.
Intracranial arachnoid cysts in children. A comparison of the eects of fenestration and shunting.
J Neurosurg 1991;74:2305.
[13] Boop F, Teo C. Congenital cysts. In: McLone D,
editor. Pediatric neurosurgery. Philadelphia: WB
Saunders; 2001. p. 48998.
[14] Becker T, Wagner M, Hofmann E, Warmuth-Metz
M, Nadjmi M. Do arachnoid cysts grow? A
retrospective CT volumetric study. Neuroradiology
1991;33:3415.
[15] Artico M, Cervoni L, Salvati M, Fiorenza F,
Caruso R. Supratentorial arachnoid cysts: clinical
and therapeutic remarks on 46 cases. Acta Neurochir (Wien) 1995;132:758.
[16] Galassi E, Gaist G, Giuliani G, Pozzati E.
Arachnoid cysts of the middle cranial fossa:
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
17
experience with 77 cases treated surgically. Acta

De Volder A, Michel C, Thauvoy C, Willems G,
Ferriere G. Brain glucose utilisation in acquired
childhood aphasia associated with a Sylvian arachnoid cyst: recovery after shunting as demonstrated
by PET. J Neurol Neurosurg Psychiatry 1994;57:
296300.
Fewel M, Levy M, McComb J. Surgical treatment
of 95 children with 102 intracranial arachnoid cysts.
Pediatr Neurosurg 1996;25:16573.
Jallo G, Morota N, Abbott R. Introduction
of a second working portal for neuroendoscopy.
A technical note. Pediatr Neurosurg 1996;24:5660.
Schroeder H, Gaab M, Niendorf W. Neuroendoscopic approach to arachnoid cysts. J Neurosurg
1996;85:2938.
Paladino J, Rotim K, Heinrich Z. Neuroendoscopic
fenestration of arachnoid cysts. Minim Invasive
Neurosurg 1998;41:13740.
Kim M. The role of endoscopic fenestration
procedures for cerebral arachnoid cysts. J Korean
Med Sci 1999;14:4437.
Ruge J, Johnson R, Bauer J. Burr hole neuroendoscopic fenestration of quadrigeminal cistern
arachnoid cyst: technical case report. Neurosurgery
1996;38:8307.
Furuta S, Hatakeyama T, Nishizaki O, Fukumoto
S. Usefulness of neuroendoscopy in treating supracollicular arachnoid cystscase report. Neurol
Med Chir (Tokyo) 1998;38:1079.
Brunori AA, Chiappetta F. Endoscopy for cysts.
J Neurosurg 1999;91:10678.
Hayashi N, Endo S, Tsukamoto E, Hohnoki S,
Masuoka T, Takaku A. Endoscopic ventriculocystocisternostomy of a quadrigeminal cistern arachnoid cyst. Case report. J Neurosurg 1999;90:11258.
Hopf N, Perneczky A. Endoscopic neurosurgery
and endoscope-assisted microneurosurgery for the
treatment of intracranial cysts. Neurosurgery 1998;
43:13307.
Wagner HJ, Seidel A, Reusche E, Sepehrnia A,
Kruse K, Sperner J. A craniospinal enterogenous
cyst: case report. Neuropediatrics 1998;29:2124.
Decq P, Brugieres P, Le Guerinel C, Djindjian M,
Keravel Y, Nguyen JP. Percutaneous endoscopic
treatment of suprasellar arachnoid cysts: ventriculocystostomy or ventriculocystocisternostomy?
Technical note. J Neurosurg 1996;84:696701.
Walker M, Petronio J, Carey C. Ventriculostomy.
In: Cheek W, Marlin A, McLone D, editors.
Pediatric neurosurgery: surgery of the developing
nervous system. Philadelphia: WB Saunders; 1994.
p. 57281.
Basic principles and equipment in neuroendoscopy

Vit Siomin, MDa, Shlomi Constantini, MD, MScb,*
a
Department of Neurosurgery, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA
Division of Pediatric Neurosurgery, Dana Childrens Hospital, Tel Aviv Medical Center, 6 Weizman Street,
Tel Aviv 64239, Israel
A fool with a tool is. . . still a fool. Lars Leksell
The advent of neuroendoscopy has had a remarkable impact on the eld of neurosurgery.
Although the use of an endoscope to diagnose and
treat various central nervous system (CNS)
conditions, particularly those conned to the
ventricular system, has been well recognized for
years [17], the story of modern neuroendoscopy
is just beginning. It has been attended by a number
of remarkable breakthroughs in optical physics,
technology, and instrumentation. Needless to say,
understanding the basic physics and instrumentation underlying todays endoscopes is essential
for safe and successful work with these delicate
instruments. This article presents a basic overview
of the technology that literally brought light into
the depths of contemporary neurosurgery.
Todays market is saturated with neurosurgical
endoscopes, with many companies producing
similar pieces of equipment. Any article attempting to comprehensively list the advantages and
disadvantages of each endoscope on the market
today will probably be somewhat outdated by
publication because of the rapid progress in this
eld. This article therefore does not even attempt
to provide a full list of all commercially available
endoscopes with their descriptions. The authors
simply intend to provide readers with a general
overview of the selected endoscopes, including
a brief explanation of their respective advantages
and disadvantages.
* Corresponding author.
E-mail address: Sconsts@netvision.il
(S. Constantini).
A brief history of optics

Roots of the modern rigid endoscope
Until the sixties, the optic chains in a lens
system were constructed of small glass lenses
interspersed with large air spaces. A British
physicist named Hopkins realized that the total
amount of light transmitted through an endoscope
is proportional to the refraction index of the
material used and that using more glass than air
would increase the amount of light by a factor of
about 2. He restructured the lens system to consist
of long glass lenses interspersed with small lens-like
air spaces [3]. These rigid rod lens scopes, with
their increased light transmission ability, form the
basis of most modern endoscopic systems (Fig. 1).
Further reduction in light loss is achieved
through a special coating of the glass surfaces to
minimize light reection. Uncoated glass surfaces
usually lose about 5% of their light through
reection. Because endoscopes consist of multiple
glass lenses, the cumulative light loss can become
signicant. The solution is to coat the glass surface
with an ultrathin layer of magnesium uoride. This
layer markedly decreases the reection and improves the optic characteristics of endoscopes and
cameras [3,8,9].
Understanding beroptics
Before the sixties, endoscopists used miniature
tungsten light bulbs inside the endoscopes. These
bulbs had two disadvantages, however: the heat
generated by the bulbs could easily burn tissues,
and the bulbs could not emit blue light waves, so
a red color would dominate the working eld.
Fortunately, beroptic cables were invented in the
doi:10.1016/S1042-3680(03)00075-5
20
V. Siomin, S. Constantini / Neurosurg Clin N Am 15 (2004) 1931
Fig 1. Conventional and rod lens systems.
sixties, a major breakthrough for endoscopic

technology. With their development, the light
source could be separated from the endoscope
and its emitted lighted transmitted via beroptic
cables to the tip of the endoscope. This avoided the
problem of tissue injury from heat and also
delivered a more natural light. Specialized cables
could also be used to conduct images to a camera.
Fiberoptic cables are made of individually coated
exible bers with an inner core of silica glass.
Fiberoptic bundles can be arranged in two ways [3]:
Coherentlythe proximal end portrays the
image exactly as it is perceived distally. A
coherent arrangement is used for visualization during surgery.
Incoherentlyber bundles transmit light
only. An incoherent arrangement is used to
transmit light to the surgical site.
Most modern beroptic endoscopes contain
a minimum of 10,000 coherent bers. Some
contain as many as 30,000 bers. A greater
number of bers is an advantage, because the resolution of beroptic endoscopes is proportional
to the number of bers in the endoscope. More
bers provide an image with a much sharper
resolution.
Although a great number of bers in bundles
used to transmit images is an obvious advantage,
it is not so gainful when it comes to transmission
of light from the source box to the scope tip.
Having many individual bers bundled together
in light source cables increases the total surface
area and leads to a loss of 30% to 40% of the light
being transmitted through all those individual
bers. Thus, contrary to popular belief, the light
that emerges from the proximal tip is somewhat
weaker than the light that enters the endoscope at
the original source [3]. Light loss is discussed in
greater detail later in this article in the section on
light sources.
Working with optic angles

The tip of the scope can be bent to dierent
angles, bringing dierent aspects of the surgical
eld into view. These angles are obviously xed in
the rigid solid lens scopes as opposed to the exible
beroptic scopes, where the tip can be manually
deected to change the angle of view o the scopes
long axis. For the rigid solid lens scopes, the most
commonly used angles are: 0 (head-on), 30 , 70 ,
and 120 . The 0 deection scopes portray only
what they are viewing head-on, minimizing the risk
of disorientation. The angled scopes are more
versatile in that they visualize areas that would
otherwise be out of view or dicult to illuminate.
The major disadvantage of angled scopes is that the
indirect image may cause the surgeon to become
disoriented. The authors encountered orientation
diculties mostly when operating on patients with
congenital lesions, such as arachnoid and porencephalic cysts. Similar problems may also be encountered in endoscopically assisted transsphenoidal
surgery and intracranial tumor removal. Steerability of beroptic endoscopes adds another
element to the problem of disorientation. Extensive
experience is necessary to get used to oblique and
around-the-corner views.
Light sources
An endoscopes light source is extremely
important, often becoming a limiting or facilitating
factor. Currently, most neurosurgeons use highintensity xenon light sources. The light is transmitted through incoherent ber bundles to the
surgical eld. Setting the light source to between
300 and 500 W provides a superior picture quality.
Other types of light sources, such as halogen, are
not used in modern endoscopy because they do not
generate a bright enough light.
Light loss in a beroptic system is a signicant

issue. As a rule, 30% of light is lost at the lamp,
because up to 30% of the surface area consists of
cladding and ller, which absorb and reect light.
Fiber mismatch (between the incoherent cable
bundles and the coherent optic ber bundles)
accounts for another 20% loss. Signicant loss
also occurs because of mismatch between the
wider transmitting cable and the thinner scopes
used in neurosurgery. Additional light is also lost
from inadvertent reection in various surfaces.
Thus, at best, only about 30% of the light
generated within the light source reaches the
distal tip of the scope [3,8]. This problem is still
awaiting a technologic solution.
Cameras
The chip camera, or charged-coupled device
(CCD), is a critically important part of any
endoscopic system. Invention of the chip camera
was a remarkable technologic achievement, allowing a signicant decrease in the size of the
endoscope combined with an improvement in
the quality of the transmitted image in comparison with previously used tube cameras. Two
kinds of CCDs are currently used in neuroendoscopy: the single-chip camera and the threechip camera.
The principle used in chip camera technology
is the same as in the microprocessor industry. The
chips consist of horizontal and vertical photosensitive elements that are arranged in intersecting
lines. The points of intersection correspond to
pixels, or picture units, that appear on the display
screen. Light reected by the object being viewed
hits each pixel, and a current is generated. Each
pixel on the display appears either brighter or
dimmer depending on the voltage of the current
signal sent to that pixel. The voltage is a function
of the brightness of the image being portrayed:
a brighter point on the original object being
viewed generates a higher voltage current, triggering a brighter image on the display screen. The
exact current voltage sent to each point is not
just an approximation or estimate; it is actually
a precise numeric representation of the energy
generated when light hits the horizontal and
vertical matrix of the camera. A digital image
incorporating a complete set of picture units
(pixels) is stored in the cameras memory [8].
Most endoscopic systems now use single-chip
cameras. A good resolution for neuroendoscopy is
available with 0.5-in cameras. When the resolu-
21
tion of the one-chip camera is not enough, the

image may be computer enhanced.
Three-chip cameras provide better picture
quality. The authors use the David-3-Chip-Camera
from Aesculap (Center Valley, Pennsylvania),
which features a resolution of more than 800
horizontal lines, compared with 500 lines in the
standard 0.5-in micro-lens-on-chip technology
camera (Fig. 2). The improved image of the
three-chip camera comes at a price, howevera
somewhat larger camera size and higher cost.
Video monitors
The monitor is an integral part of every
endoscopic system (Fig. 3). Three things should
be considered in selecting a monitor: resolution,
screen size, and cost.
A higher resolution monitor provides better
picture quality. There is no advantage to a monitor
resolution that is signicantly better than the
camera resolution, however, because the monitor
only displays the image seen and processed by the
camera. The monitor cannot improve the cameras
image. By the same token, a top-quality image
from the highest resolution camera must still be
displayed on the monitor, so a lower resolution
monitor ruins even the best camera image.
When choosing the screen size, remember that
the camera output is displayed over an area
signicantly larger than the cross section of the
optic cable. As an image is enlarged, the image
resolution and quality decrease. Screens larger
than 13 in display poorer quality images because
of reduced resolution. This becomes even a greater
problem with images generated with a beroptic
scope. The pixels that are only slightly visible
Fig 2. 3-chip CCD camera (Aesculap, Center Valley,

Pennsylvania).
22
Fig 4. 15F peel-away blunt and tapered tip introducers.
Fig 3. 100Hz monitor (Sony, New York, New York)

and 3-chip CCD camera (Aesculap, Center Valley,
Pennsylvania).
when ber scopes with reasonable resolution are

transmitted to a screen smaller than 13 in undergo
magnication with consequent enlargement of pixels when projected to larger screens. Conversely,
larger screens are useful for displaying multiple
images. Finally, note that it is often most costeective to purchase a monitor from the manufacturer that supplies the rest of the system.
Endoscopic components and tools
Peel-away catheter
Because the concept of minimally invasive
surgery lies at the heart of neuroendoscopy, every
stage of the procedure, from cannulation of the
uid space to exit, should minimize interference
with the patients anatomic structures. During an
endoscopic procedure, the surgeon may need to
insert and remove the scope many times. It is
therefore logical to create a single safe passageway that allows multiple reinsertions of the scope
along the same tract without jeopardizing surrounding brain.
Most surgeons today use disposable peel-away
catheter introducers (eg, those manufactured by
Cook [Cook Critical Care, Bloomington, IN] or

Neuroview [Integra NeuroSciences, Plainsboro,
NJ]). These may be attached to the scalp or drapes
with Steri-strips (3M Health Care, St. Paul, MN)
or staples. Both tapered and blunt-tip catheters
are available (Fig. 4). The catheter is supplied in
various diameters ranging from number 3 to
number 49 French. The most popular sizes in
neuroendoscopy fall between number 10 and
number 20 French. Number 10 French catheters
are appropriate for 2- to 4-mm diameter endoscopes, whereas number 20 French catheters
are good for the larger 6- to 8-mm endoscopes.
Larger diameter introducers are not used in
neuroendoscopy.
The disadvantage of using a peel-away introducer is that it can cause bleeding because of
injury to the choroid plexus or intraventricular
veins on insertion. Unfortunately, only Neurocares (Integra NeuroSciences, Plainsboro, NJ)
number 9 French catheter has centimeter markings. Therefore, the authors have found it helpful
to apply some bone wax 5 cm from the tip of the
introducer to prevent unnecessarily deep insertion
of the catheter and damage to intraventricular
structures.
Alternatively, one can use an obturator/operating sheath. This is a reusable metal pipe that can
only be attached with a scope holder, which can
be a slight inconvenience. The introducers and the
sheath enable multiple reinsertions of an endoscope without repeated recannulations.
Rigid endoscopes
Most neurosurgeons today still use rigid glass
endoscopes that are neither exible nor steerable.
An advantage of rigid rod lens scopes is that
smaller diameter lenses may be used, allowing the

endoscope designer to either decrease the diameter
of the whole scope, producing a smaller and
more delicate instrument that can reach more
inaccessible areas, or add more space to the
working channel, producing an instrument that
can accomplish many tasks for the surgeon. The
basic rigid glass endoscope consists of three elements: an optic system, a working channel, and an
irrigation port (Fig. 5). Table 1 summarizes the
features of the most common rigid endoscopes.
The remainder of this section discusses the
advantages and disadvantages of typical endoscopes available from various manufacturers.
The Gaab Neuro-Endoscope (K. Storz,
Tuttlingen, Germany) has a 6.5-mm operating
sheath, with a 3-mm working channel. Fiberoptic light transmission is incorporated. The
Gaab endoscope with a 2-mm operative
Hopkins telescope is a popular choice for
surgery because it allows visualization during
surgical instrumentation with specially designed rigid instruments. The 4-mm 0 , 30 ,
70 , and 120 diagnostic scopes are popular
for diagnostic orientation in the ventricular
system, basal cisterns, arachnoid cysts, or
cystic tumors. Because these diagnostic
scopes do not have a working channel, they
have room for larger diameter lenses and are
therefore able to provide exceptionally clear
visualization and orientation. Note that
Codman & Shurtle (Johnson & Johnson,
New Brunswick, NJ) also manufactures the
Gaab Neuro-Endoscope. The outer diameter
of the Codman & Shurtle endoscope is only
Fig 5. 3.2 mm rigid endoscope (K. Storz, Tuttlingen,

Germany).
23
5.8 mm, however, and the diameter of the

working channel is 1.6 mm.
The ChavantesZamorano Neuro-Endoscope
(K. Storz) has a relatively large outer diameter of 8 mm. It has one central 3-mm
working channel and two separate suction
and irrigation channels, which can also be
used for exible instruments. Better visualization can be achieved with the smaller
diameter 30 and 70 diagnostic scopes. This
endoscope also has a manometer for constant
monitoring of intracranial pressure during
surgery. The disadvantage of the ChavantesZamorano system is that the scope has
a relatively large diameter and the overall
size impedes maneuverability. Nevertheless,
many neurosurgeons use this endoscope,
especially for evacuation of intracerebral
and, particularly, intraventricular hematomas, in conjunction with stereotaxy. Note
that Codman & Shurtle also manufactures
the Chavantes-Zamorano endoscope. The
dierence is that the diameter of that endoscopes working channel is 1.9 mm.
The Auer Neuro-Endoscope (K. Storz) has
an outer diameter of 6.6 mm. This endoscope features two working channels: a
straightforward 0 telescope that is 2.9 mm
in diameter and an adapter for a Greenberg
retractor. The larger size of the Auer endoscope is a major disadvantage; however,
the excellent optics and specially positioned
irrigation and suction apertures at the tip of
the endoscope facilitate a clear view of the
eld.
The Decq Neuro-Endoscope (K. Storz) comes
in two models. The 3.5-mm 5.2-mm model is
suitable for rigid or exible instruments of
1.7 mm in diameter. The 4.0-mm 7.0-mm
model is suitable for instruments of up to
3 mm in diameter. The telescope uses the
Hopkins lens system with a 30 angle of
vision.
Aesculap produces an endoscope with a
6.2-mm operating cannula (Fig. 6). This
model includes a 2.2-mm working channel,
one diagnostic endoscope channel, one irrigation channel, and one overow channel.
This endoscope is available in two lengths:
250 mm, which is compatible with stereotactic frames, and 160 mm. The major advantage of this scope is excellent image
quality; however, the large size is a limiting
factor.
24
Table 1
Basic features of rigid endoscopes
Endoscope
Outer
diameter
Neuroview 700R
(NeuroNavigational)
5.6 mm
Aesculap
6.2 mm
MINOP
(Aesculap)
3.2 mm
4.6 mm
6.0
Gaab endoscope
(Storz/Codman & Shurtle)
6.5 mm
(5.8 mm)
Chavantes-Zamorano
(Storz/Codman)
8.0 mm
Auer endoscope
(Storz)
6.6 mm
Decq endoscope
(Storz)
Oval
3.5 5.2 mm
Oval
4.0 7.0 mm
Channels
1
1
1
1
1
1
1
1
working
irrigation
overow
scope
working
irrigation
overow
scope
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
working/scope
working/scope
irrigation
overow
working
irrigation
overow
scope
working
irrigation
overow
scope
working/scope
irrigation
overow
2
1
1
1
working
irrigation
overow
scope
1/2 working
1 irrigation
1 overow
1 scope
1/2 working
1 irrigation
1 overow
1 scope
The MINOP Neuroendoscopy System (Aesculap) includes the following (Fig. 7):
Trocars with three dierent shaft diameter
options. The 3.2-mm model has one
optic/working channel. The 4.6-mm model has three optic/working, irrigation, and
Diameter of
working
channels
Angles of
diagnostic
scopes

Remark/clinical use

2.0 mm
0 , 30 , 70
2.8 mm
2.2 mm
0 , 30
2.7 mm
2.8 mm
2.8 mm
0 , 30
Glass rod optics,

Good image,
Quite thick and
heavy
Glass rod optics,
Good image,
Quite thick and
heavy, Good for
stereotaxy
Glass rod optics,
Good image,
Short,
maneuverable
2.2 mm
2.8 mm
3.0 mm
(4.0 mm)
3.0 mm
(1.9 mm)
0 , 30 , 70 ,
120
0 , 30 , 70
0
3.0 mm
1.7 mm
30
Glass rod optics,

Good image,
Quite thick and
heavy
Glass rod optics,
Good image,
Quite thick and
heavy, Features
manometer for
ice intracranial
measurement,
Good for
evacuation of
hematomas
Glass rod optics,
Good image,
Quite thick and
heavy, Adapter
for Greenberg
retractor
Glass rod optics,
Good image
3.0
Glass rod optics,
Good image,
Quite thick and
heavy
overow channels. The 6.0-mm model has

four optic/working, additional working,
irrigation, and outow channels.
2.7-mm endoscopes that are angled to 0 and
30 . These are suitable for freehand surgery
as well as for xation by the scope holder.
25
Fig 6. 6.2 mm rigid endoscope (Aesculap, Center Valley, Pennsylvania).
Rigid instruments and electrodes

Good image quality; they are suitable for
pure ventriculostomy, endoscopic surgery, or endoscope-assisted microsurgery:
Rigid wide-angled endoscopes for endoscopeassisted cranial neurosurgery; 0 , 30 , and 70
angles of view are also available from
Aesculap.
The classic model Neuroview rigid scope,
produced by NeuroNavigational (Integra
NeuroSciences, Plainsboro, NJ), oers superior imaging characteristics. The outer diameter of the scope is 5.6 mm, with a 2.0-mm
working channel, irrigation and aspiration
ports, and a 2.8-mm scope channel. The
scopes are 2.7 mm in diameter and are
available with 0 , 30 , and 70 oset angles,
providing 80 of view.
Despite higher image quality, the classic design
multiple-use scopes have some disadvantages.
First, because the camera is directly attached to
the scope, they are quite heavy and cumbersome.
Second, the classic scopes are thicker, leaving
larger holes in the brain. Third, there is a signicant
risk of cross-contamination. Many surgeons, unfortunately, are familiar with bouts of postoperative infections secondary to cross-contamination
Fig 7. 4.6 mm rigid endoscope (Minop, Aesculap,

Center Valley, Pennsylvania).
occurring during endoscopic procedures. Finally,

because of natural wear and tear of the optical
components, the image quality deteriorates after
prolonged use. It was thus inevitable that disposable endoscopes would be introduced into practice.
Note, however, that disposable rigid endoscopes
do not use a rigid lens optic system, because rigid
lenses require a larger diameter. Instead, a 10,000to 30,000-pixel beroptic system is used. The
quality of an image obtained through multipixel
beroptic bundles is still signicantly lower than
that obtained with rigid lenses. Conversely, with
disposable scopes, the surgeon receives a brand
new set of optical components each time, which
eliminates the problem of wear and tear.
Fiberoptic endoscopes
This section discusses the exible and rigid
beroptic neuroendoscopes produced by a few
dierent companies. Table 2 compares the advantages and disadvantages of rigid versus exible
endoscopes.
Several types of exible scopes are available
from Aesculap:
A steerable scope with an outer diameter of
3.9 mm. This scope has a 1.1-mm working
channel and a 0.5-mm suction channel.
Unidirectional steerability allows 160 turns.
A steerable scope with an outer diameter of
2.9 mm. This scope does not have a working
channel. Although this scope is mainly used
for inspection purposes, a steerability of 140
up and down makes this a reasonable choice
for dissection.
A series of nonsteerable exible endoscopes
with outer diameters ranging from 0.7 to 2.3
mm. None of them, except the 2.3-mm scope
(see section on syringomyeloscope), have
working channels. This group of endoscopes
is usually used for endoscopic inspection,
endoscopy-assisted surgery, or passing
through a shunt. These nonsteerable endoscopes are superior to disposable exible
scopes in image quality because they use
higher pixel optical bers.
26
Table 2
Comparison of rigid and exible scopes
Advantages
Disadvantages
Rigid scopes
Flexible scopes
Better image
Higher resolution
Wider view
Better color
Better light transmission
Light weight of disposable scopes
Less maneuverable
Not applicable in spine
Steerability
Application in spine
For some light weight, because camera not attached to scope
Aesculap produces another endoscope originally designed by Perneczky. This instrument

is able to look around the corner while
operating with the microscope. The Perneczky endoscope consists of a beroptic
scope encased in a rigid shaft with an 80
curved tip 1.4 mm in diameter. Using this
curved tip, structures in the operating eld
that are hidden from direct microscopic view
can be inspected. The Perneczky endoscope
may be set into a desired position with a scope
holder (see section on scope holders).
Two types of exible scopes are available from
Codman & Shurtle:
Codman & Shurtle produces a steerable
neuroendoscope that is similar to the 3.9-mm
endoscope from Aesculap. Its outer diameter
is 4 mm, with a 1-mm working channel that
accepts several types of instruments and
accessories. The distal viewing tip can be
rotated 100 to 160 via a thumb control unit.
Codman & Shurtle also produces the Epic
microvision. The Epic microvision is a semiexible endoscope 1.8 mm in diameter with
a bayonet-like design. The angle of the scope
can be changed by inserting it into more rigid
cannulas with straight, 20 , or 45 angles.
Such small scopes can be used for:
Better intraoperative visualization, particularly in conjunction with an operative
microscope, when an around-the-corner
view is required
Blunt perforation of membranes when the
scopes themselves are used as dissecting
and fenestrating tools
Poor image
Pixel granules
Narrower view
Less true color
Worse light
Smaller working channel
Limited selection of scopes and instruments
Endoscope-assisted shunt placement when

they are easily passed through the ventricular catheter as described in the following
section.
Other beroptic scopes options include:
NeuroNavigational produces a few disposable exible endoscopes with outer diameters
of 1.2, 2.3, and 4.6 mm. These scopes feature
a rigid design with 10,000 pixel optical bers
for ne image resolution. The camera is
elongated and not directly connected to the
endoscope body, which signicantly decreases
the weight of the instrument and makes its
pencil-like body easy to handle. The 0.9-mm
Neuroview Fiberscope is one of the thinnest
endoscopes available today. The 1.2-mm
scope has only irrigation channels and is
designed to t into the lumen of shunt
catheters. The 1.5-mm scope has a 0.49-mm
working channel, and the 2.3-mm scope has
a 1-mm working channel and a deected tip,
which improves maneuverability and access
in conned areas. The 4.6-mm scope has two
working channels.
A few lightweight endoscopes are oered
by Clarus (Minneapolis, MN), including the
rigid beroptic Channel endoscope. It is quite
light and can be held as a pen (Fig. 8). The
Channel endoscope is available in two
lengths: 13.0 and 21.6 cm. The working
channels may be either 3.15 or 2.15 mm in
diameter. Irrigation is available through the
irrigation port, with outow through the
working channel. Other beroptic endoscopes
from Clarus include the MurphyScope and
27
improve the surgical outcome once the surgeon

is within the ventricle, because no specic place
within the ventricle has been proven superior to
any other [5].
Syringomyeloscope
Fig 8. Light-weight beroptic endoscope (Clarus,

Minneapolis, MN).
the NeuroPen (designed for shunt insertion).

Three variants of the MurphyScope are
available: 2.3-mm scopes with a bayonet-like
malleable 40 angle, a straight 50 curved
endoscope, and a 1.4-mm curved malleable
scope.
Comparison of rigid and exible endoscopes
Rigid and exible endoscopes each have
advantages and disadvantages. Table 2 highlights
some of the pros and cons of each.
Endoscopes for shunt placement
It has been hypothesized that the use of a smalldiameter endoscope as a stylet within the ventricular catheter may allow more accurate catheter
positioning within the ventricle, which should
decrease the number of proximal shunt malfunctions and revisions [5,10]. This theory has never
been proven. Nevertheless, several types of beroptic endoscopes are commercially available for
ventricular catheter placement. NeuroNavigational, Clarus, Cordis (Miami, FL), and Codman
(Johnson & Johnson, New Brunswick, NJ) all
produce semirigid lightweight scopes that have
10,000 optic bers, providing adequate visualization of intraventricular structures. The NeuroNavigational 1.2-mm Neuroview endoscope,
a disposable unit, can be used in conjunction with
several ventricular catheter modications, such as
the commonly used Innervision catheter from PS
Medical (Goleta, CA). The Neuroview endoscope
can be used as a stylet to insert the catheter. Note
that although using this technique to insert the
catheter may decrease the possibility of completely
missing the ventricle, it does not necessarily
Endoscopes can be used in surgical treatment

of syringomyelia. Aesculap produces a exible
syringomyeloscope with an outer diameter of 2.3
mm. This scope has a working channel 1 mm in
diameter, which may be suitable for 400-lm laser
bers. This endoscope is mainly used for intracavity inspection and dissection after the lesion is
accessed via a laminectomy. A syringomyeloscope
may also be useful in the treatment of chronic
subdural hematomas.
Coagulation
Good hemostasis is essential in neuroendoscopy for adequate visualization of structures as
well as safety. Hemostasis is achieved either with
cautery systems (monopolar or bipolar) or with
lasers (discussed in the next section).
Monopolar cautery is the most direct and
commonly used method to achieve hemostasis
(Fig. 9). The Bugbee Wire (USA) is probably the
simplest monopolar cautery, commonly used by
urologists as well as neuroendoscopists. The
Bugbee Wire without applied cautery current
can also be used as a probe to feel tissue or
membrane (eg, palpation of the oor of the
third ventricle before fenestration) or to pierce
a membrane (eg, third ventriculostomy, fenestration of cyst).
Fig 9. Various coagulation probes.
28
Slightly more sophisticated monopolar cauteries are available from Cook and Codman &
Shurtle. The Cook system is a number 3 French
retractable instrument that is kept within a protective sheath while passing through the working
channel. The surgeon squeezes the loop handle to
expose the tip as needed. Both round ball and
pencil-point tips are commercially available. The
Cook electrodes may be used through rigid and
exible endoscopes. Codman & Shurtle designed
the Micro Endoscopic Electrode (ME2), which
relies on a retraction mechanism similar to the
Cook cautery.
Because there can be a problem with standard
monopolar cauteries of burnt tissue adhering to
the tip of the instrument, Heilman and Cohen [2]
invented the saline torch monopolar cautery.
With this tool, the electric current is directed
through a jet of saline, allowing the surgeon to
dissect and coagulate structures without direct
contact, even when completely immersed in
cerebrospinal uid (CSF).
Bipolar coagulation is a more precise way of
achieving hemostasis with less scatter of current
compared with a monopolar cautery. The simplest
bipolar electrode is a fork electrode (eg, Aesculap
2.1-mm fork electrode). When using a fork
electrode, however, the surgeon cannot pick up
tissue, which makes its use somewhat limited.
Codman & Shurtle therefore introduced grasping bipolar forceps 2.5-mm in diameter. These
forceps are suitable for coagulation of vessels of
no more than 2 mm in diameter. NeuroNavigational oers both 1.0- and 2.4-mm exible bipolar
electrodes. They can be used through rigid and
exible scopes. An advantage of these electrodes is
a built-in irrigating channel to decrease adherence
of the tissue to the wires. Finally, Clarus
manufactures a bipolar cautery that looks like
a pencil and has a 30 tip angle. This allows the
surgeon better visibility of the cauterized object.
Fig 10. Endoscopic microsurgical instruments (Aesculap,

Center Valley, Pennsylvania).
occluded ventricular catheter overgrown by

choroid plexus (Fig. 10).
Cup forceps, two- or three-pronged forceps,
and exible loop retrievals are designed to
obtain tissue for biopsy. The design of these
instruments facilitates removal or retrieval of
tissue fragments. Some of these instruments
are intended for single use. Aesculap manufactures similar instruments in exible and
rigid variants. The malleable steering forceps
are available from Clarus.
Another helpful instrument is the Fogarty
balloon (Fig. 11). Usually surgeons use
number 2 or 3 French catheters, depending
on the diameter of the working channel. The
Fogarty balloon is used by many surgeons to
dilate the opening in the oor of the third
ventricle during an endoscopic third ventriculostomy. The advantage of this technique
is that the inated balloon compresses
Instruments
Various grasping forceps are commercially
available from dierent manufacturers:
Cook produces exible forceps designed for
use through rigid and exible scopes.
Rat tooth, alligator, and mouse tooth forceps
can be used for dissection, enlargement of an
opening, and pulling tissues. On a few occasions, the authors have used them with
endoscopic scissors to cut and remove an
Fig 11. Balloon catheters.
surrounding tissues, constricting bleeding

vessels.
In the past few years, there have been case
reports of accidental, often fatal, injuries to
the vessels of the basilar artery bifurcation
complex [6]. Recently, a new device for
perforation of the oor of the third ventricle
(no through perforator) was developed by
Grotenhuis and became commercially available from Synergetics (USA). It consists of
a 3-mm thick hollow rod with a sharp cutting
edge. After suction is applied, the oor of the
third ventricle jumps up and sticks to the
tip of the perforator, leaving the interpeduncular vessels below. The device is turned
gently to create a round 3-mm diameter hole.
The device seems to improve safety and lower
the risks involved in perforation of the oor
of the third ventricle.
Lasers
The laser beam is a form of energy used in
surgery to cut, coagulate, vaporize, and dissect
tissues. Its applications are quite similar to those
of electricity; however, the nature of the laser is
uniquely dierent from that of electricity.
Laser is an acronym for light amplication by
stimulated emission of radiation. Laser energy can
be generated in a plasma tube by a powerful
electromagnetic eld. Application of this eld to
some type of gas molecules, such as carbon
dioxide or argon, leads to their excitation.
Electrons from the excited molecules start moving
from one energy state to another. When these
molecules drop from an excited to a resting state,
a photon (a unit of energy that has a characteristic
wavelength) is released. The release of a photon
energy unit is called uorescence. Once a photon
hits a neighboring molecule, another photon is
released, the wavelength of which is in phase with
the rst photon. This process, called stimulated
emission, is self-perpetuating, because neighboring molecules continue to bombard each other
with photons. The mirrors on both tips of the
laser tube reect the emitted photons, increasing the level of movement and energy within
the tube until, nally, the beam emerges. On
emergence, the beam has the following three
features:
The beam is coherent, because all photons are
released in the same phase.
The beam is monochromatic, because the
material emits only a single wavelength.
29
The beam is collimated, because the waves

emerge parallel to each other.
As a result of these features, a lens is able to
focus the light to a ne point, with an extremely
high concentration of power density [4,8].
Biologic eects of laser
Application of laser to tissue causes an instantaneous increase in temperature, leading to
vaporization of the extracellular uid and explosion of the cell. Four concentric zones of impact
appear as a result of the interaction between laser
and tissue:
Central zoneno cells remain, only burnt
debris.
Vacuolated cells zonethe tissue is composed
of nonviable vacuolated cells that retain some
structure.
Edematous zonesome cells in this area are
dead, but most of the cells are viable with
increased intracellular water content.
Reversible ultramicroscopic changes zone
although no alterations are seen under
light microscopy, some minor histochemical
changes are present.
The rst two zones are irreversibly dead. The
latter two zones are viable and recoverable [4].
Types of medical lasers
Many lasers are available for clinical use, but
only two are commonly used in neuroendoscopy
because of their ability to work through water
and transmit through the miniature beroptic
cables:
neodymium:yttrium-aluminum-garnet
(Nd:YAG) and potassium-tetanyl-phosphate
(KTP) [4,9,11,12].
Initially, lasers were popular in neuroendoscopy, especially to make the opening in an
endoscopic third ventriculostomy [12]. After
a few instances of injury to the vessels in the
interpeduncular fossa [6], however, laser use
signicantly declined for safety reasons. Nevertheless, the Nd:YAG laser and, especially, the KTP
laser may be useful in endoscopic tumor removal
and cyst fenestration. The tissue penetration of
the KTP laser is less than with the Nd:YAG laser;
its visible light makes it easier to handle, and it is
less dependent on the tissue pigmentation, which
makes it suitable for dissection of colloid cysts.
Scope holders
Neuroendoscopy can be done either freehand
by one or two surgeons (one navigating the scope
30
Table 3
Advantages and disadvantages of a scope holder
Advantages
Disadvantages
Freehand
With rigid holder
More freedom of movement,

particularly when conguration
needs to be frequently or
continuously changes (eg, tumor
removal, inspection of
subarachnoid cisterns)
Surgeon has the use of both hands
More fatigue for surgeon

Greater risk of accidental
movements
and the other working with instruments) or using

a rigid holder. Table 3 describes the advantages
and disadvantages of each technique.
Today, there are many scope holders available,
such as the Leyla self-retaining retractor, the
exible Greenberg retractor, the Neuroview Scope
Holder, and the Unitrac pneumatic holder produced by Aesculap (Fig. 12). The authors have
found the latter system particularly helpful, even
when operating freehand, because the holder can
be used as an easily adjustable hand rest. The
Minimizes accidental movements

and tremor
More static (inexible)
Inconvenient when frequent
repositioning is needed
device uses the same pressurized gas that powers

pneumatic drills.
Summary
In sum, understanding some of the basic
principles of endoscopy and awareness of available
resources can potentially be of considerable help to
experienced neurosurgeons as well as beginners in
selection of the most appropriate tools for dierent
procedures and making cost-eective choices when
browsing through multiple commercial advertisements and purchasing new equipment.
Although numerous advantages in science and
industry have made it possible to oer a wide
variety of neuroendoscopes and tools, we believe
the major achievements in this eld are yet to
occur. This particularly refers to the development
of smaller beroptic scopes with better image
quality and three-dimensional endoscopes and to
the invention of more ecient tools for endoscopic tumor removal with the same degree of
safety as in open surgery.
References
Fig 12. Unitrac scope holder (Aesculap, Center Valley,

Pennsylvania).
[1] Cohen AR. Endoscopic ventricular surgery. Pediatr

Neurosurg 1993;19:12734.
[2] Heilman CB, Cohen AR. Endoscopic ventricular
fenestration using a saline torch.. J Neurosurg
1991;74:2249.
[3] Hulka JF, Reich H. Light: optics and television.
In: Textbook of laparoscopy. Philadelphia: WB
Saunders; 1994. p. 921.
[4] Hulka JF, Reich H. Power: electricity and laser.
In: Textbook of laparoscopy. Philadelphia: WB
Saunders; 1994. p. 3943.
[5] Levy ML, Lavine SD, Mendel E, McComb JG. The
endoscopic stylet: technical notes. Neurosurgery
1994;35:3356.

[6] McLaughlin MR, Wahlig JB, Kaufmann AM,
Albright AL. Traumatic basilar aneurysm after
endoscopic third ventriculostomy: case report [see
comments]. Neurosurgery 1997;41:14004.
[7] Mixter WJ. Ventriculoscopy and puncture of the
oor of the third ventricle. Boston Med Surg J
1923;188:2778.
[8] Nobles A. The physics of neuroendoscopic systems
and the instrumentation. In: Jimenez DF, editor.
Intracranial endoscopic neurosurgery. Park Ridge,
IL: American Association of Neurological Surgeons; 1998. p. 112.
[9] Shiau JSC, King WA. Neuroendoscopes and instruments. In: Jimenez DF, editor. Intracranial endo-
31
scopic neurosurgery. Park Ridge, IL: American

Association of Neurological Surgeons; 1998. p. 1327.
[10] Theodosopoulos PV, Abosch A, McDermott MW.
Intraoperative ber-optic endoscopy for ventricular
catheter insertion. Can J Neurol Sci 2001;28:
5660.
[11] Wharen REJ, Anderson RE, Scheithauer B, Sundt
TM. The Nd:YAG laser in neurosurgery. Part 1.
Laboratory investigations: dose-related biological
response of neural tissue. J Neurosurg 1984;60:
5319.
[12] Wood FA. Endoscopic laser third ventriculostomy
[letter, comment]. N Engl J Med 1993;329:
2078.
The anatomy of the ventricular system

David G. McLone, MD, PhD
Childrens Memorial Hospital, 2300 Childrens Plaza, #28, Chicago, IL 60614, USA
An understanding of the form and detail of the

ventricular system is an important point of
departure when beginning to learn neuroendoscopy. Knowledge of the form and detail allows
planning of the optimal entry point to enable
viewing desired structures. This information also
acts as a road map to direct the endoscopist to
targets. It is essential to know what is just beyond
the structure at the tip of the scope. Pathologic
processes, such as hemorrhage, ventriculitis, and
hydrocephalus, may alter the appearance of the
ependymal surface of the ventricle but usually do
not alter structures needed to navigate the
ventricular system. Obviously, knowledge of the
function of structures within the eld of interest is
important to you and the patient.
Detailed function of the structures forming the
ventricular systems and the consequences of
injuring them are beyond the scope of this article.
The author assumes the reader understands the
functional anatomy.
What is seen on the monitor depends on the
orientation at entry to the ventricle. Structures at
the immediate penetration point are dicult to see
because they remain behind the lens. Noting the
orientation of the patients head before draping is
critical for the endoscopists interpretation of the
structures and direction within the ventricular
system. Initially, it may be helpful to have the
patients head straight brow up or lateral to limit
the mental gymnastics. As one develops a threedimensional concept of the anatomy, orientation
becomes easier.
As with all neurosurgical procedures, complications are always a possibility. Only with the use
of this information and practice with the scopes
E-mail address: dmclone@nwu.edu
within the ventricular system will this become

a safe and useful procedure.
Development
The cerebral ventricular system at rst seems
complex; however, when understood from the
point of the developmental anatomy, it is much
simpler. Telencephalic vesicles bud from the
prosencephalon as symmetric bilateral spheres at
the extreme rostral end of the embryo. The
openings into the diencephalic vesicle, the future
third ventricle, become the foramen of Monro.
Choroid plexus develops along the dorsal raphae
of the diencephalic vesicle and splits at the
foramen, extending into each telencephalic vesicle.
From this point, the rapidly developing cerebral
hemispheres draw a group of structures that
originate near the foramen out into the ventricular
system. These structures become the endoscopic
road maps for the lateral ventricles. Structures
that originate near the foramen and end up in the
temporal horn include the caudate nucleus,
hippocampus, fornix, choroidal ssure, and choroid plexus. In lower mammals, the hippocampal
structures move posteriorly along the paramedian
with advancing phylogeny to end up in primitive
forms near the foramen and, in rodents, over the
posterior third ventricle.
As the hippocampal structures move posteriorly, the single large bundle of bers is drawn out
and trails behind as the fornix. Columns of the
fornix fuse together as they pass over the foramen
of Monro. At the anterior commissure, they again
separate and each side splits into a precommissural
tract and postcommissural tract. These two tracts
pass ventrally and laterally, one to the septal area
and the other to the mamillary body, thalamus,
and midbrain.
doi:10.1016/S1042-3680(03)00073-1
34
D.G. McLone / Neurosurg Clin N Am 15 (2004) 3338
In primates, especially man, the parietal lobe

has evolved from a structure principally for
sensory function of the face to a structure serving
intellectual functions. This has caused a large
expansion of the parietal lobe driving the temporal lobe with its contents inferiorly and anteriorly.
Rapid enlargement of the cortex of the frontal,
parietal, and temporal lobes opercularizes the
insular cortex.
The choroidal ssure with the choroid plexus
wraps around the thalamus, passing into the
temporal horn along its medial surface. The
hippocampus moves with the choroidal ssure
into the medial temporal lobe. This process
creates a lateral ventricular system that has the
form of a rams horn. The head of the caudate
nucleus begins anteriorly and laterally in the
anterior horn of the lateral ventricle. The body
and tail of the caudate wrap around the lateral
portion of the thalamus, again like a rams horn.
The venous drainage comes out of the caudate
nucleus and thalamus in the groove between the
caudate and thalamus, the thalamostriate vein,
and ends at the foramen of Monro, where it joins
the internal cerebral veins in the roof of the third
ventricle.
The corpus callosum originates as a commissural bundle anterior to the foramen of Monro
and then expands superiorly and posteriorly as
the cerebral hemispheres blossom. Obviously, this
large midline commissure cannot continue into
the temporal lobe because of diencephalic structures occupying the midline. Thus, the caudal end
of the corpus callosum, the splenium, comes to
rest at the posterior end of the third ventricle
above the habenular commissure, the pineal
gland, and the Galenic venous system.
The columns of the fornix are connected to the
corpus callosum by paired thin membranes. As
the hippocampus moves posteriorly with the
corpus callosum, these membranes are drawn
out as the septum pellucidum. This creates two
potential cavities. One, located between the leaves
of the septum, is the cavum septum pellucidum
and its posterior extension, the cavum verge.
Below the fornix, the arachnoid and choroidal
ssures are folded back over the roof of the third
ventricle, creating the second potential space, the
cavum vallum interpositum. In children with
poorly developed brains, these potential cavities
can be large spaces that can either obstruct the
ventricular system and cause hydrocephalus or
participate with the ventricles in the expansion
caused by hydrocephalus. These cavities are also
the sites where missed directed shunts or endoscopes can end up, which presents a confusing
picture to the endoscopist.
The diencephalic vesicle becomes the third
ventricle, and the mesencephalic vesicle, which is
the largest portion of the embryonic ventricular
system, becomes the relatively small aqueduct.
The fourth ventricle is created by the development of the cerebellum from the lip of the
rhombencephalon. This lip runs from side to side
in the coronal plane. Also developing along the
edge of this lip is the choroid plexus of the fourth
ventricle. The expansion of the cerebellum rolls
the choroid plexus under the caudal edge. Thus,
the choroid plexus runs from side to side from one
foramen of Luschka to the other in the inferior
roof of the fourth ventricle.
Anatomy of the lateral ventricles

In considering the lateral ventricular system, it
is helpful to use the designation of the portions of
the ventricle that was employed by neurosurgeons
when pneumoencephalography was in common
use.
Portion 1: the frontal tip of the lateral ventricle
to the anterior edge of the foramen of
Monro
Portion 2: the anterior edge of the foramen of
Monro to its posterior edge
Portion 3: from the posterior edge of the
foramen of Monro to the posterior edge of
the thalamus
Portion 4: the trigone of the lateral ventricle
from the posterior edge of the thalamus to
the beginnings of the occipital and temporal
horns
Portion 5: the occipital horn
Portion 6: the temporal horn
Portion 1 of the lateral ventricle, with the
occipital horn and the tip of the temporal horn,
are portions of the ventricle without choroid
plexus. The absence of choroid plexus and its
immediate proximity to the foramen of Monro
are the major identifying characteristics of this
portion. Medially is the septum with the septal
vein. Inferiorly is the septal area, and laterally is
the head of the caudate nucleus. The anterior limit
and the roof are formed by the anterior bers of
the genu of the corpus callosum. These bers pass
obliquely anterior and lateral to the frontal lobe;
thus, the anterior horn slopes anterior lateral.
Portion 2 of the lateral ventricle is essentially

the foramen of Monro and is the structure for
orientation within the anterior lateral ventricle.
The anterior and superior edge of the foramen is
the fornix. Above the fornix is the septum
pellucidum, which forms the medial wall of this
portion of the ventricle and extends superior to
the corpus callosum. The corpus callosum forms
the roof, which joins laterally with the head and the
beginning of the body of the caudate nucleus.
The anterior thalamus forms the oor. The
anterior inferior edge of the foramen is the septal
area. Turning in the groove between the caudate
and thalamus is the thalamostriate vein, which
enters the foramen of Monro to join the internal
cerebral vein. With the thalamostriate vein, the
choroids plexus passes into the foramen and
forms the posterior boarder.
Portion 3 of the lateral ventricle is that portion
of the ventricle located above the thalamus. Its
main identifying characteristics are that it is
the only portion with the choroid plexus on
the medial oor and it is immediately behind the
foramen of Monro. The presence of the choroid
plexus on the medial oor orients the endoscopist
when the approach is posterior from the trigone.
The body of the caudate and the roof form the
lateral wall by the corpus callosum. The thalamostriate vein is usually a prominent vessel
running in the groove between the thalamus and
caudate. The septum diminishes rapidly posteriorly so that the most caudal part of the medial
surface of portion 3 is formed by the choroid
plexus, which covers the fornix. When approaching posteriorly, the choroid plexus and the
thalamostriate vein are the road map to portions
1 and 2. The anterior thalamus is at the foramen
of Monro, and the fornix and choroid plexus are
closely applied to the thalamic surface in portion
3. The distance to the contralateral thalamus
through the septum pellucidum is short when the
contralateral ventricle is not dilated. Septostomy
directed posterior to the foramen and at a small
contralateral ventricle risks injury to the contralateral thalamus and internal capsule, especially
the genu.
Portion 4 of the lateral ventricle is the
conuence of portions 3, 5, and 6, often referred
to as the trigone. It lies behind the thalamus and
contains the glomus of the choroid plexus.
Medially, the visual radiations pass to occipital
pole, whereas ventrally and laterally is the white
matter of the occipital and temporal lobes. It is
important to be familiar with the characteristics of
35
the entry into portions 3 and 6 from this portion

because this determines whether you enter into the
frontal or temporal horns.
Portion 5 of the lateral ventricle is the occipital
pole and the most variable in size of all the
ventricular portions. Again, it is one of the
portions without a choroid plexus. Medially and
inferiorly are the visual radiations passing to the
medial occipital lobe. Superiorly and laterally are
the bers from the splenium of the corpus callosum passing to the lateral and inferior occipital lobe.
Portion 6, the nal portion of the lateral
ventricular system, is located within the temporal
lobe. When approaching the trigone posteriorly,
the temporal horn is slightly more lateral than
portion 3 and the choroid plexus is on the medial
roof rather than on the oor. Medially and
superiorly are the choroid plexus and hippocampal structures. The deep white matter tracts of the
temporal lobe form the lateral wall.
Malformations of the lateral ventricles
Many of the malformations of the cerebral
hemispheres are associated with hydrocephalus,
and ventriculostomy may be a treatment option.
Hydrocephalus is a sign that the malformation has
created an underlying obstruction causing problems with the circulation of cerebrospinal uid.
As in all good medical practice, a through
review of the patients history and physical
examination and study of the neuroimages before
the procedure are essential.
Holoprosencephaly results from a lack of division of the prosencephalon into the two telencephalic vesicles. Hydrocephalus with an expanding
large dorsal cyst may be present, requiring management of the hydrocephalus. Abnormalities of
the face, such as extreme hypotelorism, reveal the
underlying brain malformation. Children with this
malformation lack intellectual development. Holoprosencephaly is common with trisomy.
Schizencephaly is a defect in the cerebral
mantle. It is most common in the third and fourth
portions of the ventricular system but can occur in
the other portions. The ventricular system may
communicate directly with the subarachnoid
space, which is referred to as open-lipped schizencephaly. In closed-lip schizencephaly, the lips
of the cortical defect are fused. This may be an
incidental nding in normal children or associated
with seizures or intellectual delay. Hydrocephalus
is occasionally present.
36
Hydranencephaly is an extreme form of

schizencephaly in which most of the cerebral
hemispheres are absent bilaterally, precluding
intellectual development. Hydrocephalus is often
present. It is essential to distinguish this entity
from maximal hydrocephalus, because the outlook is quite dierent.
Porencephaly is a cavity within the paraventricular white matter. Most communicate with the
ventricular system. They may be congenital or
occur secondary to injury, either vascular or
traumatic. Premature infants with intraventricular
hemorrhage often have such a cavity located in
the frontal lobe, but it may be located anywhere
along the germinal matrix. Poorly controlled
hydrocephalus can cause a porencephalic cyst to
develop along the shunt tract.
Diverticula of the ventricular system are rare,
but it is essential that they be recognized for what
they are. Most often, they occur in cases of
hydrocephalus in which the supra- and infratentorial compartments are isolated from each other.
Dierential pressures allow a portion of the
medial lateral ventricle to herniate into the parasellar or quadrigeminal cisterns. These tics
usually arise along the choroidal ssure from
portions 3 and 4.
Agenesis of the corpus callosum results in
a major decrease in the amount of white matter
in each cerebral hemisphere. Because the third
ventricle is not constrained by the corpus
callosum, the fornix moves laterally, causing
the frontal tips of the lateral ventricles to
diverge. Portions 3, 4, and 5 are usually large,
so-called copocephaly, expanding to ll the
space left by the absent white matter. Only rarely
is agenesis of the corpus callosum accompanied
by hydrocephalus.
Ependymal and choroidal cysts are not uncommon within the lateral ventricles. They are
usually small and can spontaneously disappear.
Occasionally, they are large and obstruct the
ventricular pathway, causing hydrocephalus proximal to the obstruction. These cysts can be
managed by endoscopic fenestration.
Loculations of the ventricular system are rare.
What is common is for paraventricular cysts to
coalesce and expand, producing what seems to be
a loculated portion of the ventricle. When entering
these cavities, the absence of ventricular elements,
choroid plexus for intense, reveals their nature.
Biopsy of the wall reveals it to be composed of white
matter containing myelinated axons. Hydrocephalus is almost invariably present. The hydrocephalus
has usually been complicated by ventriculitis or

ventricular hemorrhage.
Anatomy of the third ventricle
A clear understanding of the structures that
comprise the limits of this ventricle is extremely
important to the neuroendoscopist. Even more
importantly, the structures just beyond the walls
of this ventricle must be clearly understood.
Anteriorly, the two fornices come close together in the midline with only a small groove between them, forming the anterior and superior
margin of the third ventricle and of the foramen
of Monro. As the columns pass ventrally, each
splits into an anterior and posterior limb over
the anterior commissure and moves laterally, with
the anterior limb ending in the septal area and the
posterior limb ending in the mamillary bodies.
Ventral to the anterior commissure is the lamina
terminalis, which ends in the optic chiasm. The superior optic chiasm actually forms a portion of the
anterior third ventricle.
The roof of the third ventricle is composed of
the choroid plexus, internal cerebral veins, and
columns of the fornix. As the most caudal portion
of the fornix, the columns move laterally, and an
interconnecting band of tissue, the mbria, forms
the posterior superior roof of the third ventricle.
Below these structures are the choroid plexus,
internal cerebral veins, and velum interpositum.
The pineal recess and the pineal gland form the
posterior limits of the third ventricle superiorly.
Below the recess are the posterior commissure and
the aqueduct of Sylvius.
The paired mamillary bodies and the hypothalamus form the oor of the third ventricle.
The oor slopes downward from the mamillary bodies into the infundibulum and pituitary
recess. The mamillary bodies are usually prominent structures that are easily seen. The oor of
the ventricle immediately in front of the mamillary
bodies is often not transparent, and structures
under it cannot be seen. Usually, when hydrocephalus is present, the oor is thinned and the
vascular structures are visible. More anteriorly,
the oor of the hypothalamus may be slightly
pigmented. Below, exterior to the third ventricle
is the posterior clinoid, and between the clinoid
and the interpeduncular fossa runs the membrane of Liliequist. If one passes through the
oor of the third ventricle in the midline anterior
to the mamillary bodies and the basilar artery
tip, the prepontine cistern is entered. As one
move more rostrally and passes through the oor

of the hypothalamus, the suprasellar cisterns
anterior to the membrane of Liliequist are entered.
The oor of the third ventricle is the hypothalamus, and injury can lead to endocrinopathies.
Moving too far forward may cause the scope to
pin the oor against the posterior clinoid.
There are three recesses that can increase
greatly in size during hydrocephalus. They are
the optic, pituitary, and pineal recesses. When
distended, the optic recess leaves the anterior
commissure as an obvious structure in the
anterior third ventricle. The pituitary recess can
become large and thinned to the point of being
almost transparent, revealing the structures in the
parasellar cisterns. The pineal recess expands into
and often lls the quadrigeminal cistern.
The lateral walls of the third ventricle are
formed by the medial surfaces of the thalami.
Passing posteriorly along the superior margin of
the medial thalamus is the stria terminalis, which
ends in the habenular nucleus and commissure.
Because this structure is often obscured by the
choroid plexus of the third ventricle, it is usually
not visible through the ventriculoscope. The
massa intermedia usually occupies the middle
portion of the third ventricle. This is a noncommissural structure of variable size connecting
the medial portions of the thalami. In the Chiari II
malformation, the massa intermedia occupies
a major portion of the third ventricle.
Some important vascular structures are located
external to the third ventricle in the subarachnoid
space. Below the posterior oor of the third
ventricle in the interpeduncular cistern are the tip
of the basilar artery and the two mesencephalic
portions of the posterior cerebral arteries. There
are also a number of smaller penetrating arteries
that branch from these main trunks. Anteriorly,
just above the optic chiasm are the anterior cerebral
arteries and the anterior communicating artery.
Running up the anterior surface of the lamina
terminalis are the paired anterior cerebral arteries.
Malformations of the third ventricle
Severe hydrocephalus can signicantly alter
the appearance of the structures at the foramen of
Monro. Even in these cases, the basic structures of
the road maps are available to locate the area for
ventriculostomy.
In the Chiari II malformation, the third
ventricle is usually remarkably dierent from
normal. A large portion is occupied by the massa
37
intermedia, but, more importantly, the oor is

quite dierent. The distance between the mamillary bodies and the pituitary recess is extremely
short, and the oor is thickened. This malformation often precludes any attempt at fenestration of
the oor, especially in infants.
Agenesis of the corpus callosum allows the
third ventricle to expand superiorly between the
cerebral hemispheres. The foramina of Monro are
displaced laterally.
Anatomy of the cerebral aqueduct
The aqueduct is the small portion of the
ventricular system that connects the third and
fourth ventricles. Superior to the aqueduct is the
quadrigeminal plate. Inferiorly is the tegmentum
of the midbrain. Nuclei of the third cranial nerve
are located bilaterally near the anterior portion of
the aqueduct.
Anatomy of the fourth ventricle
The normal fourth ventricle is a dicult target
for endoscopy. Its size and orientation do not
allow easy access. More importantly, the oor of
this ventricle is the brain stem. The sixth cranial
nerve nucleus, the tract of the seventh cranial
nerve, and the medial longitudinal fasciculus are
located supercially just below this oor. Although the cerebellum tolerates penetration,
minor trauma to the oor inicts signicant neurologic decits. In pathologic conditions in which
the fourth ventricle is enlarged, however, access is
possible.
The oor of the fourth ventricle is diamond
shaped. The anterior point begins at the aqueduct,
the lateral points hook slightly posterior to form the
lateral recesses, and the posterior point ends at
the obex. The normal oor is relatively at with
some identifying structures in it. Running in the
midline is the median sulcus, and near the midpoint, the stria medullaris runs from the sulcus
toward each lateral sulcus. The facial colliculus
is located on either side near midline just rostral
to the stria medullaris.
The walls of the fourth ventricle are formed by
the superior and inferior cerebellar peduncles. The
fourth ventricle has a pyramid shape, with the apex
tilted slightly posterior. The anterior and posterior
medullary velum forms the roof. The posterior side
of the pyramid is convex because of the cerebellar
vermis bulging inward. The vermis also splits the
apex into two recesses: the posterior superior
38
recesses. The midline apex is the vestigium. The

choroid plexus is located on the posterior surface of
the fourth ventricle and runs in the coronal plane
from one lateral recess to the other. Foramina that
drain the fourth ventricle are located bilaterally in
lateral recesses: the foramina of Luschka and, in
the midline, the foramen of Magendie. The
choroidal ssure runs with the choroid plexus from
one lateral recess to the other, separating the
inferior cerebellum from the brain stem.
Malformations of the fourth ventricle
Dandy Walker cysts represent a group of
anomalies in which a portion of or the entire
cerebellar vermis is absent and the fourth ventricle
communicates with or becomes a large cyst. When
it occurs early in development, the torcula is
located high, often near the midparietal region.
Hydrocephalus may or may not be present. In its
most severe form, it is associated with other brain
malformations, such as agenesis of the corpus
callosum, and intellectual delay is common in this
form.
Loculation of the fourth ventricle, or a trapped
fourth ventricle, like loculations in the lateral
ventricles, is usually a complication of hydrocephalus. Ventriculitis is the most common cause.
The Chiari II malformation is a malformation
that aects the entire brain. It has a major impact
on the contents of the posterior fossa. Because the
size of the embryonic posterior fossa is much too
small to contain the rhombencephalic and metencephalic structures that will develop, derivatives of
these embryonic portions of the brain extrude out
of the posterior fossa. A portion of the (occasionally the entire) cerebellum and fourth ventricle is
displaced into the cervical canal. In other individuals, the decient tentorium allows a major
portion of the cerebellum and fourth ventricle to
herniate upward into the middle fossa between the
cerebral hemispheres. When the ventricular shunt
fails or the fourth ventricle becomes trapped or
isolated, the fourth ventricle expands, increasing
the pressure on the surrounding hindbrain. This is
usually manifested as a dramatic increase in the
patients symptoms. When the isolated fourth
ventricle in the Chiari II malformation expands
superiorly, endoscopic access is available from
above in a plane parallel to the oor of the fourth
ventricle. This aords the opportunity to fenestrate or shunt the fourth ventricle endoscopically
without endangering the oor of the fourth
ventricle.
Summary
The embryology of the ventricular development of the brain assists in understanding the nal
relations between structures forming these cavities. An accurate concept of this anatomy allows
the endoscopist to maneuver within the ventricular system.
Selecting patients for endoscopic third ventriculostomy

Harold L. Rekate, MD
Pediatric Neurosurgery, Barrow Neurological Institute, 2910 North Third Avenue, Phoenix, AZ 85013, USA
At the annual meeting of the American

Association of Neurological Surgeons Meeting
in 1995, Professor Fred Epstein was called to
discuss our paper on the use of endoscopic third
ventriculostomy (ETV) as the ultimate treatment
for the slit ventricle syndrome (SVS) as presented by Dr Jonathan Baskin [1]. Dr Epstein
was surprised to learn that we were viewing all
previously shunted patients with shunt problems
as potential candidates for the procedure and
wondered aloud whether patients with normal
pressure hydrocephalus (NPH) might not be
considered candidates for ETV.
The thought process leading to the following
presentation regarding who is and who is not
a candidate for ETV was largely generated by that
discussion. Can we know beforehand who will
and who will not benet from an ETV? What are
the absolute contraindications for performing the
procedure? What are the relative contraindications? How can we analyze the risk-benet ratio
for ETV in various clinical settings? For the most
part, dening the role of ETV in the overall
management of hydrocephalus must await large,
multicenter, randomized, prospective trials. In
the interim, I will attempt to dene a rational
approach to the problem in partial answer to the
posed questions.
Anatomy and biophysics of the cerebrospinal uid
as it relates to third ventriculostomy
The credit for dening the basic pathophysiologic mechanisms in hydrocephalus goes to Walter
Dandy and his colleagues at Johns Hopkins
University for work done between the second
and fourth decades of the twentieth century.
E-mail address: Harold.rekate@bnaneuro.net
There had been many previous pathologic descriptions of the eects of hydrocephalus on
cadaveric material. Dandy and Blackfan [2] studied laboratory animals and patients with hydrocephalus by injecting supravital dyes into the
lateral ventricles of the subjects. They then performed lumbar punctures to determine whether
the dye could be recovered in the spinal subarachnoid spaces (SSASs) [3]. Based on this
information, Dandy and Blackfan classied hydrocephalus into communicating (the dye was
recoverable) and obstructive or noncommunicating (the dye was not recoverable). As a result of
these ndings, they recommended attempting to
create a communication between the third ventricle and the subarachnoid spaces by performing an
open craniotomy and initially resecting one of the
optic nerves.
By applying Dandys classication, patients
with noncommunicating hydrocephalus would be
considered candidates for ETV to create an
internal bypass for the treatment of hydrocephalus. Over the years, there has been a tendency to
equate noncommunicating hydrocephalus with
triventricular hydrocephalus, which could be
diagnosed by air studies and, subsequently, by
CT and MRI. In 1960, Ransoho and Epstein [4]
voiced their objection to the Dandy classication.
They believed that all hydrocephalus was obstructive and preferred to classify the condition into
intraventricular obstructive hydrocephalus and
extraventricular obstructive hydrocephalus. The
latter would be consistent with Dandys communicating hydrocephalus. Based on this discussion,
many patients with communicating hydrocephalus are good candidates for ETV.
With current imaging technologies, it is usually
possible to determine the actual site of the
obstruction and to plan treatment to address the
specic pathophysiologic mechanisms involved. In
doi:10.1016/S1042-3680(03)00074-3
40
H.L. Rekate / Neurosurg Clin N Am 15 (2004) 3949
our laboratory, we have measured the amount of

resistance that can be anticipated at each point of
potential obstruction. Because of the presence
of multiple ventricular catheters, we have also
attempted to measure intraventricular pressures in
various components in human beings. Fig. 1 is
a schematic diagram of the anatomy of the
cerebrospinal uid (CSF) pathways in a patient
with aqueductal stenosis as a hydraulic analogue to
an electrical circuit. In normal animals, pressure
dierentials can be measured only at the level of
ow of CSF from the cortical subarachnoid spaces
(CSASs) into the superior sagittal sinus (SSS) [5,6].
For CSF to be absorbed at the level of the SSS,
intracranial pressure (ICP) must be 5 to 7 mm Hg
higher than the pressure in the SSS. Fig. 2 is
a schematic diagram of the physical eect of an
ETV. Essentially, the procedure is an internal
bypass between the third ventricle and the interpeduncular cistern in the CSASs.
Given this paradigm, all patients with obstructions between the third ventricle and the
CSASs are potential candidates for ETV. Not

only can ETV treat hydrocephalus caused by
aqueductal stenosis, but it is at least of theoretic
benet in obstruction of the outlet foramina of
the fourth ventricle and to blockage of CSF ow
at the level of the basal cisterns between the
SSASs and the CSASs. Only when CSF ow is
obstructed at the level of the arachnoid villi or
venous ow from the sagittal sinus is restricted
do patients fail to benet from ETV. These forms
of distal obstruction represent absolute contraindications to ETV.
With contemporary neuroimaging techniques,
it may or may not be possible to determine the
actual site of obstruction to the ow of CSF. The
more denitive the diagnosis of the site of
obstruction, the better is the physicians ability
to predict the outcome of ETV when used to treat
patients with hydrocephalus. Some patients with
hydrocephalus may have an obstruction at more
than one point along the CSF pathway. In spina
bida, hydrocephalus caused by a Chiari II
Fig. 1. Schematic hydraulics of the cerebrospinal uid system with aqueductal stenosis. (Courtesy of Barrow
Neurological Institute.)
41
Fig. 2. Schematic diagram of the eect of performing an endoscopic third ventriculostomy on the circuit diagram of
cerebrospinal uid ow. (Courtesy of Barrow Neurological Institute.)
malformation is associated with as many as four

dierent sites of potential obstruction to the ow
of CSF [7]. Based on patients clinical manifestations at shunt failure, all these potential sites of
obstruction actually occur and any patient may
have more than one site of obstruction.
Patients with hydrocephalus caused by intraventricular or subarachnoid hemorrhage (SAH)
for whatever cause (prematurity, trauma, or
aneurysmal bleeding) usually have a CSF absorptive diculty at the level of the basal cisterns
(SSASs to CSASs). They could have an obstruction at the level of the arachnoid granulations,
however, which would not respond to ETV. Patients who suer signicant infections may develop
hydrocephalus. The most common location is also
the basal cisterns. Patients with an infection may
have an obstruction at the level of the outlet
foramina of the fourth ventricle or at the level of
the arachnoid granulations, however.
Absolute contraindications for endoscopic third

ventriculostomy
Communicating hydrocephalus: obstruction to
the absorption of cerebrospinal uid
ETV should not be performed if CSF is already
owing unimpeded between the ventricles and the
interpeduncular cisterns. It also should not be done
if it is technically impossible to manipulate the
endoscope within the lateral and third ventricles to
a position that allows the oor of the third ventricle
to be visualized. These are the only two absolute
contraindications for performing the procedure. In
the rst instance, ow between the ventricles and
the CSASs is not impeded. Bypassing the aqueduct
of Sylvius, outlet foramina of the fourth ventricle,
and the basal cisterns oers no advantage, because
the point of obstruction is downstream from the
internal bypass. As can be seen from a review of
Fig. 2, this condition occurs only if the arachnoid
42
granulations are obstructed or the pressure of the

SSS is increased markedly.
As an isolated event, obstruction of the
arachnoid granulations is quite rare. Pathologically, congenital absence of these structures is the
cause of at least some cases of benign familial
megalencephaly or external hydrocephalus. Radiologically, this type of obstruction shows mild
to moderate ventriculomegaly and distention of
the CSASs [810]. The terminal absorptive mechanisms can be clogged by blood or by inammatory or tumor cells. This clogging is usually
accompanied by thickening and scarring of the
arachnoid in the basal cisterns. Although the CSF
absorptive failures in acute SAH and meningitis
may be related to an obstruction at this level, the
hydrocephalus usually results from an obstruction
between the SSASs and CSASs, a condition that is
amenable to ETV.
Hydrocephalus caused by obstruction of
absorption to CSF at the level of the arachnoid
villi is never an all-or-none phenomenon. It is
pressure dependent in that the presence of the
particulate matter changes the pressure ow
characteristics of the valvular mechanism already
present. Under normal conditions, the arachnoid
villi (microscopic structures) contained within the
arachnoid granulations (macroscopically visible)
function as a dierential pressure valve between
the CSASs and SSS with a closing pressure of 5 to
7 mm Hg (70100 mm H2O). In adults with
closed sutures and fontanels, clogging of the
arachnoid villi usually increases ICP, with a minimal increase in ventricular size. In the context of
traumatic or aneurysmal SAH, the usual indication for shunting is the failure to be able to
wean the patient from an external ventricular
drain in the presence of minimal ventriculomegaly. The shunt can usually be removed later in
the course of a patients life, but ventricular size
may increase signicantly with symptoms at the
time of shunt failure [11]. In this case, the point of
obstruction usually is at the level of the basal
cisterns and no longer at the level of the
arachnoid granulations.
In infants with neonatal intraventricular hemorrhage or meningitis, clogging of the arachnoid
villi increases both the CSAS and ventricular
volume. In this situation, the open fontanels and
distensibility of the sutures allow CSF to accumulate, because ICP cannot increase to the point
that it will overcome the valvular mechanisms in
the arachnoid villi. This condition responds to
increasing ICP by wrapping the head with an ace
bandage. ICP then increases to a point where CSF

can be absorbed [12].
The second point of obstruction leading to the
type of communicating hydrocephalus that
cannot be aected by ETV involves obstruction
to outow of venous blood from the SSS. In
adults, marked increases in pressure in the SSS
lead to pseudotumor cerebri rather than to
hydrocephalus [13,14]. In infants with open
fontanels and distensible sutures, marked increases
in pressure in the SSS lead to marked ventriculomegaly associated with concomitant increases in
the volume of the CSASs. At presentation, it may
be dicult, if not impossible, to discern that the
hydrocephalus is caused by increased venous
pressure. This form of hydrocephalus can lead to
secondary obstruction of the aqueduct as the brain
stem is displaced inward and the temporal horns of
the lateral ventricles displace the midbrain medially [15]. This form of hydrocephalus has been well
studied in the context of achondroplasia and
craniofacial syndromes [1620]. In the case of
Crouzons and Pfeiers syndromes, the hydrocephalus may not be evident until a formal cranial
remodeling operation has been performed. These
patients share the same syndrome in that their
ventricles do not dilate to abnormally large
volumes at the time of shunt failure. They show
signs of overtly increased ICP without ventriculomegaly, a condition that has been referred to as
normal volume hydrocephalus [21].
Unresponsive ventricles: ventricles too small or
too distorted to manipulate an endoscope safely
Performing an ETV requires that the endoscope
be placed in the lateral ventricle and manipulated
through the foramen of Monro to the oor of the
third ventricle, where the hole can be made. The
smaller the ventricles are, the more dicult it
becomes to perform the manipulations needed to
fenestrate the oor of the third ventricle. If the size
of the ventricles is normal or smaller than normal
after chronic shunting, it may be impossible to
manipulate the endoscope safely to the appropriate
location to perform the fenestration. There are two
important circumstances in which third ventriculostomy is contraindicated in this context. Marked
enlargement of the massa intermedia of the third
ventricle is a well-recognized concomitant of the
Chiari II malformation in the context of spina
bida. In these situations, the walls of the third
ventricle are often closely opposed because of
the size of the massa intermedia despite the
43
Fig. 3. MRI of a patient with spina bida showing enlargement of the massa intermedia and a small third ventricle
despite marked enlargement of the lateral ventricles. (A) Sagittal T2-weighted MRI demonstrating the enlarged massa
intermedia in a small third ventricle despite massive enlargement of the lateral ventricles. (B) Axial T2-weighted MRI
demonstrating the dierence in size between the lateral and third ventricles.
enlargement of the lateral ventricles (Fig. 3). The

third ventricle can be markedly distorted in
patients with a Chiari II malformation. Even when
the third ventricle can be cannulated, the anatomic
landmarks can be so distorted that it may be
impossible to nd a point on the oor of the third
ventricle to make the fenestration.
The second context in which ETV is contraindicated because it would be unsafe to manipulate
the endoscope occurs in patients with normal
volume hydrocephalus as described previously.
These patients have smaller than normal ventricles,
and the ventricles do not expand at the time of
shunt failure. These patients should not undergo
ETV for two reasons. First, the procedure is
associated with a high risk of damaging the
columns of the fornix, mamillary bodies, cerebral
peduncles, or hypothalamus. Second, ETV in this
context is unlikely to be of benet to the patient.
We have performed iohexol ventriculography in 31
such patients, and 28 showed free communication
from the ventricle to the interpeduncular cistern
[22]. In the 3 patients who did not show such
communication, treatment of their hydrocephalus
was complicated by a signicant infectious process
after the original shunt had been placed.
Some nonresponsive ventricles have resulted in

surprising observations. In one patient whose
hydrocephalus was associated with a pineal tumor
originally treated in infancy, overt shunt failure
occurred despite no change in ventricular volume.
Ventriculography revealed that the dye injected
through the shunt into the lateral ventricle rapidly
owed into the interpeduncular cistern area (Fig. 4).
The hydrocephalus of patients with normal
volume hydrocephalus is related to increased
venous pressures, and they are not candidates
for ETV for both reasons.
Selection of candidates for endoscopic third

ventriculostomy as an initial procedure in the
treatment of hydrocephalus
Late occlusion of the aqueduct of Sylvius
Hydrocephalus caused by occlusion of the
aqueduct of Sylvius is most common in infants
who manifest overt hydrocephalus as newborns or
whose head circumference rapidly increases after
birth. This indication would seem to be ideal for
performing an ETV, because occlusion of the
aqueduct obstructs CSF ow between the third
44
Fig. 4. (A) Sagittal MRI of a child with hydrocephalus associated with a pineal region tumor showing communication
between the lateral ventricles and spinal subarachnoid spaces despite the presence of the tumor. (B) CT of the basal
cisterns after iohexol has been injected into the lateral ventricle showing the ow of contrast into the basal cisterns, thus
conrming communication in the cerebrospinal uid pathways.
and fourth ventricles. Theoretically, CSF should

be left in the interpeduncular cistern in communication with the arachnoid granulations, which
should be normal. Creating this internal bypass
between the third ventricle and the cistern should
normalize CSF dynamics.
Aqueductal stenosis is often diagnosed by
MRI only when triventricular hydrocephalus is
present and the size of the fourth ventricle is
normal. Triventricular hydrocephalus does not
always mean that the underlying cause of the
hydrocephalus is aqueductal stenosis. The phenomenon of communicating hydrocephalus causing dilatation of the temporal horns of the lateral
ventricle and compression of the midbrain from
both sides leading to a secondary occlusion of the
aqueduct of Sylvius has been discussed previously.
In these patients, shunting opens the aqueduct of
Sylvius to CSF ow again [15]. Depending on the
initial site of obstruction, the patient may or may
not be a candidate for ETV.
Regardless of the site of obstruction to CSF
ow, infants are usually poor candidates for
a third ventriculostomy, presumably because the
ventricles fail to respond after third ventriculostomy despite the blockage at the aqueduct. In
these cases, the ventricles are quite large; unlike
shunting, which literally sucks CSF from the
brain, ETV normalizes CSF absorption. The
latter process requires intraventricular pressure

to increase 5 to 7 mm Hg greater than atmospheric pressure, which may be impossible in
babies with open anterior fontanels. Some authors
suggest that these infants should not be candidates
for ETV because of their poor response rate
[23,24]. Other authors quote low rates of shunt
independence after ETV but believe that the
benet is great enough and the risks of the
procedure are low enough to justify ETV in
infants in an attempt to avoid placing a shunt [25].
Aqueductal stenosis is diagnosed in older
children and adults in two contexts. It is dicult
to establish treatment criteria for and to assess the
outcome of treatment in patients who develop
symptomatic hydrocephalus later in life and have
extremely large heads. Obviously, these patients
have had severe ventriculomegaly from infancy.
Oi et al [26] have termed this condition longstanding overt ventriculomegaly of the adult
(LOVA). These patients usually have no overt
symptoms after undergoing an imaging study for
a seizure or minor head injury. Frequently, they
have no symptoms related to their hydrocephalus.
When asked, some patients identify daily headaches, clumsiness, or poor memory that may or
may not have changed recently.
How to treat these patients and, indeed,
whether to treat these patients are controversial
issues. Some authors strongly believe that these

patients will benet from intervention, usually
ventricular shunting [27]. Data showing improvement in these patients are hard to nd, and such
patients seldom have high ICP. My approach to
these patients is to obtain formal neuropsychologic studies to determine whether they exhibit
problems with higher cognitive function. If so, it is
important to determine whether the deterioration
is recent or long standing. I then provide patients
with the pertinent information and attempt to
help them reach a treatment decision. A signicant decrease in the size of the lateral ventricles
after shunting or ETV is unusual. Shunts can be
shown to be working manometrically, but it may
be dicult to determine whether an ETV has
produced the best outcome. Imaging studies, at
least cine MRI through the oor of the third
ventricle and preferably the injection of iodinated
contrast material into the lateral ventricles to
trace its ow into the basal cisterns, is needed to
conrm that CSF dynamics have been maximized.
There is one important caveat if the decision is
made to attempt to treat LOVA with ETV.
Occasionally, the head is so large that some
neuroendoscopes are too short to reach the oor
of the third ventricle. The distance from the skull
to the oor of the third ventricle can be measured
on the console of the CT or MRI scanner. A
selection of endoscopes should be available to
obviate this problem.
In the second clinical syndrome, older patients
develop hydrocephalus caused by aqueductal
stenosis from the presence of severely increased
ICP. Patients often seek treatment from an
ophthalmologist for some form of ophthalmoplegia, such as bilateral sixth nerve palsies or
Parinauds syndrome. On examination, the patient
is found to have bilateral papilledema. This
phenomenon can occur as an isolated event with
obstruction caused by fusion of the walls of the
aqueduct of Sylvius as occurs in some cases of
neurobromatosis 1. This overt form of late
aqueductal stenosis is more likely to result from
small benign tumors (tectal gliomas) of the
aqueduct. These are true tumors, but their prognosis in terms of propensity to grow or disseminate
is usually good [28]. A direct attack on these
tumors, even for a biopsy, is seldom warranted.
This acute adult form of hydrocephalus from
aqueductal stenosis is perhaps the most compelling and most satisfying condition to treat with
ETV. The success rate for this particular subset of
patients is 75% to 80%. The rate of signicant
45
morbidity, including endocrinopathy, memory

decits, and hemiparesis, has been reported to
be 3%. After 1 year, less than 1% are still
troubled by complications of ETV [24]. In this
situation, the risk-benet ratio favors ETV over
internal shunting.
Hydrocephalus associated with newly diagnosed
brain tumors, particularly of the posterior fossa
This area of study is evolving, and the use of
ETV before tumor removal is discussed more than
it is reported in the literature. Posterior fossa
tumors usually manifest with hydrocephalus and
acutely increased ICP. The blockage of the CSF
ow meets the criteria for ETV in that CSF ow
between the third ventricle and interpeduncular
cistern is obstructed. Conceptually, performing an
ETV in this context is similar to placing an
internal shunt before tumor resection. It relieves
symptoms in most patients and thus allows a direct
attack on the tumor at a later stage. It shares with
the placement of an internalized shunt the
possibility, or even the probability, that it will
not be needed after the tumor has been removed.
It shares with shunts the real possibility of leading
to an upward herniation syndromethe herniation of the superior cerebellar vermis through the
tentorium, leading to compression of the midbrain. Typically, contemporary pediatric neurosurgeons begin patients on high doses of
dexamethasone at the time of diagnosis to
stabilize their condition. When the tumor is
attacked directly, an external ventricular drain is
placed before the lesion is removed.
The role of ETV in the overall management of
children with brain tumors aecting the CSF
pathways is evolving. Soon a series of publications
will support its more or less routine use. We await
the results of those studies before applying this
technique generally. If neurosurgeons elect to
perform an ETV before removing a posterior
fossa brain tumor, they must carefully review the
anatomy as it appears on sagittal MRI scans. In
this situation, the brain stem is often pushed
anteriorly against the clivus. The distance between
the clivus and basilar artery is short, which makes
the procedure more risky than it would be
otherwise. The procedure should be performed
by an experienced neuroendoscopist and can
be made signicantly safer by using frameless
stereotaxy.
If tumor resection is not essential to the
patients outcome, the neuroendoscope becomes
46
a much more exciting tool. Adolescent male

patients with a pineal tumor that enhances
diusely, hydrocephalus, and no markers of
malignancy (elevated levels of a-fetal protein or
human chorionic gonadotrophin) are likely to
have a pineal germinoma. Ideally, these patients
should be managed by using an endoscope to
perform a third ventriculostomy to obtain a biopsy
of the tumor. Denitive treatment can then be
managed using conformal radiation therapy. This
approach is associated with minimal rates of
morbidity and high rates of success.
Hydrocephalus associated with intraventricular
and subarachnoid hemorrhage
Whether SAH results from diuse craniocerebral trauma or from the rupture of an aneurysm
or arteriovenous malformation (AVM), patients
are often left with a signicant CSF absorption
problem. Treatment is usually improved substantially by the widespread use of external
ventricular drains. In these contexts, the most
common indication for shunting is failure to wean
the patient from the ventriculostomy without high
levels of ICP. The ventricles do not dilate. ETV is
contraindicated based on the two reasons stated
previously. The ventricles are small enough to
make performing ETV problematic. The presumed point of obstruction is probably the
arachnoidal granulations distal to the interpeduncular cisterns. Therefore, these patients are unlikely to benet from the procedure. They respond
to treatment using either ventricular shunting or
lumbar shunting. If it is possible to temporize long
enough, they may not need a shunt at all.
If the inammatory process associated with
SAH is allowed to continue long enough, the
primary point of obstruction becomes the basal
cisterns. CSF ow is then occluded between the
SSASs and CSASs. This process also follows
bacterial or fungal meningitis. Under the Dandy
classication system, these forms of hydrocephalus would be considered communicating hydrocephalus. Consequently, patients would be
considered candidates for ventricular shunting or
lumboperitoneal shunting. The obstruction to
ow is upstream from the interpeduncular cistern,
however, and patients may indeed respond to
ETV. In fact, the success of the procedure has
been highest in this group.
This point of obstruction causes most cases of
NPH [29]. In a small series, three of four patients
with NPH had excellent outcomes after treatment
[30]. This approach represents an exciting area for

future research.
Special case: hydrocephalus in infancy
ETV is much less likely to be successful when
used to treat infants than when used to treat
hydrocephalus later in life. Consequently, a number of authors have concluded that an age less
than 6 or 12 months should be considered
a contraindication to the use of ETV in the
management of hydrocephalus. A number of
reasons may underlie this failure. The rst relates
to the ability to dene the point of obstruction in
these babies with severe hydrocephalus. Because
of the presence of open fontanels and unfused
sutures, the degree of hydrocephalus is greater
than when hydrocephalus occurs later in life.
Furthermore, it is dicult to interpret the scans to
determine the actual point of obstruction. The
same patient also may have multiple points of
obstruction, especially when hydrocephalus is
associated with the Chiari II malformation (spina
bida cystica). In the latter case, there are four
dierent points of obstruction, and more than one
site can be obstructed at once [7]. When hydrocephalus occurs later in life, obstruction is less
likely to occur at more than one site. It is also
easier to determine where the actual point of
obstruction is based on imaging studies.
The second reason relates to the actual physics
of the system. When the anterior fontanel is
opened widely and the sutures are splayed, the
intracranial compartment is essentially in communication with and has the same pressure as
atmospheric pressure. Natural absorption of CSF
depends on ICP being at least 5 mm Hg greater
than sagittal sinus pressure [5,31,32]. It is often
easier for the size of the head to expand than to
maintain ICP greater than 5 mm Hg. In addition,
Laplaces law implies that the larger ventricles are,
the less distending force is needed to maintain
them at that size.
Although ETV is less likely to be successful in
these children than in older children, good
outcomes are obtained in some babies. The
procedure also may be safer to perform in infants
than in older children or adults. The oor of the
third ventricle is more diaphanous in infants than
it is likely to be in older children. The basilar
artery can be identied with certainty and thus
protected. The ventricular membrane is so fragile
that it can usually be opened with normal
irrigation.
Theoretically, it is better to have communicating than noncommunicating hydrocephalus. When

a patient with hydrocephalus related to aqueductal
stenosis experiences a shunt failure, no mechanisms are available to compensate for the increased
ICP that accompanies failure of the shunt. If the
point of obstruction is distal, a signicantly longer
time is likely to elapse before the situation becomes
critical.
Endoscopic third ventriculostomy for programmed
shunt removal
Chronically shunted patients may suer from
severe and incapacitating headaches caused by
SVS. The percentage of patients suering from this
problem is controversial, and some investigators
doubt its existence. If the threshold for making this
diagnosis includes not only the classic triad of
severe intermittent headaches lasting 10 to 90
minutes, smaller than normal ventricles on imaging
47
studies, and a slowly relling ushing mechanism

but the need to discontinue activity or be brought
home from school at least two times per month,
about 15% of chronically shunted older children
and adults suer from this condition [33]. Based on
chronic monitoring of ICP, we have identied four
dierent pathophysiologic mechanisms responsible for SVS: intermittent proximal obstruction,
intracranial hypertension with smaller than normal ventricles and a failed shunt (normal volume
hydrocephalus), intracranial hypertension with
a working shunt (cephalocranial disproportion),
and migraine in shunted patients [34].
Originally, these patients were managed with
shunt revision using higher resistance valves and
devices that retarded siphoning. Recently, we
have oered these patients the opportunity to
assess the possibility of becoming shunt independent [1]. These patients are admitted to the
hospital for externalization of their shunt, or their
shunt is removed and replaced with an external
Fig. 5. Algorithm for managing shunt-related diculties with a shunt removal protocol. (Courtesy of Barrow
Neurological Institute.)
48
Box 1. Contraindications to endoscopic third ventriculostomy

Absolute contraindications
1. Third and lateral ventricles too small to allow safe manipulation of the endoscope within
the ventricular system
2. Tumor or other mass lesion obstructing the surgeons access to the floor of the third
ventricle
3. Proven communication between the CSF that is in the third ventricle with CSF within the
interpeduncular cistern
Relative contraindications
1. Multicompartment hydrocephalus
2. Thickened floor of third ventricle
3. Chiari II malformation
4. Age less than 1 year
5. Distortion of third ventricular anatomy
CSF = cerebrospinal fluid.
ventricular drain. Under controlled conditions in

the intensive care unit, their shunt is occluded or
raised to a higher level and a scan is obtained.
About two thirds of these patients become
symptomatic with marked ventricular dilatation.
These patients undergo ETV, and 70% tolerate
shunt removal. Patients who develop marked
intracranial hypertension without ventricular distention undergo cisternographic assessment. In
my practice, 28 of 31 patients have been shown to
have communicating hydrocephalus and have
been treated with lumboperitoneal shunts employing valve systems. In chronically shunted patients,
ventricular distention at the time of shunt failure
is probably sucient to recommend ETV (Fig. 5).
Summary
ETV using contemporary instrumentation has
been used for more than 50 years, but its use has
become widespread only in the last 10 to 15 years.
Randomized prospective trials comparing ETV
with shunts are needed before denitive statements can be made about the role of the former in
managing the many forms of hydrocephalus. The
absolute and relative contraindications for the use
of ETV in the management of hydrocephalus are
shown in the Box 1 on this page. It is important
not to presume that a specic radiographic or
clinical feature would prevent a patient from
responding to this rather new procedure without
testing the hypothesis. Patients should be given as
much information as possible regarding the risks
and benets of ETV so they can participate in the

decision-making process.
When should the role of ETV in the management of hydrocephalus be discussed with a patient? At the initial diagnosis of hydrocephalus,
the patient or family should be informed of this
potential alternative to shunting for the management of hydrocephalus. I also believe that patients
with working shunts who are being followed
chronically should be informed about ETV as
a potential treatment option when their shunt
fails. Every shunt failure or infection should be
viewed as an opportunity to explore the possibility
that the patient could become shunt independent.
References
[1] Baskin JJ, Manwaring KH, Rekate HL. Ventricular
shunt removal: the ultimate treatment of the slit
ventricle syndrome. J Neurosurg 1998;88:47884.
[2] Dandy W, Blackfan K. An experimental and
clinical study of internal hydrocephalus. JAMA
1913;61:22167.
[3] Dandy W, Blackfan K. Internal hydrocephalus. An
experimental, clinical and pathological study. Am J
Dis Child 1914;8:40682.
[4] Ransoho J, Epstein F. Proceedings: avoidance of
shunt dependency. J Neurol Neurosurg Psychiatry
1975;38:4101.
[5] Olivero WC, Rekate HL, Chizeck HJ, et al.
Relationship between intracranial and sagittal sinus
pressure in normal and hydrocephalic dogs. Pediatr
Neurosci 1988;14:196201.

[6] Rekate HL. Circuit diagram of the circulation of
cerebrospinal uid. 1989. Pediatr Neurosurg 1994;
21:24852.
[7] Rekate H. Neurosurgical management of the
newborn with spina bida. In: Rekate H, editor.
Comprehensive management of spina bida. Boca
Raton, FL: CRC Publishers; 1991. p. 126.
[8] Gilles F, Davidson R. Communicating hydrocephalus with decient dysplastic parasagittal arachnoidal granulations. J Neurosurg 1971;35:4216.
[9] Barlow CF. CSF dynamics in hydrocephaluswith
special attention to external hydrocephalus. Brain
Dev 1984;6:11927.
[10] Akaboshi I, Ikeda T, Yoshioka S. Benign external
hydrocephalus in a boy with autosomal dominant
microcephaly. Clin Genet 1996;49:1602.
[11] Rekate H, Nulsen FE, Mack H, et al. Establishing
the diagnosis of shunt independence. Monogr
Neural Sci 1982;8:2236.
[12] Epstein F, Hochwald GM, Ransoho J. Neonatal
hydrocephalus treated by compressive head wrapping. Lancet 1973;1:6346.
[13] Karahalios DG, Rekate HL, Khayata MH, et al.
Elevated intracranial venous pressure as a universal
mechanism in pseudotumor cerebri of varying
etiologies. Neurology 1996;46:198202.
[14] Pare LS, Batzdorf U. Syringomyelia persistence
after Chiari decompression as a result of pseudomeningocele formation: Implications for syrinx
pathogenesis: report of three cases. Neurosurgery
1998;43:9458.
[15] Nugent G, Al-Mefty O, Chou S. Communicating
hydrocephalus as a cause of aqueductal stenosis.
J Neurosurg 1979;51:8128.
[16] Francis PM, Beals S, Rekate HL, et al. Chronic
tonsillar herniation and Crouzons syndrome.
[17] Saint-Rose C, LaCombe J, Pierre-Kahn T, et al.
Intracranial venous sinus hypertension: cause or
consequence of hydrocephalus in infants? J Neurosurg 1984;60:72731.
[18] Pierre-Kahn A, Hirsch JF, Renier D, et al.
Hydrocephalus and achondroplasia. A study of 25
observations. Childs Brain 1980;7:20519.
[19] Cinalli G, Renier D, Sebag G, et al. Chronic
tonsillar herniation in Crouzons and Aperts
syndromes: the role of premature synostosis of the
lambdoid suture. J Neurosurg 1995;83:57582.
[20] Nishihara T, Hara T, Suzuki I, et al. Third
ventriculostomy for symptomatic syringomyelia
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
49
using exible endoscope: case report. Minim Invasive Neurosurg 1996;39:1302.

Engel M, Carmel P, Chutorian A. Increased
intraventricular pressure without ventriculomegaly
in children with shunts: normal volume hydrocephalus. Neurosurgery 1979;5:54952.
Rekate HL, Wallace D. Lumboperitoneal shunts in
children. Pediatr Neurosurg 2003;38:416.
Jones RF, Kwok BC, Stening WA, et al. Neuroendoscopic third ventriculostomy. A practical
alternative to extracranial shunts in non-communicating hydrocephalus. Acta Neurochir Suppl
(Wien) 1994;61:7983.
Teo C, Jones R. Management of hydrocephalus by
endoscopic third ventriculostomy in patients with
myelomeningocele. Pediatr Neurosurg 1996;25:
5763.
Buxton N, Macarthur D, Mallucci C, et al. Neuroendoscopic third ventriculostomy in patients less
than 1 year old. Pediatr Neurosurg 1998;29:736.
Oi S, Shimoda M, Shibata M, et al. Pathophysiology of long-standing overt ventriculomegaly in
adults. J Neurosurg 2000;92:93340.
Larsson A, Stephensen H, Wikkelso C. Adult
patients with asymptomatic and compensated
hydrocephalus benet from surgery. Acta Neurol
Scand 1999;99:8190.
Pollack IF, Pang D, Albright AL. The long-term
outcome in children with late-onset aqueductal
stenosis resulting from benign intrinsic tectal
tumors. J Neurosurg 1994;80:6818.
Di Rocco C, Di Trapani G, Maira G, et al.
Anatomo-clinical correlations in normotensive hydrocephalus. Reports on three cases. J Neurol Sci
1977;33:43752.
Mitchell P, Mathew B. Third ventriculostomy in
normal pressure hydrocephalus. Br J Neurosurg
1999;13:3825.
Ransoho J, Shulman K, Fishman R. Hydrocephalus: a review of etiology and treatment. J Pediatr
1960;56:399411.
Cutler R, Page L, Galicich J. Formation and
absorption of cerebrospinal uid in man. Brain
1968;91:70720.
Hyde-Rowan MD, Rekate HL, Nulsen FE. Reexpansion of previously collapsed ventricles: the slit
Rekate HL. Classication of slit-ventricle syndromes using intracranial pressure monitoring.
Techniques of endoscopic third ventriculostomy

Douglas Brockmeyer, MD
Primary Childrens Medical Center, 100 North Medical Drive, Suite 2400,
Salt Lake City, UT 841131100, USA
Modern techniques of endoscopic third ventriculostomy (ETV) are based on the concept of
establishing a natural conduit for cerebral spinal
uid (CSF) ow through the oor of the third
ventricle. Through the years, a wide variety of
techniques have been used as a means to this
end and have included both open and closed
approaches. The relatively recent application of
endoscopic technology to intraventricular surgery
has allowed neurosurgeons to perform third
ventriculostomies in a minimally invasive fashion,
however. Advances in third ventriculostomy technique have been based on a detailed understanding of third ventricular anatomy, surgical
trajectories, and improved instrumentation. The
goal of this article is to discuss these issues in
detail and to point out the relevant risks and
known complications associated with them.
open and closed [512]. Notable among these

series is the Toronto experience as reported by
Homan et al [13], which describes a percutaneous
third ventriculostomy technique in the management of noncommunicating hydrocephalus.
Signicant experience with closed stereotactic
techniques was reported by Kelly in 1991 [14].
During this period, neurosurgeons began taking
advantage of smaller and smaller endoscopes;
eventually, the routine use of beroptic or rodlens ETV was accepted. Multiple studies were
published in the 1990s reecting considerable
success with endoscopic techniques [1427]. The
current state of the art in ETV obviously reects
signicant new concepts brought out by these
pioneering neurosurgeons.
Relevant anatomy
History
An understanding of the current state of ETV
is not complete without an appreciation of its
history. Briey, in 1922, Walter Dandy originated
the concept of third ventriculostomy by performing a craniotomy and fenestrating the lamina
terminalis for the treatment of hydrocephalus [1].
This was quickly followed by Mixter [2], who, in
1923, used a urologic endoscope to puncture the
oor of the third ventricle, thus ushering in the era
of ETV at an early stage. In 1936, Stokey and
Scarrf [3] and, in 1951, Scarrf [4] described their
own procedures for third ventriculostomy. In the
1970s and 1980s, a host of authors described
various techniques for third ventriculostomy, both
E-mail address: douglas.brockmeyer@hsc.utah.edu
A brief review regarding the anatomy relevant

to ETV is useful. The recognition of critical
landmarks and structures is vital to the overall
success of ETV. In addition, complications may
be avoided when important anatomy is identied
early and respected throughout the procedure.
Choroid plexus
The choroid plexus lies along the oor of the
lateral ventricle in the choroidal ssure and is
oriented in an anterior/posterior direction. Early
recognition of the choroid plexus within the
lateral ventricle is a powerful navigational tool
because its anterior extent leads to the foramen of
Monro and the third ventricle. Thus, if ones
initial endoscopic trajectory does not lead directly
to the foramen of Monro, the choroid plexus can
act as a road map to lead one to the proper site
quickly and eciently. Even in patients with
doi:10.1016/S1042-3680(03)00066-4
52
D. Brockmeyer / Neurosurg Clin N Am 15 (2004) 5159
distorted ventricular anatomy, such as those with

spina bida, the choroid plexus remains in the
choroidal ssure and still leads to the third
ventricle. In patients with large ventricles, this
point is not much of an issue, but when the
ventricular size is small, early recognition of the
choroid plexus is sometimes the only navigational
guide a neuroendoscopist has.
this problem are to be aware of the orientation of

your endoscopes camera in relation to the crosssectional area of the tip and to make appropriate
adjustments in the approach trajectory. In addition, if one guides the scope in a handheld fashion,
exquisite tactile feedback can be obtained and the
surgeon could be alerted to hanging up the edge
of the endoscope on the fornix.
Fornix
The fornix forms the superior and anterior
margin of the foramen of Monro. It is important
that every eort is made to avoid its injury during
endoscopic neurosurgery. Because of its location,
however, it can easily be injured during passage of
the endoscope from the lateral ventricle into the
third ventricle. The risk of injury is multiplied when
multiple passes through the foramen of Monro are
made. For that reason, the number of passes of the
endoscope through the foramen of Monro should
be kept to an absolute minimum. In addition,
it is important to recognize that the endoscopes
camera port is only a small percentage of the crosssectional area of the endoscope tip. For that
reason, it is easy to assume that if one passes the
endoscope easily through the foramen of Monro,
the fornix is not injured. This may not be the case,
however, because the edge of the endoscope tip
near the light source or working chamber may
inadvertently cause injury. Potential solutions for
Hypothalamus
The paired hypothalami form the lateral walls
of the third ventricle. The supraoptic and paraventricular arcuate nuclei are the structures that
are at the most risk during third ventriculostomy
(Fig. 1). Injury to these structures may have
signicant endocrinologic consequences. Although
some evidence suggests that the supraoptic
nucleus is related mainly to vasopressin (antidiuretic hormone) and the paraventricular nucleus
to oxytocin, both hormones are found in each
nucleus. Therefore, surgical trajectories to the
third ventricular oor must be planned with the
idea that hypothalamic damage must not occur.
Although surgical trajectories are presented in
a later section, this concept is important when
discussing hypothalamic anatomy. These issues
are especially important when spina bida patients with distorted hypothalamic anatomy undergo ETV.
Fig. 1. Hypothalamic nuclear anatomy adjacent to the third ventricle. Note that the supraoptic and paraventricular
nuclei are in close proximity to the third ventricle.
53
Third ventricular oor
Lillequists membrane
The oor of the third ventricle is essentially

a thinned out portion of the hypothalamus and
can potentially have functioning hypothalamic
nuclei (eg, supraoptic and paraventricular nuclei)
within it. As noted previously, this consideration is
especially important in patients with spina bida.
The traditional boundaries of the third ventricular
oor include the mamillary bodies posteriorly, the
walls of the hypothalamus laterally, and the optic
chiasm/infundibular recess anteriorly. Within
this area is a relative safe zone where the third
ventricular oor may be entered (Fig. 2). This
consists of the area just anterior to the midway
point between the mamillary bodies and the infundibular recess. If one penetrates the third ventricular oor posterior to this point, the basilar
artery tip or proximal portion of the posterior
communicating artery may be encountered. If the
oor is penetrated anterior to this point, the
clivus is encountered . Obviously, if a choice has to
be made, entering slightly more anterior to the
midpoint is the safer than entering more posterior
to the midpoint. Usually, within the midportion
of a thinned-out third ventricular oor, there is
a translucent bluish-appearing area that corresponds to a safe zone in which to make the initial
opening. How the opening is actually made is
discussed in later in this article.
Lillequists membrane is an arachnoid plane

that contains within it the basilar artery complex
and separates the posterior fossa arachnoid
cisterns from the suprasellar cisterns. Once a neuroendoscopist penetrates through the oor of the
third ventricle, Lillequists membrane is frequently
encountered and hides the basilar artery complex
and prepontine cistern from view. It is important
that Lillequists membrane be opened to have
a successful third ventriculostomy. The goal is to
communicate the uid of the third ventricle to the
prepontine cistern area. A clear view of the basilar
artery complex, pons, pontine perforators, and
clivus must be obtained before an ETV can be
considered successful. Failure to recognize this
anatomic point may lead to endoscopic failures.
Fig. 2. Line drawing depicting the oor of the third

ventricle and the endoscopic safe zone, where initial
dissection during endoscopic third ventriculostomy
should begin.
Endoscopic approaches to the third ventricle

Trajectories
For a typical patient with enlarged ventricles
undergoing ETV for the rst time, a standardized
trajectory to the third ventricle should be used.
One such trajectory that yields excellent results
consists of an entry site at the intersection of the
coronal suture and the midpupillary line, with the
trajectory of the endoscope slightly medial and
oriented in line with the external auditory meatus
in an anterior/posterior direction (Fig. 3A). This
approach yields an endoscopic trajectory toward
the foramen of Monro and into the oor of
the third ventricle. The consistency and excellent
results of this approach cannot be overemphasized.
A standardized approach is particularly important when previously shunted patients become
candidates for ETV. The existing shunt entry site
or ventricular catheter trajectory is sometimes
positioned so that a safe approach to the third
ventricular oor is not feasible. In this situation,
a separate burr hole, and sometimes a separate
incision, must be made in the standard location,
and the standard approach should be employed. It
is also possible in this situation to use a exible
endoscope to navigate into the third ventricle. In
this authors experience, however, orientation
may be dicult, and the acute curve of the scope
may injure brain structures or preclude passing
instruments safely down the working chamber.
The endoscope trajectory can also be changed
based on the preoperative ventricular size. If the
54
Type of endoscope to use

There are two basic types of neuroendoscopes
that are available: rod-lens and beroptic. These
scopes may be further subdivided into rigid,
semirigid, and exible. Understandably, there is
no single endoscope that is adaptable enough
to be superior in all applications and situations.
Therefore, the neuroendoscopist must be exible
in his or her choice of endoscopes. For instance, if
one encounters an adolescent patient with lateonset aqueductal stenosis and large ventricles,
a good argument could be made for being as
minimally invasive as possible, and a small, 1-mm,
semirigid endoscope could be used for the entire
procedure. In that case, the tip of the scope is used
to perform the fenestration. The optics of the
1-mm endoscope are sucient to allow positive
identication of critical structures and rapid and
safe performance of the third ventriculostomy.
A school-aged spina bida patient with distorted anatomy in the third ventricular oor
requires maximum visibility and adaptability,
however. This situation requires either a rod-lens
system or one of the larger beroptic scopes. In
either case, anticipating the anatomy of the
patient and maximizing visibility must be weighed
against potential risks. For the experienced endoscopist, the proper endoscope choice may vary
from patient to patient.
Techniques of endoscopic third ventriculostomy
Fig. 3. Line diagrams depicting trajectories to third

ventricle. (A) Traditional trajectory toward the third
ventricle oor beginning at the midpupillary line. (B)
Modied trajectory because of the small ventricle size.
Note that the entry point is moved slightly medial
compared with the standard trajectory.
ventricles are generous, the traditional trajectory

described previously may be used. If the ventricles
are small, however, the burr hole placement
should be slightly medial to the midpupillary line,
facilitating an easier approach into a narrow third
ventricle (see Fig. 3B).
In discussing techniques of endoscopic third

ventriculostomy, it is logical to begin with the
simplest technique available and then to proceed
toward more complex or sophisticated procedures. The list that is included here is by no means
meant to be all-inclusive; however, it is meant to
cover most of the more commonly used techniques of ETV and serves primarily as a guide or
reference for further reading.
Endoscope tip
Perhaps the easiest and most straightforward
way to perform the fenestration for third ventriculostomy is using the tip of the endoscope
itself (Fig. 4). Multiple case series have documented its safety and ecacy over time [15,19
21,26,28,29]. In essence, the tip of the endoscope is
used as a dissecting tool to penetrate the oor of
the third ventricle and Lillequists membrane.
Obviously, the smaller the tip of the endoscope,
the easier the dissection becomes. This techniques
55
Fig. 4. (A, B) Line drawing depicting the use of the endoscope tip to perform the fenestration through the oor of the
third ventricle.
critical drawback is that during the moments of

dissection, ones visualization is completely obscured by the tissue in front of the scope. The lack
of visualization is made up for by tactile feedback
during manual opening of the third ventricular
oor. Sometimes, a windshield wiper technique
may be used to gently dissect the oor of the third
ventricle and Lillequists membrane, much like
arachnoid dissection in other parts of the brain.
With practice, this technique is a safe and eective
way of performing a large opening of the oor of
the third ventricle. In addition, the scope tip can
be placed into the prepontine cistern to inspect the
surrounding anatomy. Obviously, this technique
has a rather steep learning curve, and the risk of
basilar artery injury is a very real threat during
this procedure. Only endoscopists condent with
their dissecting technique using the tip of the
endoscope should employ it.
A variation of this technique was reported by
Wellins et al [30]. They describe preloading
a ventricular catheter over the endoscope. Once
the ETV is performed, the ventricular catheter is
used to dilate the opening.
Dissection/balloon dilatation
Perhaps the most widespread and safest
technique to open the oor of the third ventricle
is to pass some type of semiblunt dissector down
the working channel of the endoscope rst and
then to open a small hole in the oor of the third
ventricle (Fig. 5A, B). Next, the dissector is
withdrawn, and a 2- or 3-French balloon-tipped
catheter (or similar device, such as a wire stone
extractor [30]) is passed through the endoscope
and through the initial fenestration in the oor of
the third ventricle (see Fig. 5C). The balloon is
slowly inated at the site of the fenestration to

enlarge the opening gradually and gently (see Fig.
5D) [15,20,25]. Once the balloon has been opened
to its widest diameter, it is deated and removed
and the endoscope is placed through the opening
to inspect and conrm that Lillequists membrane
has been opened (see Fig. 5E). If not, the procedure is repeated until Lillequists membrane is
open. Several endoscope manufacturers supply
small dissecting tools that are ideal for making
the initial opening. The authors preferred technique is to use the tip of an endoscopic bipolar
coagulator.
This technique, used either with a rod-lens or
beroptic endoscope, is safe, simple, and eective.
The oor of the third ventricle is under direct
visualization during the entire act of fenestration.
For these reasons, the dissector/balloon dilatation
technique and its variations deserve to be at
the top of the list for recommended fenestration
techniques. One notable exception is placing
a fully inated balloon-tipped catheter all the
way through the third ventricle and pulling it back
through the membrane in a fully inated fashion.
It is quite easy to injure the hypothalamic
nuclei or to cause signicant bleeding with this
technique.
Saline jet
Saline jet irrigation has been used in performing ETV and cyst wall fenestration. The technique
involves deploying and then directing a strong jet
of saline irrigation in the region of the proposed
dissection. The uid forced from the saline jet
atraumatically performs the dissection, at least
theoretically. Although this technique has its
merits and probably deserves more widespread
56
Fig. 5. (AE) A series of line drawings depicting the use of an endoscope, blunt dissector, and Fogarty catheter to
perform endoscopic third ventriculostomy (see text for details).
use, a saline jet apparatus is not readily available

to most operating rooms and would need to be
specically requested. Further work and investigation of this technique are probably warranted
before more widespread use is recommended.
Saline torch
Manwaring [31] has previously described the
use of the saline torch. As a result of the fact that
a potentially dangerous form of active energy is
used during the fenestration, most neuroendoscopic surgeons have moved away from this
technique toward blunt dissection of the third
ventricular oor.
Laser energy
The application of
type of active energy
mentioned here only
When using a laser
laser energy or some other

source to perform ETV is
to discourage its use [32].
or other type of thermal
energy source to perform the fenestration of the

third ventricular oor, absolutely no tactile feedback is available and spread of the energy may
easily injure or rupture critical neurovascular
structures. Several signicant complications, some
reported and some unreported, have occurred
using this type of technique [33]. Until safer
sources of energy are developed to provide for
third ventricular fenestration, these types of
techniques should be avoided.
Stereotactic approaches
Over the years, several authors have reported
closed and endoscopic stereotactic approaches
to third ventriculostomy [11,13,14,19]. One such
approach describes guiding a rigid rod-lens scope
through the lateral and third ventricle using a
predetermined stereotactic trajectory [19]. Although this type of technique has its merits, it
does not allow much room for exibility, a key
factor that may represent a critical margin of
safety in ETV. In addition, because of the extra
expense of the preoperative stereotactic scan,
extra operative time for stereotactic registration,
and extra expense for the stereotactic apparatus,
using a stereotactic technique is most likely not
a cost-eective way to perform this procedure.
Further cost-benet analysis of this procedure
would be important.
Doppler ultrasound
In 1998, Schmidt [34] described the use of
a micro-Doppler technique to identify the basilar
apex before performing the fenestration in the
third ventricular oor. In his report, two third
ventricular oors were mapped out before fenestration with a high degree of accuracy of
predicting the position of the basilar artery.
Although perhaps not necessary in all cases, this
type of technique may be extremely useful in
performing third ventriculostomies in patients
with thick opaque oors or in those patients with
distorted anatomy (eg, spina bida).
Complications
Vascular
The most important vascular complication to
avoid during ETV is injury to the basilar artery
and its nearby branches. Two reports have
documented injury to the basilar artery or the
57
posterior cerebral artery by the tip of the neuroendoscope or a laser [33,35]. In one case,
a pseudoaneurysm developed at the site of arterial
injury, which later required further surgery and
trapping [35]. Obviously, a careful look at the
preoperative MRI to identify the position of the
basilar artery in the prepontine cistern is extremely helpful in potentially avoiding vascular injury.
In addition, placing the initial third ventricular
oor fenestration in the safe zone also may
prevent an arterial catastrophe.
If major arterial injury does occur, several
techniques may help in avoiding disaster. One is
the preplacement of a Sheath dilator in the lateral
ventricle, through which the endoscope is passed.
If arterial injury does occur, expression of blood
through the Sheath dilator may help to decompress the intracranial compartment during the
episode of rapid bleeding before arterial spasm. In
addition, the ventricular cavity should be irrigated
copiously with saline irrigation for perhaps as
long as 20 or 30 minutes. Placement of an external
ventricular drain and performance of a postoperative CT scan are mandatory. It is important to
remember that although major arterial injuries
may occur, most can be avoided by careful
planning and meticulous technique.
This type of major arterial bleeding should be
distinguished from the small amount of bleeding
that frequently occurs from the oor of the third
ventricle once the fenestration is performed. This
bleeding is typically capillary in nature and
subsides after 1 to 2 minutes of continuous irrigation. Further bleeding almost always requires
only more patience and irrigation. If the CSF
is not clear enough to see with an endoscope, the
procedure should be abandoned, an external
ventricular drain placed, and an emergency head
CT scan obtained.
Occlusive hydrocephalus
The occurrence of acute occlusive hydrocephalus by blockage of the foramen of Monro with
the endoscope during continuous irrigation has
been described previously [36]. Recognition and
prevention of this potentially lethal complication
are important. The size of the endoscope versus
the size of the foramen of Monro should be
judged during surgery; if the endoscope plugs the
foramen of Monro, the irrigation should be shut
o. This problem should be in the endoscopists
mind if sudden vasomotor instability occurs
during ETV.
58
Endocrine
Because the third ventricular oor is in direct
continuity with the hypothalamus, endocrine
dysfunction may occur after ETV. In 1996, Teo
et al [37] reported four endocrine complications in
55 ETVs. Their complications included diabetes
insipidus, increase in appetite, loss of thirst, and
amenorrhea. The rst two complications resolved
completely over time. In a separate report, Teo
and Jones [38] described transient endocrine
complications in 2 of 69 spina bida patients
who underwent ETV. In contrast, reports of other
large ETV case series [15,25] have reported no
endocrine complications.
Obviously, any patient who undergoes ETV is at
potential risk for damaging critical neuroendocrine
structures, such as the supraoptic or paraventricular nuclei. Although it is dicult to prove with the
existing data, it makes sense that patients who are
at the most risk for endocrine complications during
ETV are probably those with spina bida or a thick
third ventricular oor. As expected, the treatment
of an endoscope-caused endocrine complication
is supportive and hopefully transient. Therefore,
careful attention to the regional anatomy as well as
willingness to stop a procedure if the anatomy is
unfavorable may help to avoid this complication.
References
[1] Dandy WE. An operative procedure for hydrocephalus. Johns Hopkins Hosp Bull 1922;33:18990.
[2] Mixter WJ. Ventriculoscopy and puncture of the
oor of the third ventricle. Boston Med Surg J
1923;188:2778.
[3] Stokey B, Scarrf JE. Occlusion of aqueduct of
Sylvius by neoplastic processes with a rational
surgical treatment for relief of resultant obstructive
hydrocephalus. Bull Neurol Inst 1936;5:34877.
[4] Scarrf JE. Treatment of obstructive hydrocephalus
by puncture of the lamina terminalis and oor of
the third ventricle. J Neurosurg 1951;8:20413.
[5] Avman N, Kanoplat Y. Third ventriculostomy by
microtechnique. Acta Neurochir Suppl (Wien) 1979;
28:58895.
[6] Brocklehurst G. Trans-callosal third ventriculochiasmatic cisternostomy: a new approach to
hydrocephalus. Surg Neurol 1974;2:10914.
[7] Guiot G. Ventriculocisternostomy for stenosis of
the aqueduct of Sylvius. Puncture of the oor of the
third ventricle with a leucotome under television
control. Acta Neurochir (Wien) 1973;28:26489.
[8] Jaksche H, Lowe F. Burrhole third ventriculocisternostomy. An unpopular but eective procedure for
the treatment of certain forms of occlusive hydrocephalus. Acta Neurochir (Wien) 1986;79:4851.
[9] Natelson SE. Early third ventriculostomy in myelomeningocele infantsshunt independence? Childs
Brain 1981;188:2778.
[10] Reddy K, Fewer HD, West M, Hill NC. Slit ventricle
syndrome with aqueduct stenosis: third ventriculostomy
as denitive treatment. Neurosurgery 1988;23:7569.
[11] Sayers M, Kosnik E. Percutaneous third ventriculostomy: experience and technique. Childs Brain
1976;2:246.
[12] Vries J. An endoscopic technique for third ventriculostomy. Surg Neurol 1978;9:1658.
[13] Homan HJ, Harwood-Nash D, Gilday DL.
Percutaneous third ventriculostomy in the management of non-communicating hydrocephalus. Neurosurgery 1980;7:31321.
[14] Kelly PJ. Stereotactic third ventriculostomy in
patients with nontumoral adolescent/adult onset
aqueductal stenosis and symptomatic hydrocephalus. J Neurosurg 1991;75:86573.
[15] Brockmeyer DB, Abtin K, Carey L, Walker ML.
Endoscopic third ventriculostomy: an outcome
analysis. Pediatr Neurosurg 1998;28:23640.
[16] Buxton N, Macarther D, Mallucci C, Punt J,
Vloeberghs M. Neuroendoscopic third ventriculostomy in patients less than 1 year old. Pediatr
Neurosurg 1998;29:736.
[17] Dalrymple S, Kelly P. Computer-assisted stereotactic third ventriculostomy in the management of
noncommunicating hydrocephalus. Stereotact Funct
Neurosurg 1992;59:10510.
[18] Drake J. Ventriculostomy for treatment of hydrocephalus. Neurosurgery 1993;4:65766.
[19] Goodman RR. Magnetic resonance imagingdirected stereotactic endoscopic third ventriculostomy. Neurosurgery 1993;32(6):10437.
[20] Goumnerova L, Frim D. Treatment of hydrocephalus with third ventriculocisternostomy: outcome
and CSF ow patterns. Pediatr Neurosurg 1997;27:
14952.
[21] Jones RF, Stening WA, Brydon M. Endoscopic
third ventriculostomy. Neurosurgery 1990;26:8691.
[22] Kunz U, Goldman A, Bader C, Waldbaur H,
Oldenkott P. Endoscopic fenestration of the 3rd
ventricular oor in aqueductal stenosis. Minim
Invasive Neurosurg 1994;37:427.
[23] Oka K, Yamamoto M, Ikeda K, Tomonaga M.
Flexible endoneurosurgical therapy for aqueductal
stenosis. Neurosurgery 1993;33:23643.
[24] Rieger A, Rainov NG, Sanchin L, Schopp G,
Burkett W. Ultrasound-guided endoscopic fenestration of the third ventricular oor for non-communicating hydrocephalus. Minim Invasive Neurosurg
1996;39:1720.
[25] Sainte-Rose C, Chumas P. Endoscopic third ventriculostomy. Tech Neurosurg 1996;1:17684.
[26] Teo C, Jones R. Management of hydrocephalus by
5763.

[27] Tuli S, Alshail E, Drake J. Third ventriculostomy
versus cerebrospinal uid shunt as a rst procedure
in pediatric hydrocephalus. Pediatr Neurosurg 1999;
30:115.
[28] Grant JA, McLone DG. Third ventriculostomy:
a review. Surg Neurol 1997;47:2102.
[29] Wellins JCI, Bagley CA, George TM. A simple and
safe technique for endoscopic third ventriculocisternostomy. Pediatr Neurosurg 1999;30:21923.
[30] Wong TT, Lee LS. A method of enlarging the
opening of the third ventricular oor for exible
endoscopic third ventriculostomy. Childs Nerv Syst
1996;12:3968.
[31] Manwaring K. Endoscopic ventricular fenestration.
In: Manwaring K, Crone K, editors. Neuroendoscopy. New York: Mary Ann Liebert; 1992. p. 7990.
[32] Wood FA. Endoscopic laser third ventriculostomy.
N Engl J Med 1993;329:2078.
[33] McLaughlin MR, Wahlig JB, Kaufmann AM,
Albright AL. Traumatic basilar aneurysm after
[34]
[35]
[36]
[37]
[38]
59
endoscopic third ventriculostomy: a case report.

Neurosurgery 1997;41:14003.
Schmidt R. Use of microvascular Doppler probe to
avoid basilar artery injury during endoscopic third
ventriculostomy. Technical note. J Neurosurg 1999;
90:1567.
Abtin K, Thompson BG, Walker ML. Basilar
artery perforation as a complication of endoscopic
third ventriculostomy. Pediatr Neurosurg 1998;28:
3541.
Handler MH, Abbott R, Lee M. A near-fatal
complication of endoscopic third ventriculostomy:
case report. Neurosurgery 1994;35:5257.
Teo C, Rahman S, Boop FA, Cherny B. Complications of endoscopic neurosurgery. Childs Nerv
Syst 1996;12(5):24853.
Teo C, Jones R. Management of hydrocephalus by
myelomeningocele. Pediatr Neurosurg 1996;25(2):
5763.
Complications of third ventriculostomy

Marion L. Walker, MD
Primary Childrens Medical Center, 100 North Medical Drive, Salt Lake City, UT 84113, USA
Improvements in instrumentation and growing

experience with neuroendoscopy have made it
possible to perform endoscopic third ventriculostomy (ETV) safely in a selected group of patients.
The complication rate for ETV now approaches
the rate for ventriculoscopy in general, approximately 6% to 8%, and is similar to the expected
infection rate of shunts [1,2]. Although the
exposure to signicant neurologic complications
is higher than the risk of a single shunt operation,
the possibility of achieving long-term shunt independence with ETV is preferable in good-risk
patients when compared with the cumulative
morbidity of multiple shunt procedures [36].
The collective morbidity for a ventriculoperitoneal
shunt is dicult to quantify; however, a recent
prospective multicenter study of children undergoing their initial shunt insertion found that 31%
had shunt obstruction, 3% had overdrainage, and
8% had infection within 1 year [1]. There is a 60%
failure rate at 2 years.
Many of the complications of ETV are
transient in nature. Skilled management of intraoperative problems can keep these complications
limited to the intraoperative or perioperative
period. At least one author has suggested categorizing complications as clinically insignicant
and clinically signicant [2]. Although this
system may reect the long-term outcome of the
patient, the most important consideration, it does
not emphasize the potentially devastating consequences of seemingly minor intraoperative
events. This article seeks to dene the most
common potential complications of ETV and to
identify factors in the pre-, intra-, and postoperative periods that can minimize their occurrence or eects.
Complications of endoscopic third ventriculostomy

Cerebrospinal uid leak
Although rarely reported, cerebrospinal uid
(CSF) leakage can delay wound healing and
increase the risk of infection. The possibility of
a CSF leak may be minimized by using the
smallest appropriate endoscope (especially in
patients with large ventricles) and by minimizing
the dural opening. A layered closure of the scalp is
necessary. More importantly, however, a CSF
leak is frequently a sign of failure of the third
ventriculostomy, and close observation for this
possibility is required.
Pneumocephalus
Often considered a minor or insignicant
complication, pneumocephalus can delay postoperative recovery and can be associated with
headache, nausea, and vomiting. Entrapment of
air at the time of surgery can interfere with direct
visualization of the anatomic landmarks that are
essential to performing a third ventriculostomy
safely. This can be minimized by keeping the
patients head in a midline anatomic position with
the burr hole at or near the most superior point,
by carefully ushing all irrigation lines of air
bubbles, and by irrigating gently while introducing the endoscope. Attention should be paid to
minimizing CSF loss, especially in the early stages
of the procedure. Decompression of the ventricular system too rapidly can contribute to the
formation of a subdural hematoma. Nitrous oxide
should not be used for anesthesia during ETV
because of the potential for formation of tension
pneumocephalus.
Ventriculitis
E-mail address: marion.walker@hsc.utah.edu
Fever after ETV is not uncommon. In most

cases, it is caused by residual blood in the
doi:10.1016/S1042-3680(03)00070-6
62
M.L. Walker / Neurosurg Clin N Am 15 (2004) 6166
ventricular system. In this setting, fever is selflimited and lasts 24 to 48 hours. In some cases,
postoperative fever has been attributed to raising
the temperature of the CSF in the third ventricle
when employing cautery or laser energy, indirectly
heating the hypothalamus. Fever may also herald
infection and should be monitored closely with
analysis and culture of the CSF [7,8]. Fukuhara
et al [9] report multiple cases in which preexisting
ventricular shunt hardware was found to harbor
infection after ETV. We believe that it is important to remove all shunt hardware after ETV
whenever possible.
Careful and thorough sterilization of the
endoscopic equipment and the use of perioperative antibiotics can minimize the occurrence of
infection. Prompt diagnosis and treatment of
ventriculitis are essential.
Subdural hematoma
Subdural hematoma has been reported after
ETV [9]. Signicant ventricular dilatation with
a thin cortical mantle is a risk factor for subdural
hemorrhage. Eorts should be made to avoid
rapid drainage of large quantities of CSF, and lost
CSF should be replaced with lactated Ringers
solution.
Injury to periventricular structures
The oor of the third ventricle is not a membrane but a part of the hypothalamus. It is
apparently safe to puncture the oor of the third
ventricle but only when it is thinned as a result of
the pressure of hydrocephalus. Amenorrhea, diabetes insipidus, loss of thirst, and increased
appetite have been reported after ETV [2,10,11].
Irrigation solution can also be the cause of
complications. Generous irrigation with normal
saline solutions has been associated with disturbance of electrolytes and hypothalamic dysfunction. Late arousal and postoperative confusion
have also been noted [2]. These complications are
caused by trauma to sensitive hypothalamic
structures comprising the walls and oor of the
third ventricle.
Several strategies can minimize injury to sensitive brain structures in the periventricular region.
Careful preoperative evaluation may exclude patients with a narrow third ventricle; some authors
have suggested a minimum third ventricular width
of 7 to 10 mm [12,13]. Third ventriculostomy can
be performed in small third ventricles, but this
increases the risk of injury and should be
attempted only in unusual circumstances and by

experienced endoscopists. As experience has
grown, the range of patients suitable for third
ventriculostomy has widened. A number of
surgeons have performed third ventriculostomy
in patients with slit ventricle syndrome, with good
results [14,15]. Modern stereotactic systems and
small endoscopes have added a signicant margin
of safety to the procedure when it is performed in
a narrow ventricular system.
Positioning of the patient is crucial; we advocate
keeping the head in a midline position to simplify
the surgeons visualization of the anatomic landmarks. In a small space, such as the third ventricle,
disorientation can quickly lead to injury of the
hypothalamic structures or the fornix.
Planning of the incision and placement of the
burr hole also play important roles in avoiding
parenchymal injury; a narrow third ventricle
demands a more medially placed burr hole, with
a slightly more vertical trajectory (Fig. 1).
Knowledge of the relevant anatomy is crucial
because it allows the trajectory of the endoscope
to be visualized before surgery and the procedure
to be mentally rehearsed before making the
incision. Once the patient is draped, the surgeon
loses access to most of the external landmarks that
dene the trajectory of the endoscope and
reorientation may be dicult.
During surgery, emphasis must be placed on
controlled ecient movement of the endoscope
with constant identication of anatomically relevant structures. When using small endoscopes
without a working channel, we place a peel-away
Fig. 1. An artists illustration of the proper approach

angle for endoscopic third ventriculostomy. The burr
hole is made at the coronal suture in the midpupillary
line. This provides an appropriate angle of approach
through the foramen of Monro to the oor of the third
ventricle.
sheath through the foramen of Monro under

direct visualization; this prevents injury to the
structures surrounding the foramen during removal and reinsertion of the endoscope and
allows passage of instruments into the operative
area without risk of injury to adjacent structures.
Familiarity with the particular endoscope in
use is essential. When conditions conspire to make
visualization or perforation of the oor of the
third ventricle dicult, there should be a low
threshold for abandoning the procedure. This is
considered minimally invasive surgery; another
attempt can be made at a later time, or a shunt
can be placed.
Bradycardia/asystole
Several authors have reported transient bradycardia or asystole when perforating the oor of
the third ventricle or while enlarging the ventriculostomy [2,7,9,16]. This is presumably caused by
a direct mechanical eect on the hypothalamus or
as the eect of irrigation. Irrigation uid can be
trapped in the third ventricle if the endoscope lls
the foramen of Monro, and the addition of
a seemingly insignicant volume may cause
a dramatic increase in pressure. Handler et al
[16] reported a ventricular arrhythmia progressing
to asystole while irrigating with the endoscope in
the third ventricle. This was believed to be
consistent with expansion of the third ventricle
while the endoscope prevented egress of uid
through the foramen of Monro. Prevention of
these intraoperative complications lies in limiting
the rate of irrigation, especially in the third
ventricle, and in leaving one port of the endoscope
open for the egress of uid. Management consists
of ceasing irrigation and withdrawing the endoscope until the patient is hemodynamically stable.
63
continued irrigation. Brisk bleeding may also be

encountered after injury to intraventricular veins
and to perforating arteries and bridging veins in
the interpeduncular cistern. Brisk bleeding requires signicant irrigation and may require
abandonment of the procedure. A pathway for
CSF to exit is crucial in this circumstance.
Without a doubt, the most important and
potentially devastating complication of ETV is
injury to the basilar, posterior cerebral, or
posterior communicating arteries. After arterial
injury, major hemorrhage, vasospasm, pseudoaneurysm, and delayed subarachnoid hemorrhage
have been reported, with signicant morbidity
and mortality (Fig. 2) [1719]. Although the
reported incidence is low, this author is aware of
other as yet unreported cases. Because the
consequences of arterial injury are so profound,
it is essential to inform patients and their parents
that major arterial injury is a real possibility.
Avoidance of hemorrhage is of utmost importance. Careful patient evaluation and selection
may exclude patients at high risk or at least alert
the surgeon to the specic risk factors. MRI and
MR angiography are useful not only to establish
the third ventricular anatomy but may demonstrate the location of the basilar artery in relation
to the oor. During the procedure, the anatomy of
the oor of the third ventricle must be carefully
analyzed. The oor may appear opaque, and in
Vascular complications
Minor bleeding that is easily controlled with
irrigation is common during ETV. Teo et al [2]
reported clinically insignicant hemorrhage in 6
of 55 endoscopic ventriculostomies. Of 98 cases,
Hopf et al [8] reported venous bleeding in 3 cases,
arterial bleeding in 1 case, and 1 case in which
there was bleeding from a bridging vein during the
opening. Fukuhara et al [9] reported intraventricular hemorrhage necessitating placement of
External Ventricular Drainage in 2 of 89 cases.
Small vessels may be encountered within the
oor of the third ventricle. This type of bleeding
can often be controlled with patience and
Fig. 2. Anteroposterior view of a basilar artery angiogram demonstrating a pseudoaneurysm of the posterior
communicating artery at the junction with the basilar
artery. This aneurysm created by injury to the basilar
artery during endoscopic third ventriculostomy was
successfully treated.
64
some cases, especially in myelomeningocele patients, the walls of the hypothalamus may be
partially or entirely fused. If an area of thinning of
the oor can be identied, the procedure can be
performed; the risk, however, is increased with
abnormal anatomy, and under such conditions,
only an experienced endoscopist should attempt
the procedure. A microvascular Doppler probe
has been employed successfully in some cases to
pinpoint the location of the basilar artery through
an opaque oor [20,21].
The management of most intraoperative hemorrhage consists of irrigation and patience, as
previously noted. Electrocautery is rarely helpful
and is certainly dangerous in cramped ventricular
spaces. In certain instances where the source of
bleeding is in the oor of the third ventricle at the
site of fenestration, a balloon catheter can be used
to apply hemostatic pressure to the bleeding tissue.
In our reported case of basilar artery injury,
the patient made a full recovery [17]. We believe
this was in large part a result of the fact that we
had positioned the tip of the peel-away sheath
within the third ventricle, allowing the arterial
blood to exit the sheath instead of lling the
ventricular system. An external ventricular catheter should be placed if any signicant intraventricular hemorrhage has occurred. A cerebral
angiogram should be obtained within 1 to 3 days
of any signicant hemorrhage to rule out a pseudoaneurysm.
Intraoperative bleeding is dicult to quantify
as to its signicance, and there seem to be diering
opinions in the literature regarding what constitutes reportable bleeding. It is essential for
the further renement of this technique that
hemorrhagic complications be reported.
dependent risk factors for ventriculostomy failure

and that nearly half of ETV failures occurred within
2 weeks of the procedure, with only 3 of 89 cases
failing more than 10 months after ETV [7]. It has
been suggested that delayed failures are most often
a result of closure of the ventriculostomy and can be
managed by repeating the procedure [23].
Basic principles of complication avoidance
Patient selection
As experience has grown and the technique has
been rened, third ventriculostomy has been
performed successfully and safely in patients with
obstructive hydrocephalus associated with aqueductal stenosis, posterior fossa masses, tectal plate
tumors, myelomeningocele, and slit ventricle syndrome (Fig. 3). Relative contraindications, however, such as postinfectious and posthemorrhagic
hydrocephalus, still exist. It is up to the surgeon to
match his or her level of experience, skill, and
condence with the patients anatomy and to
inform the patient and family of the risks accordingly. For example, it may be appropriate in certain
circumstances to oer ETV to a patient with
hydrocephalus secondary to prematurity and
intraventricular hemorrhage. The family should
be aware that the procedure has a signicantly
Failure of third ventriculostomy

The success of third ventriculostomy lies in
large part in selecting patients whose CSF
physiology can respond favorably to the procedure. Initially, the procedure was performed
primarily for late-onset aqueductal stenosis, with
good results [22]. As familiarity with the technique
has grown, the indications have broadened and
patients with noncommunicating hydrocephalus
resulting from early-onset aqueductal stenosis,
posterior fossa masses, tectal plate tumors, and
myelomeningocele have been included as well as
younger patients [7,2327].
A recent review found that a history of shunt
infection and postoperative meningitis were in-
Fig. 3. An endoscopic view of the oor of the third

ventricle in a patient with myelomeningocele. Almost
complete fusion of the oor of the third ventricle is
present. The mamillary bodies are not identiable. There
is thinning of the oor in two areas. Successful
endoscopic third ventriculostomy was accomplished
through the most posterior area of thinning.
less chance of success, however. In these patients,

preoperative MRI should play a large role in
determining the patients suitability for the
procedure.
65
especially in previously shunted patients whose

shunts have been removed. With the exception of
infants, we prefer to leave an external ventricular
drain in place and monitor the patient in the
intensive care unit after surgery.
Proper training
Manipulating the endoscope is a learned skill
and one for which there can be a steep learning
curve. Familiarity with the endoscope, knowledge
of the relevant anatomy, and preoperative visualization of the trajectory are important factors in
avoiding complications. We are now in an era
when computer modeling of surgical procedures is
possible, and this resource will play an increasingly important role in training for ETV.
Positioning
Proper positioning of the patient is crucial to
maintain orientation during the procedure. The
head should be in the midline position and slightly
elevated. The burr hole should be placed with the
nal trajectory in mind, and the anatomy and
trajectory should be visualized before the patient
is covered by surgical drapes (see Fig. 1).
Endoscopic equipment
The endoscope with the smallest diameter
appropriate to the patient should be chosen. The
method of perforation of the oor is a subject on
which there are many diering opinions. The most
popular method is to use a pointed (but not sharp)
instrument to open the oor and then to expand
the opening with a balloon. We rmly believe that
laser, cautery, and other forms of energy should
not be employed because they signicantly increase the risk of arterial injury [17,18].
Intraoperative management of complications
Irrigation and patience are the two most
important methods for achieving hemostasis
during ETV. The surgeon should also remember
that ETV is a minimally invasive procedure and
should be abandoned if conditions (eg, bleeding,
unfavorable anatomy) conspire to make fenestrating the oor too dangerous. If signicant bleeding
occurs, an external ventricular drain should be left
in place at the close of the procedure.
Postoperative management
In our experience, intracranial pressure remains elevated for 24 to 48 hours after ETV,
Summary
As experience with ETV grows, the procedure
will be performed by an increasing number of
neurosurgeons. Although the technique has been
greatly rened since its advent almost a century
ago, todays neurosurgeon must never forget that
this seemingly simple procedure holds the potential for a number of devastating complications.
Appropriate training and experience are important to the success of ETV and for avoiding
complications It is imperative that surgeons
continue to report their experience with the
complications of ETV so that the procedure can
continue to be made as safe as possible.
References
[1] Drake J, Kestle J, Milner R, Cinalli G, Boop F,
Piatt J, et al. Randomized trial of cerebrospinal
uid shunt valve design in pediatric hydrocephalus.
Neurosurgery 1998;43:294303.
[2] Teo C, Rahman S, Boop FA, Cherny B. Complications of endoscopic neurosurgery. Childs Nerv
Syst 1996;12:24853.
[3] Guzelbag E, Ersahin Y, Mutluer S. Cerebrospinal
uid shunt complications. Turk J Pediatr 1997;39:
36371.
[4] Iskandar B, Tubbs S, Mapstone T, Grabb P,
Bartolucci A, Oakes W. Death in shunted hydrocephalic children in the 1990s. Pediatr Neurosurg
1998;28:1736.
[5] Lazare J, Peacock W, Holly L, Halen JV, Wong
A, Olmstead C. Multiple shunt failures: an analysis
of relevant factors. Childs Nerv Syst 1998;14:2715.
[6] Lee JY, Wang KC, Cho BK. Functioning periods
and complications of 246 cerebrospinal uid shunting procedures in 208 children. J Korean Med Sci
1995;10:27580.
[7] Brockmeyer D, Abtin K, Carey L, Walker M.
Endoscopic third ventriculostomy: an outcome
analysis. Pediatr Neurosurg 1998;28:23640.
[8] Hopf NJ, Grunert P, Fries G, Resch KD,
Perneczky A. Endoscopic third ventriculostomy:
outcome analysis of 100 consecutive procedures [see
comments]. Neurosurgery 1999;44:795806.
[9] Fukuhara T, Vorster S, Luciano M. Risk factors for
failure of endoscopic third ventriculostomy for
obstructive hydrocephalus. Neurosurgery 2000;46:
11009.
66
[10] Drake JM. Ventriculostomy for treatment of hydrocephalus. Neurosurg Clin North Am 1993;4:65766.
[11] Lowry D, Lowry D, Berga S, Adelson P, Roberts
M. Secondary amenorrhea due to hydrocephalus
treated with endoscopic ventriculocisternostomy.
Case report. J Neurosurg 1996;85:114852.
[12] Homan HJ, Harwood-Nash D, Gilday DL.
Percutaneous third ventriculostomy in the management of noncommunicating hydrocephalus. Neurosurgery 1980;7:31321.
[13] Jones R, Stening W, Brydon M. Endoscopic third
ventriculostomy. Neurosurgery 1990;26:8691.
[14] Baskin JJ, Manwaring KH, Rekate HL. Ventricular
[15] Reddy K, Fewer HD, West M, Hill NC. Slit
ventricle syndrome with aqueduct stenosis: third
ventriculostomy as denitive treatment [see comments]. Neurosurgery 1988;23:7569.
[16] Handler MH, Abbott R, Lee M. A near-fatal
complication of endoscopic third ventriculostomy:
case report. Neurosurgery 1994;35:5258.
[17] Abtin K, Thompson B, Walker M. Basilar artery
perforation as a complication of endoscopic third
ventriculostomy. Pediatr Neurosurg 1998;28:3541.
[18] McLaughlin M, Wahlig J, Kaufman A, Albright A.
Traumatic basilar aneurysm after endoscopic third
ventriculostomy: case report. Neurosurgery 1997;
41:14003.
[19] Schroeder HW, Warzok RW, Assaf JA, Gaab MR.
Fatal subarachnoid hemorrhage after endoscopic
third ventriculostomy. Case report. J Neurosurg
1999;90:1535.
[20] Cartmill M, Vloeberghs M. The use of transendoscopic Doppler ultrasound as a safety-enhancing measure during neuroendoscopic third
ventriculostomy. Eur J Pediatr Surg Suppl 1999;1:
501.
[21] Schmidt R. Use of microvascular Doppler probe to
avoid basilar artery injury during endoscopic third
ventriculostomy. J Neurosurg 1999;90:1568.
[22] Hirsch J, Hirsch E, Sainte-Rose C, Renier D,
Pierre-Kahn A. Stenosis of the aqueduct of Sylvius.
Etiology and treatment. J Neurosurg Sci 1986;30:
2939.
[23] Cinalli G, Sainte-Rose C, Chumas P, Zerah M,
Brunelle F, Lot G, et al. Failure of third ventriculostomy in the treatment of aqueductal stenosis in
children. J Neurosurg 1999;90:44854.
[24] Cinalli G, Salazar C, Mallucci C, Yada JZ, Zerah
M, Sainte-Rose C. The role of endoscopic third
ventriculostomy in the management of shunt
malfunction. Neurosurgery 1998;43:13239.
[25] Jones R, Kwok B, Stening W, Vonau M. The
current status of endoscopic third ventriculostomy
in the management of non-communicating hydrocephalus. Minim Invasive Neurosurg 1994;37:
2836.
[26] Jones R, Kwok B, Stening W, Vonau M. Third
ventriculostomy for hydrocephalus associated with
spinal dysraphism: indications and contraindications. Eur J Pediatr Surg 1996;6:56.
5763.
Results of endoscopic third ventriculostomy

Mark R. Iantosca, MDa,*, Walter J. Hader, MD, FRCS(C)b,
James M. Drake, FRCS(C)c
a
Connecticut Childrens Medical Center, 100 Retreat Avenue, Suite 705,

Hartford, CT 061062565, USA
b
Division of Neurosurgery, University of Calgary, Alberta Childrens Hospital,
Calgary, Alberta, Canada
c
Hospital for Sick Children, 555 University Avenue, Toronto,
Ontario M5G 1X8 Canada
The recent resurgence of interest in ventriculostomy has arisen out of dissatisfaction with the
complications and long-term outcomes of conventional cerebrospinal uid (CSF) shunting
systems [1,2]. Initial attempts at ventriculostomy
via open craniotomy and subsequent percutaneous uoroscopic and CT-guided techniques have
largely been replaced by modern endoscopic
procedures with reduced morbidity [3,4]. Unfortunately, sucient long-term follow-up data
necessary for direct comparison with outcomes of
conventional shunting (CS) procedures are currently unavailable. Clinical decisions regarding
use of endoscopic third ventriculostomy (ETV)
are based on the results of studies with mean
follow-up intervals of less than 5 years. Results of
ETV are most closely associated with the etiology
of hydrocephalus encountered as well as with the
clinical and radiographic features of the individual patient.
Factors aecting outcome
Hydrocephalus etiology
Third ventriculostomy is designed to treat
noncommunicating hydrocephalus with patent
subarachnoid spaces and adequate CSF absorption. It is not surprising, therefore, that the results
of this procedure are most strongly inuenced by

the etiology of hydrocephalus. Box 1 depicts
hydrocephalus etiologies divided according to
reported success rates of ETV. Patients with
acquired aqueductal stenosis or tumors obstructing third ventricular outow have demonstrated
the highest success rates, exceeding 75% in
carefully selected series of patients [410]. Previously shunted patients with or without myelomeningocele, tumors, or cystic abnormalities
leading to fourth ventricular outow obstruction
(ie, arachnoid cyst, Dandy-Walker malformations)
and patients with congenital aqueductal stenosis
have shown an intermediate response [6,7,9,11
13]. Further study of this intermediate group is
likely to identify subgroups with higher success
rates. Infants developing hydrocephalus after
hemorrhage or infection or with associated myelomeningocele (without prior shunting procedure)
have demonstrated a poor response to ventriculostomy [3,5,1316]; despite limited reports of
success in such patients [7,11,12,1720], they are
more controversial candidates for this procedure.
The procedure is not advisable in patients who
have undergone prior radiation therapy because
of the extremely poor response rates, altered
anatomy (ie, thickened third ventricular oor),
and increased risk of bleeding [3,7,11].
Aqueductal obstruction by stenosis or tumor
E-mail address: miantos@ccmckids.org
(M.R. Iantosca).
Blockage to CSF ow at the level of the cerebral aqueduct by anatomic or postinammatory

septations/membranes or by extrinsic compression
doi:10.1016/S1042-3680(03)00067-6
68
M.R. Iantosca et al / Neurosurg Clin N Am 15 (2004) 6775
Box 1. Endoscopic third

ventriculostomy success rates by
hydrocephalus etiology
High success rates (>75%)
Acquired aqueductal stenosis
Tumor, cyst, or infectious lesion
obstructing third ventricular outflow
Tectal, pineal, thalamic, or
intraventricular tumor
Arachnoid cyst
Toxoplasmosis
Shunt malfunction in patient with
obstructive etiology
Intermediate success rates
(approximately 50%)
Tumor obstructing fourth ventricular
outflow
Myelomeningocele (previously shunted,
older patients)
Congenital aqueductal stenosis
Cystic abnormalities obstructing fourth
ventricular outflow
Arachnoid cysts
Dandy-Walker malformation
Previously shunted patients with
difficulties
Slit ventricle syndrome
Recurrent or intractable shunt infections
Recurrent or intractable shunt
malfunctions
Low success rates (<50%)
Myelomeningocele (previously
unshunted neonatal patients)
Posthemorrhagic hydrocephalus
Postinfectious hydrocephalus (excluding
aqueductal stenosis of infectious
etiology)
secondary to tumor, benign cyst, or vascular

lesions in a surrounding structure (tectum, pineal,
posterior thalamus, and superior cerebellum) is
currently the best-studied and most successful
indication for ETV. Success rates generally exceed
75% for this select group of patients [410]. In
a recent large clinical series, including a wide age
range of patients (3 weeks to 77 years, n = 95)
Hopf et al [9] reported an ETV success rate
greater than 80% for postinammatory and
idiopathic aqueductal stenosis and a 95% success
rate for aqueductal stenosis caused by benign
space-occupying lesions. ETV was less successful

in patients with progressive neoplastic lesions,
primarily posterior fossa tumors and posthemorrhagic (intraventricular or subarachnoid) cases
(64% and 63%, respectively) [9]. Similar results
have recently been reported in a series analyzing
213 children with similar etiologies by Cinalli et al
[4] encompassing stereotactic ventriculostomy and
ETV. The less favorable outcomes associated with
posthemorrhagic and posterior fossa tumors and
cysts likely reect the multifactorial causes of
hydrocephalus in these patients [4,9]. Studies of
ETV for other hydrocephalic etiologies are limited
to smaller series and require further investigation
to establish overall success rates.
Clinical features
Box 2 depicts preoperative clinical and radiographic criteria frequently cited for improving
outcome and limiting morbidity of ETV. A recent
study by Gangemi et al [10] oers the only
detailed assessment of the contribution of presenting symptoms and signs to procedure success.
Clinical presentations indicative of elevated intracranial pressure (ICP) were associated with the
highest rate of ETV success (83%), followed by
Parinauds syndrome (77%) and macrocephaly
(71%). Other presentations (memory disturbances, incontinence, and focal decits with associated tumors) exhibited success rates of
approximately 50% or less. Patient age and
history of prior shunting procedure are the most
frequently cited clinical features associated with
outcome.
Age
There seems to be a signicant association
between increasing patient age and a more favorable outcome of ETV in multiple pediatric studies
[69,11,12,15,16,21]. These studies frequently implicate a decreased pressure gradient across the
arachnoid granulations (secondary to open sutures) or age-related immaturity of patient CSF
absorptive capacity as reasons for this phenomenon. Evidence suggests that reduced success rates
are associated with the age at which hydrocephalus initially developed as well as with the age at
the time of ventriculostomy [7,21]. The series
by Jones et al [21] reports that only 8 of 25
ventriculostomies were successful in patients less
than 6 months of age, with 8 of 17 successful
procedures in patients developing hydrocephalus
69
Box 2. Favorable clinical and radiographic features for endoscopic third ventriculostomy
Clinical
Etiology of hydrocephalus in high or intermediate success group (see Box 1.)
Age greater than 6 months at time of hydrocephalus diagnosis
Age greater than 1 year at time of procedure (controversial)
No prior radiation therapy
No history of hemorrhage or meningitis
Previously shunted, particularly for high success group etiology (see Box 1.)
Radiographic
Clear evidence of ventricular noncommunication
Obstructive pattern of HCPQ
Aqueductal anatomic obstruction
Absence of aqueductal flow on T2-weighted MRI or cine study
Favorable third ventricular anatomy
Width and foramen of Monro sufficient to accommodate endoscope
Rigid > 6 mm
Flexible > 4 mm
Thinned floor of third ventricle
Downward bulging floor draped over clivus
Basilar posterior to mamillary bodies
Absence of structural anomalies impeding procedure
Arteriovenous malformation (AVM) or tumor obscuring third ventricular floor
Enlarged massa intermedia
Insufficient space between mamillary bodies/basilar and clivus
Basilar artery ectasia
before this age. Most of the successes in these

groups were in previously shunted patients [21].
Several studies show success rates at or below
50% in patients less than 2 [6,7,9] or 1 [8,15,16]
year of age, regardless of etiology. Results are
even poorer in patients less than 6 months of age
[12,21]. Teo and Jones [12] demonstrated statistically signicant (P = 0.005) lower success rates in
patients less than 6 months of age versus older
patients (12.5% vs. 80%). Open third ventriculostomy has also been shown to have a dramatically decreased success rate in patients less than 6
months of age [18,22]. The results of these studies
are seemingly contradicted by the recent report of
Cinalli et al [4], who demonstrated no dierence in
failure rates of ETV between patient groups older
or younger than 6 months, concluding that patient
age should no longer be considered a contraindication to the procedure.
Despite the lower success rates and increased
morbidity that they report in this younger age
group, several authors still advocate an initial
attempt at ventriculostomy in these patients so as
to avoid the lifelong complications of shunting

[8,15,16]. These authors argue that the successes
are spared the long-term morbidity, mortality,
and economic burden of repeated evaluations and
hospitalizations for shunt evaluation. Barlow and
Ching [23] recently published an analysis of
anticipated cost reduction comparing ETV and
CS for treatment of obstructive hydrocephalus at
their institution, estimating a reduction of 74
hospital days and nine surgeries annually. This
study, however, assumes a success rate of 80%,
with no subsequent admissions or procedures for
complications or failures among the successful
procedures. Insucient long-term data currently
exist to support this argument in younger patient
groups, particularly given the often-cited reduced
success rates and increased rates of complications.
ETV in this younger age group has also been
advocated as a primary treatment to remove
intraventricular hemorrhage or purulent CSF,
assuming that this will result in a more suitable
environment for shunt insertion [15,24,25].
Comparison of this procedure with standard
70
temporizing measures, such as ventriculostomy,

ommaya placement, and ventriculosubgaleal shunting is needed to determine comparative ecacy
and safety.
Interestingly, a recent report by Tisell et al [26]
demonstrates an increased failure rate of ETV
in older adults with aqueductal stenosis. These
delayed failures were hypothetically attributed to
decreasing CSF absorptive capacity as a result of
age-related loss of arachnoid villi and decreasing
resistance to transependymal CSF resorption over
time. Only 50% of ETVs were successful in this
series, with almost all failures occurring after
an initially favorable response lasting between
1 month and 1 year [26]. An earlier report by
Kelly [5] demonstrated a 94% success rate in
a population of adults with the same etiology
treated with stereotactic CT-guided ventriculostomy. Interestingly, most patients in Kellys series
were previously shunted (11 of 16), whereas only 3
of Tissell et als 18 patients had existing shunts.
Clearly, further delineation of the pathophysiology of age-related changes in CSF absorption,
and development of studies to quantify this
capacity, would greatly benet clinicians attempting to identify appropriate patients for ETV.
Previous shunting
Multiple studies have demonstrated a trend
toward more successful ventriculostomy outcome
in patients with existing shunt systems [6,12,21].
Teo and Jones [12] demonstrated a statistically
signicant dierence in success rates for previously shunted patients with myelomeningocele
(84%) versus those never shunted (29%;
P = 0.002). In fact, third ventriculostomy has
been found to be useful in the treatment of
intractable shunt infections and malfunctions and
even slit ventricle syndrome refractory to other
treatments [5,11,24,2730]. Other series of ventriculostomy in previously shunted patients have
been less promising [13,31]. These contradictory
results may reect the mix of patients in small
series. High success rates in patients with aqueductal obstructive etiologies present a strong
argument in favor of ETV for patients in this
group who present with shunt malfunction. Improved outcome in previously shunted patients
with other etiologies is commonly attributed to
increased CSF absorptive capacity; however, the
eect of the shunt itself on CSF absorption is
dicult to distinguish from the eect of increased
age in these patients.
Preoperative assessment of CSF absorptive

capacity has been advocated by some authors
[32,33], but these techniques have not been widely
accepted [7,21]. Because the patency of CSF
pathways is one of the assumptions on which
the procedure is based, a preoperative CSF
absorption study would be invaluable in identifying patients who would likely benet from
ventriculostomy. The observed delay between
operation and decrease in ventricular size (see
section on assessing outcome) suggests that CSF
absorption may increase slowly after ventriculostomy, however. Therefore, preoperative assessment of CSF absorptive capacity may fail to
identify all appropriate patients [7]. Until an
accurate functional predictor of this capacity is
practical, radiographic assessment of the obstructive pattern of hydrocephalus by MRI is likely to
remain the preoperative diagnostic procedure of
choice.
Anatomy
Preoperative MRI optimally demonstrates all
relevant anatomic features and should be obtained for all proposed third ventriculostomy
patients. Initially, conrmation of noncommunicating hydrocephalus of favorable etiology should
be established by the pattern of ventricular
dilatation. Anatomic obstruction of CSF pathways between the aqueduct and the fourth
ventricular outow foramina may be visible on
T2- or T1-weighted contrast images. Additionally,
T2-weighted and cine images should reveal no
evidence of the normally present aqueductal CSF
ow pattern.
Once the patients suitability for the procedure
has been established, the neurosurgeon must
clarify the details of third ventricular anatomy
that are likely to aect morbidity. First, the width
of the third ventricle and diameter of the foramen
of Monro must be sucient to accommodate
the endoscope of choice. Additionally, the thickness of the third ventricular oor and anatomy of
the proposed puncture site must be assessed in
relation to vital structures, particularly the basilar
artery and its branches. A downward bulging
third ventricular oor draped over the clivus has
been cited as a prerequisite for success of this
procedure in the past, but others have not found
this to be necessary [3,7]. Ultimately, the surgeon
must be satised that there is no structural lesion
(ie, tumor, AVM) or anatomic variation that
would render the procedure unduly dicult or
hazardous. In cases of doubt, it is reasonable to

visualize the oor of the third ventricle and to
abandon the procedure if the oor is unsuitable.
Reported procedure abort rates as a result of
unfavorable anatomy or intraoperative bleeding
range from 0% to 26% (Table 1).
Assessing outcome
ETV has yielded a higher success rate with
lower morbidity and mortality than earlier
methods of third ventriculostomy. Mortality rates
for open ventriculostomy procedures varied between 5% and 27%, with success rates from 37%
to 75% [3,5,18,31,34,35]. Percutaneous radiographic and, later, CT-guided techniques reduced
this mortality rate to 2% to 7%, with a 44% to
75% rate of shunt independence [3,5,14,19,31,36].
Most recent studies using modern endoscopic
techniques and equipment, with or without
stereotactic CT or MRI guidance, have reported
low morbidity (3%12%) and extremely rare
mortality, with success rates greater than 75%
for carefully selected patient groups. Table 1
depicts the results of recent ETV studies with
success rates by etiology where these data are
available. The current challenge is to dene
appropriate indications and devise objective
measures for pre- and postoperative assessment
of these patients.
The goal of ETV and, to date, the best
objectively quantiable measure of a successful
outcome is shunt independence. Vague outcome
measures, such as successful, improved, or
favorable, are used in a number of studies
[12,21]. Only two reports have critically compared
modern ETV techniques with CS [37,38]. SainteRose [38] reported no statistically signicant
dierences in measures of neurologic, endocrinologic, social, or behavioral outcomes in a group of
68 patients treated for aqueductal stenosis with
either ETV (n = 30) or CS (n = 38). Tuli et al [37]
demonstrated no dierence in failure rate between
ETV (n = 32, 44% failure rate) and CS (n = 210,
45% failure rate) in a prospectively followed group
of patients with aqueductal stenosis or tumor. The
long-term outcomes of CS, including multiple
procedures for obstruction, infection, and overdrainage, have been well documented. Failure
rates range from 30% to 47% over 1 year to
63% to 70% over 10 years [3942]. Additional
long-term studies of ETV and analytic comparison
to established CS outcomes are necessary to clarify
clinical decisions in these patients.
71
One of the most confusing aspects of outcome

evaluation in ETV patients is the failure of the
ventricles to return to normal size. Many reports
detail a gradual decrease in the ventricular size
over months to years after surgery, with resolution
of periventricular edema and increased extracerebral spaces coinciding with clinical improvement
[58,31,38]. One series of patients treated with
either CSF shunts or ETV showed no dierence in
intellectual outcome despite enlarged ventricles in
the ETV group [38]. Radiographic evaluation of
suspected closure has been confounded by the
presence of persistently enlarged ventricles.
Studies detailing the serial measurements of
multiple radiographic indices of ventricular size
after surgery show that the third ventricular size
responds more quickly (usually within 3 months)
than the lateral ventricular size (up to 2 years)
[43,44]. Additionally, third ventricular size seems
to correlate most closely with outcome in these
patients [8,43,44]. A recent study limited to adult
patients showed no relation between ventricular
size reduction and outcome [26]. Kulkarni et al
[45] have demonstrated a greater reduction in
ventricular size after successful procedures versus
treatment failures in 29 children.
MRI and Doppler ultrasound studies documenting CSF ow through a patent ventriculostomy are currently the most useful and widely
used postoperative radiographic indices. MRI
detection of T2-weighted ow void around the
ventriculostomy has been correlated with clinical
outcome in ETV [4648]. This observation has
proven most helpful in conrming ventriculostomy patency after surgery, particularly in
patients with persistent ventriculomegaly [47,48].
Kulkarni et al [45] have recently demonstrated
a statistically signicant relation between evidence
of postoperative aqueductal CSF ow on MRI
and clinical success. Multiple studies have reported long-term patency of third ventriculostomy by means of cine phase contrast (PC) MRI
studies. A report by Cinalli et al [4] included 15
patients with conrmed long-term (>10 years)
patency of their ventriculostomy on cine PC MRI
(median = 14.3 years, range: 1017 years). Unfortunately, several recent studies in adults and
children have demonstrated a lack of correlation
between postoperative MRI ndings and outcome
[15,16,26,37]. Quantication of ow velocity
through ventriculostomies has also been achieved
by PC MRI and Doppler ultrasound [49,50].
Hopefully, methods of quantifying ventriculostomy function will allow neurosurgeons to rene
72
Table 1
Outcome of endoscopic third ventriculostomy by etiology
Morbidity
Jones et al / 1990 [11]

Jones et al / 1992 [6]
Jones et al / 1994 [7]
Sainte-Rose and Chumass / 1996 [8]
Teo and Jones/ 1996 [12]
Goumnerova and Frim / 1997 [47]
Baskin et al / 1998 [30]
Brockmeyer et al / 1998 [13]
Buxton et al / 1998 [16]
Buxton et al / 1998 [15]
Tuli et al / 1998 [37]
Gangemi et al / 1999 [10]
Cinalli et al / 1999 [4]
Hopf et al / 1999 [9]
Tisell et al / 2000 [26]
(4 M17 Y)
(PNB78 Y)
N/R
5.3 Y
11 Y (1 W32 Y)
11.2 Y (2 D36 Y)
17.3 Y (1.549 Y)
8.1 Y (1 D29.5Y )
3.7 M (0 M10 M)
2 M (011 M)* All PNB
8.1 Y (0 M18 Y)
31 Y (7 D81 Y)
NR (1 M18 Y)
36 Y (3 W77 Y)
48 Y (1780 Y)
8%
7%
5%
N/R
3%
9%
12%
6%
15%
21%
N/R
12%
N/R
6%
N/R
Procedure
abort
rate
Follow-up
mean
(range)
16%
9%
6%
N/R
9%
N/R
N/R
26%
0.4%
11%
N/R
N/R
27 M (3 M7 Y)
N/R
1.8 Y
32 M (1 l17 Y)
17 M (7 l44 M)
(21.4 M)
24.2M (15 l69 M)
N/R (6 l42 M)
(6 l42 M)
N/R (1 Y11 Y)
28 M (12 l54 M)
2.1 Y (4 D9 Y)
26 M (3 l71 l)
37 M (3 M5 Y)
6%
2%
N/R
Success rate (%) by

etiology (n)
Tumor/AS
Dysraphic
PHH/PIH
59%
68%
80%
81%
40% (5)
40% (10)
52% (21)
0% (2)
(17)
(19)
(25)
(82)
Other (n-etiology)
72% (69)
70% (20)
59%
57%
43%
66%
91%
86%
83%
50%
(34)
(7)
(7)
(32)
(110)
(119)
(82)
(18)
0% (1)
50% (16)
0% (2)
0% (7)
10% (10)
30% (10)
100% (2-IDO)
64% (22-SVS)
20% (5-IDO)
63% (8)
Abbreviations: AS aqueductal stenosis; D, days; IDO, idivpathie hydrocephalues; m, months; N/R, not reported; MMC, myelomeningocele, PHH post hemorrhagic
hydrocephalus; PIH post infectious hydrocephalus; PNB, Premature newborn, SC; shunt complications (intractable shunt infection/malfunction); SVS, slit-ventricle
syndrome; W, weeks; Y, years.
* All patients born prematurely: mean age at birth was 31.9 weeks (range: 2636 weeks).
Author/Year
Age mean
(range)
the techniques necessary to improve outcome

further. Radiographic conrmation may also help
to dene indications with more subjective outcomes, for example, the observation of less
fulminant shunt malfunctions in patients after
ventriculostomy [11].
The authors use postoperative MRI imaging
with sagittal T2-weighted images or cine MRI to
conrm ventriculostomy ow and assess ventricular size 3 to 6 months after surgery. The absence
of a ow void, although occasionally present in
patients with good clinical outcome, was present
in 12 of 13 patients presenting with delayed
treatment failure in the series by Cinalli et al [4].
Increasing head circumference, signs and symptoms of elevated ICP, or evidence of CSF egress at
the ventriculostomy incision site generally heralds
early failures. Because of the risk of delayed
failure of ETV, and the potential mortality if
this is unrecognized, we recommend that these
patients use medical alert bracelets or wallet
identication cards. A low threshold of suspicion
for reimaging these patients should be followed,
especially for the rst 5 years after ETV. Families
and caregivers are given the same education as
for patients with shunts regarding symptoms of
failure. Education and regular follow-up are
important to counteract the false sense of security
that may arise in the absence of a shunt.
Failure of endoscopic third ventriculostomy
Studies of ETV are currently limited to studies
with an average of less than 5 years of follow-up.
These studies have documented early (<6 months)
and late (>1 year) postoperative failures. Early
failures are more frequently reported in patients
less than 6 months of age [4,15,16]. Buxton et al
[15,16] have reported a mean time to failure in
patients younger than 1 year of age of 1.36
months. Multiple authors have reported late
failures, where the ventriculostomy closes, sometimes years after surgery [4,7,9,24,26]. Obstruction
of the ventriculostomy by a gliotic scar or arachnoid membranes has been described [31,51]. Rare
cases of sudden death after late failure of ETV
have been reported [52].
Although the long-term results of ETV are
largely unknown, studies that include long-term
follow-up on stereotactic ventriculostomy cases
are likely applicable to ETV. Cinalli et als recent
report [4] included 94 cases of stereotactic ventriculostomy with an average of 6.32 years of
follow-up (range: 20 days to 17.4 years) and 119
73
case of ETV with an average of 2.1 years of

follow-up (range: 4 days to 9 years). The overall
functional ventriculostomy rate for the study was
72% at 6 years, without signicant dierences in
the long-term results between the stereotactic and
endoscopic groups [4]. Cine PC MRI studies on 15
patients up to 17 years after stereotactic ventriculostomy conrmed long-term patency (range: 10
17 years, median = 14.3 years) [4]. No failures
were reported in either group after 5 years of
follow-up. Additionally, 7 of 9 patients failing
ventriculostomy with radiographically proven
obstruction of the stoma had successful repeat
ventriculostomies. Long-term data on ETV outcomes are only now beginning to emerge, and
comparison to known outcomes of standard
shunting procedures will be instrumental in dening the indications for initial treatment with
ETV and the management of treatment failures.
Summary
ETV is emerging as the treatment of choice for
aqueductal stenosis caused by anatomic, inammatory, and selected neoplastic etiologies. The
technique has also proven useful in the pathologic
diagnosis and treatment of these conditions
[5356]. Long-term results of this procedure and
comparison to standard shunting procedures are
necessary to dene indications for patients with
pathologic ndings in the intermediate response
groups. Development of new studies for preoperative assessment of CSF absorptive capacity
and quantitative postoperative measures of
ventriculostomy function would be invaluable
additions to our ability to assess candidates for
this procedure and their eventual outcome. Further study and technical renements will, no
doubt, lead to many more potential uses for these
procedures in the treatment of hydrocephalus and
its associated etiologies. The challenge for neurosurgeons will be to dene the operative indications
and outcomes, while rening techniques for safely
performing these useful procedures.
References
[1] Sainte-Rose C, Homan HJ, Hirsch JF. Shunt
failure. Concepts Pediatr Neurosurg 1989;9:720.
[2] Hoppe-Hirsch E, et al. Late outcome of the surgical
treatment of hydrocephalus. Childs Nerv Syst 1998;
14(3):979.
[3] Drake JM. Ventriculostomy for treatment of
hydrocephalus. Neurosurg Clin North Am 1993;
4(4):65766.
74
[4] Cinalli G, et al. Failure of third ventriculostomy in

the treatment of aqueductal stenosis in children.
J Neurosurg 1999;90(3):44854.
[5] Kelly PJ. Stereotactic third ventriculostomy in
patients with nontumoral adolescent/adult onset
aqueductal stenosis and symptomatic hydrocephalus [see comments]. J Neurosurg 1991;75(6):
86573.
[6] Jones RF, et al. Neuroendoscopic third ventriculostomy. In: Manwaring KH, Crone KR, editors.
Neuroendoscopy, vol. 1. New York: Mary Ann
Liebert; 1992. p. 6377.
[7] Jones RF, et al. The current status of endoscopic
third ventriculostomy in the management of noncommunicating hydrocephalus. Minim Invasive
Neurosurg 1994;37(1):2836.
[8] Sainte-Rose C, Chumas P. Endoscopic third ventriculostomy. Tech Neurosurg 1996;1:17684.
[9] Hopf NJ, et al. Endoscopic third ventriculostomy:
outcome analysis of 100 consecutive procedures [see
comments]. Neurosurgery 1999;44(4):795806.
[10] Gangemi M, et al. Endoscopic third ventriculostomy for hydrocephalus. Minim Invasive Neurosurg 1999;42(3):12832.
[11] Jones RF, Stening WA, Brydon M. Endoscopic
third ventriculostomy. Neurosurgery 1990;26(1):
8692.
myelomeningocele. Pediatr Neurosurg 1996;25(2):
5763.
[13] Brockmeyer D, et al. Endoscopic third ventriculostomy: an outcome analysis. Pediatr Neurosurg
1998;28:23640.
[14] Jaksche H, Loew F. Burr hole third ventriculocisternostomy. An unpopular but eective procedure
for treatment of certain forms of occlusive hydrocephalus. Acta Neurochir (Wien) 1986;79:4851.
[15] Buxton N, et al. Neuroendoscopy in the premature
population. Childs Nerv Syst 1998;14(11):64952.
[16] Buxton N, et al. Neuroendoscopic third ventriculostomy in patients less than 1 year old. Pediatr
Neurosurg 1998;29(2):736.
[17] Natelson SE. Early third ventriculostomy in
meningomyelocele infantsshunt independence?
Childs Brain 1981;8(5):3215.
[18] Patterson RH Jr., Bergland RM. The selection of
patients for third ventriculostomy bases on experience with 33 operations. J Neurosurg 1968;29(3):
2524.
[19] Sayers MP, Kosnik EJ. Percutaneous third ventriculostomy: experience and technique. Childs Brain
1976;2(1):2430.
[20] Vries JK, Friedman WA. Postoperative evaluation
of third ventriculostomy patients using 111inDTPA. Childs Brain 1980;6(4):2005.
[21] Jones RF, et al. Neuroendoscopic third ventriculostomy. A practical alternative to extracranial
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
shunts in non-communicating hydrocephalus. Acta

Akdemir H, et al. Failure of open third ventriculostomy for shunt infections in infants. Pediatr
Neurosurg 1999;30:3059.
Barlow P, Ching HS. An economic argument in
favour of endoscopic third ventriculostomy as
a treatment for obstructive hydrocephalus. Minim
Jones RF, et al. Third ventriculostomy for shunt
infections in children. Neurosurgery 1993;32(5):
85560.
Mallucci C, et al. Neuroendoscopic third ventriculostomy: the rst line of treatment for blocked
ventriculoperitoneal shunts? Eur J Pediatr Surg
1997;7(Suppl 3).
Tisell M, et al. How eective is endoscopic third
ventriculostomy in treating adult hydrocephalus
caused by primary aqueductal stenosis? Neurosurgery 2000;46(1):10411.
Yamamoto M, et al. Percutaneous exible neuroendoscopic ventriculostomy in patients with shunt
malfunction as an alternative procedure to shunt
revision. Surg Neurol 1994;42(3):21823.
Perlman BB. Percutaneous third ventriculostomy in
the treatment of a hydrocephalic infant with
aqueduct stenosis. Int Surg 1968;49(5):4438.
Reddy K, et al. Slit ventricle syndrome with
aqueduct stenosis: third ventriculostomy as denitive treatment [see comments]. Neurosurgery
1988;23(6):7569.
Baskin JJ, Manwaring KH, Rekate HL. Ventricular
ventricle syndrome. J Neurosurg 1998;88(3):47884.
Homan HJ, Harwood-Nash D, Gilday DL.
Percutaneous third ventriculostomy in the management of noncommunicating hydrocephalus. Neurosurgery 1980;7(4):31321.
Foltz EL. Treatment of hydrocephalus by ventricular shunts [letter]. Childs Nerv Syst 1996;12(6):
28990.
Pudenz R, Foltz E. Hydrocephalus: overdrainage
by ventricular shuntsa review and recommendations. Surg Neurol 1991;35:20012.
Dandy WE. Diagnosis and treatment of strictures
of the aqueduct of Sylvius (causing hydrocephalus).
Arch Surg 1945;51:114.
Scar JE. Evaluation of treatment of hydrocephalus. Results of third ventriculostomy and endoscopic cauterization of choroid plexuses compared with
mechanical shunts. Arch Neurol 1966;14(4):38291.
Forjaz S, Martelli N, Latuf N. Hypothalamic
ventriculostomy with catheter. Technical note.
J Neurosurg 1968;29(6):6559.
Tuli S, Alshail E, Drake J. Third ventriculostomy
versus cerebrospinal uid shunt as a rst procedure
in pediatric hydrocephalus. Pediatr Neurosurg
1998;30:115.

[38] Sainte-Rose C. Third ventriculostomy. In: Manwaring KH, Crone KR, editors. Neuroendoscopy, vol.
1. New York: Mary Ann Liebert; 1992. p. 4762.
[39] Sainte-Rose C, et al. Mechanical complications in
shunts. Pediatr Neurosurg 1991;17:29.
[40] Di Rocco C, Marchese E, Verladi F. A survey of the
rst complication of newly implanted CSF shunt
devices for the treatment of nontumoral hydrocephalus. Cooperative survey of the 19911992
Education Committee on the ISPN. Childs Nerv
Syst 1994;10(5):3217.
[41] Albright AL, Haines SJ, Taylor FH. Function of
parietal and frontal shunts in childhood hydrocephalus. J Neurosurg 1988;69:8836.
[42] Piatt JH Jr. Cerebrospinal uid shunt failure: late is
dierent from early. Pediatr Neurosurg 1995;23:
1339.
[43] Oka K, et al. The radiographic restoration of the
ventricular system after third ventriculostomy.
[44] Schwartz TH, et al. Third ventriculostomy: postoperative ventricular size and outcome. Minim
[45] Kulkarni AV, et al. Imaging correlates of successful
endoscopic third ventriculostomy. J Neurosurg
2000;92(6):9159.
[46] Jack CR Jr., Kelly PJ. Stereotactic third ventriculostomy: assessment of patency with MR imaging.
AJNR Am J Neuroradiol 1989;10(3):51522.
[47] Goumnerova LC, Frim DM. Treatment of hydrocephalus with third ventriculocisternostomy: out-
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
75
come and CSF ow patterns. Pediatr Neurosurg

1997;27(3):14952.
Wilcock DJ, et al. Neuro-endoscopic third ventriculostomy: evaluation with magnetic resonance
imaging. Clin Radiol 1997;52(1):504.
Lev S, et al. Functional analysis of third ventriculostomy patency with phase-contrast MRI velocity measurements. Neuroradiology 1997;39(3):
1759.
Wilcock DJ, Jaspan T, Punt J. CSF ow through
third ventriculostomy demonstrated with colour
Doppler ultrasonography. Clin Radiol 1996;51(2):
1279.
Hirsch JF, et al. Stenosis of the aqueduct of Sylvius.
Etiology and treatment. J Neurosurg Sci 1986;
30(12):2939.
Hader W, et al. Death following late failures of
third ventriculostomy in children: a report of 3
cases. J Neurosurg 2002, in press.
Ellenbogen RG, Moores LE. Endoscopic management of a pineal and suprasellar germinoma with
associated hydrocephalus: technical case report.
Drake J. Neuroendoscopy tumour biopsy. New
York: Mary Ann Leibert; 1992. p. 1039.
Ferrer E, et al. Neuroendoscopic management of
pineal region tumours. Acta Neurochir (Wien)
1997;139(1):1221.
Oka K, et al. Endoneurosurgical treatment for
hydrocephalus caused by intraventricular tumors.
Childs Nerv Syst 1994;10(3):1626.
Neuro-oncologic applications of endoscopy

Charles Teo, MBBS, FRACSa, Peter Nakaji, MDa,b,*
a
Centre for Minimally Invasive Neurosurgery, Prince of Wales Hospital/University of New South Wales,
Barker Street, Randwick, NSW 2031, Sydney, Australia
b
Division of Neurological Surgery University of California at San Diego Medical Center, San Diego, CA, USA
No other subspecialty of neurosurgery exemplies the advantages of endoscopy more than

neuro-oncology. From a pure technical standpoint, visualization is the sine qua non of surgery.
If a surgeon cannot see what he is doing, he is
ineective and dangerous. Cranial neurosurgery
in particular is a constant struggle against poor
visualization. In an attempt to minimize operative
trauma, the surgeon aims to limit the size of the
exposure and to avoid vigorous brain retraction.
Meanwhile, the tumor/brain interface is often
hard to distinguish, and the tumor often insinuates itself behind or between structures that
cannot be sacriced. Critical structures, such as
cranial nerves, vessels, and eloquent brain tissue,
may prohibit direct visual access to some parts of
the tumor. At the same time, in the case of the
more benign tumors that occur in the cranial
cavity, total excision is vital to the patients
survival [1,2]. Complete removal of an acoustic
neuroma is nearly tantamount to cure. Other
extra-axial tumors, such as craniopharyngiomas
and meningiomas, also must be removed as
completely as possible to reduce the incidence of
tumor recurrence. Because low-grade gliomas
respond poorly to radiotherapy and chemotherapy, prognosis is directly related to the degree
of tumor resection. Ependymomas are also historically insensitive to adjuvant therapies, and
total macroscopic removal is paramount [3]. Any
tool that improves visualization, thereby oering
E-mail address: charles.teo@neuroendoscopy.info
(C. Teo).
patients a better surgical outcome, should be

embraced by the neuro-oncologic surgeon.
The endoscope is such a tool. It enhances the
surgeons view by increasing illumination and
magnication [4,5]. It allows the surgeon to view
tumor remnants, such as those hidden behind the
tentorial edge, a cranial nerve, or eloquent brain
tissue. Once the tumor is removed, the surgeon
can use the endoscope to assess the degree of
resection. Often, the same surgery can be
performed through a smaller craniotomy by using
the endoscope, in keeping with the concept of
minimally invasive yet maximally eective surgery
[6]. Adjunctive procedures, such as third ventriculostomy and septostomy, can be performed
through the same access to manage related
problems, such as secondary hydrocephalus.
Endoscopic tumor removal or cerebrospinal uid
(CSF) diversion may allow patients to avoid shunt
placement. Finally, the endoscope is an excellent
teaching tool. The anatomic denition and unique
angles of view available to the endoscope help
residents with their understanding of operations
and help to illuminate anatomicopathologic concepts underlying neuro-oncologic surgery.
Applications
Applications for the endoscope are limited
only by ones imagination. Initially used for
intraventricular surgery alone, the endoscope has
also become an invaluable tool in the surgical
management of extra-axial pathologic processes.
It is being increasingly applied as an adjunct to the
removal of masses in the subarachnoid space.
Within the ventricle, the endoscope is a versatile
tool for diagnosing and treating tumors and
associated problems of CSF circulation.
doi:10.1016/S1042-3680(03)00068-8
90
C. Teo, P. Nakaji / Neurosurg Clin N Am 15 (2004) 89103
We discuss the neuro-oncologic applications of

endoscopy under the following headings:
1.
2.
3.
4.
5.
Ventriculoscopy
Management of secondary hydrocephalus
Endoscopic tumor biopsy
Endoscopic intraventricular tumor resection
Endoscope-assisted microsurgery
Ventriculoscopy
The prognosis of some primary intracranial
tumors is dependent on the presence or absence of
ependymal spread of tumor. Patients with primitive neuroectodermal tumors, for example, fall
into the high-risk group rather than the low-risk
group if there is evidence of spinal or ventricular
ependymal tumor involvement. Although MRI is
reliable in the detection of ependymal tumor
spread in most cases, some patients may have
ependymal spread without radiologic evidence [7].
Ventriculoscopy can be more sensitive than MRI
with little added morbidity. Through a frontal or
parietal burr hole, one can access the lateral
ventricle, examine the surface, document any
ndings with color photography, and even biopsy
suspicious areas. This can be performed with a 10minute general anesthetic, which is substantially
less time than is needed for an MRI examination.
Furthermore, if present, denitive treatment of
CSF obstruction can be achieved by either third
ventriculostomy or tumor resection at the same
sitting.
Fig. 1 shows endoscopic pictures of the lateral
ventricle of a patient with a pineal primitive
neuroectodermal tumor who had no imaging
evidence of ependymal tumor involvement. The
patient had an endoscopic third ventriculostomy
(ETV) and, thereafter, chemotherapy. The response to chemotherapy was well documented by
ventriculoscopy, thereby allowing the patient to
be a candidate for high-dose chemotherapy and
autologous bone marrow transplantation.
Management of secondary hydrocephalus
The management of secondary hydrocephalus
is a controversial area of pediatric neurooncology. There are several dierent treatment
paradigms. A popular option would be to place

a temporary external CSF drain as a primary
procedure, remove the obstruction, and then, as
a secondary procedure, shunt those patients who
continue to require CSF drainage. Another
option is to insert a permanent shunt as
a primary procedure before tumor resection. A
third option would be to perform an ETV either
primarily or secondarily. The arguments in favor
of the early management of the hydrocephalus
are as follows:
1. There are some circumstances in which the
surgeon cannot denitively treat the primary
cause of obstruction, and CSF diversion is the
only option. These patients may present with
symptoms of raised intracranial pressure and
impending brain herniation.
2. The surgeon may need more time for preoperative assessment (eg, the patient with
a bleeding diathesis, the patient who needs
staging of his or her tumor).
3. The primary CSF diversionary procedure
may be denitive. An example of this is the
patient with a tectal glioma, in whom
symptoms are usually secondary to hydrocephalus and not the primary tumor.
4. On rare occasions, MRI reveals pathology
previously hidden by the hydrocephalus
(Fig. 2). The patient in Fig. 2 presented with
severe intracranial hypertension and was
thought to have hydrocephalus from aqueductal stenosis. After ETV, it became clear
that the obstruction was secondary to a pineal
region tumor.
5. Some surgeons believe that operative conditions for the denitive surgery are better
after prior drainage of CSF.
The arguments against initial treatment of the
hydrocephalus are as follows:
1. Many patients do not have a permanent
CSF ow problem; therefore, some shunts
are placed unnecessarily. Of course, this is
not a concern when temporary CSF diversion is undertaken using external ventricular
drainage.
2. If the obstruction is caused by a posterior
fossa tumor, there is potential for upward
c
Fig. 1. (A) The ependyma is frosted with plaque-like tumor that was not seen on MRI. (B) After conventional
chemotherapy, the clinical response is documented ventriculoscopically. (C) After high-dose chemotherapy and
autologous bone marrow transplantation, the ependymal surface is clear of visible tumor.
91
92
Fig. 2. (A) The patient is an 11-year-old boy with profound hydrocephalus of unclear etiology. (B) After an endoscopic
third ventriculostomy and resolution of the hydrocephalus, an occult pineal region tumor is now apparent.
herniation with decompression of the supratentorial CSF spaces.

3. There is a risk of infection with external
drainage.
4. There is a risk of spreading tumor cells to the
peritoneum with extracranial CSF diversion.
ETV is a reasonable option if the surgeon
decides to treat the secondary hydrocephalus
before denitive treatment of the primary tumor.
This operation can be done through a standard
frontal burr hole similar to that which would be
made for insertion of an external ventricular drain
(Fig. 3). Similarly, CSF samples can be taken at
the time of surgery, and one may also explore the
third ventricle, taking biopsies if necessary. There
is no risk of ongoing infection as can be seen with
external drainage; no risk of seeding as has been
documented with ventriculoperitoneal shunting;
and no risk of upward herniation, which can
potentially happen with any extracranial diversionary procedure. Finally, ETV may be the
denitive treatment if the obstruction is caused
by a tumor that does not require removal, such as
a tectal plate tumor.
Endoscopic septum pellucidotomy may also be

the denitive treatment for some types of hydrocephalus. Fig. 4 shows the MRI scan of an elderly
lady who presented with headaches and disorientation. She was found to have a cystic tumor of
the third ventricle causing obstruction of only
one of the lateral ventricles. The provisional diagnosis was a craniopharyngioma, and she underwent endoscopic cyst fenestration and drainage,
followed by septum pellucidotomy. The postoperative scans showed complete resolution of
the unilateral hydrocephalus and shrinkage of
the cystic component of the biopsy-proven
craniopharyngioma.
Technical notes
The technique of ETV is described in elsewhere in this issue. There are a few precautions
when performing ETV for hydrocephalus secondary to a tumor. First, the anatomy may be
altered. A tumor like a pontine glioma may
distort the oor of the third ventricle and displace
the basilar artery forward so that the safe zone to
penetrate the oor is a submillimetric area just
93
Fig 3. (A) The endoscopic third ventriculostomy is performed in the third ventricular oor just behind the mamillary
bodies. The close relation to the basilar apex is illustrated. (B) In chronic hydrocephalus, the third ventricular oor is
thin and visualization of the mamillary bodies is easy. (C) In acute hydrocephalus, it may not be possible to see the usual
landmarks. The oor is thicker, and a sharp technique may be necessary. In this instance, more experience is required to
identify the proper location for the ventriculostomy safely.
behind the dorsum sella and just in front of the

basilar bifurcation. Using a blunt technique to
create the stoma would be prudent in such a case.
Second, hydrocephalus resulting from tumor
obstruction may be relatively acute in onset.
When this is the case, the oor of the third
ventricle is often opaque and nonattenuated. This
makes penetration more dicult and invariably
requires a sharper technique without visualization
of the underlying neurovascular structures. This
blind fenestration through a thick oor can be
hazardous and should not be undertaken by the
novice endoscopist. Third, ETV combined with
biopsy of posterior third ventricular tumors
cannot be done through the standard ETV burr
hole. The ideal trajectory for the anterior third
ventricular oor is just anterior to the coronal
suture, whereas that for the posterior third
ventricle is just posterior to the hairline or 8 cm
behind the nasion and 3 cm from the midline.

The foramen of Monro is sometimes expanded
enough to allow access to both parts of the third
ventricle through a single burr hole placed
between the coronal suture and the hairline. This
may be assessed before surgery by close inspection of the MRI.
A comment should be made about the use of
exible berscopes. It is true that with a rigid
endoscope, it is dicult to perform an ETV and
biopsy of a posterior third ventricular tumor
through the same burr hole. Nonetheless, we
believe that a rigid endoscope is preferable to
using a exible scope insofar as it can be
impossible to be sure where the back of the
exible endoscope is. The risk of complications in
most hands is consequently higher with exible
endoscopes. It is our personal practice not to use
exible endoscopes in the head.
94
Fig. 4. Cystic craniopharyngioma in an elderly woman

who presented with high intracranial pressure. The cyst
was fenestrated endoscopically, and a septum pellucidotomy was performed. The patients symptoms resolved, and she did not require shunting.
Endoscopic tumor biopsy

There are no studies showing that endoscopic
tumor biopsy is any better than stereotactic needle
biopsy. Anecdotally, however, there seem to be
denite advantages to this technique. Endoscopic
biopsy should be considered when the tumor is
either within the ventricle or at least presenting to
the ventricular surface. The advantages are as
follows:
1. Direct visualization of the tumor allows more
accurate and safer sampling. A region for biopsy can be chosen under endoscopic vision,
and vessels can be avoided.
2. The specimen obtained is larger and not
subjected to as much mechanical artifact. It
does not need to be sucked up a needle or
manipulated.
3. Any resultant bleeding can be stopped by
either coagulation or packing under direct
visualization.
4. If the tumor is relatively avascular, it may be
removed totally by endoscopic techniques.
5. Other procedures can be performed at the
same operation (eg, ETV, septum pellucidotomy).
Examples of tumors that are typically
approachable endoscopically are colloid cysts,
subependymal giant cell astrocytomas, other
Fig. 5. Biopsy of a tumor in the posterior wall of the

third ventricle with grasping forceps. Bleeding after such
a biopsy can usually be managed with irrigation alone.
low- (including tectal gliomas) and high-grade

gliomas [Fig. 5], central neurocytomas, subependymomas, and choroid plexus tumors and cysts
[8]. Most of these tumors are relatively avascular,
and hemorrhage is rarely a problem. The burr
hole is made so that the scope enters the ventricle
as far from the tumor as possible and so that the
scope is directly viewing the tumor rather than
peering from around a corner. The distal approach allows the surgeon to orient himself by
identifying normal anatomic structures before
encountering the abnormal anatomy. Because
most of the distal part of the scope is within the
ventricle, it also allows the surgeon to move the
scope in multiple directions more freely without
damaging the surrounding normal brain. Starting
from farther away means that only relatively small
excursions are necessary within the normal brain
to visualize the entire target. Many surgeons are
hesitant to employ endoscopy for tumors because
of the fear of bleeding. Almost all bleeding can be
controlled with irrigation alone until the bleeding
stops. If bipolar and monopolar instruments
are available, they can be used, but they are
often ineective. Patience and copious irrigation
are the keys.
For tumors that are primarily intra-axial, there
are few advantages of endoscopic biopsy over
standard stereotactic techniques. Some extra-axial
tumors that require biopsy, such as those that
might be seen around the circle of Willis and
suprasellar region, are well suited to this technique, however. For these tumors, stereotactic
techniques are too dangerous and standard open
craniotomy is extremely invasive. Endoscopic

biopsy allows a minimally invasive approach with
excellent visualization and safety. The opening is
made through a supraorbital brow incision and
a small frontal craniotomy. The suprasellar region
is then approached using standard microsurgical
technique, and the scope is introduced when the
cisterns have been opened. Biopsies should be
taken with straight or angled cup forceps passed
along the scope and not down a working channel.
The instruments should always be passed ahead of
the endoscope so that their progress can be
observed. Blind passage of the instruments into
the endoscopes eld of view risks damaging
structures located behind the tip.
Endoscopic intraventricular tumor resection
Not all intraventricular tumors should be
approached endoscopically. The ideal tumor for
endoscopic consideration has the following characteristics:
1. Moderate to low vascularity
2. Soft consistency
3.
4.
5.
6.
95
Less than 2 cm in diameter [9]

Associated secondary hydrocephalus
Histologically low grade
Situated in the lateral ventricle
Clearly, from this list of desirable features,

almost all patients with colloid cysts are appropriate candidates for this technique. The results
with these tumors are often excellent, and in at
least one study, endoscopic resection was clearly
superior to microsurgery (Figs. 6 and 7) [10].
Other tumors that are well suited to total
endoscopic removal are subependymal giant cell
astrocytomas around the frontal horn of the
lateral ventricle, other low-grade gliomas that
are exophytic into the ventricles, central neurocytoma, small choroid plexus tumors, and the
purely intraventricular craniopharyngioma.
There are few articles in the neurosurgical
literature on the application of endoscopy for the
removal of intraventricular tumors. Most of the
experience with endoscopic tumor resection has
specically addressed the removal of colloid cysts
[1012]. Whichever technique is used, patient
Fig. 6. Pre-endoscopic (A) and postendoscopic (B) removal of a colloid cyst. Note resolution of the hydrocephalus and
preservation of the normal anatomy.
96
Fig. 7. (A) Initially narrow view of the foramen of Monro. The choroid plexus is to the left of the image, with the septal
vein located superiorly. (B) The choroid plexus is coagulated with monopolar cautery. The cyst is now clearly visible as
a gray structure through the foramen. (C) View of the foramen of Monro after removal of the colloid cyst. The fornix is
seen to the right of the image, the septal vein superiorly, the thalamostriate vein to the lower left, and the choroid plexus
to the upper left.
selection is controversial. For tumors in the

vicinity of the foramen of Monro, the presence
of hydrocephalus is an absolute indication for
some form of surgical intervention. Headache in
the absence of hydrocephalus can be a sign of
intermittent obstruction of the foramen of Monro
or some other part of the ventricular system, and
patients should be informed of the possibility of
insidious onset of hydrocephalus and sudden
death. Another indication for surgery would be
radiologic progression either in size or enhancing
characteristics. This is particularly true for giant
cell subependymal astrocytomas that are located
in the frontal horn of the lateral ventricle.
Whatever tumor is approached, some basic
principles apply. An approach trajectory that

avoids eloquent structures but allows a good view
of the tumor is essential. Most intraventricular
tumors are dealt with in the same fashion. Once
good visualization is achieved, the outside of the
tumor is coagulated with either monopolar
electrocautery or a laser. Copious irrigation is
used to clear blood and debris and to prevent too
much heat from building up inside the ventricle. If
there is a cyst, it is opened and drained. The
contents are removed via suction or piecemeal.
The remaining wall is coagulated and removed
piecemeal. Hemostasis is obtained with copious
irrigation. The scope is withdrawn while inspecting the tract for intraparenchymal bleeding. A
small wick of Gelfoam is placed in the cortical

opening to prevent CSF leakage from the ventricle.
Endoscopic colloid cyst removal
The colloid cyst is the prototypical intraventricular tumor; as such, we discuss its removal in
some detail. The colloid cyst represents the ideal
intraventricular tumor that can be completely
resected via the endoscopic approach. The goals
of surgery are to extirpate totally the cyst wall, to
avoid spillage of the cyst contents, to minimize
trauma to the adjacent structures (ie, fornices,
walls of the third ventricle), to spare venous
structures, and to minimize cortical damage along
the approach route. The rst two goals are more
dicult with the endoscope than with open
surgery, whereas the last three are actually easier
to achieve. The lower morbidity and shorter
operating time form the rationale for the endoscopic approach. In practice, complete cyst
drainage and removal can be better accomplished
endoscopically than microsurgically once the skill
is mastered. The choice of endoscopy as the
treatment modality should depend on (1) proper
patient selection, (2) adequacy of the equipment,
and (3) adequacy of the surgeons experience and
preparation.
Patient selection
Almost all patients with colloid cysts are good
candidates for the endoscopic approach. We
encourage endoscopic removal as the preferred
rst procedure, with the option of craniotomy if
endoscopy is unsuccessful. In fact, a wide range of
management strategies and therapeutic options
are available to treat patients with colloid cysts.
These include observation, shunting of the lateral
ventricles, stereotactic aspiration, open transcortical or transcallosal resection, and endoscopic
removal. For a patient who presents nonemergently, such as with an incidentally discovered
colloid cyst, it is important to discuss all options.
Nonetheless, given the risk of acute CSF obstruction, herniation, and death associated with these
masses, we recommend denitive treatment by the
least morbid route, which is via the endoscope.
The main technical factor that aects colloid
cyst removal is the density of the cyst contents.
This can best be estimated using noncontrast CT.
Hypodense or isodense contents are usually fairly
liquid and can be removed through the small
apertures of the endoscope without diculty [13].
A small suction catheter attached to a 10- or 20mL syringe may help in this regard. Hyperdense
97
cysts may have more tenacious contents, which

can prove more dicult to remove with the
endoscope. Cup forceps may need to be used to
remove the contents of these denser cysts piecemeal. On MRI, high signal on T1-weighted
imaging correlates with higher cholesterol content [14], thicker consistency, and more dicult
removal.
Technique
The patient is positioned straight supine with
the head up 45 . A single burr hole is placed on
the nondominant side 8 cm behind the nasion and
approximately 7 cm lateral to the midline. This is
farther lateral than previously published by the
senior author (C.T.), a modication that has
allowed better removal of the component of the
cyst in the roof of the third ventricle. A strip shave
is made for the incision, which is 4 cm long and
coronally oriented over the site of the burr hole.
Preparations should be made to preserve the
option of converting to an open craniotomy, with
a craniotomy tray and microscope available. If the
ventricles are small, frameless stereotactic guidance is used to guide the placement of the
endoscope sheath and trocar. Otherwise, freehand
cannulation of the ventricle with a ventricular
needle precedes placement of the sheath.
A 2- to 5-mm 30 rigid endoscope with at least
one working channel is placed down into the
ventricle. The boundaries of the foramen of
Monro are appreciated: the fornix above and
anteriorly, the choroid plexus and the septal
and thalamostriate veins exiting posteroinferiorly,
and the anterior nucleus of the thalamus located
below. The foramen of Monro is normally 0.3 to
0.8 mm in diameter. In the presence of a colloid
cyst, it is variably expanded. The cyst is attached
to the roof of the anterior third ventricle, and
its anterolateral surface usually bulges into or
through the visible part of the foramen. Every
eort is made to preserve all normal structures in
removing a colloid cyst. Generally, this can be
accomplished; however, three structures may be
sacriced, if necessary, to provide widened access
to the tumor, in descending order of preference:
choroid plexus, the thalamostriate vein, and the
ipsilateral fornix. The rst structure is sacriced
with impunity, but sacrice of the latter two
should only be considered if all other options have
been exhausted. If choroid plexus obstructs the
view of the cyst, it can be coagulated with
a monopolar probe and dissected free from the
foramen without diculty. No morbidity attends
98
this maneuver. The thalamostriate vein may be

large and present an obstacle. Every eort should
be made to preserve the vein. Past authors have
warned of potentially devastating consequences
from sacricing a thalamostriate vein, but this has
not been borne out in subsequent reports by other
authors or in our own experience. The primary
reason not to sacrice the vein is that it is far more
dicult to do this safely using the endoscope than
microsurgically. The bleeding from the vein can be
daunting. Fortunately, sacrice of the vein rarely
becomes necessary. Last, although it is not
preferable, a thinned-out ipsilateral fornix may
need to be sacriced to provide adequate access to
a large colloid cyst. Permanent short-term memory loss is sometimes a consequence of this
maneuver. Other structures that may be pitfalls
in this case are the thalamus, the internal capsule, and the head of the caudate, all of which
must not be harmed. The internal capsule runs
in the groove between these other two structures; damage to this structure during endoscopic procedures has resulted in contralateral
hemiparesis.
Once exposed, the surface of the cyst is gently
devascularized with electrocautery. Although a
laser may also be used for this role, the results are
not superior to monopolar electrocautery despite
the greater expense and complexity. The wall of
the cyst is then pierced with the monopolar probe,
and the contents are removed piecemeal or with
suction attached to a hand syringe once inside the
cyst. After it is emptied, the cyst wall is coagulated
and removed piecemeal. Care is taken not to
provide excessive traction on the roof of the third
ventricle, because severe bleeding from the internal cerebral veins can result. Other authors
have written that a small amount of the cyst wall
may have to be left on the third ventricle roof and
the columns of the fornix, but this has at times
been associated with recurrence. In the senior
authors personal series, complete removal was
possible in 24 of 26 cases; the remaining two
patients underwent transcortical removal of the
residual. From the anterolateral approach described here, the 30 endoscope aords a view of
the roof of the third ventricle that can help to
conrm complete removal of the cyst wall.
It has not been necessary to place drains at the
time of surgery or to maintain them after surgery.
Most authors with experience in this area have
used drains in their early cases and, over time,
moved away from using them. A drain may be
considered in a patient in whom the risk of sudden
decompensation because of postoperative hydrocephalus is believed to be high. Ultimately, it is up

to the surgeon to make an individual assessment
on a case-by-case basis. In the senior authors
personal series, no patient has required placement
of a permanent ventricular shunt.
Variations in technique
Endoscopic removal of colloid cysts using
a exible berscope has been described. Although
this approach is possible, it oers no advantage
over rigid endoscopy, has a substantially higher
learning curve, requires more coordination with
an assistant, and is associated with greater risk.
This approach is absolutely contraindicated for
any surgeon not already familiar with the
berscope.
The technique described previously is a oneportal technique. We have found this approach
adequate in most cases. From the anterolateral
approach, both fornices and the roof of the third
ventricle can be seen from one side. A two-portal
ipsilateral technique has been described by
Jimenez [15], but we have not found this to add
much to the procedure and believe that the addition of more portals goes against the goals of a
minimally invasive approach.
Complication avoidance
Copious irrigation is important throughout the
procedure to maintain good visualization and
dissipate heat generated by the electrocautery. We
use 20-mL syringes to provide pulsed handinjected irrigation. This allows good feedback as
to how much pressure is being generated as well as
rapid adjustment of irrigation depending on local
conditions. This method does necessitate frequent
changing of syringes throughout the case. It is
important that a good path for egress of CSF be
maintained so that uid does not build up under
pressure during the procedure. Herniation has
been described as a result of induced hydrocephalus in this setting. A tube attached to the outlet
port, which can be raised or lowered, permits the
pop-o pressure to be varied, allowing the
amount of ventricular dilatation to be controlled.
Lactated Ringers solution is used as the irrigation
solution. It is slightly hypo-osmolar (276 mOsm)
but is physiologically more like CSF than normal
saline. Some authors have described postoperative
confusion for 24 to 48 hours after procedures in
which normal saline (308 mOsm) is used as the
irrigant. There is no widely commercially available CSF substitute at present.
Bleeding from the colloid cyst or choroid

plexus should be addressed with monopolar
coagulation. All other bleeding should be addressed with irrigation. Almost all bleeding stops
with irrigation alone if the irrigation is continued
long enough; this may mean continuous irrigation
for as long as 20 minutes. The value of copious
irrigation cannot be overemphasized. With patience, almost any degree of bleeding can controlled in this manner. If this is not adequate, in
some cases, bleeding vessels may have to be
sacriced with the use of electrocautery. In cases
of generalized ooze, a balloon may be inated
gently in the foramen of Monro or the endoscope
itself may be used to provide focal compression.
Other techniques for controlling dicult bleeding
include irrigation with cool saline (which causes
vasoconstriction), draining the CSF to perform
dry eld coagulation, and raising the height of
the drain until the intracranial pressure exceeds
venous pressure. If nothing stops the bleeding, the
procedure may be need to be converted to an open
craniotomy. In our experience, this last step has
never actually been necessary.
Conclusions
The endoscopic removal of colloid cysts is one
of the most satisfying of neuroendoscopic procedures. Once some experience is gained, the
technique is faster and less morbid than the open
approach. Skills obtained working on these
tumors provide the foundation for the techniques
that allow the removal of other intraventricular
tumors.
The principles discussed here can be applied
to any mass that meets the indications for endoscopic removal. The following points reiterate
the techniques applicable to the removal of intraventricular tumors in general:
1. Have the correct instrumentation available.
Essential tools are a pair of grabbing forceps
and scissors, a good assistant or a scopeholding device so you can work with two
hands, a coagulation device (either monopolar or bipolar), a means of irrigating, and
straight and 30 angled scopes.
2. Make the ventricular entry as far away from
the tumor as possible. The further the tip of
the scope is from the burr hole, the smaller
is the angle required to displace the end of
the scope a given distance. This allows the
surgeon to angle the rigid scope in dierent
directions without causing as much brain
99
injury as would otherwise occur. It also gives

the surgeon a chance to view the tumor from
a distance, an important point when there is
excessive bleeding. A good amount of ventricle between the entry point and the tumor
means the scope can be pulled back from the
bleeding site without exiting the ventricle.
3. Maintain the ventricular volume to ensure
a workable operative eld. This is achieved
with continuous irrigation. Of course, it is
imperative to make sure there is an unimpeded egress of uid so as to prevent
iatrogenic intracranial hypertension. A channel on the endoscope leading to an open port,
serving as a pop-o valve, may be useful in
this regard.
4. Manage hemorrhage with the following
techniques: copious irrigation usually clears
the operative eld of blood and allow the
surgeon to continue without using the other
techniques. It is important to keep the
temperature of the uid close to body
temperature and the composition as close to
CSF as possible. Lactated Ringers solution is
preferred because it is the closest to isotonic
of the relatively available uids. Irrigation
settles almost all bleeding eventually, although it may take upward of 20 minutes of
continuous lavage in some cases. Using the
scope or an instrument to tamponade the
bleeder can be eective, especially when
combined with head-up elevation and irrigation. The next technique is to attempt direct
coagulation with either a bipolar or monopolar instrument. This technique can be quite
a challenge given the obscuration of the eld
with blood, the mobile vessels, the presence of
adjacent vital structures, and the diculty of
nding the exact bleeding point with a single
instrument. Finally, if all these techniques
fail, the ventricle can be drained of CSF. This
allows more eective bipolar and monopolar
coagulation without clouding of the operative
eld. The obvious downside is the collapse of
the ventricular walls, resulting in constriction
of the operative eld. This can be partially
oset by injecting air into the ventricle, which
keeps the walls apart if there is a good seal
around the working channels and around the
brain/scope interface.
5. Finally, it is imperative to keep in mind the
patients best interests. Large and vascular
intraventricular tumors are dicult to remove by pure endoscopic techniques. The
100
operation can become long and tedious with

more blood loss and risk to surrounding
eloquent brain tissue than a routine microsurgical approach. Do not hesitate to abandon the endoscope if you feel more condent
using a more familiar technique. Always
remember the dictum: minimally invasive
and maximally eective.
Endoscope-assisted microsurgery
Endoscope-assisted microsurgery is the area of
neuroendoscopy that has received the least
attention by neurosurgeons but whose potential
value to the neuro-oncologic surgeon is the
greatest. Microsurgery itself evolved to maximize
visualization and minimize retraction; endoscopy
allows the neurosurgeon to move another step
further toward achieving these goals. With endoscopy, previously inaccessible or poorly accessible
tumors located in the skull base, within narrow
cavities, deep to key vascular or neural structures,
or around corners in the intracranial space can be
clearly visualized; once visualized, they can be
resected. The introduction of the endoscope has
the potential to revolutionize the approach to
certain tumor types by allowing safe radical
removal. These techniques are particularly applicable to a wide range of troublesome tumors,
including but not limited to sellar tumors, including pituitary tumors, craniopharyngioma, and
Rathkes cleft cysts; clival chordomas; pineal
lesions, intraparenchymal tumors near the brain
stem or cranial base; acoustic neuromas; and
anteriorly or centrally located posterior fossa
tumors [16,17]. At the point at which dissection
must halt because of the failure of the microscope
to provide an adequate view, the endoscope
should be brought into play.
A summary of the advantages of the endoscope as an adjunct to microsurgery includes the
following:
1. Better denition of the normal and pathologic anatomy. The endoscope can be used
to clarify the anatomy before and during
tumor removal. Key neural or vascular structures can be identied and thereby spared.
This may be particularly important when
working around or within the brain stem,
between small perforating vessels, or between
the cranial nerves.
2. Identication of portions of the tumor
located behind or adherent to vital structures.
Some portions of tumor that are apparently

invasive into the brain have brain/tumor
interfaces that can be identied when visualized at more direct angles than is possible
with the operating microscope alone.
3. Minimization of retraction. The authors
seldom employ any xed retractors even in
the approach to extremely deep lesions. The
endoscope allows narrow corridors to be
used, reducing the need to displace sensitive
structures.
4. Assessing adequacy of tumor removal. At the
conclusion of the procedure, endoscopic inspection allows a more accurate estimation of
the completeness of resection or, if complete
resection cannot be achieved, documentation
of where and how much residual tumor
remains.
5. As a teaching tool. The endoscope oers
a superior and often novel view of the
anatomy, which can be benecial to residents
understanding of the surgical approach.
Furthermore, the operating surgeon and
the student share the same view, which is
not generally true even with an operating
microscope.
The benet contributed by adding endoscopy
to a traditional craniotomy cannot be overemphasized. Tumors frequently extend at acute
angles to the cranial base or to the cortical
surfaces along which the traditional surgical
approach is made (Fig. 8). Although these
avenues are inaccessible to the microscope, which
requires a direct line of sight, they are ideal for
endoscopy. The degree of retraction required can
frequently be lessened substantially. Traditionally,
all microsurgical approaches can be conceptualized as forming the shape of a cone, with the base
on the surface of the head and the working space
at the apex. With the use of angled endoscopes,
this limitation can be overcome. At the tip of the
cone of visualization and illumination available to
the microscope, 360 of additional view can be
obtained. When working around the brain stem
and cranial nerves, the corridor available to the
microscope is often narrow, because extensive
retraction is frequently not an option. The
endoscope allows the surgeon to obtain the
maximal access possible via the spaces naturally
present in the extra-axial compartment.
The endoscope has allowed a conceptual
change in the approach to low-grade or benign
tumors. A specic example of this point may be
101
Fig. 8. Preoperative (A) and postoperative (B) views of a tumor approached using a purely microsurgical approach.
Note the location of residual tumor. This residuum would still be accessible with the endoscope.
illustrated by the case of craniopharyngioma:

failure to achieve a complete resection ultimately
proves disastrous for the patient. In the past, some
surgeons have discouraged radical debulking of
the hypothalamic portion of these tumors because
of the unacceptable side eects attended by
damage to this vital structure. In fact, poor
visualization is what limits adequate and safe
removal of tumor. With the endoscope, the occult
parts of the tumor can be removed under vision,
making this part of the tumor as accessible as the
directly visualized portion (Fig. 9). Angled instruments allow access to the endoscopically visualized parts of the tumor. Because the prognosis for
a number of tumors is a function of the adequacy
of tumor removal, the use of the endoscope in this
fashion to obtain the best removal possible and to
document that removal is highly recommended.
Using this rationale, the authors encourage an
aggressive approach to the management of lowgrade tumors, such as pilocytic astrocytomas,
pleomorphic xanthoastrocytomas, and dysembryoplastic neuroepithelial tumors, as well as
so-called benign tumors, such as craniopharyngiomas and meningiomas.
Thus far, there are no studies documenting the
superiority of the endoscopic approach to tumor
removal. Some reports have begun to document
its use in transsphenoidal surgery; surgeons
familiar with the improved view oered in this
venue can easily imagine the same benets applied
within the rest of the cranial cavity.
The endoscope is a powerful ally, but as is the
case with all tools, the surgeon should become
fully familiar with it before using it with patients.
Practice is required to develop the visuomotor
skills necessary to guide the tip in and out of
narrow spaces safely, to watch the video image
while still respecting supercial structures along
the shaft, and to work with other instruments
while maintaining good visualization with the

endoscope. Familiarization with the endoscopic
perspective and a review of the pertinent microsurgical anatomy are essential before using the
endoscope on patients. Used properly, it is an
invaluable adjunct to traditional microsurgery.
Technical notes
1. Guide the endoscope in and out of the eld
under direct vision. The most dangerous
aspect of using the endoscope is the risk of
impacting structures while introducing the
endoscope. It is important to guide the
endoscope by viewing it along the length of
its barrel rather than by watching the image
on the screen. After placing the endoscope
into the working area, it is essential to
continue to mind the shaft: if the scope is
not xed, small barely noticed movements at
the tip can be the result of larger excursions at
the back of the scope, which potentially can
have disastrous consequences. An experienced assistant or a microscope with a headup endoscope display can help in this regard.
2. Use a xed endoscope holder so that once the
endoscope is in place, the surgeon can work
with both hands. This allows the surgeon to
use more complex instruments and also
prevents the endoscope from drifting against
vital structures located supercially along the
operative corridor.
Summary
Neuro-oncology, in all its aspects, provides an
ideal venue for the application of endoscopy.
The main obstacle to its use has been neurosurgeons lack of familiarity with the techniques
102
Fig. 9. (AC) These MRI scans show a midbrain tumor in a 5-year-old girl. The tumor was removed through a small
keyhole craniotomy and a transcallosal subchoroidal approach to the third ventricle. Once the microscope demonstrated
what appeared to be complete removal, the endoscope was passed into the tumor cavity and more tumor was removed.
(DF ) These MRI scans show complete tumor removal. The pathologic nding was a benign xanthoma. There has been
no recurrence after 4 years of follow-up.
and their advantages. As the neuro-oncologic

surgeon uses the endoscope more, endoscopy will
take its rightful place in the surgeons armamentarium. The advantages of improved visualization of intraventricular pathology, better
management of tumor-related hydrocephalus,
less morbid biopsies, and minimally invasive
removal of intraventricular tumors are invaluable
adjuncts to traditional tumor management.

Furthermore, endoscopy is the logical next step
for surpassing the limitations of traditional
microsurgery. Endoscopy is still in its infancy.
Rigorous application of the technology is increasingly allowing us to provide our patients the
most maximally eective and minimally invasive
surgery possible.
103
Fig. 9 (continued )
References
[1] Wallner KR, Gonzales M, Sheline GE. Treatment
of oligodendrogliomas with or without post-operative irradiation. J Neurosurg 1988;68:6848.
[2] Garcia DM, Fulling KH. Juvenile pilocytic astrocytomas of the cerebrum in adults. A distinctive
neoplasm with favorable prognosis. J Neurosurg
1985;63:3826.
[3] Good CD, Wade AM, Hayward RD, Phipps KP,
Michalski AJ, Harkness WF, et al. Surveillance
neuroimaging in childhood intracranial ependymoma: how eective, how often, and for how long?
J Neurosurg 2001;94(1):2732.
[4] Perneczky A, Fries G. Endoscope-assisted brain
surgery:. part Ievolution, basic concept, and
current technique. Neurosurgery 1998;42:21925.
[5] Teo C. Endoscopic-assisted tumor and neurovascular procedures. Clin Neurosurg 2000;46:51525.
[6] Perneczky A, Muller-Forell W, van Lindert E, Fries
G. Keyhole concept in neurosurgery. Stuttgart:
Thieme; 1999.
[7] Ebinger F, Bruehl K, Gutjahr P. Early diuse
leptomeningeal primitive neuroectodermal tumors
can escape detection by magnetic resonance imaging. Childs Nerv Syst 2000;16(7):398401.
[8] Oka K, Kin Y, Go Y, et al. Neuroendoscopic
approach to tectal tumors: a consecutive series.
J Neurosurg 1999;91(6):96470.
[9] Gaab MR, Schroeder HWS. Neuroendoscopic
approach to intraventricular lesions. J Neurosurg
1998;88:496505.
[10] Lewis AI, Crone KR, Taha J, et al. Surgical

resection of third ventricle colloid cyst. Preliminary
results comparing transcallosal microsurgery with
endoscopy. J Neurosurg 1994;81:1748.
[11] Abdou MS, Cohen AR. Endoscopic treatment of
colloid cysts of the third ventricle: technical note
and review of the literature. J Neurosurg 1998;89:
10628.
[12] Rodziewicz GS, Smith MV, Hodge CJ. Endoscopic
colloid cyst surgery. Neurosurgery 2000;46:655
62.
[13] El Khoury C, Brugieres P, Decq P, CossonStanescu R, Combes C, Ricol F, et al. Colloid
cysts of the third ventricle: are MR imaging
patterns predictive of diculty with percutaneous
treatment? AJNR Am J Neuroradiol 2000;21(3):
48992.
[14] Maeder PP, Holtas SL, Basibuyuk LN, Salford LG,
Tapper UA, Brun A. Colloid cysts of the third
ventricle: correlation of MR and CT ndings with
histology and chemical analysis. AJR Am J Roentgenol 1990;155(1):13541.
[15] Jimenez DF, editor. Intracranial endoscopic neurosurgery. Park Ridge, IL: American Association of
Neurological Surgeons; 1998. p. 132.
[16] Abdullah J, Caemart J. Endoscopic management of
craniopharyngioma: a review of 3 cases. Minim
Invasive Neurosurg 1995;38:7984.
[17] Cohen AR, Perneczky A, Rodziewciz GS, et al.
Endoscope-assisted craniotomy: approach to the
rostral brainstem. Neurosurgery 1995;36:112830.
Index
Note: Page numbers of article titles are in bold face type.
A
Aesculap neuroendoscopes, 2326
Age factors, in third ventriculostomy outcome,
7476
Aqueduct of Sylvius, occlusion of
CSF entrapment in, 8485
third ventriculostomy for, 4951, 7374
Cerebellum
anatomy of, 4243
developmental anatomy of, 35
vermis of, agenesis of, 43
Cerebral aqueduct, anatomy of, 42
Cerebral arteries, injury of, in third
ventriculostomy, 6970
Arachnoid cysts, 917

clinical presentation of, 10
complications of, 16
diagnosis of, 1011
frameless stereotactic systems for, 1314
intraventricular, 1011
laterally projecting, 13
locations of, 10
pathology and pathogenesis of, 910
shunting of, 15
suprasellar, clinical presentation of, 10
treatment of, 4, 1116
Cerebrospinal uid
hydraulics of, 4547
leakage of, in third ventriculostomy, 67
Arachnoid granulations, obstruction of, 48

Asystole, in third ventriculostomy, 69
Choroid plexus
anatomy of, 37, 5758
endoscopy of, 3
Auer Neuro-Endoscope, 2324
Choroidal cysts, of lateral ventricle, 3940
Choroidal ssure, developmental anatomy of, 34
Basilar artery, in third ventricle

anatomy of, 41
injury of, in third ventriculostomy, 63, 6970
Clarus neuroendoscopes, 26
Biopsy, of tumor, endoscopic, 4, 100101

Bleeding. See Hemorrhage.
Channel neuroendoscope, 26
Chavantes-Zamorano Neuro-Endoscope, 2324
Chiari II malformation, 4344
fourth ventricle malformation in, 4344
third ventricle malformation in, 42
third ventriculostomy contraindicated in, 4849
Chip cameras, for endoscopy, 21
Coagulation, endoscopic equipment for, 2728

Codman & Shurtle neuroendoscopes, 26
Colloid cysts, endoscopic removal of, 4, 103106
Bradycardia, in third ventriculostomy, 69
Communicating arteries, posterior, injury of,

in third ventriculostomy, 6970
Brain, tumors of. See Tumor(s), brain.
Computed tomography, in arachnoid cysts, 11
Bugbee Wire, for hemostasis, 27
Cook & Codman, cauterization equipment of,

2728
C
Cameras, for endoscopy, 21
Catheter, peel-away, for endoscopy, 22
Cauterization, equipment for, 2728
Copocephaly, 39
Corpus callosum
agenesis of, 39, 42
developmental anatomy of, 3435
doi:10.1016/S1042-3680(03)00100-1
106
Index / Neurosurg Clin N Am 15 (2004) 105108
scope holders, 29
syringomyeloscope, 27
video monitors, 2122
Craniopharyngioma, versus arachnoid cyst, 1112

CSF. See Cerebrospinal uid.
Cyst(s)
arachnoid. See Arachnoid cysts.
choroidal, of lateral ventricle, 3940
colloid, endoscopic removal of, 4, 103106
Dandy Walker, 43
ependymal, of lateral ventricle, 3940
solitary, drainage of, 8889
F
Fenestration
of arachnoid cyst, 1216
of multiloculated ventricles, 8990
Fiberoptic endoscopes, 1920, 2526
Fogarty balloon, for endoscopy, 28
Foramen of Monro, anatomy of, 3536
Dandy classication, of hydrocephalus, 45
Forceps, for endoscopy, 28
Dandy Walker cysts, 43
Fornix, anatomy of, 58
Decq Neuro-Endoscope, 2324
Fourth ventricle
anatomy of, 4243
CSF entrapment in, 8486, 89
malformations of, 4344
Diverticula, of lateral ventricle, 39
E
Electrocauterization, equipment for, 2728
Embryology, of ventricular system, 3336
G
Gaab Neuro-Endoscope, 2324
Endocrine disorders, in third ventriculostomy, 64
Endoscopy
equipment for, 1931
for arachnoid cysts, 917
for cancer, 95109
for hydrocephalus, with loculated ventricles,
8393
history of, 17
principles of, 1931
third ventriculostomy. See Third
ventriculostomy, endoscopic.
ventricular anatomy for, 3344
Headache
in arachnoid cyst, 10
in slit ventricle syndrome, shunt removal in,
third ventriculostomy for, 5354
Ependymal cysts, of lateral ventricle, 3940

Equipment, endoscopic, 1931
cameras, 21
beroptic, 1920, 2526
for arachnoid cyst endoscopy, 1215
for coagulation, 2728
for CSF entrapment, 8788
for shunt placement, 27
for third ventriculostomy, 6061, 71
history of, 1920
instruments, 2829
lasers, 29, 6263
light sources for, 2021
optic angles in, 20
peel-away catheter, 22
rigid, 19, 2226
Hematoma, subdural, in third ventriculostomy, 68

Hemorrhage
in arachnoid cyst endoscopy, 16
in third ventriculostomy, 63, 6970
intraventricular, hydrocephalus in, third
ventriculostomy for, 52
subarachnoid, hydrocephalus in, third
ventricular, CSF entrapment in, 8586
Hippocampus, developmental anatomy of, 33
Holoneural canal dilatation, in hydrocephalus, 85
Holoprosencephaly, 38
Hydranencephaly, 3839
Hydrocephalus
communicating, third ventriculostomy
contraindicated in, 4748
Dandy classication of, 45
double compartment, 85
etiology of, third ventriculostomy outcome
and, 73, 78
hydromyelic, 85
in
in
in
in
in
in
aqueduct of Sylvius occlusion, 4951

arachnoid cyst, 10
brain tumors, 5152
fourth ventricle malformations, 4344
infants, third ventriculostomy for, 5253
intracranial hemorrhage, third
in lateral ventricle malformations, 3840
in third ventricle malformations, 4142
isolated compartments in, 8393
multiloculated, 8990
postinammatory, 8586
shunt-related, 8385
solitary cysts in, 8889
treatment of, 8691
occlusive, in third ventriculostomy, 63
secondary, endoscopic treatment of, 9699
third ventriculostomy for. See Third
ventriculostomy.
Hypertension, intracranial, endoscopic treatment
of, 3
Hypothalamus
injury of, in third ventriculostomy, 64, 68
I
Infant(s), hydrocephalus in, third ventriculostomy
for, 5253
Infections, CSF entrapment in, 8586
107
M
Magnetic resonance imaging
in aqueduct of Sylvius occlusion, 50
in arachnoid cysts, 1011
in third ventriculostomy outcome assessment,
77, 79
Mamillary bodies, anatomy of, 40
Microsurgery, endoscopic-assisted, for tumors,
106107
MINOP Neuroendoscopy System, 2325
MurphyScope, 26
Myeloscopy, 4
N
Neuroendoscopy. See Endoscopy.
NeuroNavigational endoscopes, 26
NeuroPen endoscope, 26
Neuroview endoscope, 2526
O
Occipital pole, anatomy of, 38
P
Pneumocephalus, in third ventriculostomy, 67
Porencephaly, 39
Inammation, CSF entrapment in, 8586

Instruments, for endoscopy, 2829
Intracranial hypertension, endoscopic treatment

of, 3
Rigid endoscopes, 19, 2226
Intraventricular hemorrhage, hydrocephalus in,

Seizures, in arachnoid cyst, 10
Lasers, for endoscopy, 29, 6263
Light sources, for endoscopy, 2021
Shunt
CSF entrapment due to, 8385
endoscopic equipment for, 27
existing, third ventriculostomy outcome
with, 76
for arachnoid cyst, 15
removal of, third ventriculostomy for, 5354
Loculations
of fourth ventricle, 43
of lateral ventricle, 40
Slit ventricle syndrome

entrapped CSF in, 8385
shunt removal in, third ventriculostomy for,
5354
Longstanding overt ventriculomegaly of the adult,

5051
Subarachnoid hemorrhage, hydrocephalus in,

Lateral ventricle
anatomy of, 3638
Schizencephaly, 38
108
clinical features and, 7478

factors aecting, 7374
failure, 79
saline jet irrigation in, 6162
saline torch use in, 62
stereotactic, 63, 79
technique for, 6063, 6869, 9899
training for, 71
trajectories for, 5960
with narrow ventricle, 6869
stereotactic, versus endoscopic procedure, 79
Subdural hematoma, in third ventriculostomy, 68

Syringomyeloscope, 27
T
Temporal horn, anatomy of, 38
Thalamostriate vein, anatomy of, 37
Thalamus, anatomy of, 41
Third ventricle
endoscopy contraindicated in, 4849
small, endoscopy contraindicated in, 4849
Third ventriculostomy
endoscopic
anatomic considerations in, 5759, 7678
approaches to, 5960
balloon dilatation in, 61
complications of, 6364, 6772
asystole, 69
avoidance of, 7071
bradycardia, 69
cerebrospinal uid leak, 67
failure, 70
intraoperative management of, 71
periventricular injury, 6869
pneumocephalus, 67
subdural hematoma, 68
vascular, 6970
ventriculitis, 6768
dissection in, 61
equipment for, 60, 71
failure of, 70, 79
for arachnoid cyst, 12
for secondary hydrocephalus, 9699
goals of, 77
history of, 34, 57
patient selection for, 4555, 7071
anatomic considerations in, 4547
biophysical considerations in, 4547
contraindications, 4749
in aqueduct of Sylvius occlusion, 4951
in hemorrhage, 52
in infant hydrocephalus, 5253
in programmed shunt removal, 5354
in tumors, 5152
positioning for, 68, 71
postoperative management in, 71
results of, 7381
anatomic considerations in, 7678
assessment of, 77, 79
Trigone, anatomy of, 38

Tumor(s), brain, endoscopy in, 95109
history of, 4
in biopsy, 100101
in intraventricular tumor resection, 100106
in microsurgery, 106107
in secondary hydrocephalus, 5152, 9699
in ventriculoscopy, 96
U
Ultrasonography, in third ventriculostomy, 63
V
Ventricle(s). See also specic ventricles.
anatomy of, 3344
cerebral aqueduct, 42
developmental, 3336
fourth ventricle, 4244
lateral ventricle, 3640
third ventricle, 4041, 59
loculated
multiloculated, 8990
postinammatory, 8586
shunt-related, 8385
solitary cysts in, 8889
treatment of, 8691
fourth, 43
lateral, 40
Ventriculitis, in third ventriculostomy, 6768
Ventriculomegaly, longstanding, of the adult,
5051
Ventriculoscopy, for tumor evaluation, 4, 96
Ventriculostomy, third. See Third
ventriculostomy.
Video monitors, for endoscopy, 2122

04.01 - Clinical Neuroendos PDF

Загружено:

Сведения о документе

Исходное описание:

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

04.01 - Clinical Neuroendos PDF

Загружено:

Авторское право:

Доступные форматы

CLINICAL NEUROENDOSCOPY

The Endoscopic Management of Arachnoidal Cysts

Basic Principles and Equipment in Neuroendoscopy

The Anatomy of the Ventricular System

NUMBER 1 JANUARY 2004

Selecting Patients for Endoscopic Third Ventriculostomy

Techniques of Endoscopic Third Ventriculostomy

Complications of Third Ventriculostomy

Results of Endoscopic Third Ventriculostomy

Endoscopic third ventriculostomy is emerging as the treatment of choice for aqueductal

Loculated Ventricles and Isolated Compartments in Hydrocephalus: Their

Trapped cerebrospinal fluid spaces can, on occasion, complicate the management of

Neuro-oncologic Applications of Endoscopy

THE CLINICS ARE NOW AVAILABLE ONLINE!

Neurosurg Clin N Am 15 (2004) ix

Neuroendoscopy has come of age over the last

look toward the future there is also an article on

Neurosurg Clin N Am 15 (2004) 17

The insertion of tubes into the body for either

E-mail address: Rabbott@bethisraelny.org

nitrate [4]. In 1870, Kussmaul demonstrated using

R. Abbott / Neurosurg Clin N Am 15 (2004) 17

a telescope containing a series of lens housed in an

type of bundle could be used to transmit light

R. Abbott / Neurosurg Clin N Am 15 (2004) 17

As the use of endoscopes for diagnostic

after, Putnam [19] reported on the eectiveness of

R. Abbott / Neurosurg Clin N Am 15 (2004) 17

a myeloscope, to visualize dorsal nerve roots of the

[60,61]. This has allowed for a much narrower

R. Abbott / Neurosurg Clin N Am 15 (2004) 17

technical standards in neurosurgery. New York:

tive hydrocephalus in the pediatric population:

R. Abbott / Neurosurg Clin N Am 15 (2004) 17

[43] Auer LM. Ultrasound stereotaxic endoscopy in

R. Abbott / Neurosurg Clin N Am 15 (2004) 17

[86] Wahlig JB Jr, Welch WC, Weigel TL, Luketich JD.

Neurosurg Clin N Am 15 (2004) 917

The endoscopic management of arachnoidal cysts

Arachnoidal cysts are a not uncommon lesion

* Department of Neurosurgery, INN, Beth Israel

R. Abbott / Neurosurg Clin N Am 15 (2004) 917

Arachnoidal cysts typically arise within normal

Other symptoms at presentation point to the

R. Abbott / Neurosurg Clin N Am 15 (2004) 917

Mouse sign; and do not contain calcication

R. Abbott / Neurosurg Clin N Am 15 (2004) 917

R. Abbott / Neurosurg Clin N Am 15 (2004) 917

With regard to the use of a second instrument

scope was pointing posterior with the tip in the

R. Abbott / Neurosurg Clin N Am 15 (2004) 917

on MRI or CT. Before performing the surgery, such

establish registration. Also, more than once, I have

R. Abbott / Neurosurg Clin N Am 15 (2004) 917

as the target; this is quite helpful, because the goal

Fig. 3. Second recurrence of quadrigeminal cyst after

scope is removed, and a proximal catheter of

R. Abbott / Neurosurg Clin N Am 15 (2004) 917

a parasagittal cyst, and the other 5 had middle

R. Abbott / Neurosurg Clin N Am 15 (2004) 917

experience with 77 cases treated surgically. Acta

Neurosurg Clin N Am 15 (2004) 1931

Rigid instruments and electrodes

Aesculap produces another endoscope originally designed by Perneczky. This instrument

Endoscope-assisted shunt placement when

The beam is collimated, because the waves