Introduction
Thoracic disk herniation (TDH) is a rare condition, representing < 1% of all disk
herniations in most series.[1 ]
[2 ] The unusual symptoms of this condition may challenge early diagnosis. Routine use
of magnetic resonance imaging (MRI) has increased the diagnostic rate of TDH.[3 ]
[4 ] Anterior approaches to the thoracic spine have been established as the standard
procedure for appropriate treatment of TDH because they permit direct visualization
of the ventral spinal cord and require less manipulation of the neural tissues in
the spinal canal.[5 ]
[6 ]
[7 ] Thoracoscopic surgery of the spine has been reported in studies performed since
the early 1990s.[8 ]
[9 ]
The development of surgical approaches to the spine has expanded the options for managing
complex spinal pathologies in a minimally invasive manner. In the last decade, the
extensive use of the microscope, endoscope, and two-dimensional (2D) or three-dimensional
(3D) navigation systems has assisted surgeons in performing complex surgical procedures
in a safer and less aggressive manner.[6 ]
[9 ]
[10 ]
[11 ]
[12 ]
[13 ]
The trend toward the use of minimally invasive procedures with endoscopic visualization
of the thoracic cavity in thoracic spine surgery has evolved.[11 ]
[14 ]
[15 ] Both experienced and less experienced endoscopic spinal surgeons have often found
it very difficult to develop a new set of visuomotor skills unique to endoscopic procedures
and to understand 3D anatomy while performing a 2D imaging procedure.[11 ] We believe that introduction of an image guidance system would make these procedures
safer and more precise. We report the results of 10 patients who underwent diskectomy
for TDH using video-assisted thoracoscopic surgery (VATS) assisted by an O-arm-based
navigation system and describe the surgical technique.
Material and Methods
This study describes the surgical technique of VATS assisted by O-arm navigation for
the treatment of TDH and reports the clinical results of 10 patients who underwent
surgery at our institution from September 2008 to April 2009 using this new technique.
The inclusion criterion was the presence of soft TDH manifesting with myelopathy or
radiculopathy despite conservative treatment. Patients with hard disk herniation,
ossified posterior longitudinal ligaments, and/or concomitant moderate to severe bony
spurs were excluded.
Case 1
A 45-year-old male patient presented with gait disturbance and a dull, aching radiating
pain through the right and left buttocks and lower extremities; he had been experiencing
these symptoms for the past 3 years. He also complained of impotence and numbness
of the left flank and lateral calf. Physical examination revealed abnormally brisk
deep tendon reflexes. The Babinski reflex and the ankle clonus signs were positive
on both sides. MRI revealed cord compression caused by disk protrusion at the T6–T7
level ([Fig. 1 ]). The computed tomography (CT) scan showed soft disk herniation. TDH with myelopathy
was diagnosed, and a thoracic diskectomy using VATS assisted by O-arm-based navigation
was performed. Three portals were created ([Fig. 2 ]), and a partial diskectomy and decompression of the spinal canal was performed.
The annulus was gently removed, thus exposing the posterior longitudinal ligament.
The disk fragment was found in the epidural space, and the margin of the bony decompression
was confirmed by the O-arm ([Fig. 3 ]). After the operation, the patient experienced transient difficulty in voiding.
He was discharged on day 17 after resolution of his voiding difficulty. Postoperative
radiologic scans showed good decompression and no evidence of instability ([Fig. 1 ]). At the most recent follow-up performed 12 months after the surgery, he could walk
without any loss of balance. The impotence and radiating pain and numbness along both
legs had greatly improved, although mild radiating pain persisted in the right buttock.
Fig. 1 (A) Preoperative sagittal and axial magnetic resonance imaging (MRI) showing disk
protrusion causing compression at T6–T7. (B) Postoperative sagittal and axial MRI
show a well-decompressed thoracic disk with a restoration of the thecal sac shape.
Fig. 2 (A) An intraoperative picture of the three portals created for the video-assisted
thoracoscopic surgery showing the left side up position, with the cranial side on
the right side of the image. The large portal is used for suction; the other two are
working portals. (B) Postoperative wounds were 2 cm long.
Fig. 3 The intraoperative O-arm (A) axial, (B) coronal, and (C) sagittal images enable the
surgeon to determine the depth and the location of the drilling and show a decompressed
bony margin.
Case 2
A 63-year-old woman presented with gait disturbance, increased deep tendon reflexes,
back pain, bilateral buttock pain, and radiating lower extremity pain. Radiologic
investigations revealed a herniated disk at the T9–T10 level, with cord signal changes
([Fig. 4 ]), and degenerative spondylolisthesis with stenosis at levels L3–L5 and S1. We planned
a two-stage operation and performed a thoracic diskectomy first. Her postoperative
MRI showed good decompression of the T9–T10 disk along with bony removal around the
disk ([Fig. 4 ]). At the 6-month follow-up, the radiating buttock pain showed improvement, and she
was able to walk with better balance. However, the radiating pain in her lower extremity
was not resolved. A subsequent lumbar lesion surgery was planned.
Fig. 4 (A) Preoperative sagittal and axial magnetic resonance (MR) images show disk protrusion
at T9–T10 with cord signal changes. (B) Postoperative sagittal and axial MR images
show a well-decompressed thoracic disk.
Eight other patients with similar symptoms underwent the same operation. [Table 1 ] presents additional patient information.
Table 1
Demographics of video-assisted thoracoscopic surgical patients using intraoperative
O-arm-based navigation[a ]
Case no.
Sex
Age
Thoracic disk level
Intraoperative bleeding, mL
Total operation time, min
Time spent in navigation and O-arm, min
Follow up duration, mo
Hospital stay(day)
1
M
45
T6–T7
250
340
30
12
16
2
F
63
T9–T10
590
405
35
6
8
3
M
56
T10–T11
420
330
20
5
5
4
M
33
T12–L1
250
330
30
17
4
5
F
45
T3–T4
840
200
40
19
5
6
M
54
T6–T7
310
360
20
12
13
7
M
56
T12–L1
420
320
30
6
7
8
M
77
T11–T12
450
330
30
13
6
9
M
46
T9–T10
360
220
35
8
5
10
F
44
T12–L1
220
310
25
11
7
a All diagnoses were thoracic disk with myelopathy.
Surgical Technique
Under general double-lumen endotracheal anesthesia, the patient is placed in a left
or right side up lateral decubitus position with nerve intramuscular monitoring. The
first step in surgical navigation with the O-arm (Medtronic Sofamor Danek, Memphis,
TN) involves a registration scan of the thoracic spine. The O-arm is an intraoperative
cone beam CT scan that provides 3D visualization. The data are reformatted by the
workstation into coronal, axial, and sagittal 3D images of the spinal anatomy and
transferred via a network connection to StealthStation navigation on a Treon system
(Medtronic Sofamor Danek, Memphis, TN). The dynamic reference frame (DRF) is usually
attached to the patient's iliac bone. However, in cases with upper thoracic level
involvement, the DRF is attached to the more proximal operating table to reduce mismatching
error. The coordinates of three to six intraoperatively identifiable anatomical landmarks
of the level to be decompressed are obtained for use in the matching procedure during
the surgery ([Fig. 5 ]). The infrared camera is positioned facing the receiver, close to the operative
field. This system is operated by a technician under the surgeon's guidance. The image
guidance probe (IGP) is inserted through the portals in the chest wall to guide the
surgical procedure to obtain a multiple planar or a 3D view.
Fig. 5 (A) The operation room setting showing the navigation monitor, the dynamic reference
frame, and the O-arm. (B) The dynamic reference frame on the table.
An accuracy check, a crucial step in this procedure, is then performed to verify the
quality of the matching. The surgeon has to determine whether the matching accuracy
is acceptable for safe navigation by comparing the position of the instrument in the
operative field with the displayed position of the instrument in the CT image on the
monitor. If the accuracy is insufficient, the matching procedure is repeated.
After the localization of the disk level to be operated on using the IGP and the navigation
system, three or four endoscopic portals are inserted for the VATS on the lateral
chest. The first port is usually placed perpendicular to the main lesion. If the procedure
involves the use of an end-viewing 30- or 60-degree thoracoscope, the first port can
be positioned above or below the perpendicular level. The ideal position of the port
insertion is determined depending on the location of the target disk level identified
on intraoperative O-arm imaging and navigation in the StealthStation. After careful
insertion of the first port to avoid violating pulmonary tissues, the second and third
ports are inserted ventrally, making a triangular configuration on the ventral side.
This layout allows the optic to be inserted through any port to obtain the optimum
view at any time. Care is taken throughout the procedure to protect the collapsed
lung by using an endoscopic fan retractor.
The first step in the thoracoscopic procedure is pleural exploration involving the
identification of the major organs around the main lesion (e.g., the heart, great
vessels, sympathetic chain, etc.), followed by opening of the parietal pleura at the
appropriate spinal level using navigation. The second step involves the resection
of 1.0 cm of the articulated rib. A partial diskectomy and partial vertebral body
resection are performed using a high-speed drill in both cranial and caudal directions,
until the normal dura margin is exposed. A Kerrison punch and pituitary forceps are
then used to gently remove the ruptured disk fragments by cautious probing and removing
the posterior longitudinal ligament, to achieve full decompression of the cord. Total
decompression is verified when the pulsation of the spinal cord is visualized. At
this point, if some bleeding is encountered, it can be managed by using Avitene (MedChem
Products, Woburn, Massachusetts, United States), Gelfoam (Pfizer, Inc., New York,
United States), and bone wax, but the use of bipolar coagulation is highly discouraged.
The use of bipolar cauterization in this step may cause thermal injury to the spinal
cord.
The O-arm enables precise determination of the margin of bone work and thus avoid
injury to the spinal cord. Because the motion and view in thoracoscopy is limited
compared with those in an open thoracotomy, the exact position of the drill is difficult
to determine. The O-arm enables the surgeons to precisely track surgical instruments
in relation to their anatomy (i.e., accurately detect the depth and direction of the
drill trajectory when approaching the target). After satisfactory decompression is
achieved and the pulsation of the spinal cord is visualized, another scan of the thoracic
spine using the O-arm is performed to confirm the margin of the bone work and the
efficacy of the decompression, giving the surgeon intraoperative feedback. If the
margin of the decompression is found to be insufficient, the drilling and bone work
can be extended immediately. [Table 2 ] shows the workflow and the time needed for each step.
Table 2
O-arm workflow
Procedures
Time, min
Draping
3
Moving O-arm into surgical field/ensure camera visibility
6
Scanning
1
Parking O-arm
3
Uncovering sterile drape
1
Ready to go
14
Finally, with endoscopic guidance, a chest tube is placed through the posterior port,
at which time the anesthesiologist reinflates the lung.
Results
Our patients included seven men and three women, with a mean age of 53.5 years (range:
35–77 years). [Table 1 ] shows the distribution of the affected levels in the patients. Gait disturbance
and back pain were the most common symptoms. Upper motor neuron abnormalities were
present in all patients. The average duration of the symptoms was 2.8 years (range:
1–5 years).
The average operation time was 326.9 minutes (range: 200–405 minutes), and the average
blood loss during the surgical procedure was 441 mL (bleeding range: 250–840 mL).
The average additional time required for the image guidance surgery using O-arm-based
navigation was ∼ 29.4 minutes (range: 20–40 minutes).
There were no complications during the surgical procedure or the immediate postoperative
period. However, we experienced moderate intraoperative bleeding due to segmental
vessel injury in one case (case 5 in [Table 1 ]). Successful ligation was performed without any postoperative neurologic sequelae.
After the surgery, all patients' symptoms as well as walking improved. The mean follow-up
period was 11.3 months (range: 5–19 months).
Discussion
With the advances in biotechnology and the development of specialized instruments
such as improved microendoscopes, digital video equipment, and percutaneous systems,
minimally invasive spinal surgery (MISS) has rapidly evolved and gained popularity
in both spinal fusion and nonfusion surgery.
Recently, another new technique using computer-assisted surgery (CAS) for acquiring
access in surgical decompression in spinal disease including cervical and thoracic
disorders was shown to be useful[12 ]
[13 ]
[16 ] because it enables real-time feedback to the surgeon about the anatomical location
of the thoracic spine. This type of surgery can help the surgeon map anatomical distances
and safe trajectories, thereby decreasing the risk of neural damage. Therefore CAS,
in addition to being used in conventional open spinal surgeries, is also highly successful
in other types of spinal surgery.
Therefore, we attempted to combine MISS with CAS to treat cervical disk herniation[13 ] and TDH, and termed this procedure computer-assisted minimally invasive spinal surgery (CaMISS).
This introduction of CaMISS has not been well evaluated, and the surgical field in
CaMISS is very narrow and limited. However, with the help of intraoperative O-arm
and navigation in the StealthStation (Medtronic Sofamor Danek, Memphis, TN), the surgeon
can determine the exact boundary of surgical decompression in a narrow operative field.[17 ]
Accurate intraoperative visualization of spinal anatomy is a crucial element in enabling
thoracoscopic spine surgery. Precision is critical because of the unfamiliar thoracic
anatomy such as the costochondral junction complex, narrow spinal canal, and proximity
of bone margin to spinal cord. The success of the thoracoscopic approach depends on
high-quality images so the surgeon can work within this complex environment. However,
accurate knowledge of spinal anatomy totally depends on the surgeon's experience and
the ability to visualize anatomical structures three-dimensionally.
Navigation in the StealthStation combines specialized surgical hardware with computer-assisted
software that allows tracking of the location of an instrument in the surgical field,
and a continuous update of this location within 3D planes (coronal, axial, and sagittal)
is acquired through intraoperative O-arm scanning. This virtual navigation allows
the surgeon to navigate using multiple CT views simultaneously and avoid unwanted
radiation exposure unlike conventional fluoroscopic navigation. Furthermore, with
the aid of intraoperative CT, the surgeon can evaluate the spinal anatomy, correct
the surgical path, and assess for correct instrumentation placed intraoperatively.
Moreover, if an unfortunate change of the patient's position occurs, navigation re-registration
is more feasible and fast with the O-arm.
The accuracy of the image-guidance system may be affected by the technical accuracy
of the system, the registration process conducted by the surgeon, the voxel size of
the scan data, distortion of the image data, and intraoperative events that may result
in errors. The O-arm allows both initial scanning and intraoperative registration
of images. A major advantage of this system is that the surgeon can scan as many times
as necessary to achieve adequate orientation during decompressive surgery or implantation.[10 ]
[11 ]
[12 ]
[13 ]
[16 ]
Several different approaches have been reported in the literature to treat TDH that
can be classified into three categories: posterior, posterolateral, and anterior approaches.
The first two approaches allow indirect decompression of the spinal cord and therefore
carry the inadvertent cord injury risks. However, the anterior approach allows better
visualization and direct decompression of the spinal cord, with the possible risks
of pulmonary compromise. The choice of specific approach can be influenced by various
factors including the disk consistency, the location in the canal, clinical presentation,
the level of the pathologic lesion, and the surgeon's familiarity with the selected
technique. There is no gold standard, and each approach has its own unique advantages
and disadvantages. VATS is one of the important minimally invasive approaches to access
anterior thoracic pathologies with several advantages including reduced postoperative
pain and complication rates and faster recovery times. It is not yet familiar to most
spine surgeons owing to its technically demanding performance, steep learning curve,
and the need for special endoscopic instruments that are relatively expensive. Therefore,
the posterior approach is still the more universally used procedure to treat TDH.
Many novel techniques have recently been introduced to overcome known disadvantages
of the anterior approach including the minimally invasive transthoracic approach.[18 ]
[19 ] With the help of a specially designed minimal invasive retractor system and combining
it with thoracoscopy techniques, one could achieve safe access and excellent surgical
views even without any rib resection. However, unfamiliar anatomy and unique characteristics
of the thoracic spine with even smaller incisions may confound novice spine surgeons.
Furthermore, the combination with thoracoscopy can aggravate the difficulties. The
average spine surgeon requires additional training to acquire the skills to perform
the procedure effectively.
A thoracoscopic procedure largely simulates an open thoracotomy procedure in decompressing
the spinal canal. Since the early reports of thoracoscopic disk surgery, the indications
of this technology have extended to thoracic vertebral reconstruction, as well as
deformity correction. Using thoracoscopic techniques, anterior column reconstruction
can be performed to provide effective loadbearing. In addition, thoracoscopic surgery
provides small thin incisions without the need for rib resection or rib retractors,
unlike other anterior approaches such as costotransversectomy or the transpedicular
approach. Other advantages of this procedure include treatment of multisegmental abnormality
through the same portals without the need for additional rib resection, excellent
intraoperative visualization of the abnormality, direct anatomical access to the anterior
structures of the thoracic spine,[9 ] avoidance of diaphragm division, and significantly reduced chest wall injury. The
VATS procedure is known to provide significant pain reduction, better cosmesis, lower
perioperative morbidity, and an earlier return to normal activity.[2 ]
[6 ]
[8 ]
The disadvantages of the thoracoscopic spinal procedure are related to the dependence
of the bidimensional endoscopic visualization on a video monitor; the working portals
can significantly restrict the surgeon's movements in relation to the instruments,
and the larger instruments require more dexterity.[8 ]
[9 ] However, these disadvantages of thoracoscopic surgery can be decreased with the
help of CaMISS based on O-arm navigation.
A lack of adequate training and a steep learning curve are the most common reasons
for the limited use of thoracoscopic procedures. The absence of a 3D image of the
surgical field and the lack of convenient guidance methods during endoscopic surgery
can increase operating time; this does not encourage less experienced surgeons to
use this technique.[10 ]
We believe the use of 3D image real-time guidance during VATS in thoracic spine would
make the surgery safer and optimize spinal cord decompression. The biggest advantage
of using surgical navigation is easier orientation in a limited operation field. Navigation
enables a surgeon to find the target vertebral level, pedicle margin, and thecal sac
more easily in a lateral decubitus position of the patient. Because both the motion
and view of the thoracoscope is limited compared with those in an open thoracotomy,
the exact margin removed by the drill is difficult to determine. If a protruded disk
is on the midline, the diskectomy margin should go over the midline, and the opposite
normal dural margin should be identified; this procedure is not easily performed through
a thoracoscope. The O-arm allows the surgeons to check the depth of the working space
and the direction of the trajectory in relation to the target. This provides important
intraoperative feedback. Moreover, ideal portal placement is crucial in thoracoscopic
surgeries because the surgeon's working view is greatly limited, and with poor portal
placement, operating through the smaller portals with long surgical instruments can
be difficult. We used IGP and the navigation system to decide the ideal portal position
before the first port insertion depending on target disk location to overcome such
difficulties. With the help of intraoperative O-arm imaging and navigation in the
StealthStation, we could avoid significant difficulty due to poor portal placement
in all cases.
O-arm assisted navigation system has several great advantages over conventional spinal
navigations. The image acquiring process to navigation set-up time is much shorter
in the O-arm-based system, and the patient does not have to change position or location,
which greatly decreases accuracy because the spinal column is mobile. Furthermore,
O-arm-assisted navigation systems do not need point-to point surface matching, like
conventional spinal navigation, a factor that increases their use for further surgery
without landmarks.
As mentioned earlier, the VATS procedure requires an extremely steep learning curve
for most spine surgeons to perform proper thoracoscopic techniques such as establishing
proper orientation under the unfamiliar angled endoscopic views and operating through
the smaller portals with long surgical instruments far from the target pathology.
Furthermore, as with any other surgical techniques, there is a separate learning curve
for using the O-arm-assisted navigation system. Surgeons have to practice the use
of the planning modules to interpret CT image projections to identify anatomical landmarks
and match them with the operation fields. Moreover, to identify the exact location
of pathology, they have to learn to judge the intraoperative quality of the navigation
accuracy. We had a little experience with conventional VATS or O-arm-assisted navigation
technique previously, which increased the time required for disk removal as compared
with others. We experienced a decrease in operative time from a mean of 360 minutes
in the first three cases to a mean of 280 minutes in the last three cases. We hope
further experience with thoracoscopic techniques will permit us to shorten the operative
time.
The most common obstacles that surgeons encounter during VATS is intraoperative bleeding
due to segmental or intercostal vessel damage. We experienced a case of segmental
vessel injury during which intraoperative blood loss exceeded 800 mL. Ligation of
the segmental vessels is encouraged, although time consuming, to eliminate possibilities
of catastrophic intraoperative or postoperative bleeding. To avoid such drawbacks
during operation, it is important to identify and preserve the segmental vessels at
the midportion of the vertebral body.
Although the surgical time may be longer than most conventional procedures and cost
is a limiting factor in the wide use of this technique, the advantages of ensuring
safety and determining the right trajectory, accompanied with appropriate bone decompression,
compensate for these shortcomings. The surgeon is allowed real-time navigation of
the unfamiliar thoracic anatomy, with the advantage of mapping anatomical distances
and trajectories without fear of neurologic injury. O-arm-based navigation increases
the accuracy of conventional VATS in a minimally invasive manner, leading to successful
surgical outcomes; a longer-term follow-up is warranted.
