Introduction
Neuronavigation is a technology recently added to neurosurgical procedures for planning
craniotomies and real-time pinpointing of intracranial lesions and their relationship
with anatomical structures. The available systems use magnetic resonance imaging (MRI)
or computed tomography (CT)-scan images taken before the procedure.[1] For subcortical lesions, such references may become inaccurate because of brain
displacement caused by gravity and cerebrospinal fluid (CSF) drainage. This phenomenon,
brain shift, may cause the surgeon to err, which is one of the limitations of this
method. The present study aimed at describing a simple method to minimize the effects
of brain shift in approaches to subcortical lesions, providing the surgeon with greater
safety and confidence in the use of neuronavigation.
Method
We used the technique hereby described to operate five patients with subcortical neoplasms.
We performed the craniotomy according to planning, aided by neuronavigation. We used
two neuronavigator models (Medtronic, Minneapolis, Minnesota, USA; BrainLab, Munich,
Germany
), in which the record is obtained by correlating landmarks on the patient's skin
and skull, and we used a handpiece tip to obtain the point-to-point correspondence.
Neuronavigation was initially used to plan the extension of the craniotomy, keeping
the subcortical lesion as the central target. After removing the bone flap, we anchored
the dura to the bone, using the conventional technique. We then used neuronavigation
to plan the dura mater opening around the projection of the lesion. After exposing
the cortical surface and before opening the arachnoid over the area chosen for corticectomy,
we anchored the arachnoid to the dura, close to the edges of the exposed area, using
2 or 3 6–0 mononylon sutures under microscopy ([Fig. 1]). Next, we used neuronavigation on the surface to identify the ideal point for corticectomy
and the trajectory to address the subcortical lesion. Then, we opened the arachnoid
at the point chosen for corticectomy. We checked the trajectory of the approach in
real time using neuronavigation, until we reached the subcortical lesion. Throughout
the five cases in which we used this approach, the cortical surface remained close
to the dura mater, anchored by the arachnoid sutures. All the lesions were located
bearing a good correlation between the handpiece tip inserted in the desired brain
area and what the navigation system displayed.
Fig. 1 Frontoparietal craniotomy. The suction tip points to the arachnoid anchorage suture
to the dura-mater, close to the craniotomy edge.
Discussion
Neuronavigation is a technology already incorporated in most neurosurgery services.
It assists the surgeon in planning craniotomies and in the precise location of deep
lesions and their relationship with anatomical landmarks. It is also useful in providing
maximum safety in lesion resections.
Neuronavigation uses the same principles as stereotaxis, based on the Cartesian coordinate
system. Any point in the brain can be reached by measuring the lesion to be addressed
in the horizontal, frontal and sagittal planes. The neuronavigation system provides
precise surgical guidance, paralleling the measurements obtained with MRI or CT scan
data from the patient. This reconstruction is displayed on a computer workstation
console, placed next to the surgical field. Thus, the resonance images displayed on
the screen become maps, correlating points on the scanned image with corresponding
locations within the brain in real time.
In subcortical lesions, the neuronavigation system assists in choosing the corticectomy
location, in planning and executing the trajectory and, finally, in the precise location
of the lesion.
However, the method has limitations. The navigation system handpiece tip used by the
surgeon at a certain point in the brain correlates with MRI or CT images obtained
prior to the surgical procedure. However, during the surgical procedure, the brain
moves, either because of gravity or due to relaxation caused by the drainage of CSF
(brain shift). Regular systems do not automatically correct for this brain shift and,
therefore, the target to be reached. This displacement causes an error in the correlation
between the structure shown on the console display and the exact brain region pointed
out by the surgeon, making the navigation method inaccurate.
The error caused by brain shift is progressive and depends on several other factors
such as head position, cerebral edema, bleeding, use of brain retractors, removal
of the lesion and the effects of anesthetic drugs, diuretics or mechanical ventilation.
This correlation accuracy between the neuronavigation and the target structures was
studied by Zinreich et al in a replica of the skull made of plastic, and subsequently
in three patients.[2] In the model, the average error between the image and the actual location was between
1 and 2 mm. In surgical cases, the display error between the point shown on the system
display and the true anatomical location varied from 0.3 to 2.2 mm.
The error caused by the brain shift was estimated by Golfinos et al[3] as being < 2mm in 92% of the navigation systems using MRI images, and in 82% when
CT scans were used. However, over time, the method becomes even more inaccurate. Errors < 2mm
can be found in 77% of cases using MRI and in 62% with CT scan images.
There are several technological solutions to correct the brain shift effects, updating
the data obtained in real time. Among them are the electromagnetic systems and those
based on ultrasound. However, these systems are not widely available.[4]
The accuracy of neuronavigation systems can be increased by using intraoperative MRI
images, which provide real-time data, assessing the effects of brain shift during
surgery and documenting residual lesions.
Nimsky et al used intraoperative MRI to quantify and correct brain shift.[5] There was great variability in measurements, reaching 24 mm of cortical surface
shifting and > 3 mm in the margins of deep lesions. Of the 64 study participants,
intraoperative MRI scans were performed using a 0.2-T device, with an open setup,
to correct and update navigation data.
In another study,[6] Nimsky et al described the use of a microscope integrated with the MRI machine in
the operating room. The patients were placed on the MRI machine table with their heads
fixed at ∼ 1.5 m from the center of the magnetic field by a head support made of compatible
ceramics. For the acquisition of intraoperative images, the table was slid to the
center of the resonance in < 30 seconds, enhancing the acquisition of intraoperative
images to correct for the brain shift.
Brain shifting can be quite distinct according to the brain region of interest. Furthermore,
it is a dynamic process. Nabavi et al developed a software (3D Slicer) that makes
a quantitative analysis of the degree and direction of the brain shift.[7] For this, they used a 0.5 T MRI device, vertically open, which enables the acquisition
of multiple intraoperative images without the need to move the patient. With this
method, they managed to optimize image acquisition for computational analysis of brain
deformations during the surgical procedure.
Marreiros et al assessed a method for measuring and compensating for brain shift in
an experimental study in rabbits.[8] For this, they used superficial blood vessels as reference points. The vessels were
displayed by means of infrared cameras aligned with the central point of the vessels
seen in the preoperative magnetic resonance angiogram. There was a good correlation
between the volume shift seen in the system and the real state of the brain, with
∼ 4 mm displacements.
Following the same rationale, Jiang et al proposed a method to compensate for the
brain shift that integrates vessel and groove sightings, using a phase-change three-dimensional
(3D) measurement, which provides direct 2D image pixels and their corresponding 3D
points.[9] The brain surface measurements were tracked noninvasively in five pig brains, assessing
the shifting caused by gravity or by direct compression.
Sun H et al developed a method capable of estimating the three-dimensional shape of
the brain surface in real time.[10] They compared the shape with the cortical surface preregistered in the MRI. Cortical
displacement was used to guide the whole brain model by updating the volumes obtained
by preoperative MRI imaging. According to the authors, the accuracy of this system
was ∼ 1mm.
In a two-part experimental study, Reinertsen et al compared the brain shift correction
from intraoperative ultrasound images and preoperative MRI.[11] In the first part, they tested the technique using realistic simulation, comparing
the results with previously measured brain shifts. In the second part of the study,
they obtained MRI and ultrasound images from a pva-cryogel phantom model for three
different deformations. On average, deformations of a 7.5 mm magnitude were corrected
to 1.6 mm by the ultrasound recording, and to 1.07 mm using the MRI data.
The methods described to compensate for the effects of brain shift are sophisticated
and are not available in most services.
Our arachnoid anchoring method, hereby described, is simple and easy to perform. The
arachnoid anchorage sutures keep the cortical surface closer to the dura mater, decreasing
the cortical surface displacement caused by gravity and CSF drainage. It can be used
to approach subcortical lesions, minimizing the effects of brain shift vis-à-vis the
location of the lesion. Its great merit is to facilitate the location of the lesion,
avoiding errors in trajectory and distance from the surface caused by the brain shift.
It does not prevent the effects of brain dislocation caused by the progressive exeresis
of the lesion; thus, it is not useful for assessing the degree of lesion resection.
Conclusion
Arachnoid anchorage sutures to the dura near the edges of the cortex area exposed
by the craniotomy constitute a simple method to minimize brain displacement (brain-shift)
in craniotomies for subcortical lesions, optimizing the usage of the neuronavigation
system.
