Echoplanar BOLD fMRI in a Patient with Neoplasm in Motor Cortex

Presurgical Mapping of Regional Brain Activity using Echoplanar BOLD fMRI
 
Mark S. George, MD; Diana J. Vincent, PhD; Donna Roberts, MS; 

Daryl Bohning, PhD; Sunil Patel, MD; Cris Vera, MD; 

Joel Cure, MD; Joe Horton, MD; Jeremy Young, MD

Departments of Radiology, Psychiatry, Neurology and Neurosurgery,
 Medical University of South Carolina, Charleston; Psychiatry Department, 
Ralph H. Johnson Veterans Affairs Medical Center, Charleston

INTRODUCTION
Neurosurgeons have long sought ways to determine the consequences of resecting brain tissue:
that is, of understanding the function of brain tissue around areas that require operation and
removal. This information is vitally important both to plan the extent of surgical resection
as well as for accurate presurgical assessment of the risks and benefits of surgery. Through
the years, various techniques have been developed to this end. They include somatosensory
evoked potential (SEP) mapping, or temporarily anesthetizing a hemisphere using
amobarbital, a procedure known as the Wada test. More recently, PET and SPECT scans have
been used, however these are often not helpful because of radiation dosimetry limits,
limiting the number of scans in a patient, and thus the ability to make statements about
regional brain function within an individual. Thus, a non-invasive method for determining
brain function, with good spatial resolution, would be a powerful addition to neurosurgeons
armamentarium.

Recently, echoplanar Blood Oxygenation Level Dependent (BOLD) MRI scanning has been
used to detect local brain activity. By weighing the images to be more sensitive to
hemoglobin, these MR scans can detect subtle changes in the ratio of oxy to deoxy
hemoglobin. Brain regions that are more active in a task have an initial anaerobic moment,
and send out signals causing the brain to flood that region with oxyhemoglobin. By
alternately having someone perform a task, and then rest, one can detect these subtle
differences in regional brain activity, as measured by oxyhemoglobin, associated with a
particular task .

Several groups are now using this technique in selected presurgical patients. For
example, researchers at the Medical University of Wisconsin have scanned epilepsy patients
before elective surgery, determining with high accuracy the regions involved in language. We
have been performing these epilepsy language studies as well at MUSC. Below is an example
of a slightly different activation paradigm, designed to map motor cortex activity.

CASE DESCRIPTION
A 30 year old white male was in good health until he suffered a focal seizure originating in
his left leg, with secondary generalization. An MRI scan showed a hypointense (T1)
unenhancing lesion in the right parasagittal posterior frontal lobe.
Prior to surgery, we attempted to define his motor cortex using the following method. The
patient was scanned in a Picker 1.5 Tesla EDGE scanner equipped with whole body high
performance gradients, where 15 5 mm thick coronal structural (FSE) images were initially
acquired. We then acquired BOLD T2* weighted images (TE 40, TR 3000, FOV 25.5 cm, slices
5mm thick, no interslice interval) over 6 minutes, with the subject alternating between rest
and performing a simple foot movement task with his left foot (alternately moving the foot
in a cross-like pattern). Twenty 15 slice volumes were thus acquired per minute (x 6 = 120
volumes x 15 = 1800 total images in 6 minutes). These BOLD-weighted images were
transferred to a SUN SPARCstation and realigned for movement (AIR, Roger Woods). Moving
images were compared with rest, using a paired t-test (FIDAP, Maisog and Haxby, NIMH).
Areas of significant activation (p<0.001) were then superimposed onto the structural scan,
acquired in the same planes.

A small area of increased signal is apparent just above the tumor, on the medial
surface. Interestingly, there is also activity in the left hemisphere (ipsilateral to
movement). This activity is even more significant than the right hemisphere activity (in
terms of statistical significance and size, or number of pixels activated).

Based entirely on this scan, it appeared that the left hemisphere cortex had an
important role in ipsilateral left foot voluntary movement. We thus expected the patient to
have good left leg motor function post-surgery. Further, we reasoned that although bilateral
motor activation is commonly seen, especially when complex tasks are performed, it would
be unusual in an adult to have this degree of ipsilateral brain activation during a motor task,
unless the tumor had been there for many years and was slow-growing, thus allowing the
cortex to remap. We thus reasoned, albeit tentatively with this novel

At time of surgery, cortical mapping was performed with electrical stimulation which
confirmed that the motor area was just behind the tumor. The tumor was removed. One week
post-op the patient was walking normally. Pathological examination of the tumor revealed it
to be a low-grade glioma.

DISCUSSION
This clinical example demonstrates the potential uses of this powerful new tool. However,
a note of caution is needed. The use of echoplanar BOLD fMRI as a presurgical tool is still in
its infancy, with many issues poorly understood or still unresolved. We know that head
movement can greatly degrade the quality of the scans, limiting these studies to cooperative
alert patients who are able to follow commands while remaining otherwise still. We are
also unclear about the extent of 'effort' or practice effects on the signals obtained , the
degree of the signal that comes from capillary activity versus draining veins, and issues
about possible artifacts from the lesion tissue itself .

Clearly, much work needs to be done in refining this technique in order to fully
understand its role, and limitations, in presurgical planning. However, examples such as
this hint at the potential of this non-invasive method.

Return to Functional Neuroimaging