EARLY CAROTID ENDARTERECTOMY AFTER ISCHEMIC
STROKE IMPROVES DIFFUSION/PERFUSION MISMATCH
ON MAGNETIC RESONANCE IMAGING:
REPORT OF TWO CASES
Satish Krishnamurthy,
M.D., M.Ch.
Pennsylvania State
University–Milton S. Hershey
Medical Center, Hershey,
Pennsylvania

David Tong, M.D.
Stanford University Medical
Center, Stanford, California

Kevin P. McNamara, M.D.
Pennsylvania State
University–Milton S. Hershey
Medical Center, Hershey,
Pennsylvania

Gary K. Steinberg, M.D.,
Ph.D.
Stanford University Medical
Center, Stanford, California

Kevin M. Cockroft, M.D.
Pennsylvania State
University–Milton S. Hershey
Medical Center, Hershey,
Pennsylvania
Reprint requests:
Kevin M. Cockroft, M.D., Division
of Neurosurgery, Pennsylvania
State University–Milton S. Hershey
Medical Center, PO Box 850, MC
H110, Hershey, PA 17033-0850.
Received, November 7, 2001.
Accepted, August 12, 2002.

OBJECTIVE AND IMPORTANCE: The functional magnetic resonance imaging techniques of diffusion-weighted imaging and perfusion-weighted imaging allow for ultraearly detection of brain infarction and concomitant identification of blood flow
abnormalities in surrounding regions, which may represent brain “at risk.”
CLINICAL PRESENTATION: We report two patients with acute ischemic stroke associated with ipsilateral high-grade carotid stenosis. The first patient, a 64-year-old
woman with a remote history of ischemic stroke and a vertebral artery aneurysm,
presented with worsening of her preexisting right hemiparesis. The second patient,
another 64-year-old woman with known multiple intracranial aneurysms and bilateral
high-grade internal carotid artery stenosis, was admitted for the elective microsurgical
clipping of an enlarging giant left carotid-ophthalmic artery aneurysm. Postoperatively,
she developed right hemiparesis and mild aphasia. Both patients showed progressive
worsening of their neurological deficits in the setting of small or undetected diffusionweighted imaging abnormalities and large perfusion-weighted imaging defects.
INTERVENTION: After prompt carotid endarterectomy, symptoms in both patients
resolved or improved. Follow-up magnetic resonance imaging scans demonstrated
resolution or significant improvement in the perfusion abnormalities in both patients.
CONCLUSION: Carotid endarterectomy in the setting of diffusion-weighted/perfusionweighted imaging mismatch can lead to improvement in cerebral perfusion as evidenced by resolution of the perfusion-weighted imaging lesion. Diffusion/perfusion
magnetic resonance imaging may be useful in identifying patients with severe neurological deficits but without large territories of infarction who may safely undergo early
surgical revascularization.
KEY WORDS: Acute stroke, Carotid endarterectomy, Carotid stenosis, Diffusion-weighted magnetic
resonance imaging, Perfusion-weighted magnetic resonance imaging
Neurosurgery 52:238-242, 2003

W

DOI: 10.1227/01.NEU.0000039562.07785.A8

hereas computed tomography may
not provide an accurate analysis of
infarct size for several days, newer
magnetic resonance imaging (MRI) techniques
allow ultra-early detection of brain infarction.
Diffusion-weighted imaging (DWI) can rapidly detect ischemic brain lesions, and a complementary technique, perfusion-weighted
imaging (PWI), can identify blood flow abnormalities (1, 14). Such acute information may
prove valuable in selecting patients for early
surgery and predicting future neurological
outcomes. We describe two patients with progressive ischemic symptoms leading to severe
neurological deficits. Both patients had small

238 | VOLUME 52 | NUMBER 1 | JANUARY 2003

www.neurosurgery-online.com

or nonexistent lesions as revealed by DWI but
large lesions as revealed by PWI. After early
carotid endarterectomy, neither patient had a
significant complication. Both patients subsequently demonstrated resolution of their
DWI/PWI mismatch with improvement in
neurological function.

CASE REPORTS
Patient 1
A 64-year-old woman with a remote history
of ischemic stroke and Hunt and Hess Grade
IV subarachnoid hemorrhage from a right vertebral artery aneurysm presented with wors-

www.neurosurgery-online.com

DIFFUSION/PERFUSION MISMATCH AND CAROTID ENDARTERECTOMY

ening of her preexisting right lower extremity weakness. At
baseline, her right lower extremity strength was graded as 4/5
and she was able to walk with a cane. At presentation, she
barely had antigravity strength and had considerable difficulty walking. The patient’s medical history also was remarkable for hypertension and non-insulin-dependent diabetes
mellitus. A DWI scan obtained on her first hospital day did
not reveal evidence of a new acute infarct. A magnetic resonance angiogram obtained at the same time showed highgrade (⬎90%) stenosis of the left internal carotid artery (ICA)
just distal to the bifurcation (Fig. 1A). The patient’s right lower
extremity paresis fluctuated initially; then it progressively
worsened on hospital Day 5 to the point that she no longer had
antigravity strength and could not walk. A follow-up DWI
scan obtained on hospital Day 5 revealed a small ischemic
lesion in the left anterior cerebral artery distribution (Fig. 1B).
However, a concomitant PWI scan revealed a much larger
region of low perfusion in the same hemisphere (Fig. 1C). On
the basis of the observed DWI/PWI mismatch, which suggested only a relatively small area of infarcted tissue, early
surgical revascularization was recommended. Postoperatively, the patient’s neurological deficit improved to baseline.
A follow-up MRI scan obtained 1 week later demonstrated

normal perfusion in the left hemisphere (Fig. 1D). There was
no change in the diffusion abnormality.

Examination Technique
All MRI scans for Patient 1 were obtained using a 1.5-T GE
Signa MRI scanner (General Electric, Milwaukee, WI) at the
Stanford University Medical Center. DWI was performed using echo planar imaging (EPI) technique. Multislice wholebrain images were obtained in 12 slices: TR, 6000 milliseconds;
TE, 110 milliseconds; field of view, 24 cm; slice thickness, 5
mm with a 2.5-mm skip between slices; image size, 128 ⫻ 128.
The b values were 0 and 741. DWI images were acquired in x,
y, and z planes, with and without an inversion pulse (fluidattenuated inversion recovery sequence). EPI diffusion images
were processed to generate average (trace) apparent diffusion
coefficient maps. The b values (0 and 741) were measured
twice to improve the accuracy of the apparent diffusion coefficient determination. PWI was performed using the same EPI
MRI equipment after the injection of 30 ml of gadolinium (0.2
mmol/kg). PWI acquisition values were TR, 2000 milliseconds
and TE, 40 milliseconds and 35 time points. Other parameters
were the same as for DWI. The total imaging time was approximately 3 minutes. Perfusion images were processed to
generate maps of time to bolus peak (TTP), which is the time
from the start of the scan to the peak bolus effect, and regional
cerebral blood volume, which is the integral of the area under
the bolus transit curve. Approximately 400 individual scans
were acquired within 3 minutes. The total scanning time for
the entire protocol was 15 to 20 minutes.

Patient 2

FIGURE 1. Patient 1. A, cervical magnetic resonance angiogram demonstrating greater than 90% stenosis of the left ICA. B, DWI scan obtained
5 days after the onset of symptoms, showing a small ischemic lesion in the
left anterior cerebral artery distribution. C, PWI scan from the same
series, revealing a much larger region of low perfusion in the left hemisphere. D, follow-up PWI scan obtained 1 week after emergent carotid
endarterectomy, showing normal (symmetric) perfusion to the left
hemisphere.

NEUROSURGERY

A 64-year-old woman was admitted for elective clipping of
an enlarging giant left carotid-ophthalmic artery aneurysm.
Her medical history was remarkable for coronary artery disease, a previous myocardial infarction, and a previous coronary angioplasty, as well as multiple intracranial aneurysms
and bilateral high-grade (80–90%) ICA stenoses. Preoperative
neurological examination revealed a left visual field defect.
Intraoperatively, temporary occlusion of the left ICA in the
neck and suction decompression of the aneurysm were required. During this temporary occlusion procedure, a significant decrement in somatosensory evoked potentials from the
patient’s right arm (decreased amplitude ⬎50%) was noted.
This improved somewhat toward the end of the procedure,
but it did not return to baseline. Postoperatively, the patient
had mild expressive aphasia and mild right hemiparesis (right
pronator drift with approximately 4⫹/5 strength). A computed tomographic scan obtained on the first postoperative
day revealed no new infarcts. A postoperative cerebral angiogram obtained the same day demonstrated complete occlusion
of her giant aneurysm without evidence of occlusion or stenosis of the adjacent intracranial carotid. In addition, there
was no change in her high-grade cervical ICA stenosis (Fig.
2A). Her examination results initially fluctuated, then progressed to hemiplegia and a dense expressive aphasia on the

VOLUME 52 | NUMBER 1 | JANUARY 2003 | 239

KRISHNAMURTHY ET AL.

second postoperative day. A repeat
computed tomographic scan obtained
at that time revealed a small hypodensity in the left caudate (Fig. 2B). A
subsequent MRI scan obtained the
same day did not demonstrate a significant diffusion defect (Fig. 2C), although her numerous aneurysm clips
generated considerable metallic artifact. PWI revealed decreased perfusion throughout the entire left hemisphere (Fig. 2D). On the basis of her
deteriorating clinical condition and
the DWI/PWI mismatch, she underwent emergent left carotid endarterectomy later that day. Postoperatively, she had no immediate change
in her neurological function. A repeat
MRI scan obtained 3 days after endarterectomy revealed a profound improvement in the perfusion of the left
cerebral hemisphere and no new diffusion defect (Fig. 2E). During the
next few days, the patient’s aphasia
and hemiparesis gradually improved.
At the time of discharge, 19 days after
her first operation, the patient had no
demonstrable weakness on her right
side.

FIGURE 2. Patient 2. A,
common carotid angiogram
obtained the day after initial
surgery, demonstrating highgrade stenosis of the left ICA.
B, computed tomographic
scan obtained on the second
postoperative day, revealing a small hypodensity in the left caudate (solid arrow). The patient’s giant basilar
artery bifurcation aneurysm also is visible (open arrow). C, DWI scan obtained the same day, revealing no
significant diffusion defect. D, concomitant PWI scan showing decreased perfusion throughout the left hemisphere.
E, follow-up PWI scan obtained 3 days after emergent carotid endarterectomy, showing a profound improvement
in perfusion to the left hemisphere.

Examination Technique
All MRI scans for Patient 2 were obtained using a 1.5-T
Picker Eclipse MRI scanner (Picker International, Inc., Cleveland, OH) at Pennsylvania State University–Milton S. Hershey
Medical Center. DWI was performed using spin echo (EPI)
technique. Multislice whole-brain images were obtained in 19
slices: TR, 5629 milliseconds; TE, 110 milliseconds; field of
view, 24 cm; slice thickness, 5 mm with a 1-mm gap between
slices; image size, 128 ⫻ 128. The b values were 0 and 741. DWI
images were acquired in x, y, and z planes. EPI diffusion
images were processed to generate apparent diffusion coefficient maps. PWI was performed using field echo (EPI) technique after injection of 15 ml of gadolinium (0.1 mmol/kg).
PWI acquisition values were TR, 1000 milliseconds and TE, 60
milliseconds and 40 time points. Other parameters were 128 ⫻
128 matrix size and 5-mm slice thickness with a 5-mm gap for
a total of 10 slices. The total imaging time was approximately
56 seconds. PWI scans were processed to generate maps of
TTP and regional cerebral blood volume. The total scanning
time for the entire protocol was 15 to 20 minutes.

DISCUSSION
DWI is of established value in detecting acute cerebral
ischemia (3, 4). Lesion volumes visualized on early DWI scans
in stroke patients have a strong correlation with final infarct

240 | VOLUME 52 | NUMBER 1 | JANUARY 2003

volumes and clinical neurological outcomes (2, 13, 15). Serial
DWI scans obtained during the first several days after stroke
onset frequently reveal progressive enlargement of the lesion.
The eventual lesion volume as revealed by DWI correlates
closely with the final lesion volume as revealed by T2weighted MRI scans, which in turn correlates closely with the
final volume of infarcted brain tissue (2, 13, 15).
PWI is a complementary technique to DWI. It uses a rapidly
injected bolus of contrast agent to create a qualitative map of
brain perfusion (3, 10, 11). A variety of hemodynamic maps
can be created by use of PWI. In animal models, lesions can be
observed with PWI immediately after vessel occlusion, and
they resolve rapidly after successful thrombolysis or reperfusion (8, 16). In studies of acute stroke patients, the volume of
the early lesion revealed by PWI correlates more closely with
the acute neurological deficit, which suggests that PWI provides a more accurate estimate of the volume of dysfunctional
brain tissue than does DWI (2, 13).
In both of the patients described herein, a large lesion revealed by PWI was associated with a severe neurological
deficit and little or no lesion on DWI scans. Undeniably, both
of these cases are complex. Whereas interpretation of neurological changes in the first patient was complicated by her
preexisting deficits, the second patient’s situation was confounded by her recent intracranial surgery and general anes-

www.neurosurgery-online.com

DIFFUSION/PERFUSION MISMATCH AND CAROTID ENDARTERECTOMY

thesia. Clearly, however, the traditional preference would
have been to offer delayed carotid surgery to both of these
patients. Yet, prompt surgical revascularization was followed
by improvement in each patient’s neurological deficit. Neither
patient had a hemorrhagic complication as a result of early
surgery, and repeat MRI scans in both patients revealed resolution of their perfusion deficits. The primary benefit of
DWI/PWI in this situation is the ability to differentiate a
severe neurological deficit caused by a large completed infarction (DWI) from a similar deficit associated with a large,
potentially reversible area of ischemia (PWI).
Whether DWI/PWI will be useful in selecting patients who
will benefit from early surgery remains to be learned. It is difficult to extrapolate, from only two cases, specific DWI/PWI risk
factors that might predict a poor outcome and/or hemorrhagic
conversion after early carotid endarterectomy. In addition, it is
impossible to conclude that early surgical intervention in cases of
significant DWI/PWI mismatch actually leads to improved neurological outcome. Without baseline perfusion data, it is impossible to ensure that the decreases in blood flow, presumably
associated the PWI changes, were the cause of the neurological
deficit. It is interesting, however, that the second patient had
bilateral high-grade ICA stenosis but there remained a substantial difference between the perfusion images of the two hemispheres (Fig. 2C). This suggests that an acute alteration in cerebral
hemodynamics between the two hemispheres may have occurred.
Furthermore, the present report does not include a quantitative measurement of cerebral blood flow or cerebral blood volume. PWI is a relative test of perfusion, and the PWI measurements used herein are based on a TTP analysis, which may
reflect the duration of collateral blood flow pathways more than
actual tissue perfusion. PWI maps usually are generated with at
least a 1- to 2-second delay in perfusion between hemispheres. In
our patients, a TTP delay of 1 second was used, because this was
the standard at the time these patients were studied. However,
this short delay may lead to an overestimation of the tissue at
risk. Recently, Neumann-Haefelin et al. (9) reported that only
severe perfusion deficits (TTP delays ⱖ6 s) correlated with final
infarct volume. In the future, it may be important to view perfusion with various TTP delays when deciding whether significant brain remains at risk. Although a quantitative measure of
cerebral blood flow may be ideal, techniques to acquire such
information, such as positron emission tomography or xenon
blood flow tomography, may not be readily available and therefore may be impractical for widespread use. Thus, several groups
are investigating alternative methods of generating quantitative
blood flow data from MRI and correlating these data with other,
more traditional measures of cerebral blood flow and cerebral
blood volume (5–7, 12).

CONCLUSION
Carotid endarterectomy in the setting of DWI/PWI mismatch can lead to improvement in cerebral perfusion as evidenced by resolution of the PWI lesion. Further study will be
needed to determine whether DWI/PWI techniques are useful

NEUROSURGERY

in identifying patients with high-grade carotid stenosis and
ipsilateral acute stroke who may be appropriate candidates for
early carotid endarterectomy despite severe neurological deficit. In the future, functional MRI techniques using DWI and
PWI may be helpful in differentiating which of these patients,
if any, will derive improved neurological outcomes after early
surgery.

REFERENCES
1. Albers G: Expanding the window for thrombolytic therapy in acute stroke:
The potential role of acute MRI for patient selection. Stroke 30:2230–2237,
1999.
2. Baird A, Benfield A, Schlaug G, Siewert B, Lovblad K, Edelman R, Warach
S: Enlargement of human cerebral ischemic lesion volumes measured by
diffusion-weighted magnetic resonance imaging. Ann Neurol 41:581–589,
1997.
3. Fisher M, Albers G: Applications of diffusion-perfusion magnetic resonance
imaging in acute ischemic stroke. Neurology 52:1750–1756, 1999.
4. Fisher M, Pritchard J, Warach S: New magnetic resonance techniques for
acute ischemic stroke. JAMA 274:908–911, 1995.
5. Kikuchi K, Murase K, Miki H, Kikuchi T, Sugawara Y, Mochizuki T, Ikezoe
J, Ohue S: Measurement of cerebral hemodynamics with perfusionweighted MR imaging: A comparison with pre- and post-acetazolamide
133Xe-SPECT in occlusive carotid disease. AJNR Am J Neuroradiol 22:248–
254, 2001.
6. Kim J, Lee S, Shin T, Kang K, Choi P, Kim J, Gong J, Choi N, Lim B:
Correlative assessment of hemodynamic parameters obtained with T2*weighted perfusion MR imaging and SPECT in symptomatic carotid artery
occlusion. AJNR Am J Neuroradiol 21:1450–1456, 2000.
7. Lythgoe D, Ostergaard L, William S, Cluckie A, Buxton-Thomas M,
Simmons A, Markus HS: Quantitative perfusion imaging in carotid artery
stenosis using dynamic susceptibility contrast-enhanced magnetic resonance imaging. Magn Reson Imaging 18:1–11, 2000.
8. Muller T, Haraldseth O, Jones R, Sebastiani G, Godtliebsen F, Lindboe C,
Unsgard G: Combined perfusion and diffusion-weighted magnetic resonance imaging in a rat model of reversible middle cerebral artery occlusion.
Stroke 26:451–457, 1995.
9. Neumann-Haefelin T, Wittsack H, Wenerski F, Siebler M, Seitz R, Modder
U, Freund H: Diffusion- and perfusion-weighted MRI: The DWI/PWI mismatch region in acute stroke. Stroke 30:1591–1597, 1999.
10. Patel M, Siewert B, Warach S, Edelman R: Diffusion and perfusion imaging
techniques. Magn Reson Imaging Clin N Am 3:425–438, 1995.
11. Sorensen A, Tievsky A, Ostergaard L, Weisskoff R, Rosen B: Contrast agents
in functional MR imaging. J Magn Reson Imaging 7:47–55, 1997.
12. Teng M, Cheng H, Kao Y, Hsu L, Yeh T, Hung C, Wong W, Hu H, Chiang
J, Chang C: MR perfusion studies of brain for patients with unilateral carotid
stenosis or occlusion: Evaluation of maps of “time to peak” and “percentage
of baseline at peak.” J Comput Assist Tomogr 25:121–125, 2001.
13. Tong D, Yenari M, Albers G, O’Brian M, Marks M, Moseley M: Correlation
of perfusion and diffusion-weighted MRI with NIHSS score in acute (⬍6.5
hour) ischemic stroke. Neurology 50:864–870, 1998.
14. Warach S, Chien D, Li W, Ronthal M, Edelman R: Fast magnetic resonance
diffusion-weighted imaging of acute stroke. Neurology 42:1717–1723, 1992.
15. Warach S, Dashe J, Edelman R: Clinical outcome in ischemic stroke predicted by early diffusion-weighted and perfusion magnetic resonance imaging: A preliminary analysis. J Cereb Blood Flow Metab 16:53–59, 1996.
16. Yenari M, de Crespigny A, Palmer J, Roberts S, Schrier S, Albers G, Moseley
M, Steinberg G: Improved perfusion with rt-PA and hirulog in a rabbit
model of embolic stroke. J Cereb Blood Flow Metab 17:401–411, 1997.

Acknowledgments
We thank Christina Moorhead for assistance in preparation of the manuscript.
This work was supported in part by funding provided by Bernard and Ronni
Lacroute and the William Randolph Hearst Foundation (GKS).

VOLUME 52 | NUMBER 1 | JANUARY 2003 | 241

KRISHNAMURTHY ET AL.

COMMENTS

K

rishnamurthy et al. highlight the issue of early endarterectomy after small strokes with ongoing perfusion mismatch. They performed early endarterectomy and suggest
that, in this subpopulation, they definitely would perform
early surgery when defined by radiological criteria such as
these.
In our institution, both of these patients would have been
offered early endarterectomy. Our tendency has been to move
quickly in treating patients with small strokes and high-grade
stenosis. The concept outlined in this report is not particularly
controversial. However, this report raises the question of how
such an approach might be used in patients with moderatesize infarcts and ongoing large perfusion/diffusion mismatch.
In such cases, the risk of hemorrhage certainly is increased for
early endarterectomy, yet there is the ongoing prospect of
additional and sizable tissue loss.
We recently experienced a case similar to this latter scenario,
in which an early endarterectomy was performed in the setting of a moderate-sized infarct as revealed by diffusionweighted imaging. However, fluctuating symptoms (hemiplegia) and a large additional area of perfusion defect prompted
us to proceed with endarterectomy. The feared hemorrhage
into the stroke area occurred, but the eventual outcome was
better than we would have anticipated if the area of perfusion
mismatch had converted to frank infarction. In the end, we
thought that we had walked a fine line between two very
difficult clinical scenarios.
This report highlights the increasing amount of data made
available by magnetic resonance imaging regarding cerebral
blood flow and areas of tissue at risk. Additional compilation
of results in these very specific scenarios is warranted to allow
us to understand the overall risk-to-benefit ratio of our
interventions.
Bob S. Carter
Christopher S. Ogilvy
Boston, Massachusetts

I

n this article, the authors report two quite different cases
that have the common thread of endarterectomy performed in a patient with a new or increasing neurological
deficit and a diffusion/perfusion mismatch. The lack of
large diffusion deficits suggests that a fixed stroke was not
present. The patients improved both in clinical course and
perfusion after endarterectomy. This is consistent with
present thinking regarding acute thrombolysis after neurological deficit when a favorable perfusion/diffusion mismatch exists. As such, it is an interesting although not
unique contribution.
The major questions here are twofold. First, can a
perfusion/diffusion mismatch be reversed by a revascularization procedure such as endarterectomy? Not surprisingly, this is true, and it is observed in these patients. The
larger questions remain: is there a safe time to proceed with
an endarterectomy or other reperfusion technique, and is
perfusion/diffusion mismatch useful in determining the
safety of this procedure? This could be addressed if we
knew the conditions that would predict worsening after an
intervention, such as the diffusion-weighted imaging conditions that would predict which patients will have hemorrhagic conversion of a stroke area if endarterectomy is
performed. These questions need to be answered to determine whether the timing of surgery needs to be restricted,
and if so, in which cases. This article does not answer that
question. As a result, it does not add information as to the
best timing of surgery for a patient with an infarct, but it
does, as the title suggests, provide information regarding
the reversal of perfusion mismatch.
The results in these patients cannot be generalized to all
patients undergoing endarterectomy. The article increases
our understanding of perfusion/diffusion mismatch and
the anatomic relationship to stenosis ipsilateral to the mismatch. However, our knowledge regarding safe timing of
endarterectomy after neurological deficit has not been
increased.

Gaudi’s workshop in the Sacred Family Cathedral, showing innumerable plaster models.

Robert J. Dempsey
Madison, Wisconsin