Although many articles have been published on the diagnosis and
imaging of carotid stenosis, very few articles have been dedicated
to a common-sense evaluation of patients with carotid vascular
disease. This article will review the current literature and
present a common-sense evaluation of patients with carotid vascular
disease. An algorithm also is provided as a useful guide to the
proper work-up of patients with carotid stenosis (figure 5).
The three most useful imaging modalities for carotid vascular
disease, carotid duplex ultrasound, magnetic resonance angiography
(MRA), and angiography, will be discussed, including their
strengths, weaknesses, and accuracy. Each modality's most common
pitfalls will be outlined as well as how to avoid them. A brief
discussion of computed tomography (CT) angiography also is
Data from the multicenter study of the surgical and medical
treatment for carotid stenosis (North American Symptomatic Carotid
Endarterectomy Trial [NASCET]) demonstrated a definite benefit of
carotid endarterectomy for symptomatic patients with a 70% to 99%
stenosis of the internal carotid artery.1-3,19 The NASCET studied
symptomatic patients with ipsilateral carotid stenosis of 70% to
99%. The 30-day postoperative risk for stroke or death was 5.8%.
The cumulative risk of ipsilateral stroke at two years was 9% in
the surgical group and 26% in the medical group. In the 70% to 99%
stenotic group, the benefit from surgery was in direct proportion
to the severity of the stenosis. The difference in stroke-free
survival at two years was 26% for patients with 90% to 99%
stenosis, 18% for those with 80% to 89% stenosis, and 12% for those
with 70% to 90% stenosis.1,3
According to the National Institute for Neurological Disorders
and Stroke, patients with asymptomatic 60% to 99% carotid stenosis
who underwent carotid endarterectomy had a 55% decrease in the
incidence of stroke or death within five years, compared to
medically treated patients. This was the case only at medical
centers that had perioperative morbidity and mortality rates of
less than 3%.1
Carotid duplex sonography is an excellent screening study for
carotid stenosis. It is inexpensive, noninvasive, and fairly
accurate, and it can be performed at the patient's bedside. Carotid
duplex ultrasound combines high-resolution real-time grey-scale
imaging with Doppler, providing information on physiologic blood
flow and luminal stenosis.
Ultrasound is more accurate than angiography and MRA in
evaluating the atherosclerotic plaque morphology, regardless of
whether the plaque is soft or hard, has intraplaque hemorrhage, or
is ulcerated. Furst et al demonstrated that Doppler sonography is
accurate in demonstrating ulcerated plaques in patients with
high-grade or critical stenosis, with a specificity of 95.3% and an
accuracy of 94%.1,4
Edwards et al concluded that conventional angiography identified
only 60% of the ulcers demonstrated at surgery.5 An ulcerated
plaque also can be a source of distal embolism in patients who do
not have a critical stenosis.9
There are many drawbacks to the use of carotid ultrasound; some
of the more common ones are discussed below.6,12,24,38
- The most common pitfall of carotid ultrasound is that it is
very technologist-dependent. Only highly qualified, experienced
sonologists should be performing the procedure (figure 2).
- There is more (inter- and intraobserver) variability with
ultrasound than with MRA or angiography in the measurement of
- Because of the technical difficulty with Doppler ultrasound,
only dedicated high-quality color-flow equipment should be
- A vessel may be absent or of anomalous origin. The internal
carotid usually arises posterolaterally to the external, but
there can be at least a 5% variability in the normal anatomic
relationship. Thus, one actually may be visualizing the external
carotid artery, confusing it with the internal carotid or vice
versa. The two vessels have different wave forms. However, if
there is a stenosis or tandem lesion, this can change the normal
wave-form pattern and make it difficult to differentiate the
internal carotid from the external carotid.
- As an ultrasound pulse encounters dense material, such as
calcification or bone, most of the energy is reflected back
towards the transducer with shadowing distal to the dense
material; this obscures the normal anatomy or stenosis, which may
cause under- or overestimation of the degree of stenosis.
- At times, ultrasound is unable to determine accurately
whether there are tandem lesions. Tandem lesions are additional
stenotic lesions that are either proximal or distal to the
original stenosis. Tandem lesions occur in 2% of patients who
have significant stenosis of the cervical portion of the internal
carotid artery.1,2 Tandem lesions occur most commonly in the
region of the cavernous segment of the internal carotid artery
(ICA) or the horizontal segment of the middle cerebral artery or
in the common carotid artery.34 The effect of the distal stenosis
depends on the degree of its severity. The effect of a second
stenosis is additive if each of the two lesions is capable of
reducing flow. If one lesion is a critical stenosis, then that
single vessel will govern the degree of the flow.1,2,7 There is
debate in the literature about the significance of tandem
In a study by Schuler et al, there were no strokes and one
death in patients with stenosis limited to the bifurcation. In
patients with distal tandem lesions, 11.4% had intraoperative
or perioperative strokes and four operative deaths.1,10 At
times, a tandem lesion can be suggested by ultrasound if there
is a proximal lesion because the velocity measurements may not
correspond to the measured stenosis, as the proximal lesion
alters the velocity measurements. Obviously, MRA and
conventional angiography can detect tandem lesions within the
common carotid, cavernous segment of the carotid, or large
intracranial branch vessels.
- The internal and external carotid vessels often are very
tortuous; when this occurs, it is difficult to obtain good axial
orientation of the vessel. The tortuosity of the vessel will
change the velocity measurements just as a sharp angulation in a
high-pressure pipe will change the velocity across the area of
angulation (figure 3).
- A soft plaque can mimic flowing blood; this can be minimized
by performing good color-flow measurements and obtaining good
visualization of the vessel.
- Velocity measurements can increase in hypertensive patients.
Patients with decreased cardiac output can have lower velocity
measurements. Cardiac arrhythmias, aortic valvular lesions, and
cardiomyopathies all can effect velocity measurements.1,2
Velocity measurements can increase in the contralateral vessel by
shunting of blood through the involved vessel.1,2
- The vertebral arteries can be difficult to visualize with
ultrasound because of their small size and deep location. The
vertebral may have decreased flow or be occluded, which may be
responsible for the patient's symptoms.
- MRA is more accurate than ultrasound in the diagnosis of
carotid stenosis.1,2,17,21 In a study by Anderson and colleagues,
ultrasound misdiagnosed 20% of cases, which altered surgical
management, compared to only 5% misdiagnosed with MRA.21 In some
studies, MRA also is superior to ultrasound in diagnosing an
In certain situations, ultrasound can be more accurate than
angiography in diagnosing stenosis. The atherosclerotic plaque is
associated with arterial wall expansion.9 Angiography measures only
the patent lumen and cannot evaluate the vessel wall; ultrasound,
however, can evaluate the vessel wall. If the plaque is concentric,
it may be difficult to evaluate the degree of stenosis accurately,
especially if there is some normal luminal dilation of the vessel.
This same pitfall also is encountered when comparing coronary
angiography with intraarterial ultrasound.
Magnetic resonance angiography
Magnetic resonance angiography is an excellent screening and
diagnostic examination for carotid vascular disease, stenosis, or
occlusion.13,15,29,32 Magnetic resonance angiography has gained
acceptance rapidly, with stronger, more uniform magnets, improved
surface coils and software, and with experience.
MRA has certain advantages over ultrasound and angiography.
However, MRA also has certain pitfalls, which will be discussed.
Because the physics of MRA is complicated, there will be only a
In MRA, a magnetic field is applied to flowing blood. A
radiofrequency pulse is then applied, which changes the alignment
of the protons of the flowing blood within the selected slice. When
the protons begin to realign with the magnetic field, they emit a
radiofrequency pulse that is used to generate an image. The signal
from flowing blood is dependent on the MRA technique used and can
vary with different MRA parameters, including repetition time (TR),
echo delay time (TE), and other MRA parameters.
There are two main MRA techniques: time-of-flight (TOF) and
phase-contrast imaging.11,29 In time-of-flight imaging, the protons
of flowing blood have full magnetization on the entering slice. TOF
imaging can be two-dimensional (2D) or three-dimensional (3D). In
3D TOF imaging, an extra phase-encoding gradient is added, and
volumetric measurement can be performed. In 2D imaging, selected
slices are obtained and stacked together like a roll of coins to
generate an image (figures 1 to 4).
Phase-contrast MRA measures changes in the phase spin
magnetization along the magnetic field gradient. 2D TOF has greater
intrinsic contrast than 3D, is less sensitive to motion, and has
greater sensitivity for differentiating occlusion or slow flow
(figure 4). 3D TOF imaging tends to have higher intrinsic
resolution because it obtains volumetric measurements rather than
thin-section images with 2D imaging. 3D also is less susceptible to
areas of turbulent flow that occur at the carotid bulb and areas of
Because phase-contrast MRA is more sensitive to slow flow than
TOF imaging, it can be used to help differentiate severe stenosis
from occlusion and to evaluate tortuous vessels; however, it has
less resolution than TOF. One disadvantage of phase-contrast MRA
versus TOF imaging is that areas of slow- or high-velocity flow can
cause signal dropout. Thus, one may overestimate the degree of
Most recent studies evaluating MRA report sensitivities and
specificities approaching those of angiography in the diagnosis of
carotid stenosis. A new MRA technique, multiple overlapping
thin-section angiography (MOTSA), found sensitivity and specificity
of 90% to 95% for stenosis of equal to or greater than 70%.1,16
Recent studies using multiplanar reconstruction technique, which
requires selection of the angle that best delineates the stenosis,
have an almost 100% correlation with the degree of pathologic
stenosis. Studies by De Marco et al and others found no significant
error in the measurement of stenosis when MRA was compared to
conventional angiography using this technique.20 Multiple papers
claim 100% accuracy in the diagnosis of carotid occlusion with
The biggest pitfall of MRA is that it overestimates stenosis.
This may occur in greater than 10% of patients and it varies,
depending on different articles. However, it often does not change
patient management. Increased experience, review of source data, 2D
axial techniques sensitive for flow, and the new MRA techniques
previously described can help prevent overestimation of
stenosis.1,2,20,21,40 Using phase-contrast MRA in combination with
time-of-flight imaging can help reduce the overestimation of
stenosis. Also, the computer can rotate the image in 360° and can
help determine whether the visualized degree of stenosis is a
projectional finding or a real stenosis. This is one advantage over
If duplex ultrasound and MRA are performed together and the
patient has severe carotid stenosis and both studies agree, then
100% accuracy can be achieved.2,17,22,23 This is remarkable, as
most imaging studies can never hope to approach 100% accuracy.
Many centers now perform MRA alone in patients who are evaluated
for surgery without any false-positive results.33 As described
previously, MRA is superior to ultrasound in delineating carotid
stenosis. Other advantages already described are the ability to
detect tandem lesions and the easier visualization of the vertebral
arteries and the intracranial vasculature. MRA also is excellent
for the evaluation of carotid dissection.
Another rare pitfall of MRA is the possibility of diagnosing a
complete occlusion when, in fact, there is an extremely high-grade
stenosis with very slow flow through the area of stenosis.17,21,30
Again, this pitfall can be avoided by evaluating source data and
axial images and performing phase-contrast imaging (figure 2).
Also, correlation with ultrasound and the patient's symptoms could
help avoid this pitfall.
MRA and ultrasound are complementary modalities, as duplex
ultrasound has certain advantages over MRA. Thus, the two
modalities should be used in conjunction with each other, and the
pitfalls of each modality should be known, evaluated, and applied
to each individual patient and circumstance.
Angiography often has been described as the gold standard for
evaluating carotid stenosis or occlusion. But what is a gold
standard? Oftentimes in imaging, certain studies are called the
gold standard. Thus, if the gold standard indicates one finding, it
often is accepted as fact.
However, this often is not the case. Ninety percent correlation
of angiography with MRA does not mean that angiography is 10% more
accurate than MRA or ultrasound. In the series of 94 carotid
arteries studied by Litt et al, interobserver agreement for grading
conventional carotid angiograms was only 72%.1,8 In the study by
Chikos et al, the intraobserver agreement was only 82%, and the
interobserver agreement was only 74%.1,18
Angiography is an invasive procedure, and complications include
hemorrhage, infection, arteriovenous fistula, pseudoaneurysm,
dissection, embolic stroke, and death, with a potential total
complication rate of up to 4%.25,31 There also is the risk of an
allergic reaction to the contrast, and the patient may not be able
to cooperate with the study after the procedure has begun.
Angiography results also vary depending on whether digital
angiography or film screen angiography is performed. The
Asymptomatic Carotid Study Group reported at least a 1.2% incidence
of carotid-associated infarction (National Institute of Health
Clinical Advisory, September 28, 1994). The incidence of
catheter-related infarction or complication rates from angiography
often are reported from institutions where dedicated fellowship
training invasive radiologists are performing multiple daily
examinations. Obviously, the complication rate will be higher in
the community hospital setting. Angiography complication rates
almost double in patients with ischemic disease, compared to those
studied for other indications. Also, the risk will vary depending
on the age of the patient and other risk factors.
Angiography has other pitfalls. The degree of stenosis may vary
depending on the orientation and projection performed. Thus,
stenosis may be under- or overestimated, based on the projection.
The increased number of projections can help avoid this pitfall,
but the longer the catheter is in place (either in the patient's
groin or in the carotid artery), the greater the complication rate.
Angiography also may be difficult if there are very tortuous
overlapping vessels. Also, as previously described, angiography
only measures the actual patent lumen. The vessel wall itself
cannot be evaluated. Thus, stenosis may be under- or
As described previously, angiography is less sensitive than MRA
or ultrasound in evaluating ulcerated or hemorrhagic plaques. In
some patients, angiography is more accurate than ultrasound or MRA
in evaluating the degree of stenosis because of the previously
described pitfalls of MRA and ultrasound.
Computed tomography angiography
Computed tomography angiography requires the combination of a
helical scanner and intravenous contrast enhancement. Multiple
studies claim a sensitivity and specificity approaching MRA
results. Marks recently reviewed several papers that demonstrated a
sensitivity of 82% to 100% and a specificity of 94% to 100% for the
detection of surgically treatable stenosis.41
CT angiography has certain advantages over MRA. It is not as
susceptible to flow artifact and patient motion. Also,
claustrophobic patients and patients with pacemakers can be imaged.
However, CT angiography does require contrast injection, and it is
limited in patients with severe concentric calcifications. Because
its results parallel those of MRA, it may be able to replace MRA in
the provided algorithm (figure 5).
What is the most practical approach for the evaluation and
diagnosis of carotid vascular disease? There is controversy in the
literature. However, if one can obtain almost 100% accuracy with
the combination of MRA and ultrasound, why proceed with angiography
and its risk of complications (figure 5)? De Marco et al recently
suggested that noninvasive studies can replace conventional
angiography.42 (Please see the algorithm that has been provided,
which can help make a more accurate assessment of patients with
carotid vascular disease.)
There is no one test that has 100% accuracy. Angiography, duplex
ultrasound, and MRA all have certain advantages and disadvantages.
The important thing is to recognize the advantages, disadvantages,
and pitfalls and how to avoid and overcome them.
As with any procedure in medicine, one has to assess the
risk-benefit ratio. Ultrasound is an excellent screening study; it
is accurate, inexpensive, and quick, and it can be performed at the
bedside. Ultrasound should be the screening study of choice. Its
advantages include its ability to evaluate the plaque morphology
and correlate this with the patient's symptoms. If the ultrasound
is normal or demonstrates mild disease, then why proceed with MRA
or angiography? There are few false-negative cases (figure 5).
If there are technical problems with ultrasound or a discrepancy
between the velocity measurements and the measured stenosis and if
there is moderate to severe disease, then proceed with MRA (figure
At certain institutions, surgery is performed on the basis of
ultrasound findings only. If there is a technically good study and
the degree of the visualized stenosis corresponds with the velocity
measurements, then ultrasound is fairly accurate. This depends on
the institution and the degree to which the surgeons feel
comfortable with the procedure.36,37,39
If there is a discrepancy between the ultrasound and the MRA, it
may be better to repeat one of the two studies, depending on which
type of pitfall was encountered. The visualization of the MRA 2D
axial images, 3D rotating images, phase-contrast MRA, or even the
IV injection of gadolinium can be performed to overcome these
pitfalls rather than proceeding to angiography. Angiography should
be reserved for cases in which the MRA and ultrasound still
disagree on repeat imaging (figure 5).
Conventional angiography or CT angiography can be performed in
patients in whom MRA cannot be performed due to pacemakers or
metallic implants. Conventional angiography or CT angiography also
can be indicated in cases in which subtotal occlusion is suspected
but not demonstrable on ultrasound and MRA.
Duplex ultrasound, MRA, CT angiography, and conventional
angiography should be used in conjunction with each other, and the
pitfalls of each should be known, evaluated, and applied to each
individual case. AR
1. Naidich TP, Righi AM: Neurovascular imaging. Radiol Clin
North Am 33(1):117-166, 1995.
2. Rosovsky MA, Litt AW, Lamparello PJ, et al: MR angiography of
the central nervous system. Magn Reson Imag Clin North Am
3. North American Symptomatic Carotid Endarterectomy Trial
Collaborators: Beneficial effect of carotid endarterectomy in
symptomatic patients with high-grade carotid stenosis. N Engl J Med
4. Furst H, Hartl WH, Jansen I, et al: Color-flow Doppler
sonography in the identification of ulcerative plaques in patients
with high-grade carotid artery stenosis. Am J Neuroradiol
5. Edwards JH, Kricheff II, Riles T, et al: Angiographically
undetected ulceration of the carotid bifurcation as a cause of
embolic stroke. Radiology 132:369-373, 1979.
6. Anderson CM, Haacke EM: Approaches to diagnostic magnetic
resonance carotid angiography. Semin Ultrasound CT MRI
7. Beckmann CF, Levin DC, Kubicka RA, et al: The effect of
sequential arterial stenosis on flow and pressure. Radiology
8. Litt AW, Eidelman EM, Pinto RS, et al: Diagnosis of carotid
artery stenosis: Comparison of 2DFT time-of-flight MR angiography
with contrast angiography in 50 patients. Am J Neuroradiol
9. Wasserman BA, Haacke EM, Debia LI: Carotid plaque formation
and its evaluation with angiography, ultrasound, and MR
angiography. JMRI 4:515-527, 1994.
10. Schuler JJ, Flanigan DP, Lim LT, et al: The effect of
carotid siphon stenosis on stroke rate, death, and relief of
symptoms following elective carotid endarterectomy. Surgery
11. Ross JS, Masaryk TJ, Ruggier PM: Magnetic resonance
angiography of the carotid bifurcation. Top Magn Reson Imag
12. Taylor DC: Current noninvasive diagnosis of carotid artery
stenosis: Indications and limitations. Can Soc Vasc Surg
13. Kramer J, Wimberger D, Haimberger K, et al: Stenosis of the
extracranial carotid artery. Magnetresonanz-lnstitut.
Universitatsklinik fur Radiodiagnostik, Wien. Wien Klin Wochenschr
14. Mittl RL Jr, Broderick M, Carpenter JP, et al:
Blinded-reader comparison of magnetic resonance angiography and
duplex-ultrasonography for carotid artery bifurcation stenosis.
Stroke 25:4-10, 1994.
15. Wesbey GE, Bergan JJ, Moreland SI, et al: Cerebrovascular
magnetic resonance angiography: A critical verification. J Vasc
Surg 16:619-632, 1992.
16. Blatter DD, Bahr AL, Parker DL, et al: Cervical carotid MR
angiography with multiple overlapping thin-slab acquisition:
Comparison with conventional angiography. Am J Roentgenol
17. Mattle HP, Kent KC, Edelman RR, et al: Evaluation of the
extracranial carotid arteries: Correlation of magnetic resonance
angiography, duplex ultrasonography, and conventional angiography.
J Vasc Surg 13:838-845, 1991.
18. Chikos PM, Fisher LD, Hirsch JH, et al: Observer variability
in evaluating carotid artery stenosis. Stroke 14:885-892, 1983.
19. Turnipseed WD, Kennell TW, Turski PA, et al: Combined use of
duplex imaging and magnetic resonance angiography for evaluation of
patients with symptomatic ipsilateral carotid stenosis. J Vasc Surg
20. De Marco JK, Nesbit GM, Wesbey GE, et al: Prospective
evaluation of extracranial carotid stenosis: MR angiography with
maximum-intensity projections and multiplanar reformation compared
with conventional angiography. AJR 163:1205-1212, 1994.
21. Anderson CM, Saloner D, Lee RE, et al: Assessment of carotid
artery stenosis by MR angiography: Comparison with x-ray
angiography and color-coded Doppler ultrasound. Am J Neuroradiol
22. Lustgarten JH, Solomon RA, Quest DO, et al: Carotid
endarterectomy after noninvasive evaluation by duplex
ultrasonography and magnetic resonance angiography. Neurosurgery
23. Levy C, Laissy JP, Raveau V, et al: Carotid and vertebral
artery dissections: Three-dimensional time-of-flight MR angiography
and MR imaging versus conventional angiography. Radiology
24. Haynes B, Thorpe K, Taylor W, et al: Poor performance of
ultrasound in detecting high-grade carotid stenosis (abstract). Can
J Surg 35:446, 1992.
25. Kido DK, Barsotti JB, Rice LZ, et al: Evaluation of the
carotid artery bifurcation: Comparison of magnetic resonance
angiography and digital subtraction arch aortography.
Neuroradiology 33:48-51, 1991.
26. Heiserman JE, Drayer BP, Fram EK, et al: Carotid artery
stenosis: Clinical efficacy of two-dimensional time-of-flight MR
angiography. Radiology 182:761-768, 1992.
27. Huston J III, Lewis BD, Wiebers DO, et al: Carotid artery:
Prospective blinded comparison of two-dimensional time-of-flight MR
angiography with conventional angiography and duplex US. Radiology
28. Hobson RW, Weiss DG, Fields WS: Efficacy of carotid
endarterectomy for asymptomatic carotid stenosis. N Engl J Med
29. Vool TJ, Heinzinger K, Juergens M, et al: Multiple-slab MR
angiography of the A. carotis interna, a preoperative comparative
study: Strahlenklinik und Poliklinik, Universitatsklinikum Rudolf
Virchow, FU Berlin, Rofo Fortschr Geb Rontgenstr Neuen Bildgeb
Verfahr 162 (5):404-411, 1995.
30. Laster RE Jr, Acker JD, Halford HH III, et al: Assessment of
MR angiography versus arteriography for evaluation of cervical
carotid bifurcation disease. Am J Neuroradiol 14:681-688, 1993.
31. Mani R, Eisenberg R, McDonald E, et al: Complication of
catheter cerebral arteriography: Analysis of 5,000 procedures. Am J
Roentgenol 131:861-865, 1978.
32. Nauven Bui L, Brant-Zawadzki M, Verghese P, et al: Magnetic
resonance angiography of cervicocranial dissection. Stroke
33. White RD, Obuchowski NA, Van Dyke CW, et al: Thoracic aortic
disease: Evaluation using a single MRA volume series. J Comput
Assist Tomogr 186(6):843-854, 1994.
34. Roederer GO, Langlois YE, Chan ARW, et al: Is siphon disease
important in predicting outcome of carotid endarterectomy? Arch
Surg 118:1177-1181, 1983.
35. Geuder JW, Lamparello PJ, Riles TS, et al: Is duplex
scanning sufficient evaluation before carotid endarterectomy? J
Vasc Surg 9:193-201, 1989.
36. Goodson SF, Flanigan B, Bishara RA, et al: Can duplex
scanning supplant arteriography in patients with focal carotid
territory symptoms? J Vasc Surg 5:551-557, 1987.
37. Kenagy JW: Comparison of duplex scanning and contrast
arteriography: A community hospital experience. J Vasc Surg
38. Jacobs NM, Grant EG, Scheillinger D, et al: Duplex carotid
sonography: Criteria for stenosis, accuracy and pitfalls. Radiology
39. Ricotta JJ, Holen J, Schenk E, et al: Is routine angiography
necessary prior to carotid artery endarterectomy? J Vasc Surg
40. Pan XM, Anderson CM, Reilly LM, et al: Magnetic resonance
angiography of the carotid artery combining two- and
three-dimensional acquisitions. J Vasc Surg 16:609-618, 1992.
41. Marks MP: Computed tomography angiography. Neuroimag Clin
North Am 6(4):899-909, 1996.
42. De Marco KJ et al: Can noninvasive studies replace
conventional angiography in the preoperative evaluation of carotid
stenosis? Neuroimag Clin North Am 6(4):911-929, 1996.