Echo-Enhanced Transcranial Color-Coded Duplex

ORIGINAL
Sonography in the Diagnosis of Cerebrovascular
RESEARCH Events: A Validation Study
A. Kunz BACKGROUND AND PURPOSE: Transcranial color-coded duplex sonography (TCCD) is a diagnostic
G. Hahn technique for evaluation of intracranial arteries in patients with acute stroke. Echo-enhancing contrast
agents (EEAs) are necessary to visualize intracranial vessels in up to 30% of patients because of
D. Mucha
limited acoustic bone windows. In this study, we assessed the diagnostic efficacy of echo-enhanced
A. Müller TCCD (eTCCD) in correlation with the gold standard, digital subtraction angiography (DSA).
K.M. Barrett
METHODS: We prospectively evaluated all patients with eTCCD who subsequently underwent DSA for
R. von Kummer evaluation of cerebrovascular symptoms over a 24-month period. We administered Levovist as an
G. Gahn EEA. Two blinded reviewers analyzed all eTCCD findings and correlated them with DSA.

RESULTS: We included 132 consecutive patients (40 women, 92 men; mean age, 58 ⫾ 14 years) with
164 datasets: 24/164 had normal findings, 98/164 had abnormalities of extracranial carotid arteries,
32/164 had abnormalities of intracranial arteries, and 21/164 had abnormalities in vertebrobasilar
circulation as determined by DSA. For eTCCD, we found a sensitivity of 82% (95% confidence interval
[CI]: 75%–90%), a specificity of 98% (95% CI: 90%–100%), a positive predictive value of 99% (95%
CI: 94%–100%), and a negative predictive value of 75% (95% CI: 64%– 85%); 7/164 (4%) examina-
tions were inconclusive because of insufficient bone windows. The interobserver agreement was
almost perfect (␬ value, 0.92; 95% CI: 0.87– 0.97).
CONCLUSION: eTCCD provides high diagnostic validity for the status of the major intracranial arteries.
In particular, a normal vessel status reliably assessed by an experienced sonographer could supersede
further imaging procedures. In patients with acute ischemic stroke not eligible for established angio-
graphic techniques, eTCCD may be useful as an alternative imaging technique.

D uring the last decade, transcranial color-coded duplex

sonography (TCCD) has been established as an alterna-
tive diagnostic technique for evaluation of the basal intracra-
color-coded duplex examination of the extracranial carotid system
and the vertebral arteries as well as a transcranial Doppler sonography
(TCD) of the major arteries of the circle of Willis and of the vertebro-
nial arteries. TCCD may help direct thrombolytic therapy in basilar arteries. This imaging is part of a routine stroke work-up for all
patients with acute ischemic stroke who are not eligible for patients admitted to the stroke unit of our institution. The 2 physi-
standard imaging techniques such as MR angiography (MRA) cians (A.K., G.H.) who performed all eTCCD examinations for this
or CT angiography (CTA).1 A major limitation of TCCD is the study were not aware of the results of this routine stroke work-up and
absorption and dispersion of the sonographic signal intensity had only been informed if a patient was assigned to DSA. Before
by the bony skull. Evaluation of the major arteries of the circle evaluation, all patients provided informed consent for participation
of Willis is not possible in up to 30% of patients because of in the study.
insufficient transtemporal bone windows.2-6
In the present study, we sought to determine the diagnostic eTCCD
validity of echo-enhanced TCCD (eTCCD) as a routine tech- The eTCCD examinations were performed before DSA by 2 experi-
nique in a large number of patients. Digital subtraction an- enced physicians (A.K., G.H.). A complete eTCCD examination in-
giography (DSA) was used as the gold standard correlative cluded evaluation of the anterior and vertebrobasilar circulations.
imaging procedure. The transtemporal approach was used to assess the anterior circula-
tion.7 The major arteries in the vertebrobasilar circulation were in-
Materials and Methods
sonated by using the suboccipital access through the foramen mag-
We prospectively included all patients admitted to our neurology
num.8 We used SH U 508A (Levovist, Schering, Berlin, Germany), an
department for evaluation of cerebrovascular symptoms who under-
intravenous galactose and palmitic acid– based echo-enhancing con-
went DSA. Patients with an acute stroke qualifying for thrombolytic
trast agent (EEA) for our study. Before EEA administration, an unen-
therapy were not included in this study during their acute phases.
hanced TCCD was performed to identify the optimal acoustic bone
Before the eTCCD examination, all included patients had a complete
windows. For the complete eTCCD examination, 2 doses of Levovist
(300 mg/mL, 4 g per dose) were given intravenously via an antecubital
Received September 11, 2005; accepted after revision January 25, 2006. access. The EEA was delivered continuously at an infusion rate of 900
From the Departments of Neurology (A.K., G.H., G.G.) and Neuroradiology (D.M., A.M., mg/min by using an infusion pump (Medrad Pulsar Sonography In-
R.v.K.), University of Technology Dresden, Dresden, Germany; and the Department of
jection System, Medrad, Indianola, Pa).
Neurology (K.M.B.), Mayo Graduate School of Medicine, Mayo Clinic, Jacksonville, Fla.
The following arterial branches were assessed bilaterally via the
A. Kunz and G. Hahn contributed equally to this study.
transtemporal bone windows: C1 segments of the internal carotid
Please address correspondence to: Georg Gahn, MD, Department of Neurology, University
of Technology Dresden, Fetscher Str 74, 01307 Dresden, Germany; e-mail: gahn@ arteries (ICAs); A1 and A2 segments of the anterior cerebral arteries
rcs.urz.tu-dresden.de (ACAs); M1 and M2 segments of the middle cerebral arteries (MCAs);

2122 Kunz 兩 AJNR 27 兩 Nov-Dec 2006 兩 www.ajnr.org

and P1 and P2 segments of the posterior cerebral arteries (PCAs), the vertebrobasilar circulation. Each examination was categorized to all
anterior communicating arteries (AComAs), and the posterior com- eligible subgroups. For all subgroups, the validating criteria were
municating arteries (PComAs). Using the suboccipital approach, we recalculated.
evaluated the V4 segments of both the vertebral and the basilar arter- 2) Validation of eTCCD by Analyzing the Distinct Intracranial
ies. The assessed arterial segments were found to be either normal or Arterial Segments. Another objective of this study was to assess the
abnormal according to the Doppler-flow patterns. Criteria for abnor- validity of eTCCD to detect abnormalities in distinct arterial segments
malities were increased mean flow velocity, retrograde flow, vessel- of the intracranial circulation. Therefore, both reviewers separately
occlusion signals, or turbulent flow patterns. assessed the status of all major arterial segments visualized by eTCCD.
The sonographic examinations were performed with an Acuson After comparing these results with the DSA findings, we calculated
128XP/4 (Siemens, Berlin and Munich, Germany) equipped with a the sensitivity, specificity, positive predictive value, and negative pre-
2.0/2.5-MHz transcranial sector scanner device. All eTCCD examina- dictive value for all arterial segments. If an arterial segment was not
tions were performed by using B-mode, pulsed-wave Doppler, and examined either by eTCCD or by DSA, it was classified as “not done.”
color velocity Doppler imaging mode. We used a standard insonation In cases in which the arterial segment was visualized by DSA but not
depth of 100 mm for the transtemporal approach and a depth of 120 by eTCCD, it was classified as “not visible.”
mm for the suboccipital approach. The focus of the sample volume
was adjusted to the appropriate depths of the individual arterial seg- Results
ments. During administration of Levovist, the pulse repetition fre- During the 24-month study period, 132 consecutive patients
quency was increased to minimize aliasing and blooming artifacts. All were enrolled (40 women, 92 men; mean age, 58 ⫾ 14 years;
eTCCD examinations were stored on super-VHS videotape for later range, 19 – 86 years). For geographic reasons, all included pa-
off-line analysis. tients were white; 121/132 patients (92%) underwent evalua-
tion of the anterior circulation and 99/132 patients (75%) un-
DSA derwent evaluation of the vertebrobasilar circulation by
All patients underwent DSA as the correlative imaging procedure. eTCCD. Eighty-eight patients had evaluation of both anterior
DSA was performed and interpreted by 3 experienced neuroradiolo- and posterior circulations by eTCCD. During the study pe-
gists (R.v.K., D.M., A.M.) blinded to the results of the eTCCD exam- riod, no patient who underwent DSA at our institution was
inations. When possible, all patients underwent 4-vessel DSA. Crite- excluded from the study and no patient refused participation.
ria for abnormalities in DSA were stenoses of the extracranial arteries No serious side effects (ie, anaphylactic reactions) were ob-
of at least 50% or any other pathologies that were found served after administration of Levovist. The mean time inter-
intracranially. val between the eTCCD and DSA examination was 3 ⫾ 3 days
(median, 1.5 days). One hundred four of 132 patients under-

BRAIN
Data Analysis and Statistics went DSA and eTCCD evaluation once, and 26/132 patients
All eTCCD examinations were separately evaluated by 2 experienced were evaluated twice (24 following unilateral percutaneous
reviewers (A.K., G.H.). At the time of evaluation, they were blinded to transluminal stent protected angioplasty of ICA stenosis and 2
the patient’s clinical diagnosis and the results of the extracranial du- following embolization of an arteriovenous fistula). Another

ORIGINAL RESEARCH
plex and TCD studies and the DSA. To correlate the findings of the patient was evaluated 3 times because of bilateral ICA stent-
eTCCD examinations with the DSA results, we used 2 different protected angioplasty in 2 successive DSA procedures. This
paradigms: patient was examined before and after each intervention. Fi-
1) Diagnosis-Based Validation of eTCCD. On the basis of the nally, 1 patient with a unilateral symptomatic high-grade
analysis of the color-coded flow pattern and the Doppler spectra in all MCA stenosis was evaluated 5 times. The lesion was not ame-
arterial segments that could be analyzed by eTCCD, each reviewer nable to stent implantation and, therefore, was treated with
created a distinct eTCCD diagnosis of the intracranial arterial status angioplasty several times. Overall, we assessed 164 datasets
for each eTCCD examination. According to the corresponding DSA consisting of an eTCCD and the correlative DSA-examination.
results, the eTCCD diagnoses were classified as either true-negative, Subjective improvement in visualization of the intracranial
true-positive, false-negative, or false-positive. If the bone window was arteries was noted during the administration of Levovist in
insufficient to allow a conclusive diagnosis, the examination was clas- most patients (data not shown).
sified as not evaluable. On the basis of this categorization, we calcu- For the assessment of neurologic deficits of the patients
lated the sensitivities, specificities, and positive and negative predic- included, we used the National Institutes of Health Stroke
tive values as the validating criteria for each reviewer’s evaluation. Scale (NIHSS) score. The NIHSS score ranged between 0 and
Additionally, we determined the interobserver agreement between 19. Of all 132 patients, 119 had a NIHSS score of less than 8,
the 2 reviewers by using kappa statistics. Agreement was considered which translates to a slight-to-moderate functional deficit.
moderate if ␬ was between 0.41 and 0.6, substantial if ␬ was between Eight patients were very moderately (NIHSS score, 8 –11)
0.61 and 0.8, and almost perfect if ␬ was between 0.81 and 1.0.9,10 compromised; 7 of them had pathologies in the extracranial
Thereafter, the discordant eTCCD examinations were re-evaluated in parts of the carotids. Five patients had more severe neurologic
a second consensus review, and the validating criteria were deficits on admission (NIHSS score, ⬎ 11). Among these pa-
recalculated. tients, 3 had pathologies in the intracranial cerebral arteries
To further specify the validity of eTCCD in relation to the location (Table 1).
of cerebrovascular abnormalities, we categorized all examinations ac- Table 2 shows the clinical spectrum of the patients in-
cording to the DSA diagnosis into the following 4 subgroups: normal cluded. Most patients (102/132) had cerebral ischemia or pre-
findings, abnormalities in the extracranial anterior circulation, ab- sented with reversible ischemic stroke symptoms. Eight pa-
normalities in the intracranial circulation, and abnormalities in the tients presented with an asymptomatic critical ICA stenosis,

AJNR Am J Neuroradiol 27:2122–27 兩 Nov-Dec 2006 兩 www.ajnr.org 2123

 Table 1: Distribution of the clinical status (assessed by NIHSS on admission) of all patients enrolled and of the patients in different
subgroups with pathologic finding
Pathologies on DSA
NIHSS on All Patients No Extracranial Intracranial VA
admission (n ⫽ 132) (n ⫽ 24) (n ⫽ 72) (n ⫽ 27) (n ⫽ 21)
⬍4 79 16 43 15 11
4–7 40 6 21 8 9
8–11 8 1 7 1 0
ⱖ12 5 1 1 3 1
Note:—NIHSS indicates National Institutes of Health Stroke Scale score; DSA, digital subtraction angiography; VB, vertebrobasilar. Numbers of patients qualifying for the defined NIHSS
ranges are shown.

Table 2: Clinical spectrum of the patients enrolled Table 3: Diagnosis-based validation of all eTCCD-examinations
(n ⴝ 164) in correlation to DSA
n
Embolic stroke 55 Reviewer 1 Reviewer 2 Consensus
Transient ischemic attack 19 NE 4 (1–7) 4 (1–7) 4 (1–7)
Low flow stroke 12 Sensitivity 82 (75–90) 80 (72–88) 82 (75–90)
Asymptomatic ICA stenosis 8 Specificity 96 (88–100) 98 (90–100) 98 (90–100)
Dissection 7 PPV 98 (92–100) 99 (93–100) 99 (94–100)
Intracerebral hemorrhage 7 NPV 75 (63–84) 73 (62–83) 75 (64–85)
Aneurysm 4 Note:—PPV indicates positive predictive value; NPV, negative predictive value; NE, not
Vascular malformation 4 evaluable; eTCCD, echo-enhanced transcranial color-coded duplex sonography. All values
are shown in percent. Numbers in parentheses represent 95% confidence intervals.
Stenosis/occlusion of basilar artery 4
Subarachnoid hemorrhage 3
Subclavian steal 3 Table 4: Agreement between reviewers 1 and 2 for all
Lacunar stroke 2 eTCCD-examinations (n ⴝ 164) using ␬-statistics
Other diagnoses 5 Reviewer 2
Note:—ICA indicates internal carotid artery.
Reviewer 1 NE FN FP TN TP Total
NE 6 0 1 0 0 7
for elective percutaneous transluminal stent-protected angio- FN 0 17 0 0 1 18
plasty. One patient with a subarachnoid hemorrhage due to an FP 0 0 0 2 0 2
TN 1 0 0 53 0 54
aneurysm at the bifurcation of the right M1 segment was as-
TP 0 3 0 0 80 83
signed to both clinical entities. Five patients underwent DSA Total 7 20 1 55 81 164
to rule out cerebrovascular origin of neurologic symptoms. Note:—NE indicates not evaluable; FN, false negative; FP, false positive; TN, true
Of all 164 DSA examinations, 24 (15%) had no abnormal- negative; TP, true positive; eTCCD, echo-enhanced transcranial color-coded duplex sonog-
raphy. The ␬-value is 0.92 (95%-confidence interval: 0.87– 0.97).
ities, 98 (60%) had abnormalities of the extracranial carotid
arteries, 32 (20%) had abnormalities in the intracranial circu-
lation, and 21 (13%) had abnormalities in the vertebrobasilar occlusion, 2 patients with a vertebrobasilar arteriovenous fis-
circulation. Among the patients with abnormalities of the ex- tula, and 3 patients with a stenosis of the subclavian artery and
tracranial carotid arteries, most had unilateral (46 patients) or steal phenomenon. A large number of patients included in our
bilateral (18 patients) ICA stenoses or occlusions. Also repre- study had multiple pathologies of the brain-supplying arteries.
sented in this subgroup were unilateral ICA dissections (4 pa- In summary, a larger part of patients (64/132) included in
tients), ICA aneurysms (2 patients), and common carotid ar- our study underwent DSA for the evaluation of stenoses or
tery occlusion (1 patient). Additionally, 1 patient with stenosis occlusions of the carotid arteries and subsequently underwent
of the brachiocephalic trunk was included in this subgroup. In percutaneous transluminal stent-protected angioplasty.
the subgroup with abnormalities of the intracranial circula- Diagnosis-Based Validation of eTCCD. The sensitivities,
tion, most patients had stenoses or occlusions of intracranial specificities, positive predictive values, and negative predictive
arteries, with the MCA being the most commonly affected values are shown in Table 3. Both reviewers were unable to
vessels (15 patients). Other locations of intracranial stenoses reach a conclusive diagnosis in 7/164 examinations (4%) be-
or occlusions were the intracranial ICA (6 patients), the ACA cause of insufficient bone windows. The mean age of those 7
(2 patients), and the PCA (1 patient). One patient had a uni- patients (3 women, 4 men) was 71 ⫾ 8 years, and they were
lateral aneurysm located at the bifurcation of the right M1 significantly older (P ⬍ .001) than the other patients. Assess-
segment, 1 patient had an aneurysm of the AComA, and 2 ment of the interobserver agreement of all 164 eTCCD exam-
patients had an arteriovenous malformation. Another 2 pa- inations revealed 8 discordant results. This refers to a ␬ value
tients presented with vasospasms following subarachnoid of 0.92 (Table 4). Because of the almost perfect agreement of
hemorrhage. The subgroup of abnormalities in the vertebro- the 2 reviewers’ results, the subsequently presented results re-
basilar circulation consisted mostly of patients with stenoses fer to the datasets obtained at the time of consensus review.
or occlusions of the extra- or intracranial vertebral arteries (14 Table 5 summarizes the validation data obtained after sub-
patients) as well as the basilar arteries (4 patients). Addition- group analysis. In the subgroup of normal findings by DSA
ally, this subgroup included 3 patients with a vertebral artery (n ⫽ 24), all patients were assessed as normal by eTCCD (spec-
dissection, 1 patient with posterior inferior cerebellar artery ificity,100%; negative predictive value, 100%). The sensitivity,

2124 Kunz 兩 AJNR 27 兩 Nov-Dec 2006 兩 www.ajnr.org

 Table 5: Diagnosis-based validation of eTCCD subgroups in correlation to DSA after consensus review
Findings Abnormalities
Normal Abnormal Extracranial Intracranial VB
(n ⫽ 24) (n ⫽ 140) (n ⫽ 98) (n ⫽ 32) (n ⫽ 21)
N.E. 4 (0–21) 4 (1–8) 6 (2–13) 0 (0–11) 0 (0–16)
Sensitivity N/A 82 (74–89) 83 (71–92) 81 (64–93) 76 (53–92)
Specificity 100 (85–100) 97 (84–100) 97 (84–100) N/A N/A
PPV N/A 99 (94–100) 98 (89–100) 100 (87–100) 100 (79–100)
NPV 100 (85–100) 64 (49–77) 76 (61–88) 0 (0–46) 0 (0–52)
Note:—VB indicates vertebrobasilar; NE, not evaluable; PPV, positive predictive value; NPV, negative predictive value; N/A, not available; eTCCD, echo-enhanced transcranial color-coded
duplex sonography; DSA, digital subtraction angiography. All values are shown in percent. Numbers in parentheses represent 95% confidence intervals.

Table 6: Visualization of the distinct arterial segments evaluable by Table 7: Validation of eTCCD in comparison with DSA by analyzing
eTCCD according to the separate assessment of the reviewers and the sensitivities, specificities, and positive and negative predictive
after consensus review values of the distinct arterial segments
Segment Reviewer 1 Reviewer 2 Consensus Segment Sensitivity Specificity PPV NPV
V4 99 99 99 M1 87 (81–94) 98 (96–99) 92 (87–97) 96 (94–98)
M1 93 94 93 M2 76 (62–87) 100 (100–100) 100 (91–100) 95 (92–97)
BA 94 88 91 C1 84 (75–91) 99 (99–100) 97 (90–100) 96 (94–98)
P1 88 91 90 A1 77 (71–84) 97 (94–99) 94 (91–98) 85 (81–89)
P2 87 91 89 A2 70 (35–93) 100 (100–100) 100 (59–100) 99 (97–100)
C1 84 83 83 P1 74 (64–83) 99 (98–100) 96 (88–99) 92 (89–95)
A1 84 81 82 P2 77 (64–88) 99 (98–100) 93 (81–99) 96 (94–98)
M2 61 50 55 PComA 91 (81–97) 79 (68–88) 78 (66–87) 92 (82–97)
A2 48 42 45 AComA 94 (86–98) 86 (42–100) 98 (92–100) 60 (26–88)
AComA 27 24 26 V4 78 (64–89) 99 (97–100) 92 (79–98) 96 (93–98)
PComA 23 25 24 BA 76 (50–93) 99 (97–100) 87 (60–98) 97 (94–100)
Note:—V4 indicates V4-segment of the vertebral artery; M1/M2, M1-/M2-segment of the Note:—PPV indicates positive predictive value; NPV, negative predictive value; PComA,
middle cerebral artery; BA, basilar artery; P1/P2, P1-/P2-segment of the posterior cerebral posterior communicating artery; AComA, anterior communicating artery; M1/M2, M1-/M2-
artery; C1–C1-segment of the internal carotid artery; A1/A2, A1-/A2-segment of the segment of the middle cerebral artery; C1, C1-segment of the internal carotid artery; A1/A2,
anterior cerebral arteries; AComA, anterior communicating artery; PComA, posterior com- A1-/A2-segment of the anterior cerebral arteries; P1/P2, P1-/P2-segment of the posterior
municating artery; eTCCD, echo-enhanced transcranial color-coded duplex sonography. All cerebral artery; V4, V4-segment of the vertebral artery; BA, basilar artery; eTCCD,
values are shown as percentages. echo-enhanced transcranial color-coded duplex sonography; DSA, digital subtraction an-
giography. The displayed values refer to the datasets obtained by both reviewers. All
specificity, positive predictive value, and negative predictive values are shown as percentages. Numbers in parentheses represent 95% confidence
intervals.
value in the subgroup of all extracranial pathologies as well as
the sensitivity and positive predictive value in the subgroup of
all intracranial pathologies are similar as to the results ob- ischemic symptoms ascribed to cerebral arteries. Several
tained for all examinations. The sensitivity in the subgroup of groups have evaluated the diagnostic reliability of unenhanced
all vertebrobasilar pathologies is slightly lower compared with as well as eTCCD in correlation with DSA, MRA, and
that of all examinations. CTA.2,3,5,6,11-17 Among these different techniques, DSA re-
Validation of eTCCD by Segmental Intracranial Artery mains the gold standard and allows the most precise reproduc-
Analysis. The percentages of visualization of the distinct arte- tion of the cerebral circulation in comparison with the real
rial segments by eTCCD are shown in Table 6. The V4 seg- anatomy of the cerebral blood vessels. To our knowledge, this
ments of the vertebral arteries, the M1 segments of the MCA, is the largest prospective study validating eTCCD findings
and the proximal part of the basilar artery could be visualized with DSA.
with the highest certainty. The basilar artery was visualized The improvement of the diagnostic potential of TCCD by
through the suboccipital transforaminal bone window to a the use of EEA has been widely accepted.3,5,6,11,16,17 In our
mean depth of 97 ⫾ 15 mm. study, we performed eTCCD examination exclusively. The ra-
Table 7 summarizes the validation of those arterial seg- tionale for this approach was the potential time delay, which
ments usually visualized with eTCCD. The displayed values might occur through an initial unenhanced TCCD examina-
refer to the datasets obtained by both reviewers. Except for the tion. In patients with acute stroke symptoms, a fast, reliable,
PComA and AComA, excellent specificities (range, and accurate finding is fundamental for the optimal treat-
97%–100%) and high sensitivities (range, 70%– 87%) were ment. Additionally, patients with stroke at their typical ages
calculated for all other segments with the best results for the have limited acoustic bone windows. Considering the gain of
M1 and C1 segments. The positive predictive values ranged time through the use of EEA at the beginning and the relatively
between 92% and 100% except for the PComA (78%) and the low risk of EEA, we consider this approach reasonable and
basilar artery (87%). The negative predictive values also justified.
mostly exceeded the 90% values, with the exception of the A1 The preponderance of men in the present study (male/fe-
segments (85%) and the AComA (60%). male ratio, 2.3) reflects the fact that more men than women
underwent DSA at our institution during the study period.
Discussion This is consistent with the higher stroke incidence in white
The objective of this study was to assess the diagnostic validity men than in white women at that age (58 ⫾ 14 years), which
of eTCCD in patients admitted to our department because of we have obtained as the mean age of our study population.18 In

AJNR Am J Neuroradiol 27:2122–27 兩 Nov-Dec 2006 兩 www.ajnr.org 2125

addition, the imbalance toward men is attributable to the fact included patient. The sonographer’s knowledge of an ex-
that a larger part of our study population underwent DSA for tracranial stenosis or occlusion could certainly increase the
evaluation of critical ICA stenosis. The prevalence of carotid sensitivity to detect alterations in the intracranial blood flow.
stenosis is higher in men than in women.19 The relatively Another group of false-negative results was from abnor-
young mean age of our patients is explained by the fact that the malities in smaller branches of the intracerebral vasculature
DSA was exclusively performed with the intention of a subse- (ie, branches distal to the M2 segments, the posterior inferior
quent intervention. cerebellar artery, or the AComA). These segments were too
Most enrolled patients were admitted with acute ischemic small to be assessed by eTCCD. This hypothesis is strength-
stroke symptoms and had slight-to-moderate neurologic def- ened by the percentages of visualization for the distinct arterial
icits (90% of patients with NIHSS score ⬍ 8). This finding is of segments displayed in Table 6. The M2 segments, for example,
practical importance because this population of patients with can be evaluated in only 55% of all eTCCD examinations. For
acute ischemic stroke has the highest probability of having a branches distal to the M2 segments, this percentage probabil-
favorable outcome.20,21 This population of patients is the pri- ity might be lower.
mary target for acute stroke therapy. Therefore, we conclude that validation of eTCCD com-
By eTCCD, we found a conclusive diagnosis in 96% (157/ pared with DSA reveals an overall high sensitivity. Pitfalls for
164) of all eTCCD examinations. The remaining 7 patients the accurate evaluation of the intracranial vasculature by
had insufficient bone windows. It has been reported that the eTCCD could be an altered blood flow pattern due to critical
rate of successful recording of blood flow signals by transcra- extracranial ICA stenoses or abnormalities in smaller arterial
nial sonography decreases with advancing age.22 The percent- branches.
age of conclusive diagnoses in our study is obviously higher Table 6 demonstrates the dependence of visualization of
than that previously reported.6,11,16,17 In contrast to those the distinct intracranial arterial segments by eTCCD on their
studies, we administered Levovist to all patients regardless of diameters. The proximal segments of the MCA, ACA, and
their bone windows. However, it has to be considered that the PCA are visible with a higher certainty than their distal
sex ratio of our study population was imbalanced, with a branches (M2, A2, and P2 segments, respectively). Addition-
strong preponderance of men. Women have a higher preva- ally, for the MCA as well as the ACA, the overall sensitivities of
lence of insufficient acoustic bone windows than men.23 Also, the distal branches are slightly lower than those of the proxi-
the relatively young mean age of the patients included in our mal segments. Together with the higher visibilities, this sug-
study could have contributed to the low number of insuffi- gests a higher diagnostic validity of the main stems of the ma-
cient bone windows.24 jor cerebral arteries, in particular the MCA. These results have
Overall, validation of eTCCD diagnosis in comparison an important implication on the efficiency of eTCCD in clin-
with DSA revealed an excellent specificity as well as positive ical practice. The MCA is the most commonly affected artery
predictive value with only 1 false-positive diagnosis of 164 in ischemic stroke syndromes. Especially, embolic strokes oc-
examinations. Furthermore, when considering only patients cur mostly in the MCA territory. Massive hemispheric infarc-
with normal findings by DSA, we were able to identify all of tion due to acute occlusion of the proximal MCA and poor
these patients by eTCCD. We conclude that 1 domain of collateral flow carries a high mortality. The only potential
eTCCD could be a highly specific identification of a normal therapy for these patients is early administration of thrombol-
intracranial arterial status. Therefore, if an experienced sonog- ysis. Although the knowledge of the patency of the MCA is not
rapher detects no abnormalities by using eTCCD in a patient required for the initiation of thrombolytic therapy in acute
with sufficient bone windows, it is reasonable not to pursue ischemic stroke, the ability to identify clot in the proximal
further imaging procedures. MCA may help in deciding to pursue thrombolytic therapy in
With regard to the reliability of eTCCD to detect abnor- individual patients. Moreover, an experienced sonographer is
malities in the intracranial arterial vasculature, we obtained an able to diagnose an MCA occlusion by eTCCD within 5
overall high sensitivity of 82%. Herein, most false-negative minutes. Because of the necessity of performing a native com-
eTCCD examinations represented a modified intracranial puted tomography of the brain and evaluating the functional
blood flow pattern because of critical extracranial ICA steno- deficit in a patient with acute stroke before initiation of
ses or occlusions. In this context, one should also consider that thrombolytic therapy, an additional eTCCD examination
DSA may not necessarily reflect the true intracranial collateral does not necessarily delay the initiation of thrombolytic ther-
flow pattern under resting conditions. One characteristic in- apy. Therefore, we propose that eTCCD is an efficient primary
tracranial collateral flow pattern in patients with critical uni- as well as follow-up imaging procedure for patients with acute
lateral ICA stenosis is prolonged cardiac cycle–independent ischemic stroke syndromes referable to the MCA territory.
cross-filling of the A1 segment ipsilateral to the stenosis. It is Thus, eTCCD is a valuable alternative bedside technique ap-
conceivable, for example, that during the DSA procedure, in- plicable to patients who are not eligible for standard imaging
jection of contrast medium at a high flow rate into the con- techniques (eg, DSA, CTA, or MRA) because of either im-
tralateral distal ICA could induce a temporary cross flow paired renal function or poor cooperation.
through the ipsilateral A1 segment. This phenomenon could In the subgroup of patients with vertebrobasilar patholo-
occur in patients with already compromised but not yet re- gies, the sensitivity for eTCCD was slightly lower. The lower
versed blood flow through the A1 segment ipsilateral to the sensitivity may be partly due to technical limitations in follow-
stenosis. This would not have been recognized under resting ing the course of tortuous arteries through the suboccipital
conditions by eTCCD. Also, in our study, none of the review- window. Only a small percentage of patients included in our
ers had been aware of the extracranial duplex status of any study had vertebrobasilar abnormalities. Also, the sonogra-

2126 Kunz 兩 AJNR 27 兩 Nov-Dec 2006 兩 www.ajnr.org

phers performing this study used the eTCCD examinations 4. Seidel G, Kaps M, Gerriets T. Potential and limitations of transcranial color-
coded sonography in stroke patients. Stroke 1995;26:2061– 66
exclusively for evaluating the vertebrobasilar vessel status. Ad- 5. Zunker P, Wilms H, Brossmann J, et al. Echo contrast-enhanced transcranial
ditional knowledge of the hemodynamics in the extracranial ultrasound: frequency of use, diagnostic benefit, and validity of results com-
vertebral arteries assessed by extracranial duplex sonography pared with MRA. Stroke 2002;33:2600 –33
6. Gahn G, Gerber J, Hallmeyer S, et al. Contrast-enhanced transcranial color-
might improve the results. However, the objective of this study coded duplexsonography in stroke patients with limited bone windows. AJNR
was to assess the validity of eTCCD based exclusively on trans- Am J Neuroradiol 2000;21:509 –14
cranial and suboccipital duplex findings. Therefore, the pre- 7. Bogdahn U, Becker G, Winkler J, et al. Transcranial color-coded real-time
sonography in adults. Stroke 1990;21:1680 – 88
sented results of vertebrobasilar pathologies are preliminary. 8. Becker G, Lindner A, Bogdahn U. Imaging of the vertebrobasilar system by
The mean transforaminal insonation depth of 97 ⫾ 15 mm transcranial color-coded real-time sonography. J Ultrasound Med 1993;12:
enables eTCCD evaluation of only the proximal two thirds of 395– 401
9. Landis JR, Koch GG. The measurement of observer agreement for categorical
the basilar artery.25,26 This is a major limitation in patients data. Biometrics 1977;33:159 –74
with a suspected basilar artery occlusion. On the basis of the 10. Landis JR, Koch GG. An application of hierarchical kappa-type statistics in the
small population of patients with suboccipital abnormalities assessment of majority agreement among multiple observers. Biometrics 1977;
33:363–74
examined in this study, we conclude that in patients with a 11. Baumgartner RW, Arnold M, Gonner F, et al. Contrast-enhanced transcranial
clinical syndrome suggestive of a vertebrobasilar disease, color-coded duplex sonography in ischemic cerebrovascular disease. Stroke
1997;28:2473–78
eTCCD alone might not be sufficient and additional imaging 12. Baumgartner RW, Mattle HP, Schroth G. Assessment of ⬎/ⴝ50% and ⬍50%
procedures should be considered. intracranial stenoses by transcranial color-coded duplex sonography. Stroke
One potential shortcoming of our study is the fact that 1999;30:87–92
13. Gahn G, Hahn G, Hallmeyer-Elgner S, et al. Echo-enhanced transcranial color-
owing to the geographic location of our neurology depart- coded duplexsonography to study collateral blood flow in patients with symp-
ment, only whites were included. The existence of racial dif- tomatic obstructions of the internal carotid artery and limited acoustic bone
ferences in the temporal bone window thicknesses and its in- windows. Cerebrovasc Dis 2001;11:107–12
14. Gerriets T, Goertler M, Stolz E, et al. Feasibility and validity of transcranial
terference with the transtemporal insonation conditions are duplex sonography in patients with acute stroke. J Neurol Neurosurg Psychiatry
known.27 This obviously limits the generalizability of our 2002;73:17–20
study. Application of our study paradigm to a racially mixed 15. Kenton AR, Martin PJ, Abbott RJ, et al. Comparison of transcranial color-
coded sonography and magnetic resonance angiography in acute stroke.
study population would be valuable. Stroke 1997;28:1601– 06
16. Nabavi DG, Droste DW, Kemeny V, et al. Potential and limitations of echo-
Conclusion contrast-enhanced ultrasonography in acute stroke patients: a pilot study.
Stroke 1998;29:949 –54
In summary, our results provide convincing evidence to sup- 17. Postert T, Braun DW, Meves S, et al. Contrast-enhanced transcranial color-
port use of eTCCD as a highly efficient imaging technique to coded sonography in acute hemispheric brain infarction. Stroke 1999;30:
1819 –26
evaluate the status of the major intracranial cerebral arteries. 18. Howard G, Howard HV. Distribution of stroke: heterogeneity of stroke by age,
We have demonstrated an excellent specificity in the identifi- race, and sex. In: Mohr P, Choi DW, Grotta JC, eds. Stroke: Pathophysiology,
cation of a normal intracranial arterial status. Our data suggest Diagnosis, and Management. Philadelphia: Churchill Livingstone; 2004:3–12
19. Mathiesen EB, Joakimsen O, Bonaa KH. Prevalence of and risk factors associ-
that if an experienced sonographer detects no abnormalities ated with carotid artery stenosis: the Tromso Study. Cerebrovasc Dis 2001;12:
by using eTCCD in a patient with sufficient bone windows, it 44 –51
is reasonable to forego further imaging procedures. In addi- 20. Adams HP Jr, Davis PH, Leira EC, et al. Baseline NIH Stroke Scale score
strongly predicts outcome after stroke: a report of the Trial of Org 10172 in
tion, we observed a high reliability in the diagnosis of proximal Acute Stroke Treatment (TOAST). Neurology 1999;53:126 –31
MCA pathologies. Furthermore, this procedure is minimally 21. Adams HP Jr, Leclerc KC, Bluhmki E, et al. Measuring outcomes as a function
invasive and easily applicable at the bedside; these features of baseline severity of ischemic stroke. Cerebrovasc Dis 2004;18:124 –29
22. Itoh T, Matsumoto M, Handa N, et al. Rate of successful recording of blood
make eTCCD a useful diagnostic tool. flow signals in the middle cerebral artery using transcranial Doppler sonog-
raphy. Stroke 1993;24:1192–95
23. Marinoni M, Ginanneschi A, Forleo P, et al. Technical limits in transcranial
Acknowledgments Doppler recording: inadequate acoustic windows. Ultrasound Med Biol 1997;
We thank Rainer Koch, PhD (Institute for Medical Biomet- 23:1275–77
24. Hashimoto H, Etani H, Naka M, et al. Assessment of the rate of successful
rics, University of Technology Dresden) for statistical advice. transcranial Doppler recording through the temporal windows in Japanese
with special reference to aging and sex. Nippon Ronen Igakkai Zasshi 199;29:
119 –22
References 25. Schulte-Altedorneburg G, Droste BW, Popa V, et al. Visualization of the basilar
1. Gahn G, von Kummer R. Ultrasound in acute stroke: a review. Neuroradiology artery by transcranial color-coded duplex sonography: comparison with
2001;43:702–11 postmortem results. Stroke 2000;31:1123–27
2. Baumgartner RW, Baumgartner I, Mattle HP, et al. Transcranial color-coded 26. Brunner-Beeg F, von Reutern GM. Color duplex ultrasonography of the intra-
duplex sonography in the evaluation of collateral flow through the circle of cranial vertebro-basilar system: improved imaging by echo signal enhance-
Willis. AJNR Am J Neuroradiol 1997;18:127–33 ment [in German]. Ultraschall Med 1999;20:83– 86
3. Gerriets T, Seidel G, Fiss I, et al. Contrast-enhanced transcranial color-coded 27. Halsey JH. Effect of emitted power on waveform intensity in transcranial
duplex sonography: efficiency and validity. Neurology 1999;52:1133–37 Doppler. Stroke 1990;21:1573–78

AJNR Am J Neuroradiol 27:2122–27 兩 Nov-Dec 2006 兩 www.ajnr.org 2127