, Volume 55, Issue 2, pp 179–185

Intracranial carotid artery disease in patients with recent neurological symptoms: high prevalence on CTA


    • Department of RadiologyAMC
    • Department of Biomedical Engineering & PhysicsAMC
  • Paul J. Nederkoorn
    • Department of NeurologyAMC
  • Leslie Bleeker
    • Department of RadiologyAMC
  • René van den Berg
    • Department of RadiologyAMC
  • Charles B. Majoie
    • Department of RadiologyAMC
Diagnostic Neuroradiology

DOI: 10.1007/s00234-012-1097-6

Cite this article as:
Marquering, H.A., Nederkoorn, P.J., Bleeker, L. et al. Neuroradiology (2013) 55: 179. doi:10.1007/s00234-012-1097-6



Novel postprocessing techniques have enabled accurate quantification of intracranial carotid atherosclerotic disease on CT Angiography (CTA). Our purpose was to estimate the prevalence of intracranial carotid artery disease, i.e., stenosis and calcium, on CTA in patients with recent neurological symptoms.


The degree of stenosis and calcium volume of 162 extracranial and intracranial internal carotid arteries (ICAs) was quantitatively measured on CTA images of 88 consecutive patients with recent neurological symptoms and extracranial ICA stenosis as screened by ultrasound. The prevalence of intracranial ICA stenosis and presence of calcium was estimated and correlated with extracranial ICA stenosis.


Intracranial ICA stenosis was observed in 83 % (95 %CI: 77–89 %) and 39 % (95 %CI: 31–47 %) for a stenosis of ≥30 % and ≥50 %, respectively. Only on the symptomatic side, a statistical significant correlation between intracranial and extracranial stenoses was observed (Pearson's r 0.32, P = 0.006). In the 37 arteries with an extracranial ICA stenosis of ≥70 %, 89 % (95 %CI: 79–99 %) and 46 % (95 %CI: 30–62 %) of the intracranial ICA showed a stenosis of ≥30 % and ≥50 %, respectively.


In our population of patients with recent neurological symptoms and extracranial stenosis as screened by ultrasound, CTA imaging resulted in a substantially higher prevalence of intracranial ICA disease than previously reported. This remarkably high prevalence of intracranial ICA disease on CTA may have important future implications for acute and preventive treatment strategies.


Intracranial arteryStenosisCTAPrevalenceAcute ischemic stroke


Carotid endarterectomy in patients with severe stenosis of the extracranial internal carotid artery (ICA) has been proven to be effective in reducing risk of recurrent stroke in large randomized trials [13]. According to current guidelines, the clinical decision for carotid endarterectomy is largely based on the degree of extracranial ICA stenosis. It has been shown that intracranial atherosclerotic disease is also an independent risk factor for recurrent stroke; however, to date, this information is commonly not included in treatment decision [12]. It has been shown that a third of the patients with a symptomatic extracranial ICA stenosis also have intracranial lesions, predominantly in the ICA [12]. This was confirmed by Takeuchi et al. [18] who showed that the carotid siphon is a predilection site for early atherosclerotic lesions.

In 1968, Hass et al. [9] already identified intracranial stenosis in 22.6 % of patients evaluated with digital subtraction angiography (DSA). For a long period, DSA was the only available diagnostic test for intracranial stenosis. In current clinical practice, non- or minimally invasive techniques such as CTA and Magnetic Resonance Angiography (MRA) are increasingly used in the workup of patients suspected of stroke. CTA has the ability to image the full 3D morphology of the whole intracranial vasculature and it is accurate in visualizing calcified plaques in the vessel wall, which may be used in the pathophysiological interpretation of intracranial artery disease.

It is, however, unclear, and should be investigated, if previously reported prevalences of intracranial stenoses diagnosed with transcranial Doppler (TCD) and DSA can be extrapolated to CTA-based studies. Recently, a high prevalence of 82 % of intracranial carotid calcium in noncontrast CT was reported [5]. Studies are needed to explore validation of these image findings, implications for prognosis, and new approaches for management of patients with recent neurological symptoms, particularly those at high risk for recurrent stroke.

The purpose of this study was to estimate the prevalence of intracranial ICA stenosis as expressed in CTA images in a population of patients with recent neurological symptoms and with a stenosis of the extracranial ICA as screened by duplex ultrasound (DUS). In addition, we determined the prevalence of calcifications and investigated correlations between intracranial and extracranial disease.

Materials and methods

Study population

In our center, the Academic Medical Center Amsterdam, patients with recent neurological symptoms Transient Ischemic Attack (TIA or stroke) suspected of having ICA stenosis are primarily screened by DUS. Using sensitive margins, the cutoff degree of extracranial stenosis for additional CTA imaging to precisely estimate the degree of stenosis is set at 30 % for men and 50 % for women, according to current guidelines [17]. These patients underwent a CT scan on either a 64-slice or four-slice scanner. All consecutive patients that underwent a 64-slice CTA scan for the evaluation of extracranial ICA stenosis between April 1, 2006 and December 31, 2008 were included in this study. Figure 1 shows a flow chart illustrating the inclusion and exclusion of patients and ICAs.
Fig. 1

Flow chart illustrating the inclusion and exclusion of patients and ICAs

Permission of the medical ethics committee was given for this retrospective analysis of anonymous patient data. No informed consent was required because no diagnostic tests other than routine clinical imaging were used in this study. Because evaluation of the images for the purpose of the current study was performed retrospectively, it could not influence clinical decisions.

Patients with a previous carotid intervention (carotid endarterectomy or stenting) and patients with CTA studies of insufficient quality were excluded. Furthermore, we excluded the images of arteries with an occlusion of the extracranial ICA and arteries with an unreliable reference diameter, precluding a valid stenosis measurement.

We classified the ipsilateral ICA of the symptomatic hemisphere as ‘symptomatic artery’ and the contralateral ICA as the ‘asymptomatic artery.’ In patients with symptoms from the posterior circulation and in patients who proved to be asymptomatic, both carotid arteries were classified as asymptomatic.

CT imaging protocol

CTA was performed with a 64-slice scanner (Brilliance 64, Philips Healthcare, Best, The Netherlands). An 18-gauge intravenous catheter was placed in the antecubital vein and 80 mL of contrast (Visipaque 320, GE Healthcare) was infused at 4 mL/s. Acquisition and reconstruction parameters were as follows: 120 kV tube voltage, 265 effective mAs, pitch of 0.765, increment 0.45 mm. The thinnest available slice thickness of the CTA images was used (either 0.9 or 1.5 mm).

Intracranial ICA measurements

In the CTA images, a central lumen line in the intracranial ICA was created by a trained observer (LB) using 3mensio software (Bilthoven, The Netherlands). The centerline was used to generate a perpendicular cross-section with a view of the artery orthogonal to the centerline. An experienced neuroradiologist (CBM), blinded for clinical information, measured the degree of intracranial ICA stenosis according to the Warfarin–Aspirin Symptomatic Intracranial Disease (WASID) trial method [7]. The degree of intracranial ICA stenosis was measured for the intracranial ICA up to the origin the posterior communicating artery (segment C7, as defined in the carotid artery classification of Bouthillier) [6]. This segment was excluded because of a natural reduction of diameter. In a previous study, these measurements were repeated by a second experienced neuroradiologist (RvdB) and the interobserver reliability was calculated [4]. In the current study, only the measurements of the first observer were used.

The volume of the calcium in the intracranial ICA was quantitatively measured using the central lumen line as determined for the stenosis measurements. After selection of the vessel segment that contained calcifications, a region of interest was automatically generated. Since the calcifications were often adjacent to the skull base and because HU values of the skull are in the same range as the calcified plaques, the region of interest was manually edited to remove the skull. A threshold value was set at 420 HU to separate the calcified voxels from the contrast enhanced lumen. Subsequently, the software automatically calculated the volume of the voxels with intensities above this threshold in the defined region of interest in cubic millimeter (mm3). The results were visually inspected and could be corrected by adjusting the volume of interest and the threshold value.

Extracranial ICA measurements

Stenosis measurements in the extracranial ICA were performed according to the method of Bartlett [2] to determine the narrowest diameter in a plane perpendicular to the center lumen line of the vessel using a caliper tool. The reference ICA diameter was measured at least 2 cm above the site of narrowing according the North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria [1].

The calcium volume of the extracranial ICA was measured according to the method of McKinney et al. [14] using a specialized software (Vitrea 2, version, Vital Images, Inc., Plymouth, MN, USA). This method consisted of carefully ‘sculpting’ 2 cm below and above the bifurcation by drawing the area intended for inclusion. Subsequently, window level and width were altered to maximize the amount of visualized calcium while not including vascular contrast (level: 250–500, width: 10–20).

Statistical analysis

An intracranial ICA stenosis was defined as a stenosis larger than the cutoff value (30 %, 50 %, or 70 %) according to Nguyen-Huynh et al. and the WASID trial [7, 15]. The mean and standard deviations of the calcium volume and degree of stenosis were determined. These measurements were performed for the symptomatic and asymptomatic arteries separately and for all arteries together. The prevalences were calculated with a 95 %CI. The presence of significant intracranial stenosis was related with the presence or absence of an ipsilateral extracranial ICA stenosis (≥70 %), and the prevalence of in-tandem stenosis, which is defined as ipsilateral stenoses in the intracranial and extracranial ICA, was also determined.

The correlations between continuous variables were calculated using Pearson's correlation coefficient (r) for normally distributed variables and Spearman's correlation coefficient (ρ) if one of the variables was not normally distributed. The significance of r and ρ was tested using a two-tailed Student's t test. We correlated intracranial ICA stenosis with extracranial ICA stenosis and with intracranial ICA calcium volume. Furthermore, the intracranial calcium volume was correlated with extracranial calcium volume.


Eighty-eight patients were included in this study. The mean age of the patients was 67 ± 12 years and 60 % of the patients were male. Seventy-nine patients (90 %) were Caucasian. The final diagnosis was ischemic stroke in 44 % of the patients, TIA in 36 %, amaurosis fugax in 16 %, 3 % of the patients were asymptomatic, and one patient had an ocular ischemic syndrome. The baseline characteristics are shown in Table 1. CTA images of 176 arteries were available. Thirteen arteries were excluded because of an occlusion of the extracranial ICA. One artery had a severe stenosis of the distal extracranial part and petrous part of the ICA and was excluded because of an unreliable reference diameter of 1.5 mm. Seventy-five of the remaining 162 arteries were classified as symptomatic.
Table 1

Baseline characteristics


N (%) patientsa

Mean age (years)

67 (±12)

Male sex

53 (60 %)



79 (90 %)


9 (10 %)

Index event

 Cerebral infarction

39 (44 %)


32 (36 %)

 Amourosis fugax

14 (16 %)

History of


32 (36 %)

 Coronary artery disease

20 (23 %)

 Peripheral artery disease

13 (15 %)

Risk factors

 Current smoker

38 (43 %)


32 (36 %)


60 (71 %)

 Diabetes mellitus

21 (24 %)


48 (55 %)

aContinuous variables are represented as mean ± standard deviation


The mean degree of stenosis was 47 % (SD: 31 %) and 45 % (SD: 16 %) of the extracranial and intracranial ICA, respectively. Table 2 shows the prevalence of stenosis for the arteries. Eighty-three percent (95 %CI: 77–89 %) and 38 % (95 %CI: 30–46 %) of the evaluated arteries showed an intracranial ICA stenosis of ≥30 % and ≥50 %, respectively. This number reduces to 7 % (95 %CI: 3–12 %) for a stenosis of ≥70 %.
Table 2

The prevalence of stenosis ≥30 % and ≥50 % in the intracranial carotid artery for the symptomatic arteries, asymptomatic arteries, and all arteries together. The 95 % CI is given between the square brackets. The prevalence is given independent of the extracranial stenosis and separated for the presence or absence of an ipsilateral extracranial ICA stenosis ≥70 %


Any extracranial stenosis

Extracranial stenosis <70 %

Extracranial stenosis ≥70 %



All arteries



All arteries



All arteries

(N = 75)

(N = 87)

(N = 162)

(N = 45)

(N = 80)

(N = 125)

(N = 30)

(N = 7)

(N = 37)

Intracranial stenosis

≥30 %

79 % (59)

87 % (76)

83 % (135)

73 % (33)

83 % (66)

79 % (99)

87 % (26)

100 % (7)

89 % (33)

[70–88 %]

[80–94 %]

[77–89 %]

[60–86 %]

[81–95 %]

[72–86 %]

[75–99 %]


[79–99 %]

≥50 %

32 % (24)

43 % (37)

38 % (63)

22 % (10)

40 % (32)

34 % (42)

47 % (14)

43 % (3)

46 % (17)

[21–37 %]

[35–53 %]

[30–46 %]

[10–34 %]

[33–55 %]

[26–42 %]

[29–65 %]

[6–80 %]

[30–62 %]

≥ 70 %

9 % (7)

6 % (5)

7 % (12)

9 % (4)

4 % (3)

6 % (7)

10 % (3)

29 % (2)

14 % (5)

[3–15 %]

[1–11 %]

[3–12 %]

[1–17 %]

[0–8 %]

[2–10 %]

[1–19 %]

[0–63 %]

[3–25 %]

Calcium volume

The mean calcium volume in the extracranial ICA was 0.10 mL (SD: 0.13 mL) the mean calcium in the intracranial ICA was 0.09 mL (SD: 0.12 mL). The calcium volumes were not normally distributed. Eighty-seven percent of the intracranial arteries showed intracranial ICA calcifications (95 %CI: 82–92 %) (Table 3).
Table 3

Prevalence of calcification in the intracranial carotid artery and the average calcium volume ± SD given for all arteries, the symptomatic arteries, and the asymptomatic arteries




All arteries

(N = 74)

(N = 87)

(N = 161)


85 % (63)

89 % (77)

87 % (140)

[77–93 %]

[82–96 %]

[82–92 %]

Mean calcium volume (mm3)

87 ± 117

95 ± 128

91 ± 122


The correlation of extracranial and intracranial ICA stenoses is shown in Fig. 2a. Table 4 shows that only on the symptomatic side was a statistically significant correlation between the extracranial and intracranial ICA stenosis observed (Pearson's correlation: 0.32).
Fig. 2

The relation between intracranial ICA degree of stenosis with extracranial ICA degree of stenosis (a) and with intracranial ICA calcium volume (b)

Table 4

The correlation coefficients of the relations of the degree of stenosis of the intracranial ICA, degree of stenosis of the extracranial ICA, the intracranial ICA calcium volume, and intracranial ICA calcium volume. Numbers displayed as bold depict significant correlations




All arteries

Intracranial and extracranial degrees of stenosis (Pearson's r)

0.32 (P= 0.006)

0.05 (P = 0.64)

0.11 (P = 0.17)

Intracranial degree of stenosis and intracranial calcium volume (Spearman's ρ)

0.42 (P= 0.004)

0.53 (P< 0.001)

0.48 (P< 0.001)

Intracranial and extracranial calcium volumes (Spearman's ρ)

0.45 (P< 0.001)

0.47 (P< 0.001)

0.46 (P< 0.001)

Table 2 shows that for patients without an extracranial ICA stenosis of ≥70 % on the symptomatic side, 22 % (95 %CI: 10–34 %) of the intracranial ICAs have an intracranial stenosis ≥50 %. For the asymptotic side, this percentage is 40 % (95 %CI: 29–51 %). The prevalence of a tandem stenosis in a symptomatic artery, i.e., a stenosis in the extracranial (≥70 %) and intracranial ICA (≥30 %) was 87 % (95 %CI: 75–99 %). The prevalence of a tandem stenosis was 47 % (95 %CI: 29–65 %) for an intracranial stenosis of ≥50 %.

Table 4 shows that the correlation between extra- and intracranial ICA calcium volumes was moderate with a correlation coefficient of 0.46 (P < 0.001). There was a moderate correlation between the degree of intracranial ICA stenosis and intracranial calcium volume; Spearman's ρ = 0.48 (P < 0.001) for all arteries, ρ = 0.42 (P = 0.004) for the symptomatic side, and ρ = 0.53 (P < 0.001) for the asymptomatic side (Fig. 2b).


In our population of patients with recent neurological symptoms, the prevalence of intracranial ICA stenosis on CTA was remarkably high, with 84 % and 39 % for ≥30 % and ≥50 % stenoses, respectively. In 87 % of the arteries, intracranial calcifications were present. Intracranial ICA stenoses were frequently observed in combination with an extracranial ICA stenosis. However, the correlation of the degree of intracranial and extracranial stenoses was weak.

The main difference with previously intracranial artery disease prevalence studies is the imaging modality: in this study, we have evaluated intracranial artery disease on CTA rather than DSA, MRA, or TCD [9, 12]. We believe that the use of CTA for the visualization of intracranial arteries contributes to this difference: with CTA, it is possible to visualize calcifications in the artery wall and CTA allows the full 3D morphological assessment of the vessel without limitations associated with projection imaging.

Differences in patient selection criteria result in differences in prevalences [20]. In this study, we included patients who were suspected of stroke and had an extracranial stenosis as established with DUS in the US (30 % for men and 50 % for women). The presence of an extracranial stenosis may very well result in a higher prevalence of intracranial stenosis. In particular, US-based studies have more general patient populations which may contribute to the lower prevalences found in these studies. In the study of Kappelle et al., a similar patient population was used: patients with recent neurological symptoms, suspected of having carotid artery stenosis. However, in the NASCET trial, the majority of patients with severe intracranial atherosclerotic disease as shown on DSA were excluded before randomization [1]. As a result, Kapelle et al. reported that only 5 % of the patients included in NASCET had intracranial stenosis, whereas this number was as high as 38 % in our study. It should be noted that on the asymptomatic side, in which no stenosis with DUS was measured as a selection, the prevalence of intracranial ICA stenosis was similar to the symptomatic side. Therefore, the bias could be present due to patient selection but not due to the selection of the evaluated artery.

The prevalence of intracranial ICA disease is also higher than previously reported CT-based studies. The prevalence of calcifications in the intracranial ICA in CT-based studies is frequently reported [5, 8, 16, 22] with prevalences ranging from 36 % in a population of patients referred for head CT in general to 82 % in a population-based study [5]. With calcifications present in 87 % of the arteries and in 92 % of the patients, we report slightly higher numbers, which may be caused by the difference in patient population. In a similar study cohort, Homburg et al. [10] quantitatively measured intracranial arterial stenosis in CTA and found intracranial carotid artery stenosis in 23 % of the patients. However, probably because of technical difficulties to reliably visualize the intracranial ICA with CTA, this segment was excluded from the analyses. Our method enabled a detailed quantitative analysis of the entire ICA including tortuous intracranial segments such as the carotid siphon. We have shown here that the inclusion of this part of the intracranial carotid artery substantially increases the prevalence of intracranial atherosclerotic disease.

In the presence of an extracranial ICA stenosis ≥70 % on the symptomatic side, we observed a prevalence of approximately 90 % and 50 % for ipsilateral intracranial ICA stenosis, for a ≥30 % and ≥50 %, respectively. Tandem lesions are strongly related to cardiovascular risk factors, and most importantly, the risk of (recurrent) stroke in patients with a combined intracranial and extracranial stenosis is significantly increased [21].

The predictive value of intracranial carotid atherosclerosis for recurrent stroke is currently still under debate. For example, Taoka et al. [19] found no correlation between the degree of calcification in the carotid siphon and the risk for recurrent stroke. However, only 15 % of the patients in the cohort in this study had a recent stroke. On the other hand, the study of Hong et al. [11] showed that the degree of calcification in the carotid siphon can be used to predict the risk of lacunar infarction. Larger and more robust studies on the risk of recurrent stroke are still needed.

It should be noted that the calcium volume score as obtained in CT images is not the actual volume of the calcified plaques. Due to the blooming artifacts, the size of the calcified plaques appears larger than it is in reality. The blooming artifacts are caused by the limited spatial resolution and blurring and create a spillover effect into adjacent voxels. Therefore, volume measurement is a combination of calcium volume, calcium intensity, and image blurring. Furthermore, the calcium volumes were obtained using intensity threshold, which may also influence the measurement. Low threshold values may include lumen voxels, whereas high threshold values may miss calcifications. Therefore, all segmented calcium volumes were visually inspected. The volume of interest and intensity threshold could be adjusted to improve the measurement.

In CTA imaging, dense and extensive mural calcifications may reduce the accuracy of measuring the degree of stenosis. DSA is still better in visualizing small-caliber vessels with a resolution of 0.4 (or less) millimeter. Previous studies have shown that CTA-based degree of stenosis measurements of intracranial arteries correlate well with DSA-based measurements. [3, 15] Furthermore, it was previously shown that the interobserver agreement of these intracranial measurements is also good (Pearson correlation coefficient: 0.78) [4].

The retrospective design of this study does not permit conclusions on the most important issue of how intracranial atherosclerosis impacts prognosis of patients with extracranial carotid atherosclerosis or risk of extracranial carotid revascularization in these patients. Therefore, further research in which patient outcome is related with intracranial artery disease is advocated.

With the technological progress of high-quality CT scanning of the whole intracranial vasculature and the advent of quantitative imaging, we took a renewed look at intracranial lesions and determined its prevalence. The difference between these observations was beyond our expectation. We are in the process of including these intracranial atherosclerotic findings in a subsequent 3–5-year follow-up study to correlate radiological carotid artery findings with the risk of recurrent stroke. We believe that CTA is accurate in detecting intracranial carotid artery disease and could, therefore, be used in future risk prognosis and treatment decision support.


In our population of symptomatic patients suspected of extracranial ICA stenosis and extracranial ICA stenosis as screened by DUS, the prevalence of intracranial carotid artery stenosis and calcification on 64-slice CTA was remarkably higher than in previous DSA-based studies. Only on the symptomatic side, intracranial stenosis was correlated with extracranial stenosis. Since intracranial atherosclerotic disease is an important and independent risk factor for recurrent stroke, the clinical consequences of the high prevalence found in this CTA-based study should be further studied.

Conflict of interest

We declare that we have no conflict of interest.

Copyright information

© Springer-Verlag Berlin Heidelberg 2012