AJR 2005; 185:772-783
© American Roentgen Ray Society
CT Neuroangiography: A Glance at the Common Pitfalls and Their Prevention
Deepak Takhtani1
1 Department of Radiology, Johns Hopkins School of Medicine, B100G Phipps
Basement, 600 N. Wolfe St., Baltimore, MD 21287.
Received September 21, 2004;
accepted after revision November 11, 2004.
Address correspondence to D. Takhtani
(dtakhta1{at}jhmi.edu).
Abstract
OBJECTIVE. The author's objective was to address several pitfalls
and missteps of CT angiography (CTA) with pictorial examples and ways to
overcome potential misinterpretations. CTA is increasingly used for the
noninvasive evaluation of the carotid and intracranial vessels. Ease of data
acquisition may belie the complexity of interpreting each individual vessel
and understanding myriad postprocessing techniques, each with its strengths
and weaknesses.
CONCLUSION. Diagnostic yield and accuracy of CT angiography are
enhanced by good data acquisition and by understanding and fully exploiting
postprocessing techniques.
Introduction
Newer-generation MDCT scanners with submillimeter detectors and
subsecond rotation capabilities have led to quantum leaps in the noninvasive
evaluation of the cerebral vasculature. Some of the MDCTs have isotropic voxel
capability that ensures the same resolution in every plane of reformations. CT
angiography (CTA) has been shown to be equal if not better than conventional
angiogram in the detection of cerebral aneurysms
[1]. In a series of 41
aneurysms measuring equal to or less than 4 mm, sensitivity of CTA for the
detection of cerebral aneurysms was higher than that of digital subtraction
angiography, with equal specificity
[2]. In another study,
aneurysms that were overlooked at the initial interpretation of CTA were
identified at a retrospective reading
[3]. This underscores the
importance of careful scrutiny of CTA and the need to be familiar with factors
that can adversely affect the interpretation. Several pitfalls of CTA have
been described, including difficulty in visualization of small arteries,
differentiating the infundibular dilatation at the origin of an artery from an
aneurysm, venous structures that can simulate or hide aneurysms, and an
inability to identify thrombosis and calcification on 3D images
[4]. CTA is better performed
with automated trigger capability to optimize the contrast bolus.
View larger version (131K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —Right arm versus left arm injection. Contrast material is in
left brachiocephalic vein arching over major arteries, which can produce
streak artifacts and obscure origin of major vessels.
|
|
CTA is also useful in diagnosing carotid stenosis and has more than a 90%
concordance rate with conventional angiogram
[5]. Plaques and ulcerations
are better visualized on the CTA than digital subtraction angiography
[6]. Accuracy of CTA depends on
many factors, including the use of proper reconstruction thickness, choosing
the appropriate postprocessing technique, perusal of the source data, and
suitable windowing. Postprocessing techniques such as maximum intensity
projection (MIP), volume rendering, and multiplanar reformation (MPR) have
strengths and weaknesses. In this pictorial essay, illustrations of pitfalls
resulting from technical factors or postprocessing methods and ways to avoid
these are presented.
View larger version (131K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —Stair-step or zebra-stripe artifact. 3D volume-rendered image
from 1-mm slices with 1.5-mm gap reconstruction parameters shows zebralike
appearance of bone and some attenuation of middle cerebral artery branches
(arrow).
|
|
View larger version (135K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B —Stair-step or zebra-stripe artifact. Image was created with
1-mm slices and 0.5-mm overlap from same raw data, and provides better
delineation of smaller vessels, smooth bone background, and better
visualization of distal vessels. Incidental note is made of right vertebral
artery stenosis (arrow).
|
|
Possible Pitfalls
Site of Injection
For the neck carotid study, it is preferable to inject from the right arm
to avoid artifacts from dense contrast in the brachiocephalic vein and
superior vena cava (Figs. 1A,
and 1B).
View larger version (97K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5A —Presumed schwannoma likely to be misinterpreted as aneurysm
in a 6-year-old girl. Maximum-intensity-projection image shows an
aneurysm-like structure in relation to posterior cerebral artery
(arrow).
|
|
View larger version (91K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 5C —Presumed schwannoma likely to be misinterpreted as aneurysm
in a 6-year-old girl. Source image clearly separates lesion (arrow)
from vessel and rules out aneurysm. Conventional catheter angiogram was
negative.
|
|
View larger version (19K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6A —Vertebral artery dissection in a 67-year-old woman.
Limitation of volume-rendered technique. Volume-rendered picture of vertebral
artery (A) shows alternate areas of narrowing and dilatation but fails
to show intimal flap (arrow, B-D) seen on curved multiplanar
reconstruction (B) and source image (C). Angiogram done later
confirms finding (D).
|
|
View larger version (93K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6B —Vertebral artery dissection in a 67-year-old woman.
Limitation of volume-rendered technique. Volume-rendered picture of vertebral
artery (A) shows alternate areas of narrowing and dilatation but fails
to show intimal flap (arrow, B-D) seen on curved multiplanar
reconstruction (B) and source image (C). Angiogram done later
confirms finding (D).
|
|
View larger version (101K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6C —Vertebral artery dissection in a 67-year-old woman.
Limitation of volume-rendered technique. Volume-rendered picture of vertebral
artery (A) shows alternate areas of narrowing and dilatation but fails
to show intimal flap (arrow, B-D) seen on curved multiplanar
reconstruction (B) and source image (C). Angiogram done later
confirms finding (D).
|
|
View larger version (76K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 6D —Vertebral artery dissection in a 67-year-old woman.
Limitation of volume-rendered technique. Volume-rendered picture of vertebral
artery (A) shows alternate areas of narrowing and dilatation but fails
to show intimal flap (arrow, B-D) seen on curved multiplanar
reconstruction (B) and source image (C). Angiogram done later
confirms finding (D).
|
|
View larger version (142K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7A —Follow-up in 54-year-old man with stents placed in the
transverse and sigmoid sinuses for the thrombosis. Volume rendering versus
curved multiplanar reformations. 3D volume-rendered image of right transverse
and sigmoid sinuses does not reveal information about its lumen or wall.
|
|
View larger version (108K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 7B —Follow-up in 54-year-old man with stents placed in the
transverse and sigmoid sinuses for the thrombosis. Volume rendering versus
curved multiplanar reformations. Curved multiplanar reconstruction through
same shows stent and areas of thrombus formation (arrow) in
transverse and sigmoid venous sinuses.
|
|
View larger version (86K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 8A —Maximum intensity projection (MIP) of circle of Willis. MIP
with 10-mm slab in axial plane (A) suggests that middle
(arrow) and posterior cerebral arteries (arrowhead) are
occluded; however, 40-mm MIP slab reformation (B) shows full extent of
arteries. Veins are also visualized. Internal cerebral veins are seen
(arrow).
|
|
View larger version (104K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 8B —Maximum intensity projection (MIP) of circle of Willis. MIP
with 10-mm slab in axial plane (A) suggests that middle
(arrow) and posterior cerebral arteries (arrowhead) are
occluded; however, 40-mm MIP slab reformation (B) shows full extent of
arteries. Veins are also visualized. Internal cerebral veins are seen
(arrow).
|
|
View larger version (94K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 9A —73-year-old woman with history of transient ischemic attack.
Curved multiplanar reconstruction image of carotid shows true lumen of
internal carotid with calcific plaques in arterial wall.
|
|
View larger version (51K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 9B —73-year-old woman with history of transient ischemic attack.
Volume-rendered 3D image fails to shows extent of stenosis, as it incorporates
calcified plaques in image and surface of artery appears blistered
(arrow).
|
|
Technical Factors
Reconstruction Slice Thickness and Slice Overlap
Raw data acquired during the scanning can be manipulated to get the desired
slice thickness and overlap. It is not possible to get a smaller slice
thickness than the detector collimation. For example, if the selected detector
collimation is 1.0 mm, one cannot reconstruct at 0.5 mm. However, it is
possible to reconstruct a thicker slice. Thickness of the reconstructed slices
has bearing on the quality of the 3D reconstructions (Figs.
2A, and
2B). Overlapping the
reconstructed slices is important to minimize the stair-step artifact. We
usually acquire reconstructed slices with 30-50% overlap. Slices with no
overlap or interslice gaps compromise the 3D processing and introduce
artifacts and false stenosis (Figs.
3A, and
3B).
Source Images
Source images should be reviewed in all cases, irrespective of the clinical
indication. As the reconstruction is done using the data from the source
images, any finding seen on the processed image should be present on the
source images. It is important to look for discordance between the
reconstructed images and the source images (Figs.
4A, and
4B). Source images are also
critical when a brightly enhancing lesion is located close to a
high-probability aneurysm site. Such strategically located enhancing lesions
like meningioma or schwannoma can mimic an aneurysm (Figs.
5A,
5B, and
5C). We also rely on the
source images to review the vessels in areas such as the base of skull, where
sculpting and 3D techniques are difficult.
Postprocessing Techniques
There are three main postprocessing techniques for CTA: MPR, MIP, and
volume rendering. MPR can be in a straight, oblique, or curved oblique plane.
Since the carotids have a twisted course, curved MPR capability is helpful.
Curved MPRs are also helpful in the cavernous segment of the internal carotid
artery (ICA), a relatively difficult area to evaluate on CTA. MPR
reconstruction can be one or more than one pixel thick and is equivalent to
taking a slice through the artery. Many vendors provide automated or
semiautomated curved multiplanar reconstruction capability in which the vessel
is traced based on the Hounsfield units (H). The automated tracing may not
work consistently as the tracer can wander off along the closely lying
contrast-filled veins and high-density bone. The author finds manual tracing
of the vessels reliable, although it takes a few extra minutes. MPRs provide
the luminal and intraluminal information better than MIP and volume-rendering
techniques, as it is akin to taking thin sequential slices through the vessel.
However, it is important to get MPRs in at least two planes to get more
accurate information. MPR is particularly useful in areas of stenosis,
visualization of the intimal flap in dissection, plaques, and intraluminal
defects (Figs. 6A,
6B,
6C,
6D,
7A, and
7B).
The MIP technique involves selection of the brightest pixels to make the
image while discarding the rest. MIP provides a 2D or 3D view of the
structures on CTA. We have found MIP useful in intracranial vessels and in
areas in which calcified plaques are present. Unlike volume rendering, MIP
does not provide "depth" in an image and the structures appear
overlapped or pruned. This has the potential to provide misleading information
(Figs. 8A, and
8B). One could vary the width
of the MIP slab to match the vessel diameter and avoid overlapping of vessels.
MIP shows the calcification and stenosis better than volume-rendered
images.
The volume-rendering technique uses all the pixels and thus provides a real
3D look. However, with this technique, it is not possible to look into the
lumen and discriminate calcium from the wall or the lumen (Figs.
9A, and
9B). Calcific plaques in the
wall give the appearance of "blisters" on the volume-rendered
images and sometimes resemble aneurysm (Figs.
10A, and
10B). The volume-rendered
technique is good for intracranial vessels, such as the middle and anterior
cerebral arteries in which calcifications are less likely and with their
location away from the bone.
View larger version (144K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 11A —75-year-old man with transient ischemic attack. Images show
importance of appropriate windowing. 3D volume-rendered image on right with
width and center of 200/180 H shows diffusely attenuated vessels.
|
|
View larger version (146K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 11B —75-year-old man with transient ischemic attack. Images show
importance of appropriate windowing. Width/center of 200/140 H shows many more
peripheral branches and no significant stenosis.
|
|
Windowing
"Tuning" the windows provides better contrast between two
adjacent structures. In the volume-rendering technique, inappropriate window
settings can create pseudostenosis, making the lesion disappear or the vessels
pruned (Figs. 11A, and
11B). A very clear background
may be one indication that windowing is too wide. One should alter the window
center or level so that the background begins to turn a little
"dirty," or the vessels stop increasing in caliber. In our
experience, a window width of 200-250 H and center/level of 140-180 H is
optimum. In case of less contrast in the vessels, the center of the window may
need to be lowered.
View larger version (84K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 13A —53-year-old woman with headaches. Venous confluence at
internal carotid artery (ICA) terminus. Confluence of veins at ICA terminus
(arrow) on source image (A) gives appearance of aneurysm
(arrow) on volume-rendered image (B).
|
|
View larger version (111K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 13B —53-year-old woman with headaches. Venous confluence at
internal carotid artery (ICA) terminus. Confluence of veins at ICA terminus
(arrow) on source image (A) gives appearance of aneurysm
(arrow) on volume-rendered image (B).
|
|
View larger version (48K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 14B —Pseudofenestration in 76-year-old woman with history of
right-sided weakness. Volume-rendered image on CT angiography shows
pseudofenestration due to deep middle cerebral vein (curved arrow)
running parallel to stenosed middle cerebral artery (straight arrow).
Basal vein of Rosenthal and posterior communicating artery are superimposed
(arrowhead).
|
|
Venous Structures
Opacification of the venous structures is inevitable on CTA, as the
circulation time from the arteries to veins is just 3-6 sec. Wetzel et al.
[7] report that the sensitivity
of CT venography in defining the cerebral veins is comparable with digital
subtraction angiography. To avoid misinterpretation from simultaneous
opacification, venous structures should be differentiated from arteries by
understanding the anatomy of some common veins. One such vein, the basal vein
of Rosenthal gets its tributaries from the deep middle cerebral, olfactory,
front-orbital, anterior-cerebral and inferior-striate veins
[8]. Some of its tributaries
converge on the ICA terminus and proximal M1 and A1 segments, causing
"kissing artifacts." The vein also follows a path parallel to the
posterior communicating artery and may be mistaken for it. One should try to
trace the vessels to their tributaries or branches on the source images to
differentiate venous from arterial structures. Tributaries to the basal vein
of Rosenthal may be mistaken for, obscure, or resemble an aneurysm in the
region of the ICA terminus or M1 segment of the MCA (Figs.
12A,
12B, and
12C).
View larger version (115K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 14C —Pseudofenestration in 76-year-old woman with history of
right-sided weakness. Maximum-intensity-projection image shows same finding.
Mild difference in density of contrast between vein (curved arrow)
and artery (straight arrow) is evident.
|
|
View larger version (154K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 15B —70-year-old man with ataxia. Basilar artery stenosis is
hidden by pontomesencephalic vein. Rotation of image shows relationship of
vein (small arrow) and stenosis in basilar artery (thick
arrow).
|
|
The confluence of venous tributaries near the ICA terminus may look like an
aneurysm (Figs. 13A,
13B, and
13C). Rarely, an artery and
vein lying close to each other may be mistaken for fenestration (Figs.
14A,
14B, and
14C). Aberrant veins may also
be mistaken for a normal vessel concealing the artery
[9] (Figs.
15A, and
15B).
View larger version (86K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 17B —CT angiography on 81-year-old man for evaluation of carotid
stenosis. Pseudothrombus in jugular veins. Sagittal reformation and axial
images show contrast material localized in dependent posterior aspect of
veins. Sagittal reformation (B) also shows continuous column of
contrast material extending from brachiocephalic to jugular vein.
|
|
View larger version (69K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 17C —CT angiography on 81-year-old man for evaluation of carotid
stenosis. Pseudothrombus in jugular veins. Sagittal reformation and axial
images show contrast material localized in dependent posterior aspect of
veins. Sagittal reformation (B) also shows continuous column of
contrast material extending from brachiocephalic to jugular vein.
|
|
Infundibulum Versus Aneurysm
At times it may be difficult to differentiate an infundibulum from a small
aneurysm, such as in cases where the branch arising from the tip is beyond the
resolution of the CTA. This is especially true if the lumen of the vessel is
less than 0.5 mm. Decreasing the window center on volume-rendered images may
help visualize a faintly opacified vessel at the tip of the aneurysm. In some
cases, careful review of the source images may be of help to differentiate an
aneurysm from the adjacent normal artery
(Fig. 16).
Venous Reflux
In some patients, contrast may reflux back into the jugular veins and other
neck veins and mix with noncontrasted blood to give an appearance of a filling
defect, suggesting a thrombus (Figs.
17A,
17B, and
17C).
References
- Karamessini MT, Kagadis GC, Petsas T, et al. CT angiography with
three-dimensional techniques for the early diagnosis of intracranial
aneurysms. Comparison with intra-arterial DSA and the surgical findings.
Eur J Radiol 2004;49
: 212-223[CrossRef][Medline]
- Villablanca JP, Jahan R, Hooshi P, et al. Detection and
characterization of very small cerebral aneurysms by using 2D and 3D helical
CT angiography. Am J Neuroradiol 2002;23
: 1187-1198[Abstract/Free Full Text]
- Jayaraman MV, Mayo-Smith WW, Tung GA, et al. Detection of
intracranial aneurysms: multi-detector row CT angiography compared with DSA.
Radiology 2004;230
: 510-518[Abstract/Free Full Text]
- Tomandl BF, Kostner NC, Schempershofe M, et al. CT angiography of
intracranial aneurysms: a focus on postprocessing.
RadioGraphics 2004;24
: 637-655[Abstract/Free Full Text]
- Moll R, Dinkel HP. Value of the CT angiography in the diagnosis of
common carotid artery bifurcation disease: CT angiography versus digital
subtraction angiography and color flow Doppler. Eur J
Radiol 2001; 39:155
-162[CrossRef][Medline]
- Randoux B, Marro B, Koskas F, et al. Carotid artery stenosis:
prospective comparison of CT, three-dimensional gadolinium-enhanced MR, and
conventional angiography. Radiology 2001;220
: 179-185[Abstract/Free Full Text]
- Wetzel SG, Kirsch E, Stock KW, Kolbe M, Kaim A, Raude EW. Cerebral
veins: comparative study of CT venography with intraarterial digital
subtraction angiography. Am J Neuroradiol1999; 20:249
-255[Abstract/Free Full Text]
- Suzuki Y, Ikeda H, Shimadu M, Ikeda Y, Matsumoto K. Variations of
the basal vein: identification using three-dimensional CT angiography.
Am J Neuroradiol 2001;l22
: 670-676[Abstract/Free Full Text]
- Teksam M, Casey S, McKinney A, Michel E, Truwit CL. Anatomy and
frequency of large pontomesencephalic veins on 3D CT angiograms of the circle
of Willis. Am J Neuroradiol 2003;24
: 1598-1601[Abstract/Free Full Text]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
