Книги по МРТ КТ на английском языке / Magnetic Resonance Imaging in Ischemic Stroke - K Sartor R 252 diger von Kummer Tobias Back
.pdf272 |
A. Thron and M. Mull |
Although this book is primarily devoted to magnetic resonance imaging, a brief description of cranial computed tomographic findings in CVST seems to be reasonable since we have learned from CCT that in imaging of brain parenchyma and cerebral veins we have to differentiate between direct and indirect signs of CVST. The direct signs prove the diagnosis by demonstrating thrombus or missing flow within a dural sinus or pial vein, the indirect signs simply raise the suspicion of CVST by demonstrating different forms of venous congestion (Chiras et al. 1985).
In CCT direct signs include the hyperdense sinus in the non-contrast-enhanced scan, the “cord sign” (hyperdense bridging vein) and the “empty triangle sign” in the contrast-enhanced CCT (Virapongse et al. 1987) (see Fig. 18.10). Within the first 2 weeks, thrombosed blood is typically hyperdense on CCT compared to brain parenchyma. Therefore, it is important to start with a non-enhanced CCT scan in patients with suspected CVST (Thron 2001). The density of the thrombus might otherwise be mistaken for a contrast-enhanced vessel lumen. After 2 weeks the thrombus will have become isodense or
hypodense to the brain parenchyma. Now the diagnosis can be made following the injection of contrast media which will show the thrombus as a filling defect of the lumen surrounded by either residual contrast-enhanced blood, the contrast enhancing meningeal wall or collateral venous channels outside the dura. This constitutes the empty triangle or empty delta sign on contrast enhanced CCT in a later stage of thrombus evolution.
Indirect signs include global and focal brain edema (see Figs. 18.6, 18.8, 18.9), intraparenchymal hemorrhages that might be solitary or multiple (Fig. 18.3) involving both grey and (preferentially) white matter and intense tentorial enhancement. Concerning the edema that is demonstrated as a hypodense area, the form and localization will typically not correspond to the classical arterial territories (see Fig. 18.8). Hemorrhages will also not suit the typical localization of parenchymal hypertensive bleeds but instead typically also expand to the cortical surface. The tentorium and the falx will appear thickened, engorged and will demonstrate pronounced enhancement that is due to dural venous collaterals.
|
|
Fig. 18.3a–e. Signal intensity of the |
|
|
thrombus and hemorrhage at dif- |
|
|
ferent stages of disease evolution. a |
|
|
T1-weighted image in the acute stage |
|
|
of CVST (< 3 days). Both hemorrhage |
|
|
(small arrows) and thrombus in the |
|
|
superior sagittal sinus appear iso- |
|
|
/hypointense. b Proton density image |
|
|
in the acute stage of CVST with |
|
|
hypointensity of hemorrhage (small |
|
|
arrows) and fresh thrombus (arrow). |
|
|
c–e In the subacute stages of CVST |
|
|
(> 3 days) the thrombus develops |
|
|
a hyperintense appearance on T1- |
|
|
weighted and T2-weighted images. |
a |
b |
(Compare with Table 18.1) |
|
c |
d |
e |
Veno-Occlusive Disorders
Table 18.1. Time dependant MR signal pattern in intraparenchymal hemorrhage and thrombosed dural sinuses. [Modified from Gomori et al. 1985; Isensee et al. 1994)
Time |
Molecule |
T1 |
T2 |
0-12 h |
Oxy Hb |
Iso-hypointense |
Hyperintense |
12-72 h |
Deoxy Hb |
Iso-hypointense |
Hypointense |
3-7 Days |
MetHb |
Hyperintense |
Hypointense |
|
intracellular |
|
|
1-4 Weeks |
MetHb |
Hyperintense |
Hyperintense |
|
extracellular |
|
|
> 4 Weeks |
Hemosiderin |
Iso-hypointense |
Hypointense |
|
|
|
|
18.4.1
Magnetic Resonance Imaging and Magnetic Resonance Angiography
The combination of these two MR techniques has become the imaging modality of first choice for the diagnosis and follow-up of CVST (Villringer et al. 1989; Vogel et al. 1994). MRI alone faces the problem that the MR signal of blood and blood products varies with clot age as is shown in Table 18.1 (Gomori et al. 1985). Therefore, a different appearance of a thrombosed vein in T1 and T2 sequences during different stages of thrombus evolution has to be taken into account (Isensee et al. 1994). The very fresh blood clot may be hyperintense in T1and hypointense in T2-weighted images during the first 12 h. In this very early phase it is very unlikely that patients with CVST are symptomatic and undergo diagnostic procedures. Afterwards the oxyhemoglobin of the acute thrombus has changed to deoxyhemoglobin which is isoto slightly hypointense to cortex on T1weighted sequences with a hypointense signal in T2-weighted images (12 h-3 days). Late acute clots (3-7 days) contain intracellular methemoglobin and are hyperintense on T1-weighted sequences and hypointense on T2-weighted images. Subacute thrombi (1-4 weeks after initial thrombosis) are hyperintense on both T1 and T2 scans due to extracellular methemoglobin. Chronically thrombosed sinuses undergo fibrosis with hemosiderin deposition and may develop extensive collaterals. It is clear that the main problem of MRI standard sequences for CVST diagnosis is the iso-hypoin- tense appearance of the acute clot (12-36 h) simulating flow. Problems may also be encountered in chronically thrombosed sinuses (Isensee et al. 1994). During the other phases of clot evolution an abnormal signal within the vessel lumen is evident
273
in at least one of the standard sequences. Bearing this consideration in mind, the sinuses are best visualized using axial and coronal sequences in which the superior and inferior sagittal sinus, as well as the transverse sinuses and the internal veins, are well imaged. Using these standard sequences, it is possible to evaluate normal anatomy or to detect anatomic variations like hypoplasia of one of the transverse sinuses.
Direct signs of CVST in MRI:
•Demonstration of an intraluminal thrombus within a dural sinus or a cerebral vein. This is easy during the time interval between 4 days and 4 weeks of thrombus age due to the high signal of methemoglobinonT1-weightedimages(Table 18.1; Fig. 18.3) It may be difficult in cases of a very fresh (< 3 days) or old (> 4 weeks) thrombus which can be organized or partially recanalized (Figs. 18.3, 18.7). Absence of the normal “flow-void” in large veins should raise suspicion of CVST, but this sign is unreliable, as it also appears with slow flow.
•Demonstration of an “empty triangle” or “delta sign” within a sinus, comparable to the finding in CT (see Figs. 18.9, 18.10) following contrastenhancement.
Indirect signs of CVST in MRI:
•Unior bilateral areas of edema that do not correspond to arterial territories (see Figs. 18.6-18.9)
•Unior bilateral hemorrhages (see Figs. 18.3, 18.8).
•Pronounced regional enhancement of the leptomeninges (falx, tentorium, convexity) due to the involvement of these structures in collateral drainage (see Fig. 18.9).
•Regional subarachnoid hemorrhage, especially if it is situated on the convexity of the brain. It may be a sign of the rare isolated cerebral vein thrombosis.
Table 18.2 summarizes the sequences and MR techniques which in our experience can be proposed (as mandatory or optional) in the diagnostic management of veno-occlusive disorders of the brain. Venous MRA can either be performed with the time-of-flight (TOF) or with the phasecontrast (PC) technique. In addition to the tomographic images, a flow sensitive gradient-echo sequence should be obtained if CVST is in question. As a fast screening examination we prefer a TOF 2D FLASH sequence (Table 18.2; Fig. 18.4, see 18.6c), oriented 90 degree to the flow direc-
274 A. Thron and M. Mull
Table 18.2. Diagnostic management of suspected CVST by |
the arterial and venous phase in coronal and sag- |
|||
MRI and MRA |
|
|
ittal direction and is useful not only in the detec- |
|
|
|
|
tion of venous drainage obstruction, but also in |
|
MRI + MRA |
Specific Parameters |
|||
the diagnosis of arteriovenous (AV) shunts. This |
||||
|
|
|
||
T1 axial |
|
|
||
|
|
is important because AV fistulae or AV malforma- |
||
T2 coronal |
|
|
||
|
|
tions are other important causes for venous drain- |
||
TOF angiogram (FLASH) |
TR/TE/FA: 23/7 ms/40°, 5 slices, |
|||
age impairment. Contrast-enhanced T1-weighted |
||||
5 mm, coronal, 1:36 min |
|
|
||
(2D FLASH) |
|
|
studies can be helpful but are not mandatory. As |
|
2D-PCA (optional) |
TR/TE/FA: 20/5.2/15°,1 slice, |
already mentioned they can demonstrate – simil- |
||
iar to the contrast enhanced CT – an “empty trian- |
||||
|
30 mm, sagittal+axial, 1:14 min |
|||
|
gle” or delta-sign (Figs. 18.9, 18.10). |
|||
3D-PCA (optional) |
TR/TE/FA : 16/6.8/10°, 200 slices, |
|||
An axial fluid-attenuated inversion recovery |
||||
|
0.8 mm, axial, 9:32 min |
|||
DWI (optional) |
TR/TE 5100/137 ms, 19 slices, |
sequence (FLAIR) is usually acquired additionally |
||
FLAIR (optional) |
5 mm, axial |
to demonstrate parenchymal involvement of CVST. |
||
|
|
Diffusion-weighted MRI in CVST has gained spe- |
||
3D contrast-enhanced MRA (optional) |
||||
cial attention in recent years because the patho- |
||||
2D dynamic contrast-enhanced subtraction MRA (optional) |
||||
physiology of diffusion abnormalities is less well |
||||
TR/TE/FA 3.5/1.0/40 coronal+sagittal |
||||
|
|
|
understood compared to arterial stroke (Sarma |
|
|
|
|
||
|
|
|
et al. 2004). The more complex pathophysiological |
|
|
|
|
process of venous congestion and infarction obvi- |
|
tion (coronal). It provides sufficient anatomical |
ously leads to both vasogenic and cytotoxic edema |
|||
details and gives reliable information whether |
(Corvol et al. 1998; Keller et al. 1999; Lövblad |
|||
there is flow (high signal) or no flow (no signal). |
et al. 2001). Three types of lesions were identified |
|||
The only information which is required for cor- |
by Mullins et al. (2004). Resolving lesions with |
|||
rect interpretation is the presence of methemo- |
elevated diffusion coefficient (vasogenic edema), |
|||
globin with a high signal on T1-weighted images. |
persisting lesions with low diffusion (cytotoxic |
|||
This substance also appears hyperintense on the |
edema, patients without seizure activity) and |
|||
FLASH image, thus simulating flow. This is one of |
resolving lesions with low diffusion (cytotoxic |
|||
several reasons why a combination of MR tomo- |
edema, patients with seizure activity). The obser- |
|||
graphic and angiographic sequences has to be |
vation of the reversibility of restricted diffusion in |
|||
postulated in CVST. 2D or 3D PC MR angiograms, |
extensive venous thrombosis was interpreted by |
|||
coded for slow flow, are established techniques for |
Sarma et al. (2004) as the existence of an “intracel- |
|||
the selective demonstration of cerebral veins and |
lular edema” which is reversible for an undefined, |
|||
should be used as additional standard sequences |
variable time. |
|||
when evaluating CVST (Fig. 18.2, 18.8--18.11). |
Our typical MR protocol for suspected CVST |
|||
These sequences create angiographic images and |
includes an axial FLAIR, axial diffusion-weighted |
|||
may facilitate image interpretation. However, loss |
MRI, coronal T1 SE and T2 TSE sequences, a coro- |
|||
of information on the maximum intensity projec- |
nal gradient echo and a 3D phase contrast venous |
|||
tion (MIP) images or pitfalls due to artefacts must |
angiogram with a total imaging time of approxi- |
|||
be taken into account (Fig. 18.10). Therefore, the |
mately 20 min. |
|||
source images always need to be included in the |
|
|||
evaluation. In the case of 3D sequences the exami- |
7.4.1.1 |
|||
nation time is considerably prolonged to about |
MRA Findings in (Benign and Idiopathic) Intracranial |
|||
10 min. Contrast-enhanced venous MRA is an |
Hypertension |
|||
advanced and costly technique (Farb et al. 2003) |
|
|||
which, on the other hand, avoids problems created |
Chronic thrombosis or only partially recanal- |
|||
by turbulent flow and improves image quality. |
ized dural sinus thrombosis may be diagnosed in |
|||
Another promising new technique is 2D dynamic |
these patients (Thron et al. 1986; Wessel et al. |
|||
(time resolved) contrast-enhanced MR subtrac- |
1987). This type is less obvious on the static MRI |
|||
tion angiography (Table 18.2; Figs. 18.5, 18.6). It is |
(Isensee et al. 1994) and requires MR venogra- |
|||
based on a single-slice T1-weighted gradient-echo |
phy, if possible in a contrast-enhanced technique. |
|||
sequence and has a temporal resolution of about |
In children, purulent mastoiditis is an important |
|||
0.34 s/image (Krings and Hans 2004). It covers |
cause of septic thrombosis (Reul et al. 1997) or |
Veno-Occlusive Disorders |
275 |
Fig. 18.4. Coronal 2D fast low angle shot (FLASH) (section). This flow-sensitive sequence offers a quick (1:30 min) screening for all major sinuses with additional anatomical information. Important: Comparison with the T1-weighted images is necessary because not only flow, but also methemoglobin appear bright
276 |
A. Thron and M. Mull |
a |
b |
c |
d |
Fig. 18.5a–d. 2D dynamic contrastenhanced MR subtraction angiography. Normal finding. Selected images from the arterial (a,b) and venous phase (c,d) are shown
inf lammatory stenosis (Isensee et al. 1992) of the ipsilateral transverse sinus (Reul et al. 1997), followed by raised intracranial pressure. A cystic developmental lesion within a sinus (Küker et al. 1997), or other space-occupying or infiltrating processes, obstructing the lumen of a dural sinus are rare causes.
In patients with idiopathic intracranial hypertension bilaterally narrowed segments in the lateral venous sinuses have been demonstrated compared to normal findings in asymptomatic volunteers using advanced techniques of MR
venography (Higgins et al. 2002, 2004; Farb et al. 2003). We have observed similar cases in recent years with unilateral sinus stenosis and contralateral hypoplasia or bilateral stenoses (see Fig. 18.2c-e).
Veno-Occlusive Disorders |
277 |
a b
c |
d |
Fig. 18.6a–d. Deep cerebral venous thrombosis in an 8-year-old girl presenting with headache, state of confusion and somnolence. a Axial FLAIR image with bilateral swelling and hyperintensity of the thalami and right-sided basal ganglia. b Coronal T2-weighted image with high signal in both thalami due to edema and/or infarction, suspicious of thrombosis of the deep venous system (internal cerebral veins, straight sinus). c On the coronal flow-sensitive Sequence (FLASH) no flow is shown in the internal cerebral veins (arrows). d 2D dynamic contrast-enhanced MR subtraction angiography. The images from the arterial phase (upper row) are normal, in the late venous phase (lower row) an area of reduced parenchymal contrast and absence of the internal cerebral veins and straight sinus are shown. (Compare with normal findings in Fig. 18.5)
278 |
A. Thron and M. Mull |
a
b
Fig. 18.7a,b. Isolated cortical vein thrombosis in a 16 y old female presenting with focal epileptic seizures. a Axial GE (T2*) images show a hypointense signal along the course of a cortical vein (arrows) which was also seen as a hyperdense structure on CT (not shown). b Axial DWI shows small areas of restricted diffusion in the corresponding brain parenchyma
Veno-Occlusive Disorders |
279 |
Fig. 18.8a–d. Thrombosis of the transverse sinuses with temporolateral congestive and hemorrhagic infarction on the right side. The lesion corresponds to the drainage territory of the vein of Labbé. a CT aspect of the partially hemorrhagic lesion. b FLAIR images showing areas of different hyperintensity. c DWI with a very inhomogeneous pattern of diffusion abnormalities. d 3D PC venography. Extensive thrombosis with incomplete occlusion mainly of the transverse sinuses (arrows). On the right side, corresponding to the lesion location, temporal cortical veins (vein of Labbé) are missing (short arrow)
a |
b |
c |
d |
280 |
A. Thron and M. Mull |
a |
b |
c |
d |
e |
Fig. 18.9a–e. Extensive thrombosis of all main dural sinuses. a Coronal T2-weighted images. Only small cortical lesions are present. The signal of the big sinuses is iso-/hypointense and does not indicate thrombotic occlusion. b Axial T1-weighted contrast-enhanced images. Intraluminal thrombus is evident in the superior sagittal sinus (arrowheads) and increased leptomeningeal enhancement can be demonstrated (arrows) indicating collateral drainage through small veins. c Axial FLAIR images. Small cortical/subcortical areas of infarction are present (arrows). d 3D PC MR venogram. Extensive thrombosis of all major dural sinuses. The drainage is restricted to superficial collateral veins. e 3D PC MR venogram after 3 months of anticoagulation. Improvement with partial restoration of flow in the superior sagittal and transverse sinuses (arrows)
Veno-Occlusive Disorders |
281 |
a
b
d
c
Fig. 18.10a–d. Superior sagittal sinus thrombosis in a 29-year-old woman. She was treated with steroids for multiple sclerosis and presented with acute headaches. a Contrast-enhanced CT, performed on admission, shows a clear “empty triangle” sign in the superior sagittal sinus as a direct sign of thrombosis. b 3D PC MR venography, performed the day after admission in addition to MRI, did not demonstrate a clear pathological finding. c,d Contrast-enhanced T1-weighted images in axial (c) and sagittal orientation reveal a long thrombus which is not completely occluding the lumen of the sinus