MRI of the head in stroke

Authors: Mikael Häggström; Authors if integrated Creative Commons article

Planning

Choice of modality

• CT of the head in stroke is generally the first investigation because of high availability. It detects about 60% of infarcts in the first 3 to 6 hours, and almost always thereafter.
• MRI of the head in stroke is generally less available, but preferred if there are no MRI contraindications, and your clinic has established procedures for its performance in acute stroke (if you do not know, it is reasonable to assume that there is not, and do CT).

Still not sure? See: CT vs MRI in stroke

How soon

According to UK guidelines, imaging should be performed immediately for people with suspected acute stroke if any of the following apply:

• Indications for thrombolysis or early anticoagulation treatment
• Current anticoagulant treatment
• Any known bleeding tendency
• Depressed level of consciousness of Glasgow coma scale (GCS) of less than 13
• Unexplained progressive or fluctuating symptoms
• Papilledema, neck stiffness or fever
• Severe headache at onset of stroke symptoms.

"Immediately" is defined as 'ideally the next slot and definitely within 1 hour.

In case of acute stroke without indications for immediate imaging, it should be performed within a maximum of 24 hours after onset of symptoms.

Conventional MRI

MR sequences usually used in acute strokes are: T1-weighted spin echo, T2-weighted fast spin-echo, fluid-attenuated inversion recovery-FLAIR, DWI for acute ischemia, MRA and PWI for penumbra. T2-weighted gradient-echo- GRE- for hemorrhage and gadolinium-enhanced T1- weighted spin-echo sequences. Typical findings in acute ischemia include: hyper-intense signal in white matter on T2-weighted images and fluid-attenuated inversion recovery images, and similar to CT findings; grey-white matter distinction loss, sulcal effacement and intravascular signal intensity changes due to occlusion. Hemorrhage is seen as abnormal blooming and while MRI is effective at detecting bleeds, thrombolytic guidelines utilize CT based evidence alone. With the onset of Diffusion weighted imaging, the conventional MR sequences play a minor role in acute stroke evaluation (Figure 7A,B,C).

• 

  Figure 7A: Fast spin echo T2-weighted fat suppressed image demonstrates increased signal intensity and effacement of the right temporal lobe, consistent with sub-acute infarct of the right middle cerebral artery.

• 

  Figure 7B: These are T1 weighted MRI images of a patient with bilateral cortical border zone infarcts (secondary to hypotension). T1 bright staining is seen along cortical grey matter suggestive of Cortical Laminar Necrosis.

• 

  Figure 7C: MRI axial FLAIR images of brain show an infarct involving left frontal lobe anterior to the Sylvian fissure,an area which corresponds to the superior division of the left middle cerebral artery.

MR angiography

High resolution imaging of the cranial vasculature can be obtained using a non-contrast 3D time-of-flight (TOF) MR angiography (MRA). The basis of this technique is that protons in tissue are saturated by repeated radiofrequency excitation and have low signal intensity but protons coming in through the vessels are unsaturated and have a high signal intensity producing flow dependent imaging. TOF can be 2D or 3D and does not expose the patient to radiation, or contrast. It is limited by both patient motion and flow artifact often resulting in over-estimation of the stenosis or occlusion. Since it obtains imaging over a short time period, it provides no data on the speed or direction of blood flow (important to study collateral supply). To overcome this, 4-D MRA (and CTA) have been developed. This involves reimaging the field of interest at several time points after the radiofrequency pulse (or contrast injection in CTA) with results similar to those of catheter angiography . Several studies show that TOF is not as accurate as other modalities such as DSA or CTA with a sensitivity of 84.2% and specificity of 84.6% . Contrast enhanced MRA (with gadolinium) is the technique used for extra-cranial vasculature imaging allowing improved anatomical details with an advantage of better diagnosis of arterial dissections to be balanced with side-effects of contrast in renally compromised patients and nephrogenic systemic fibrosis (Figure 8 A,B & C).

• 

  Figure 8A: 3-D Time Of Flight (TOF) Non contrast MR Angiography of brain shows severe right MCA proximal main stem (yellow arrow) with sparse cortical branches of right MCA compared to left (green arrow).

• 

  Figure 8B: The same case as above shows faint serpiginous high signal on T2* axial FLAIR (due to sluggish flow in the sparse cortical branches of the MCA) known as the “ivy sign” in the right perisylvian area (yellow arrows).

• 

  Figure 8C: An MR angiogram of a patient with bilateral internal carotid artery dissection showing

Diffusion-weighted imaging (DWI)

MR with diffusion is quickly becoming the gold standard in acute stroke imaging, as it can be obtained within 10 minutes at certain centers . DW MRI uses fast (echo-planar) imaging technology, and is resistant to patient motion artefacts with an imaging time of a few seconds to 2 minutes . The sensitivities and specificities are 100% and 86%, respectively, for diffusionweighted MRI versus 18% and 100% for conventional MRI . Hence, DW MRI plays an essential role in the diagnosis of acute infarctions and to rule out stroke mimics.

Figure 9:(A and B) Restricted water diffusion in the region of infarct (right middle cerebral artery) results in an increased signal intensity on diffusion-weighted imaging (A) and decreased signal on apparent diffusion coefficient imaging (B).

Principle: is based on the random (Brownian) motion of water molecules in tissues. A change in the water content of cells affects the rate of molecular diffusion in these tissues. Using T2 weighted spin-echo MR imaging sequence with 2 extra gradient pulses (equal magnitude, opposite directions) results in signal changes: greater loss of signal is seen in tissues with higher rates of diffusion and vice versa. Hence, in acute stroke (cytotoxic edema with decreased rate of molecular diffusion in the affected tissue), ischemic tissue appears brighter than normal brain tissue. DW MRI cannot be used to calculate the diffusion coefficient, which is obtained from orthogonal diffusion weighted MR images in all 3 planes and is called Apparent Diffusion Coefficient ADC. ADC maps are used because occasionally areas on diffusion imaging have a high signal due to vasogenic edema, but will appear dark on ADC map proving it is not an acute infarct. In humans restricted diffusion is seen as early as 30 minutes decreasing to reach a low at 3–5 days (hyper intense signal with reduced ADC values due to restricted diffusion), and reaches baseline at 1–4 weeks, explained by the development of vasogenic edema alongside the cytotoxic edema, replaced by gliosis which carries an increase in extracellular water (high intensity signal due to increased T2 signal: T2 shine through and variable ADC values for months). Hence DWI cannot be reliably used to estimate infarct age without the help of ADC maps. DW imaging has high sensitivity and specificity with studies showing 94% accuracy for afinal diagnosis of stroke, compared with a yield of 71%-80% when using conventional MRI (T2W/ PDW or FLAIR sequences) . DWI demonstrated infarct volumes also correlate (with statistical significance) with NIH Stroke Score and other scales . However false negatives with DWI can occur in small lacunar infarcts of the brainstem or deep nuclei . False positives may be seen in abscesses, tumors but can be easily ruled out. Diffusion imaging may show no abnormality in the setting of ischemia, hence normal DWI with altered perfusion images indicate tissue at risk, and prompt initiation of therapy (Figure 9).

Perfusion-weighted imagin (PWI)

The main advantage of Perfusion weighted imaging is its ability to image entire brain and detect the ischemic penumbra. This technique requires a method of achieving contrast for perfusion either exogenous (IV contrast agent-gadolinium) or endogenous (labeling hydrogen-1 protons in water or arterial spin labeling).

Figure 10A: Patient presents with a stroke involving left middle cerebral artery. MR perfusion shows areas of increased time to peak (TTP) which correspond to decreased flow and volume. Visual key:
- For TTP, higher scale (red) means longer TTP (compromised), and lower scale (blue) means shorter TTP (intact).
- For CBF (cerebral blood flow), higher scale (red) means faster flow (preserved), and lower scale (blue) means less flow (compromised).
- For CBV (cerebral blood volume), higher scale (red) means more volume (generally preserved), and lower scale (blue) less volume (generally compromised).

Principles: Passage of contrast in the vessels causes transient loss of signal (T2 effect). Tracking these signal change results in a time-signal curve which can create perfusion maps of cerebral blood volume and mean transit time. Using the arterial spin technique encompasses using radiofrequency to invert the spin polarity of the protons entering the image plane which measurably affects the image intensity as they perfuse the tissue. Subtracting the flow sensitive image from the flow insensitive image gives a measure of the protons perfusing that image plane, in essence a perfusion map.

PWI detects areas of reduced flow contrasted to DWI which depicts areas of injury. The difference between the DWI defect and the PWI defect is the ischemic penumbra which helps to guide further stroke therapy. This is the subject of multiple clinical trials for patients with a DWIPWI mismatch who will benefit from thrombolysis . In certain cases, DWI defect is larger than PWI defect which is seen in early reperfusion (Figure 10 A, B).

• 

  Figure 10B: TTD -time to drain, (similar to TTP), shows decreased flow in entire middle cerebral artery territory in a case of infarction. CBF is down in the lenticulo-striate area, but relatively good over the temporal lobe. CBV is increased implying that temporal lobe vascular bed has dilated effectively, thereby increasing its volume and almost normalizing temporal lobe CBF.

This is the preferred MRI sequence to rule out acute or chronic hemorrhage including a hemorrhagic transformation. It is a T2 weighted sequence sensitive to the change in local magnetic fields due to the iron in blood and degradation products. As the hemoglobin in the parenchyma becomes deoxygenated, unpaired electrons make it para-magnetic (non-uniform field) resulting in a signal loss- known as susceptibility effect seen as hypointense (dark) lesions. This sequence is used to image acute stroke patients in whom ICH is suspected or has to be ruled out for thrombolytic therapy and is found to be as accurate as CT in the detection of acute ICH and more accurate in cases of chronic ICH (Figure 11 A, B).

• 

  Figure 11 A: Arterial infarct with hemorrhagic transformation show a low signal intensity (yellow arrow) on T2* GRE (areas of bleed) in the region of infarct.

• 

  Figure 11 B: MRI axial FLAIR study shows focal hyperintensity in the region of corpus callosum (yellow arrow) which corresponds to the low signal intensity on T2* GRE (green arrow). DWI shows the corresponding focus of ischemia (red arrow).

MRI - timing and therapeutic windows

Treatment with thrombolytics is time dependent (within 3-4.5 hours) hence strokes that occur in sleep or unwitnessed pose a unique therapeutic challenge for which neuroimaging can provide decision making criteria. MRI sequences have been proposed to identify strokes in the therapeutic time window.

One imaging combination includes: DWI to detect lesions within minutes of ischemia and FLAIR which is sensitive for subacute ischemia. Hence a positive DWI with FLAIR negative mismatch (94% specific) is likely to provide a time window for safe thrombolysis . An imaging combination of DW-PWI mismatch helps identify the tissue at risk, and potential benefit from interventions even if outside the 3-4.5 hour window.

See also

• CT of the head in stroke