Imaging of Central Nervous System Ischemia

INTRODUCTION

Rapid neuroimaging is essential to acute stroke management. Multiple imaging modalities are available for use in this setting depending on the resources and systems of care established at an individual institution. In all cases, imaging must be performed as quickly as possible to allow for efficient administration of acute therapies, including IV thrombolysis (ie, with tissue plasminogen activator [tPA]) and mechanical thrombectomy. This article reviews the imaging modalities commonly used for the evaluation of ischemic stroke.

NONCONTRAST CT

Noncontrast CT of the head is the most common initial imaging modality used in the evaluation of a patient presenting with acute stroke symptoms. It is widely available and fast and has a high sensitivity for detecting intracerebral hemorrhage. Exclusion of intracerebral hemorrhage is a critical step in evaluating a patient’s candidacy for IV tPA, and a noncontrast CT alone is sufficient to make this determination. In geographic areas that utilize mobile stroke units (ambulances equipped with a CT scanner and specialized personnel), IV tPA can be administered before hospital arrival after performance and interpretation of a noncontrast head CT.¹ In addition to ruling out hemorrhage, CT can provide important information about ischemia.

Early Ischemic Changes

Early ischemic changes evident on noncontrast CT include three classic radiographic findings: obscuration of the lentiform nucleus, loss of the gray-white matter differentiation in the insula (insular ribbon sign), and effacement of the cortical sulci (cortical ribbon sign).² An example of these findings can be seen in figure 3-1. When CT is performed within 6 hours of stroke onset, these changes can be seen in approximately 60% of cases.³

IMAGE GALLERY1/13

FIGURE 3-1. Early ischemic changes evident on noncontrast head CT. A-C, Axial noncontrast CT showing early ischemic changes. Obscuration and loss of the gray-white boundary in the right insular cortex (insular ribbon sign) (A, arrow), obscuration of the right lentiform nucleus (B, arrow), and effacement of the right cortical sulci (cortical ribbon sign) (C, arrow). D-F, Axial diffusion-weighted imaging sequences from MRI performed 2 days later, confirming that the hypodense areas seen on the initial noncontrast CT represented ischemia.

Although early ischemic changes are helpful in determining the presence and extent of ischemia⁴ and have been shown to correlate with an increased risk of poor functional outcome,⁵ they should not be used to exclude patients from IV tPA in the standard treatment window of less than 3 hours from last known well time. In a study of 624 patients, early ischemic changes were not independently associated with risk of adverse outcomes after treatment with IV tPA, and the patients who were treated with IV tPA did better regardless of whether or not early ischemic changes were present.⁶

Alberta Stroke Programme Early CT Score

In contrast to IV tPA decision making, evaluation of early ischemic changes can be an important tool in determining a patient’s candidacy for mechanical thrombectomy. In this scenario, the Alberta Stroke Programme Early CT Score (ASPECTS) is commonly used. ASPECTS is a standardized quantitative 10-point grading system intended for use in acute, anterior-circulation ischemic stroke.⁷ The ASPECTS value is calculated by assessment of axial slices of noncontrast CT at two standardized levels—the ganglionic level (the level of the thalamus and basal ganglia) and the supraganglionic level (above the caudate head). At these two levels, 10 regions are identified and evaluated for evidence of early ischemic changes. These regions are illustrated in figure 3-2. If early ischemic changes are present, a score of 0 is assigned to that region. If early ischemic changes are absent, a score of 1 is assigned. These individual region scores are then summed to determine the total ASPECTS value. The maximum score of 10 indicates no evidence of early ischemic changes. Although the original ASPECTS value is calculated on noncontrast CT with 10-mm slice thickness,⁷ recent studies have suggested improved accuracy when the ASPECTS value is calculated on either the CT angiography (CTA) source images⁸ or the contrast-enhanced CT images obtained during a CT perfusion study.⁹

IMAGE GALLERY2/13

FIGURE 3-2. Axial noncontrast CT images showing the 10 specific regions to assess when calculating an Alberta Stroke Programme Early CT Score (ASPECTS). A, Ganglionic-level image showing caudate nucleus (C), internal capsule (IC), lentiform nucleus (L), insular ribbon (In), anterior middle cerebral artery (MCA) cortex (M1), lateral MCA cortex (M2), and posterior MCA cortex (M3). B, Supraganglionic-level image showing anterior MCA cortex (M4), lateral MCA cortex (M5), and posterior MCA cortex (M6).

Of the five landmark randomized controlled trials published in 2015 that demonstrated the benefit of endovascular therapy for large-vessel occlusion stroke, four used neuroimaging to exclude patients with evidence of a large ischemic core in an effort to optimize both efficacy and safety of revascularization.^10-14 Of those four, three used ASPECTS to determine this, with a required score of 6 to 7 or greater for randomization. As a result of these trials, the American Heart Association/American Stroke Association (AHA/ASA) guidelines recommend proceeding with mechanical thrombectomy for patients with large-vessel occlusion and an ASPECTS value of 6 or greater who present within 6 hours of last known well time, without additional advanced imaging.¹⁵ Additional imaging in this clinical scenario may lead to the unnecessary exclusion of patients who may otherwise benefit from treatment. The appropriateness of mechanical thrombectomy for patients with low ASPECTS values (less than or equal to 5) is uncertain and is a topic of ongoing investigation.^16-18 An example of patient selection for thrombectomy using noncontrast CT and CTA is shown in case 3-1.

CASE 3-1

A 62-year-old man with atrial fibrillation not receiving anticoagulation therapy presented after being found on the floor by his family. He was seen by family members in his usual state of health 1 hour previously. He had an initial National Institutes of Health (NIH) Stroke Scale score of 20, with examination notable for left gaze deviation, right visual field loss, right hemiparesis, and aphasia. Initial noncontrast CT was notable for a hyperdense left middle cerebral artery sign (figure 3-3). The Alberta Stroke Programme Early CT Score was determined to be 10. IV tissue plasminogen activator was administered in the CT scanner. While the infusion was running, CT angiography (CTA) was performed and was notable for a left M1 segment occlusion. He was taken for emergent mechanical thrombectomy, which resulted in Thrombolysis in Cerebral Infarction scale grade 3 reperfusion. After stabilization, he was discharged to an acute care rehabilitation facility with an NIH Stroke Scale score of 8.

IMAGE GALLERY3/13

FIGURE 3-3. Imaging of the patient in case 3-1. The initial axial noncontrast CT (A) demonstrates a hyperdense left middle cerebral artery sign. The patient’s left gaze deviation can also be seen on this image. The Alberta Stroke Programme Early CT Score (ASPECTS) was 10, as determined by visual assessment of the ganglionic (B) and supraganglionic (C) levels of the same initial axial noncontrast head CT. Axial source images (D) and coronal maximal intensity projections (E) from the CT angiogram reveal a left M1 segment middle cerebral artery occlusion.

COMMENT

This case demonstrates patient selection for thrombectomy using noncontrast CT and CTA in the early window. This practice is supported by the American Heart Association/American Stroke Association guidelines, which recommend proceeding with mechanical thrombectomy for patients with large-vessel occlusion and an Alberta Stroke Programme Early CT Score of 6 or greater who present within 6 hours of last known well time, without additional advanced imaging.

In the delayed-window thrombectomy trials,^19,20 which included patients presenting at 6 to 24 hours from last known well time, CT perfusion or MRI was used to determine the ischemic core. However, all patients randomly assigned in both studies had noncontrast CT–based ASPECTS values of 7 or greater. Although the AHA/ASA guidelines recommend adherence to the trials’ inclusion criteria (including use of advanced imaging) when selecting patients for thrombectomy beyond 6 hours from last known well time,¹⁵ some institutions use noncontrast CT and the ASPECTS value alone to select patients for endovascular therapy in the delayed window. This practice is often a result of the resources and systems of care established at a given institution. It has been both supported²¹ and refuted²² by recent literature.

Posterior-Circulation Alberta Stroke Programme Early CT Score

A limitation of ASPECTS is that it is designed for use only in anterior-circulation ischemia. As a result, a novel ASPECTS tool intended for use in the posterior circulation has been developed.²³ This score, known as posterior-circulation ASPECTS (pc-ASPECTS), is a quantitative 10-point grading system for use in suspected vertebrobasilar ischemia. Although it can be calculated from noncontrast CT, beam-hardening artifact in the posterior fossa can be limiting, and accuracy is improved when the calculation is based on CTA source images. In this score, eight distinct brain regions are identified: right and left thalamus, right and left cerebellum, right and left occipital lobes, midbrain, and pons. For early ischemic changes evident in the midbrain or pons, two points are deducted from the total score. Each of the other six regions accounts for one point. These regions and their point values are shown in figure 3-4. As with the traditional ASPECTS value, a score of 10 indicates a lack of early ischemic changes. The pc-ASPECTS value has been shown to improve the detection of ischemia and predict functional outcome.²⁴ However, given the less widespread use of pc-ASPECTS and the very high morbidity and mortality of basilar artery occlusion, most centers do not use a specific pc-ASPECTS threshold when selecting patients for mechanical thrombectomy.

IMAGE GALLERY4/13

FIGURE 3-4. Axial noncontrast CT showing the eight distinct brain regions evaluated for evidence of early ischemic changes to derive the posterior-circulation Alberta Stroke Programme Early CT Score (pc-ASPECTS). The superimposed numbers indicate the point values assigned to each region. Specifically, the midbrain and pons each account for two points in the scoring system, whereas the bilateral cerebellar hemispheres, bilateral thalami, and bilateral occipital lobes account for one point each.

Hyperdense Vessel Sign

Unilateral hyperdensity of a proximal large vessel can be seen on noncontrast CT when a thrombus is present within the lumen.²⁵ This is most commonly seen in the proximal middle cerebral artery (MCA), with a reported frequency of 30% to 40% of MCA infarctions, and is highly specific for MCA occlusion.²⁶ Similar findings can be seen with more distal MCA occlusions and in the basilar artery. Examples of hyperdense vessel signs are shown in figure 3-5. Identification of a hyperdense vessel on noncontrast CT should result in the patient’s being treated as if a large-vessel occlusion were present until proven otherwise.

IMAGE GALLERY5/13

FIGURE 3-5. Examples of hyperdense vessel signs on noncontrast CT images. A, Long thrombus in the left M1 segment appearing as a hyperdense vessel (arrow). B, Hyperdense left M2 branch in the Sylvian fissure (arrow). C, Hyperdense basilar artery (arrow). In each of these cases, the patient was confirmed to have a large-vessel occlusion by CT angiography, digital subtraction angiography, or both.

CT ANGIOGRAPHY

CTA is a result of carefully timed CT imaging obtained after administration of an IV bolus of iodinated contrast. To accommodate a high flow rate and optimize timing, the contrast should be administered through a 20-gauge or larger IV catheter in the right antecubital fossa.²⁷ In rare cases, administration of iodinated contrast can cause contrast-induced nephropathy. However, the risk of acquiring contrast-induced nephropathy from a CTA is exceedingly low, especially in patients without a history of renal impairment.^28,29 Awaiting laboratory studies to assess renal function can delay mechanical thrombectomy, which is associated with worse functional outcomes.³⁰ Thus, it is recommended to proceed with CTA before measuring the serum creatinine level in patients eligible for mechanical thrombectomy without known renal impairment.¹⁵ Rarely, a patient is truly unable to safely undergo CTA because of documented severe allergy to iodine contrast. In that circumstance, immediate MR angiography (MRA) or a direct to digital subtraction angiography (DSA) pathway is reasonable if there is a strong clinical suspicion of large-vessel occlusion.

Identification of Large-Vessel Occlusion

The images obtained by CTA capture the contrast in the lumen of the extracranial and intracranial vasculature and can be combined to create two-dimensional maximal-intensity projections and three-dimensional reconstructions. With this technique, an intraluminal thrombus will appear as a lack of contrast opacification in a given vessel segment, creating a filling defect. CTA is extremely accurate for detecting large-vessel occlusion, with sensitivity of 98.4% and specificity of 98.1%.³¹ The interoperator reliability is high, with a Pearson correlation coefficient of 0.951 measured in one study.³² As a result of its availability, efficiency, and accuracy, it has become the standard noninvasive test for identifying large-vessel occlusion.

Multiphase CT Angiography

In addition to identification of large-vessel occlusion, CTA can also be used to assess a patient’s collateral circulation. Traditionally timed CTA images underestimate collateral quality, as intraarterial contrast has not yet arrived in these collateral circulations at the time of image acquisition because of slower flow. To compensate for this, multiphase CTA images can be obtained. With this technique, two additional sets of images are acquired after the arterial time point: one at the peak venous phase and one at the late venous phase.³³ When considered together, these three sets of images obtained at three different time points allow for a more robust assessment of collateral circulation. Multiple different scoring systems are used to assign numeric values to a collateral circulation seen by multiphase CTA, typically on a 4- or 5-point scale. However, practically speaking, these scores are often dichotomized into either “good” or “poor” collaterals. Poor collateral scores have been shown in some studies to predict a poor prognosis.³⁴ Multiphase CTA was used in the ESCAPE (Endovascular Treatment for Small Core and Anterior Circulation Proximal Occlusion) trial to aid in selection of patients for mechanical thrombectomy and, in that context, demonstrated good interrater reliability.¹¹ Examples of “good” and “poor” collateral circulations are shown in figure 3-6 and figure 3-7.

IMAGE GALLERY6/13

FIGURE 3-6. Images from a 74-year-old woman who presented with sudden, witnessed onset of right-sided weakness and aphasia 1 hour before arrival at the hospital. Her initial National Institutes of Health Stroke Scale score was 19. Initial coronal (A) and axial (B) CT angiograms demonstrating a left M1 occlusion (A, arrow) with very poor collateral flow. Initial digital subtraction angiogram confirms the occlusion (C, arrow) and poor collaterals (C, oval). Subsequent digital subtraction angiogram following mechanical thrombectomy demonstrates Thrombolysis in Cerebral Infarction scale grade 3 recanalization (complete reperfusion) that was achieved after two passes (D). E, Follow-up axial noncontrast head CT showing large left middle cerebral artery infarct despite timely reperfusion.

IMAGE GALLERY7/13

FIGURE 3-7. Images from a 66-year-old man who presented with sudden onset of right-sided weakness and aphasia 3 hours before arrival. His initial National Institutes of Health Stroke Scale score was 12. Initial coronal (A) and axial (B) CT angiogram demonstrating a left internal carotid artery terminus occlusion (A, arrow) with good collaterals and retrograde filling of the left middle cerebral artery territory (B, oval). Initial digital subtraction angiogram confirms the occlusion and robust collaterals (C, oval). Subsequent digital subtraction angiogram after mechanical thrombectomy demonstrates Thrombolysis in Cerebral Infarction scale grade 3 recanalization (complete reperfusion) after one pass (D). E, Follow-up diffusion-weighted imaging MRI revealed only a very small resultant infarct (arrow).

CT PERFUSION

Like multiphase CTA, CT perfusion (CTP) imaging is performed by acquiring multiple scans over time following IV administration of iodinated contrast. The number of scans required and the resultant radiation exposure is higher for CTP compared with multiphase CTA.³⁵ The series of images follows the contrast material as it arrives in the arteries, perfuses brain tissue, and washes out through the venous system. With this information, the scanner then determines estimates of cerebral blood flow, cerebral blood volume, and mean transit time. Cerebral blood flow is the amount of blood that travels through a given brain region over time, measured as milliliters per 100 g per minute. Cerebral blood volume is the total volume of blood in a brain region, measured as milliliters per 100 g. Mean transit time is the average time it takes the blood to travel through a given brain region, measured in seconds.³⁶ Taken together, these three measures can be used to determine whether a given brain region is normally perfused, ischemic, or infarcted, as shown in table 3-1.

IMAGE GALLERY8/13

TABLE 3-1. Parameters of Perfusion Imaging Used to Differentiate Normal, Ischemic, and Infarcted Brain Regions

Rate of infarct growth is variable among individuals and is strongly dependent on the presence of collateral circulation.³⁷ In areas with poor collaterals, lack of blood flow results in irreversible metabolic and cellular failure leading to tissue infarction (infarct core). In areas with good collaterals, the tissue will be dysfunctional but not irreversibly infarcted (ischemic penumbra).³⁷ Postprocessing of CTP images creates maps that approximate the size and location of the infarct core and the ischemic penumbra, as shown in figure 3-8. When interpreting these maps, the reader is looking to identify either a mismatch between the size of the core infarct and the size of the ischemic penumbra or a mismatch between the patient’s clinical examination and the size and location of the core infarct.

IMAGE GALLERY9/13

FIGURE 3-8. CT perfusion imaging in acute ischemic stroke. A-D, Images from a single CT perfusion study. An abnormality is seen in the left parietal lobe with prolonged mean transit time (A, blue area) and decreased cerebral blood flow (B, blue area). Cerebral blood volume is decreased centrally but preserved in the periphery of the lesion (C). This pattern is consistent with an area of core infarction surrounded by ischemic penumbra. The perfusion map illustrates the infarcted tissue (D, red area) with a large area of potentially salvageable tissue (D, green area) with timely reperfusion. E-H, Images from a CT perfusion study of a different patient. An abnormality in the right hemisphere demonstrates prolonged mean transit time (E, blue area) and decreased cerebral blood flow (F, blue area). Most of this area also has decreased cerebral blood volume (G, blue area). The perfusion map shows a large core infarct (H, red area) with a relatively small surrounding ischemic penumbra (green area).

Identification of a mismatch suggests reversibility of ischemia with timely reperfusion and therefore has been used to select patients for mechanical thrombectomy beyond 6 hours from stroke onset.^19,20 Using perfusion imaging in this way results in a patient-specific “tissue clock,” as opposed to a standardized time window of eligibility for all patients. This approach of individualized patient selection for delayed mechanical thrombectomy using perfusion imaging is supported by the AHA/ASA guidelines.¹⁵

Although perfusion imaging can be helpful when performed and interpreted correctly, current perfusion techniques have many pitfalls and are affected by both conceptual issues and measurement errors that can result in overestimation of the core infarct.³⁸ There is debate as to the need for perfusion imaging in the selection of patients with acute ischemic stroke.³⁹ Several trials on the endovascular treatment of stroke due to large-vessel occlusion did not use perfusion imaging for patient selection in the early time window and showed large treatment benefit in this population.⁴⁰ In a recent multicenter observational cohort of 1530 patients, there was no difference in outcomes based on modality of imaging selection between CT and CT perfusion imaging.⁴¹

MRI

CT-based studies remain the primary imaging modalities for initial evaluation of most patients worldwide presenting with symptoms of acute stroke.⁴² However, MRI offers some advantages over CT in the evaluation of these patients, and its use as a primary modality has increased in the United States over the past 2 decades.⁴³ MRI is more sensitive and specific than CT for the detection of acute ischemia and is better at identifying stroke mimics, including infectious, inflammatory, tumoral, and traumatic conditions.^44,45

Diffusion-Weighted Imaging

Using diffusion-weighted imaging (DWI), MRI can identify cerebral ischemia as early as a few minutes after the stroke onset.⁴⁶ The sensitivity of MRI to detect ischemic lesions is about 92% when performed at the time of symptom presentation, much higher than that of CT. This sensitivity can increase up to 97.5% if perfusion imaging is included.⁴⁷ DWI is also more accurate than noncontrast CT in determining the core infarct, which usually represents the irreversibly infarcted tissue.⁴⁸ Determining the core infarct is of critical importance in the assessment of potential risk and benefit of reperfusion treatment, especially for those with unknown time of stroke onset or those presenting beyond 6 hours from symptom onset.¹⁹

Acute ischemic lesions typically demonstrate a high signal intensity on DWI, which results from the alteration of the brownian movement of water protons due to cytotoxic edema in the early phases of acute ischemic stroke. DWI images should always be interpreted in conjunction with the apparent diffusion coefficient (ADC), a quantitative measure of the water protons’ diffusion. In true restricted diffusion seen in acute ischemic stroke, the region of increased DWI signal will demonstrate low signal intensity on the ADC sequence (DWI is bright, ADC is dark). An example of this is shown in figure 3-9. In contrast, when a high signal on DWI is associated with high signal intensity on ADC imaging (DWI is bright, ADC is bright), it is T2 shine-through. T2 shine-through occurs when there is increased water content in tissue (as in vasogenic edema or cystic lesions) and is not consistent with acute ischemia. In ischemic lesions, the decreased signal intensity on ADC appears earlier than the increased signal intensity on DWI and therefore is more sensitive in detecting stroke.⁴⁹ The decreased signal intensity on ADC imaging usually persists for 1 week after stroke onset. Therefore, a dark ADC map means that the stroke is less than 1 week old.⁴⁹ A pseudonormalization of the ADC map from low to high signal intensity occurs 1 to 2 weeks after stroke onset.⁴⁹

IMAGE GALLERY10/13

FIGURE 3-9. Imaging markers of acute ischemic stroke on MRI. A, Diffusion-weighted imaging (DWI) with high signal intensity in the region of the right middle cerebral artery. B, This same region also has low signal intensity on the apparent diffusion coefficient (ADC) sequence. The result is an imaging pattern that is “bright” on DWI and “dark” on ADC, consistent with true restricted diffusion.

Although DWI is considered the most sensitive sequence to detect ischemic lesions, a small percentage (6.8%) of patients with true acute ischemic stroke have a negative DWI scan.⁵⁰ DWI-negative stroke is most frequently seen in patients with small, posterior-circulation, or hyperacute strokes.⁵⁰ It is important to be aware of the possibility of DWI-negative stroke and not exclude patients from reperfusion therapies or other stroke workup based on a negative DWI scan. The possibility of false-negative DWI decreases considerably after 3 hours. Repeat DWI is recommended in patients for whom stroke is strongly clinically suspected and with initial negative DWI that was performed within 2 hours of symptom onset.⁵¹ In addition, if there is clinical suspicion of posterior fossa ischemia and negative initial DWI, repeat MRI with coronal DWI acquisition through the posterior fossa is recommended to increase sensitivity.^51,52

T2-Weighted and Fluid-Attenuated Inversion Recovery Sequences

Another MRI finding in ischemic stroke is increased signal intensity on T2-weighted and fluid-attenuated inversion recovery (FLAIR) sequences. Although increased DWI signal intensity appears within the first few hours of stroke because of cytotoxic edema, increased signal intensity on the T2-weighted and FLAIR sequences is delayed and results from the increased concentration of water in the interstitial spaces (interstitial edema). Previous studies have indicated that FLAIR does not show signal changes within the first 4.5 hours.^53,54 This mismatch between DWI and FLAIR sequences (positive DWI, negative FLAIR) has been used to estimate the stroke onset and therefore select patients who may benefit from IV tPA. This is particularly useful in patients who awake with neurologic deficits or patients with unknown last known well times.^53,55 Specifically, patients with acute ischemic stroke of unknown onset who received IV tPA based on a DWI-FLAIR mismatch had significantly better functional outcomes than those who did not receive IV tPA in this setting.⁵³ An example of this MRI-based tPA decision making is shown in case 3-2.

CASE 3-2

A 67-year-old man with a medical history of hypertension, hyperlipidemia, diabetes, and prior strokes presented with right-sided weakness and slurred speech. He was in his usual state of health when he went to sleep the previous night at 10:00 pm, and he noticed these symptoms when he awoke in the morning at 8:00 am. The initial National Institutes of Health Stroke Scale score was 8. He had an immediate noncontrast CT that showed neither acute hemorrhage nor territorial infarction and a CTA that demonstrated patent intracerebral vessels. Subsequent MRI revealed a small area of restricted diffusion in the left corona radiata without an associated T2/fluid-attenuated inversion recovery (FLAIR) signal abnormality (figure 3-10). As a result, IV tissue plasminogen activator was administered, and he was admitted to the neurocritical care unit for post–tissue plasminogen activator monitoring.

IMAGE GALLERY11/13

FIGURE 3-10. Imaging of the patient in case 3-2. Axial MRI images show a small area of restricted diffusion in the left corona radiata (A) without an associated fluid-attenuated inversion recovery (FLAIR) signal abnormality (B).

COMMENT

This case is an example of MRI-based decision making for acute stroke with unknown last known well time. The diffusion-weighted imaging–FLAIR mismatch described is consistent with infarction less than 4.5 hours old and can therefore be used to approximate the time of stroke onset. This approach is supported by the WAKE-UP (Efficacy and Safety of MRI-based Thrombolysis in Wake-up Stroke) study and the American Heart Association/American Stroke Association guidelines.

Additional markers of ischemia on FLAIR sequences include loss of the gray-white matter differentiation with gyral swelling and sulcal effacement. Pseudonormalization of signal intensity on T2-weighted images, known as “fogging,” may occur 1 to 4 weeks after stroke, with a peak around 2 to 3 weeks. Fogging occurs as a result of infiltration of the infarcted tissue by inflammatory cells.⁴⁹ The FLAIR sequence is also useful in detecting subarachnoid hemorrhage, a contraindication to thrombolytic therapy.⁵⁶

Susceptibility-Weighted Sequences

Susceptibility-weighted or gradient recalled echo (GRE) sequences are routinely used in MRI of patients with acute stroke. These sequences are very sensitive for detecting blood products that may not be detected on other MRI sequences or even by CT. Blood products and other ferromagnetic compounds (eg, minerals, calcifications) cause distortion of the local magnetic field, resulting in loss of the MRI signal (hypointensity) with an area of blooming.⁵⁷ With the use of these sequences, MRI is as accurate as CT for the detection of hyperacute hemorrhage.⁵⁸ In addition, MRI may be more accurate than CT for the detection of hemosiderin deposits of chronic intracerebral hemorrhage, which are usually undetected by CT.⁵⁸ This ability to detect hemorrhage with high sensitivity is important for centers using rapid MRI as first-line neuroimaging in patients presenting with symptoms of acute stroke.

Vessel and Clot Imaging

On T2-weighted sequences, patent arteries with normally flowing blood usually appear dark, a finding known as a “flow void.” Therefore, lack of flow due to a vessel occlusion or thrombosis, slow flow due to stenosis, or retrograde collateral flow manifests as a lack of the normal flow void and results in increased signal intensity of the involved vessels on T2-weighted sequences. This arterial hyperintensity, best seen on FLAIR images, is called the “hyperintense vessel sign” and may be the only sign of early infarction.⁵⁶ Additionally, acute intraarterial thrombus produces susceptibility artifact and blooming on GRE or susceptibility-weighted images and is strongly suggestive of large-vessel occlusion, akin to the hyperdense vessel sign on noncontrast CT. These MRI findings of acute infarction are shown in figure 3-11.

IMAGE GALLERY12/13

FIGURE 3-11. Axial MRI findings in acute ischemic stroke. A, Hyperintense vessel signs (arrows) are seen on a fluid-attenuated inversion recovery (FLAIR) sequence as a result of slow or retrograde flow. B, The susceptibility or “blooming sign” (arrow) is seen on a susceptibility-weighted sequence as a result of high red blood cell content in an acute occlusive thrombus in an M2 branch of the right middle cerebral artery. Diffusion-weighted imaging (DWI) sequence (C) shows high signal intensity while the apparent diffusion coefficient (ADC) sequence (D) demonstrates low signal intensity, a pattern consistent with acute ischemic stroke.

The hyperdense vessel sign on noncontrast CT and the susceptibility artifact or “blooming sign” on MRI reflect the high red blood cell content of the occlusive thrombus. Absence of this sign in a patient with large-vessel occlusion may indicate a fibrin-predominant thrombus. This differentiation has clinical implications, as fibrin-rich thrombi represent a potential target for pharmacologic fibrinolysis, whereas red blood cell–rich thrombi may have a better response to stent retriever versus contact aspiration during endovascular treatment.^59,60

MR Angiography

MRA is an essential component of the acute stroke MRI protocol. It helps in determining the location and extent of vascular lesions of the head and neck such as acute occlusion, atherosclerotic disease, dissection, or fibromuscular dysplasia. One disadvantage of noncontrast MRA is the difficulty in distinguishing between stenosis and acute occlusion, as slow or turbulent flow can result in intravoxel phase dispersion, which leads to signal loss and subsequent overestimation of arterial stenosis that may appear as an occlusion on MRA.⁶¹

MR Perfusion Imaging

Analogous to CT perfusion, MR perfusion imaging uses serial consecutive imaging after contrast injection to quantify blood flow and the blood volume through the brain parenchyma. As previously mentioned, the area of restricted diffusion on DWI represents the core infarct. Therefore, the mismatch between the perfusion and diffusion abnormality represents the potentially salvageable ischemic tissue at risk for infarction.⁶²

Limitations of MRI in Acute Stroke

Despite the advantages of MRI for evaluating patients with acute stroke, it has several limitations compared with CT. MRI scanners are more expensive and less widely available than CT scanners. In addition, MRI has more contraindications and requires screening patients for ferromagnetic objects that pose safety concerns in the MRI environment. MRI takes longer to perform than CT because of the multiple sequences required. Moreover, MRI may not be feasible in patients with a diminished level of consciousness, vomiting, agitation, hemodynamic compromise, or hypoxia.⁶³

On a more technical level, a caveat of DWI in estimating the core infarct is the possibility of DWI reversibility. DWI reversibility refers to partial or complete reversal of the initial DWI signal abnormality when compared with follow-up DWI or FLAIR imaging. Given this finding, it is possible that DWI may overestimate the nonreversible ischemic core in the early hours of stroke.⁶⁴ Partial DWI reversibility has been reported to occur in 26.5% of cases in DWI-based studies.⁶⁵ Total DWI reversibility is rare and is estimated to occur in about 0.8% of cases.⁶⁴

DIGITAL SUBTRACTION ANGIOGRAPHY

DSA remains the gold standard modality to evaluate most cerebrovascular diseases. It can accurately determine the type and location of vascular lesions as well as the flow characteristics and collateral circulation in the setting of vascular occlusion or stenosis.

In patients with acute ischemic stroke, DSA is used to confirm and treat stroke due to large-vessel occlusion. After endovascular treatment of a large-vessel occlusion, the degree of resultant reperfusion is radiographically assessed. The modified Thrombolysis in Cerebral Infarction (mTICI) scale is the most commonly used scoring system to describe the degree of reperfusion achieved and has value in predicting outcomes.^66,67 A detailed description and examples of each mTICI score are shown in figure 3-12.⁶⁸ An mTICI score of 2b, 2c, or 3 is considered adequate reperfusion and a successful thrombectomy.

IMAGE GALLERY13/13

FIGURE 3-12. Modified Thrombolysis in Cerebral Infarction (mTICI) scoring system in relation to a left M1 occlusion. A score of TICI 0 indicates no reperfusion beyond the site of occlusion (arrow). TICI 1 indicates recanalization beyond the initial occlusion, but with minimal reperfusion of the distal territories. TICI 2a corresponds to recanalization with less than 50% reperfusion of the distal territories. TICI 2b reflects greater than 50% reperfusion of the distal territories. A score of TICI 2c (not shown) is sometimes used when near complete reperfusion except for a small number of distal cortical vessels occurs. A score of TICI 3 indicates complete total reperfusion of the entire territory distal to the occlusion.Reprinted from Mokin M, et al, Neurosurg Focus.68 © 2014 American Association of Neurological Surgeons.

DSA is also useful as an aid to determine the etiology of a large-vessel occlusion, which can be due to thromboembolism or intracranial atherosclerotic disease.⁶⁹ Knowing the occlusion type is important, as large-vessel occlusion due to intracranial atherosclerotic disease requires specific endovascular modalities to achieve successful recanalization⁷⁰ as well as appropriate secondary stroke prevention. Findings suggestive of large-vessel occlusion due to intracranial atherosclerotic disease include residual stenosis after thrombectomy, truncal-type occlusion (an arterial occlusion found at the middle of an artery with visible distal major branches and bifurcation site beyond occlusion), robust collateral circulation, and microcatheter first-pass effect (blood flow through the occlusion after the withdrawal of the microcatheter).⁷¹ DSA also provides valuable information about the collateral circulation that is important to maintain perfusion downstream from arterial occlusions and determine the pace of infarct evolution.⁷² In tandem occlusions or in the setting of a nonopacified carotid artery on CTA, DSA can accurately distinguish true cervical internal carotid artery occlusion from pseudo-occlusion secondary to distal thrombosis that impedes ascending blood flow.⁷³

CONCLUSION

There are multiple acceptable imaging approaches when evaluating a patient with symptoms of acute ischemic stroke. Given its wide availability, speed, and safety, CT-based imaging is the first step in the vast majority of centers. Noncontrast head CT alone is sufficient for IV thrombolysis decision making in the appropriate clinical context. CTA is extremely sensitive for detection of large-vessel occlusion and is a critical step for patients presenting with clinical symptoms consistent with this syndrome. Advanced imaging including multiphase CTA, CTP, MRI, or MR perfusion can provide additional information useful for therapeutic decision making in specific clinical scenarios described above. In all cases, it is paramount that neuroimaging be performed and accurately interpreted as quickly as possible to allow for timely reperfusion therapy for all who are eligible.

KEY POINTS

• Early ischemic changes evident on noncontrast CT include obscuration of the lentiform nucleus, loss of the gray-white boundary in the insula (insular ribbon sign), and effacement of the cortical sulci (cortical ribbon sign).
• The Alberta Stroke Programme Early CT Score (ASPECTS) is a quantitative assessment of early ischemic changes calculated by visual inspection of 10 specific neuroanatomic regions on noncontrast CT. The maximum score of 10 indicates no evidence of early ischemia.
• The American Heart Association/American Stroke Association guidelines recommend proceeding with mechanical thrombectomy for patients with large-vessel occlusion stroke and an Alberta Stroke Programme Early CT Score (ASPECTS) of 6 or greater who present within 6 hours of last known well time, without additional advanced imaging.
• The posterior-circulation Alberta Stroke Programme Early CT Score (ASPECTS) is a 10-point scale that assesses eight brain regions supplied by the vertebrobasilar system for evidence of early ischemic changes. It has been shown to improve detection of ischemia and predict functional outcome.
• The hyperdense vessel sign can be seen on noncontrast CT and is highly specific for large-vessel occlusion.
• It is recommended to proceed with CT angiography before measuring a serum creatinine level in patients eligible for mechanical thrombectomy without known renal impairment to avoid unnecessary delays in reperfusion.
• CT angiography is extremely accurate for detecting large-vessel occlusion, with sensitivity and specificity of approximately 98%.
• Multiphase CT angiography can be used to obtain a more robust assessment of a patient’s collateral circulation. Quality of collateral flow has been shown to correlate with rate of infarct growth and predict prognosis in some studies.
• CT perfusion uses three parameters to assess a given brain region: cerebral blood flow, cerebral blood volume, and mean transit time. Postprocessing software creates maps based on these measures to approximate the size and location of the infarct core and the ischemic penumbra.
• MRI is more sensitive and specific than CT for the identification of acute stroke and can detect ischemia as early as a few minutes after stroke onset.
• In true restricted diffusion seen in acute ischemic stroke, a region of increased diffusion-weighted imaging signal correlates with a region of low signal intensity on the apparent diffusion coefficient image.
• Diffusion-weighted imaging–negative stroke is rare and most frequently seen in patients with small, posterior-circulation, or hyperacute strokes. Repeat diffusion-weighted imaging is recommended if there is a strong clinical suspicion of ischemia.
• Diffusion-weighted imaging positivity appears within the first few minutes after stroke onset, whereas fluid-attenuated inversion recovery (FLAIR) signal changes take longer to develop. This mismatch has been used to estimate stroke onset and select patients for thrombolysis.
• Acute intraarterial thrombus produces susceptibility artifact and blooming on gradient recalled echo or susceptibility-weighted MRI. This finding is similar to the “hyperdense vessel sign” seen on noncontrast CT and is strongly suggestive of large-vessel occlusion.
• The modified Thrombolysis in Cerebral Infarction scale is used to describe the degree of reperfusion achieved after mechanical thrombectomy. A score of 2b, 2c, or 3 is considered successful reperfusion.