Clinically, stroke is the sudden onset of focal neurologic symptoms due to ischemia (88%), hemorrhage into the brain (9%) or hemorrhage into the subarachnoid spaces (3%).1 There are approximately 700,000 new or recurrent strokes per year in the United States, one every 45 seconds, resulting in approximately 160,000 deaths, making stroke the third leading underlying cause of death in the U.S., behind heart disease and cancer, and the leading cause of long-term disability.1 The estimated direct and indirect cost for stroke in the U.S. in 2006 was $57.9 billion.1
Noninvasive imaging plays an important role in identifying risk factors for stroke. Evaluation of the carotid arteries with invasive catheter arteriography is being replaced by duplex ultrasound (US), CT angiography (CTA), MR angiography (MRA) and time resolved contrast enhanced MRA (CE-MRA), all of which show high diagnostic accuracy for internal carotid artery (ICA) stenoses of 70%–99%.2–5 US is the most cost-effective screening method, increasingly combined with CE-MRA for complete evaluation. Measurement of reduced cerebral vascular reserve (CVR) or elevated oxygen extraction fraction (OEF) by using nuclear, CT and MR imaging techniques may further define risk but these methods are not yet in wide clinical use (Table 1).
On the basis of ready availability and high sensitivity to acute hemorrhage, non-contrast CT historically has been the preferred modality for initial imaging of acute stroke, but has lacked high sensitivity to acute ischemia and infarction. Resurgence in the use of CT for acute ischemic stroke evaluation has occurred with the increasing clinical availability of CT perfusion (CTP), resulting in high sensitivity to early perfusion deficits, detectable prior to low density abnormalities on non-contrast CT.6,7 Quantitative CTP measurements of CBF, CBV, MTT and TTP may also discriminate between infarct and salvageable “ischemic penumbra.”8,9 The wide availability of these techniques plus the ability to quickly identify acute hemorrhage and vascular lesions using CTA, have been suggested as the key advantages of CT over MR imaging for acute stroke evaluation. However, the limited volume coverage of CTP (currently restricted to a 2 or 4 cm slab, the width of the detector array), the greater risks of reaction or fluid overload from iodinated contrast materials, and the lack of a direct measure of cellular viability like diffusion mitigate these advantages over MR imaging10 (Table 1).
MR imaging diffusion weighted imaging (DWI) demonstrates diffusion “restriction” (reduced apparent diffusion coefficient, ADC), in acute cerebral ischemic injury within minutes of the precipitating ictus. Numerous direct comparisons to non-contrast CT have shown DWI to be much more sensitive to acute ischemia (80%–100% sensitivity compared to 25%–75% sensitivity for CT) with the greatest differences seen in the first 3-hours after symptom onset.11,12 The addition of blood flow parameters (CBF, CBV, MTT and TTP) obtainable with gadolinium dynamic susceptibility contrast techniques (DSC or perfusion-weighted imaging, PWI) also allows MR imaging to identify the early “ischemic penumbra” as the reduced perfusion zone outside the DWI positive zone, the perfusion-diffusion (PWI-DWI) “mismatch.”13,14 Combining this with vascular imaging (MRA) has made MR imaging an appealing tool for diagnosis and treatment monitoring of acute ischemic cerebrovascular disease. Enthusiasm for MR imaging in acute stroke evaluation has historically been dampened by the variable and confounding appearance of acute hemorrhage. Recent experience using T2* (gradient echo) imaging to detect low signal parenchymal hemorrhage and FLAIR scans to detect high signal subarachnoid blood as well as BBB disruption that may predict infarct hemorrhagic transformation have helped renew interest in MR imaging as a first-line modality in patients with acute stroke.15–18 However, there is currently insufficient widespread clinical experience to recommend MR imaging over CT for routine exclusion of intracranial hemorrhage.19 It is also important to emphasize the issue of availability of MR imaging in the therapeutic time window and potential contraindications: patients with pacemakers, cerebral aneurysm clips, ocular foreign bodies or cochlear implants and those suffering from claustrophobia, or morbid obesity (320 lbs)20 (Table 1).
Because most transient ischemic neurologic symptoms actually last for one hour or less and 50% or more show tissue injury on DWI21,22, a new definition of TIA has been proposed as “a brief episode of neurologic dysfunction presumptively caused by focal brain or retinal ischemia, typically lasting less than one hour, without neuroimaging evidence of acute infarction.”23,24 This change reflects the growing emphasis on the earliest possible diagnosis and treatment of acute ischemia and the use of MR imaging and CT for definitive diagnosis and exclusion of hemorrhage.19 The current recommended clinical practice in the U.S. is treatment with intravenous recombinant tissue plasminogen activator (rtPA), within 1 hour and no later than 3 hours after symptom onset, following the exclusion of intracerebral hemorrhage by a non-contrast CT scan.25–27 However, only 20%–25% of admissions typically arrive at the emergency department within 3 hours of symptom onset and, following appropriate exclusions, successful treatment with rtPA, without symptomatic major hemorrhage, is limited to 3%–8.5% of ischemic stroke admissions.25,28 Accordingly, current thrombolysis and clot extraction research trials are focused on expanding this time-to-treatment window with individualized, patient-specific therapy guided by the MR imaging PWI-DWI mismatch as a surrogate biomarker for the salvageable ischemic penumbra29–32, one demonstrating successful treatment as late as 9 hours after ictus.33 Because the diffusion restriction and mismatch zones may also include underperfused but metabolically stable “oligemic” tissue, BOLD-MR imaging, SPECT and PET methods may be more widely used in the future to specifically identify the treatable ischemic penumbra by elevated OEF in low CBF “misery perfusion” regions or by using specific markers of cellular oxidative stress and ischemic injury (Table 1).
As CT is almost universally available and is highly sensitive to acute intracranial hemorrhage, it remains the mainstay in emergent evaluation of suspected epidural, subdural, subarachnoid and parenchymal hemorrhage. Catheter arteriography also remains the definitive diagnostic test for the source of intracranial hemorrhage, especially subarachnoid hemorrhage (SAH), with sensitivity to all cerebral aneurysms greater than 90%, decreasing to 80% in the setting of acute SAH.34 However, recent comparisons of CTA to catheter arteriography in SAH patients have shown overall CTA aneurysm detection sensitivities of 85%–95%, declining for smaller aneurysms to approximately 50% for those under 2 mm diameter.35–37 Treatment of intracranial aneurysms following SAH is increasingly based on CTA alone38,39 and the evaluation of post-SAH vasospasm and ischemia are increasingly investigated with transcranial Doppler (TCD), CTA and CTP in place of catheter angiography and SPECT.40,41 Follow-up of clipped or coiled aneurysms for residual filling remains definitive with catheter arteriography but there is a growing interest in less invasive techniques, especially dynamic bolus CE-MRA (eg, TRICKS).42–44 Intracranial hemorrhage may also be caused by arteriovenous vascular malformations and fistulae which are currently also best assessed by catheter arteriography. Existing time-resolved CE-MRA and multi-section CTA methods do not yet match the combined high spatial and temporal resolution of catheter arteriography for primary diagnosis of arteriovenous malformations but may useful for treatment follow-up45,46 (Table 2).
Review Information
This guideline was originally developed in 1996. The last review and update was completed in 2007.
Appendix
Expert Panel on Neurologic Imaging: Robert L. DeLaPaz, MD, Principal Author, Columbia University Medical Center, New York, NY; David J. Seidenwurm, MD, Panel Chair; Patricia C. Davis, MD, Vice-Chair; James A. Brunberg, MD; Pr. Didier Dormont; David B. Hackney, MD; John E. Jordan, MD; John P. Karis, MD; Suresh Kumar Mukherji, MD; Patrick A. Turski, MD; Franz J. Wippold II, MD; Robert D. Zimmerman, MD; Michael A. Sloan, MD, MS, American Academy of Neurology.
Footnotes
This article is a summary of the complete version of this topic which is available on the ACR website at www.acr.org/ac. Practitioners are encouraged to refer to the complete version.
Reprinted with permission of the American College of Radiology.
References
- Copyright © American Society of Neuroradiology