Diffusion MRI Detects Some TIAs

Abstracts & Commentary

Sources: Kidwell CS, et al. Diffusion MRI in patients with transient ischemic attacks. Stroke 1999;30:1174-1180; Ay H, et al. Normal diffusion-weighted MRI during stroke-like deficits. Neurology 1999;52:1784-1792; Marks MP, et al. Evaluation of early reperfusion and IV tPA therapy using diffusion- and perfusion-weighted MRI. Neurology 1999; 52:1792-1798.

Diffusion-weighted magnetic resonance imaging (dwi) is now being used routinely for the evaluation of acute stroke at many centers throughout North America and Europe. Recent studies indicate that DWI may also provide clinically relevant information about transient ischemic events (TIAs), although perhaps not in every case.

Kidwell and colleagues collected DWI data on 42 consecutive patients with TIA symptoms, and found diffusion abnormalities related to neurologic symptoms in 20 cases (48%). The TIA-associated diffusion anomalies tended to be less voluminous and less conspicuous than those typically seen in stroke patients. TIAs with positive DWI findings averaged 7.3 hours in duration, while those without DWI findings lasted a mean of 3.2 hours, suggesting that diffusion abnormalities are more likely to be observed in TIAs of longer duration. Nine of the 20 patients with positive DWI scans had follow-up studies, and nearly half of those were found to have ischemic strokes in the affected region. In the remainder, symptoms and diffusion abnormalities completely resolved, indicating that an abnormal diffusion study does not necessarily herald the later development of stroke.

Ay and colleagues studied the implications of obtaining a normal DWI in the context of stroke-like deficits. They reviewed 782 consecutive imaging studies and identified 27 cases in which the DWI was normal, despite the persistence of neurologic symptoms during the scanning period. Using all available clinical and radiologic data, Ay et al concluded that conditions other than stroke were responsible for 37% of cases, while ischemic events were the most likely etiology in the remaining 63%. The nonischemic causes included migraine, seizures, brain tumor, transient global amnesia, and psychiatric disorders. They concluded that more than half of the cases of stroke-like symptoms with a normal DWI will, nevertheless, have an ischemic cause for their symptoms. Ay et al suggest that absence of DWI abnormalities in a symptomatic patient should trigger a search for other causes than ischemia. Ay et al also found perfusion-weighted MRI helped to identify some patients with ischemic events who had normal DWI.

Marks and colleagues used the combination of DWI and perfusion-weighted imaging to study 12 patients receiving recombinant tissue plasminogen activator (rTPA) for treatment of acute stroke. The imaging interval between averaged three to five hours from baseline to rescanning post-rTPA. Marks et al also scanned six patients undergoing stroke who did not receive rTPA for purposes of comparison. Six patients had normalization of the perfusion scan within 24 hours, and five of those six had received rTPA. Early signs of reperfusion were seen more frequently in patients who received intravenous rTPA than those who did not. Marks et al conclude that DWI and perfusion imaging may help to guide thrombolytic therapy as well as other aspects of acute stroke management.


DWI is sensitive to the limitations on free diffusion of water imposed by microscopic barriers in biological tissue. Freely diffusing water in CSF appears dark, while water in highly structured areas, such as white matter, often appears bright. Increased signal intensity is also observed within minutes of the onset of cerebral ischemia. Another MRI technique known as perfusion imaging (PI) is sensitive to blood flow at the capillary level. Perfusion studies can now be accomplished without injection of contrast agents in 10 minutes or less.

Large magnetic field gradients are needed to perform DWI and PI, and rapid imaging protocols such as the echoplanar technique are suitable. As such, these capabilities are not available to the thousands of neurologists using clinical scanners that are more than a few years old. This situation is likely to change as newer generations of clinical scanners are installed and older scanners are updated.

Studies of DWI in stroke and now TIA are encouraging and indicate the likelihood of an increasing clinical role in the future. The fact that not all TIAs are associated with diffusion abnormalities is hardly surprising, since some events will be too small or too transient to be detected within the resolution of these types of scans, which are typically several-fold lower in spatial resolution than standard structural MRIs. This means that clinicians do not learn that much from a negative diffusion scan, but may gain more information when the scan is positive and correlates with the neurologic symptoms. Not all diffusion abnormalities are stroke-related and, in this context, the combination of diffusion and perfusion techniques and clinical acumen may be useful in determining which events are of ischemic etiology. Serial imaging studies demonstrate that diffusion changes in the context of a TIA do not necessarily indicate irreversible ischemia. This being the case, DWI could help in acute stroke management above and beyond what can be accomplished with conventional structural MRI techniques. —nrr

Abnormalities on a diffusion-weighted MRI in a patient with TIA symptoms:

a. are an early indication of irreversible cerebral ischemia.

b. are most often observed in patients with TIAs of less than one hour duration.

c. tend to be less conspicuous and less voluminous than those seen in stroke.

d. should trigger a search for causes other than ischemia.