Peter Schellinger Discusses the Role of Diffusion- and Perfusion-Weighted Imaging for Patients with Acute Ischemic Stroke

July 12, 2010


A new AAN guideline "Assessment: The role of diffusion and perfusion MRI for the diagnosis of acute ischemic stroke" was published in the July 13, 2010, issue of Neurology (2010;75:177–185). Peter Schellinger, MD, lead author of the guideline, with the Johannes Wesling Clinical Center in Minden, Germany, spoke with José G. Merino, MD, Science Editor of, about the role of diffusion- and perfusion-weighted imaging for patients with acute ischemic stroke. What was the reason for doing this assessment? What specific questions did the panel address?

Schellinger: We wanted to determine the diagnostic accuracy of MRI, and whether its broader use for acute stroke evaluation was warranted. We looked at papers that assessed whether diffusion-weighted imaging (DWI) and perfusion-weighted imaging (PWI) are sensitive and specific in the diagnosis of acute ischemic stroke, how they compare to CT, and whether the lesion volumes on DWI or PWI correlate with initial clinical severity, final infarct size, and clinical outcome. What did you find about the value of DWI in the evaluation of patients with acute ischemic stroke? How did DWI compare to CT in this setting?

Schellinger: We found four studies that looked at the sensitivity and specificity of DWI for the diagnosis of acute ischemic stroke within 12 hours of symptom onset; one was a Class I study, and three were Class II. DWI was clearly superior for the diagnosis of acute ischemic stroke within 12 hours after symptom onset compared to CT. In the Class I study, for example, among 221 patients, the majority of readers identified patients with acute ischemic stroke three times more often using DWI compared with CT. What is the sensitivity of DWI for acute ischemic stroke? What factors affect this sensitivity? Does the sensitivity of DWI change when the time from symptom onset to scan increases?

Schellinger: I can answer these questions only partially. To assess the true sensitivity of DWI for ischemic stroke we first need to define ischemic stroke. The sensitivity of DWI is influenced by the duration and severity of ischemia, and by the size and location of the stroke. In the neuropathological sense (and this would be the best comparator) any new brain lesion caused by ischemia should be called a stroke regardless of the duration of clinical symptoms. Studies that evaluate imaging modalities, however, cannot use neuropathology as the gold standard. Studies looking at the diagnostic properties of MRI have used a clinical diagnosis of stroke at the time of discharge, made after considering all the relevant and clinical data, as the comparator. In the Class I study of 221 patients scanned within 12 hours of onset alluded to earlier, for example, the majority of readers correctly identified patients with a discharge diagnosis of acute ischemic stroke more than three times as often when reviewing the baseline DWI than the CT (in 94 versus 22 patients). In a subset of 90 patients scanned within three hours of onset, the sensitivity, specificity, and accuracy of DWI were 77 percent, 96 percent, and 86 percent, respectively. For CT, these were 16 percent, 97 percent, and 55 percent, respectively. Is PWI useful for the diagnosis of acute ischemic stroke? How can PWI help in the management of acute stroke patients? Is the role of PWI different when patients are in the thrombolytic time window?

Schellinger: Again, this is a difficult question. In general, PWI is less useful for the diagnosis of acute ischemic stroke than DWI, and our assessment showed that more studies are needed to establish the use of PWI for diagnosing acute ischemic stroke. PWI provides pathophysiological information that is crucial to understand the patient's acute problem. For example, in some stroke patients who are imaged very soon after the onset of ischemia, DWI may not detect the stroke, but PWI clearly shows the hypoperfused tissue, and MR angiography shows a vessel occlusion. Keep in mind, however, that the spatial resolution of PWI is lower, the technique is less standardized, and absolute perfusion values can only be approximated because the so-called arterial input function cannot yet be measured. When evaluating a patient, I like to look at perfusion images to see the larger picture, but I do not rely on perfusion imaging alone to make the diagnosis of stroke.

MRI-based decision making for thrombolysis (or any other novel treatment) beyond the established time windows is the focus of ongoing research, and some experts in the field have used MRI as the initial imaging modality for this indication for more than 10 years. If MRI is not immediately available, however, patients arriving within the time window for thrombolytic therapy should initially be imaged with CT, to avoid unacceptable delays to treatment. An MRI can then be done while the thrombolytics are infusing, especially if there are unanswered questions regarding vessel status and stroke etiology. What is the role of acute DWI and PWI for prognosis? Is the use of DWI and PWI for prognosis limited to clinical studies, or can they be used in this regard in the acute clinical setting?

Schellinger: To a certain extent, all the MR sequences, including PWI, DWI, gradient-echo recalled imaging, fluid-attenuated inversion recovery, MR angiography with and without contrast, and even spectroscopy provide information that can help guide clinical management and estimate the prognosis. In the assessment, we found that, in anterior-circulation strokes, baseline DWI volume probably predicts baseline clinical stroke severity and final lesion volume and is possibly accurate in predicting clinical outcome. On the basis of your assessment, what is the ideal imaging modality for patients with acute stroke?

Schellinger: In general, MRI using DWI and other sequences is the superior diagnostic tool in acute stroke patients. However, in patients who have medical contraindications to MRI or in whom the question is solely the presence or absence of brain hemorrhage, CT is an acceptable alternative. What are the barriers to wider use of MRI in the acute setting, and how may these be overcome?

Schellinger: Frankly, the main problems are availability, attitude, and costs. In addition, not every patient can have an MRI (think about pacemakers or other implants). A dedicated stroke team with emergency MRI access can investigate about 80 to 90 percent of their patients with an MRI. The first step to ensure wider access to acute stroke MRI, in my opinion, is to convince our fellow emergency physicians and radiologists and the hospital administrators that there is a superior imaging modality to CT and that this applies to most but not all patients. The availability of dedicated stroke teams and neurologists 24/7 in the emergency room and dedicated stroke teams within a comprehensive stroke service can help establish new diagnostic strategies and overcome logistical resistance. Training and experience will soften the barriers to wider use of MRI in acute stroke, as will a growing demand of the patients and their families to get state-of-the-art diagnostic studies. Last but not least, clear clinical pathways help to guide colleagues and patients; these should be reviewed on a regular and interdisciplinary basis. Do you predict that the publication of this assessment will change practice?

Schellinger: I certainly hope it will, because MRI is a very useful tool for acute stroke evaluation. Research is needed to optimize advanced MRI techniques in this setting, and to show beyond doubt that improved diagnostics also result in better clinical outcomes that justify the higher cost.