Stroke neuroimaging.

The year in neuroimaging of stroke began with a lively response to a paper published at the end of 2001 by the NINDS rt-PA Stroke Study Group regarding the controversial topic of the clinical significance of early subtle CT signs of ischemia in response to tPA therapy.1 That paper continued the series of post-hoc analyses coming from the NINDS rt-PA clinical trial data to support the position that all patients selected by the criteria used for enrollment in these trials may benefit from acute rtPA therapy, regardless of specific baseline features. This study by Patel et al1 was featured at the 2002 International Stroke Conference and engendered a passionate discussion both at the conference and in the literature2 on the use of early CT ischemia signs as potentially exclusionary criteria. The study reviewed CT films from the trial, generated from scanners of the early 1990s. After adjustment for group imbalances in several baseline clinical features, most notably clinical severity by NIH Stroke Scale, early ischemic changes were not predictive of symptomatic intracerebral hemorrhage, death at 90 days, or clinical deterioration at 24 hours. A marginal association was observed between overall 90-day outcome and early ischemic changes. When compared with the placebo patients without ischemic changes, only rtPA–treated patients without CT changes had favorable 90-day outcomes; however, within the subgroup of patients with most extensive CT changes, there did remain a treatment effect on the 90-day Rankin score. Modern scanners and analysis of the digital images are likely to be more sensitive and accurate in detecting early ischemic changes and may lead to less ambiguous conclusions. Until then, the question of whether extensive early ischemic changes on CT identify poor responders to standard rtPA therapy remains an open one. The potential of other CT measures of early ischemia measures, most notably CT perfusion, have been increasingly explored in 2002. The diagnostic value of CT angiography (CTA) source images, which reflect the distribution and concentration of contrast in the regions within a slice, have been suggested as superior to noncontrast CT and comparable to lesion volume measurements on diffusion MRI (DWI),3–5 but a comparison study with PET raised a cautionary note against overestimating the diagnostic accuracy of this approach.6 The CTA source image method gives a qualitative assessment most related to cerebral blood volume, and poor blood volume by any method is a reliable predictor of infarct, thus the method will likely prove to provide valid diagnostic information. More direct measures of cerebral blood flow with contrast CT look promising for measuring acute hemodynamic compromise and predicting infarct risk.5,7–9 Although conceptually appealing as an add-on to routine emergency noncontrast CT, the contrast perfusion CT technology is still limited by restricted slice coverage. As with any technical advance, enthusiasm and advocacy run many years ahead of validation and general application. The studies at present are few, and conclusions are preliminary. Additional technical developments and assessment of the impact of CT contrast perfusion methods on acute stroke therapeutics will need to be assessed in future years. Monitoring of patients for 2 hours after initiation of thrombolytic therapy with transcranial Doppler ultrasonography, Alexandrov and Grotta10 described reocclusion in 34% of patients with any evidence of early recanalization. This reocclusion was associated with better outcome than patients with stable occlusions but associated with worse outcome than patients with stable recanalization. This study provides a rationale for the development of supplemental therapy to improve recanalization rates over standard rtPA therapy. The coming of age of MRI diffusion and perfusion methods is evidenced by their use as the gold standard against which the above-referenced CT perfusion studies are compared. A large consecutive series of 500 acute stroke admissions evaluated by CT and DWI and retrospectively analyzed confirmed the sensitivity and specificity of DWI of 90% in the overall sample as well as the 6-hour subgroup, greater than that of CT and conventional MR.11 The first prospective, randomized comparison of CT and DWI 6 hours from onset in 50 patients confirmed the 90% accuracy and superior interrater reliability of DWI.12 The superiority to CT was most pronounced in the less experienced readers (residents), a finding of greatest relevance to the real-world use of emergency neuroimaging of stroke.12 The potential of MRI diffusion and perfusion in selecting patients for thrombolysis beyond 3 hours and evaluating the tissue effects of reperfusion was the subject of 3 preliminary yet influential studies. An open-label pilot study reported by Parsons et al13 suggested that patient selection by diffusion and perfusion MRI and MRI evaluation of response to treatment may identify the patients in whom intravenous rtPA therapy will be of clinical benefit when therapy is initiated between 3 and 6 hours. The subgroup of 16 patients with perfusion-diffusion mismatch in that study had a statistically significant degree of recanalization, reperfusion, and tissue salvage relative to untreated historical controls. It is provocThe opinions expressed in this editorial are not necessarily those of the editors or of the American Stroke Association. Received December 11, 2002; accepted December 11, 2002. From the Stroke Branch, National Institute of Neurological Disorders and Stroke, Bethesda, Md. Correspondence to Steven Warach, MD, PhD, National Institute of Neurological Disorders and Stroke, Stroke Branch, 36 Convent Drive, MSC 4129, Room 4A03, Bethesda, MD 20892-4129. E-mail [email protected] (Stroke. 2003;34:345-347.) © 2003 American Heart Association, Inc.