Rivka Rachel Colen MD, Presenter: Nothing to Disclose

William A, Copen MD, Abstract Co-Author: Nothing to Disclose

Ramon Gilberto Gonzalez MD, Abstract Co-Author: Nothing to Disclose

Pamela Whitney Schaefer MD, Abstract Co-Author: Nothing to Disclose

Perfusion-weighted MRI (PWI) provides valuable information in the emergent care of acute stroke patients, but may be underutilized because the time required to produce deconvolution-based perfusion maps may preclude immediate assessment of whether a mismatch exists between lesions on diffusion-weighted (DWI) and PWI images.

44 emergency room patients with suspected acute stroke underwent DWI and PWI using a protocol requiring less than 5 minutes of imaging time. Maps of mean time to enhancement (MTE), a non-deconvolution-based approximation of mean transit time (MTT), were prepared rapidly, usually before the patient left the scanner. Subsequently, MTT maps were prepared using a more time-consuming deconvolution-based algorithm. Two radiologists (R1, R2) reviewed both sets of images retrospectively, to determine whether either or both PWI maps showed a DWI/PWI mismatch measuring at least 20% of the volume of a DWI lesion, or 2 cm for scans with no apparent DWI lesion.

R1 and R2 found DWI lesions in 23 and 22 scans, respectively. Among unanimously DWI-negative studies, R1 and R2 found MTT abnormalities measuring at least 2 cm in 6 and 7 cases, respectively. R1 and R2 disagreed regarding whether MTT maps showed significant DWI/PWI mismatch in 4/44 cases. In 2 other cases, one or both radiologists found MTT maps uninterpretable. In 2 other cases, the radiologists agreed that an MTT abnormality was present, but one was not sure if it measured more than 20% larger than the DWI abnormality. Among the other 36 studies in which both radiologists agreed that MTT maps did or did not show a significant mismatch, the sensitivity/specificity of the rapid MTE maps for detecting that mismatch was 87%/85% for R1, and 93%/95% for R2.

PWI identifies abnormalities in a substantial number of patients with suspected acute stroke, even when DWI fails to show an infarct. Subjective assessment of whether an MTT abnormality is substantially mismatched with a DWI lesion is somewhat variable, even for optimally produced MTT maps. However, among cases in which there is interrater agreement, rapidly produced maps can provide a high level of accuracy for detection of DWI/PWI mismatch.

Colen, R, Copen, W, Gonzalez, R, Schaefer, P, Emergency Imaging of Acute Stroke Patients: Comparison of Immediately Produced Perfusion-weighted MR Images to Those Produced by Deconvolution.  Radiological Society of North America 2005 Scientific Assembly and Annual Meeting, November 27 - December 2, 2005 ,Chicago IL. http://archive.rsna.org/2005/4420418.html