Page 44

Karger_ESC London_2013

22. European Stroke Conference 2 Brain imaging A 14:40 - 14:50 Quantitative measurements of relative fluid-attenuated inversion recovery (FLAIR) signal intensities in acute stroke do not improve accuracy of estimation of lesion age B. Cheng1, M. Brinkmann2, N.D. Forkert3, A. Treszl4, M. Ebinger5, M. Köhrmann6, O. Wu7, D.W. Kang8, D.S. Liebeskind9, T. Tourdias10, O.C. Singer11, S. Christensen12, M. Luby13, C. Ger-loff14, G. Thomalla15 STIR and VISTA Imaging Investigators Klinik und Poliklinik für Neurologie, Kopf- und Neurozentrum, Hamburg, GER-MANY1, Klinik und Poliklinik für Neurologie, Kopf- und Neurozentrum, Hamburg, GERMA-NY2, 1aInstitut für ComputationalNeuroscience, Unversitätsklinikum Hamburg-Eppendorf, Hamburg, GERMANY3, 1bInstitut für Medizinische Biometrie und Epidemiologie, Unversi-tätsklinikum Hamburg-Eppendorf, Hamburg, GERMANY4, Centrum für Schlaganfallforschung Berlin, Charité-Universitätsmedizin Berlin, Berlin, GERMANY5, Klinik für Neurologie, Uni-versität Erlangen-Nürnberg, Erlangen, GERMANY6, Athinoula A MartinosCenter for Biomed-ical Imaging, Department of Radiology´, Massachusetts GeneralHospital, Harvard Medical- School, Boston, USA7, Department of Neurology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, SOUTH KOREA8, Department of Neurology, University of Cali-fornia, Oakland, USA9, 7Univ. de Bordeaux, CHU de Bordeaux, Service de NeuroImagerie Diagnostique de Théra-peutique, Bordeaux, FRANCE10, Kliniken für Neuroradiologie, Universitätsklinikum Erlan-gen- Nürnberg, Erlangen, GERMANY11, Department of Neurology, RoyalMelbourneHospital, University of Melbourne, Parkville Victoria, AUSTRALIA12, Institute of Neurological Disor-ders and Stroke, Bethesda, USA13, Klinik und Poliklinik für Neurologie, Kopf- und Neurozen-trum, Hamburg, GERMANY14, Klinik und Poliklinik für Neurologie, Kopf- und Neurozentrum, Hamburg, GERMANY15 Background The use of magnetic resonance imaging in acute stroke patients has yielded prom-ising results for selecting patients with unknown onset time, who benefit from i.v. throbolytic treatment. Recently, it was shown that a “mismatch” between visibility of an ischemic lesion on diffusion weighted imaging (DWI) and missing corresponding hyperintensities on FLAIR identifies patients with time from symptom onset ≤ 4.5 hours with high specificity but only moderate sensitivity and interrater agreement. Methods We included patients previously stud-ied in the PREFLAIR study 1 and refined visual image rating by adding a group of “subtle” FLAIR lesions to the categories previously applied. Additionally, we quantitatively measured relative FLAIR signal intensity (rSI) in three-dimensional, segmented stroke lesion volumes. A classification and regression algorithm (CART) was applied to test for optimal rSI thresholds to predict time from symptom onset ≤ 4.5 hours. Results 399 patients were included in this re-analysis. rSI of FLAIR lesions showed a moderate correlation with time from symptom onset (r=0.382, p<0.001). A FLAIR rSI threshold of <1.0721 predicted symptom onset ≤ 4.5 hours with slightly increased specificity (0.85 vs 0.78) but also slightly decreased sensitivity (0.47 vs 0.58) as compared to visual analysis. Refined visual analysis differentiating between “sub-tle” and “obvious” FLAIR hyperintensities and CART combining information from visual and quantitative analysis also did not improve diagnostic accuracy. Conclusion Use of quantitative FLAIR rSI measurements resulted in a “trade-off” of between sensitivity and specificity regard-ing the prediction of time from symptom onset ≤ 4.5 hours, but did not increase predictive val-ues overall in comparison to visual analysis. However, quantitative measures may be valuable as confirmatory methods in prospective stroke trials applying DWI-FLAIR-mismatch concept. 44 © 2013 S. Karger AG, Basel Scientific Programme


Karger_ESC London_2013
To see the actual publication please follow the link above