İnvestigation of the ability of the diffusion weighted MR imaging in differentiation of TOAST ischemic stroke subtypes

8.2 NG L ZCE ÖZET Dr Tolga YiğitkanlıINVESTIGATION OF THE ABILITY OF THE DIFFUSIONWEIGHTED MR IMAGING IN DIFFERENTIATION OFTOAST ISCHEMIC STROKE SUBTYPESPURPOSEThe aim of this study was to test the ability of diffusion weighted MR imaging indiscriminating the stroke subtypes in terms of TOAST criteria.MATERIALS AND METHODSWe retrospectively analyzed 398 patients with clinical symptoms of acuteischemic stroke and acute ischemic lesions on diffusion weighted MRI. Lesions ondiffüsion weighted imaging were classified as a single lesion, scattered lesions in onevascular territory and multiple lesions in more than one vascular territory. We alsoanalyzed patients who had internal and cortical watershed infarcts. Stroke subtype wasdetermined by the TOAST classification criteria. One stroke subtype and diffusionweighted MRI infarct pattern were assigned for each patient. We evaluated theassociation of overall 11 infarct patterns with stroke subtypes. We also evaluatedassociation of watershed infarcts with stroke subtypes. Pearson χ2 test and Fisher exacttest were used, a level of p< 0.05 was considered to be significant.RESULTSThe most common underlying stroke subtype was cardioembolism in 131(32.9%) patients. Large-artery arterosclerosis was diagnosed in 44 (11.1 %) patients,small artery occlusion in 103 (25.9 %) patients, other etiologies in 16 (4 %) patients,undetermined cause in 95 (23.9 %) patients, and two or more competing causes in 9 (2.2%) patients. The association of single corticosubcortical lesion with cardioembolism(p<0.000), single subcortical lesion < 15 mm with small artery occlusion (p<0.000),multiple lesions in the unilateral anterior circulation with large-artery aterosclerosis(p<0.000), multiple lesions in the posterior circulation with cardioembolism (p< 0.005),and multiple lesions in the bilateral anterior circulation with cardioembolism (p< 0.017)were statistically significant. The association of internal and cortical watershed infarctswith large arterosclerosis (p<0.000) were also statiscally significant.CONCLUSIONDiffusion weighted MRI is not satisfactory alone but may be useful indiscriminating the ischemic stroke subtypes.