Use of perfusion- and diffusion-weighted imaging in differential diagnosis of acute and chronic ischemic stroke and multiple sclerosis

摘要

Objective: To investigate differences in lesions and surrounding normal appearing white mattern(NAWM) by perfusion-weighted imaging (PWI) and diffusion-weighted imaging (DWI) innpatients with acute and chronic ischemic stroke and multiple sclerosis (MS).nMethods: Study subjects included 45 MS patients, 22 patients with acute ischemic stroke and 20npatients with chronic ischemic stroke. All subjects underwent T2-weighted imaging (WI), flairnattenuated inversion recovery (FLAIR), DWI and dynamic contrast enhanced PWI. Apparentndiffusion coefficient (ADC) and mean transit time (MTT) maps were generated and values werencalculated in the acute and chronic ischemic and demyelinating lesions, and in NAWM forndistances of 5, 10 and 15 mm. Fifty-three acute ischemic and 33 acute demyelinating lesions,nand 775 chronic ischemic and 998 chronic demyelinating lesions, were examined. Univariate,nmultivariate and data mining analyses were used to examine the feasibility of a prediction modelnbetween different lesion types. Correctly and incorrectly classified lesions, true positive (TP),nfalse positive (FP) and precision rates were calculated.nResults: Patients with acute ischemic lesions presented more prolonged mean MTT values innlesions (p50.002) and surrounding NAWM for distances of 5, 10 and 15 mm (all p,0.0001)nthan those with acute demyelinating lesions. In multinomial logistic regression analysis, 65 of 86nacute lesions were correctly classified (75.6%). The TP rates were 81.1% for acute ischemicnlesions and 66.7% for acute demyelinating lesions. The FP rates were 33.3% for acute ischemicnand 18.9% for acute demyelinating lesions. The precision was 79.6% for classification of acutenischemic lesions and 68.8% for prediction of acute demyelinating lesions. The logistic modelntree decision algorithm revealed that prolonged MTT of surrounding NAWM for a distance ofn15 mm (>7459.2 ms) was the best classifier of acute ischemic versus acute demyelinatingnlesions. Patients with chronic ischemic lesions presented higher mean ADC (p,0.0001) andnprolonged MTT (p50.013) in lesions, and in surrounding NAWM for distances of 5, 10 andn15 mm (all p,0.0001), compared to the patients with chronic demyelinating lesions. Datanmining analyses did not show reliable predictability for correctly discerning between chronicnischemic and chronic demyelinating lesions. The precision was 56.7% for classification ofnchronic ischemic and 58.9% for prediction of chronic demyelinating lesions.nDiscussion: We found prolonged MTT values in lesions and surrounding NAWM of patients withnacute and chronic ischemic stroke when compared to MS patients. The use of PWI is a promisingntool for differential diagnosis between acute ischemic and acute demyelinating lesions. Thenresults of this study contribute to a better understanding of the extent of hemodynamicnabnormalities in lesions and surrounding NAWM in patients with MS. [Neurol Res 2008; 30:n816–826]