The association between neurological deficit in acute ischemic stroke and mean transit time

The purpose of our study was to identify the perfusion MRI (pMRI) algorithm which yields a volume of hypoperfused tissue that best correlates with the acute clinical deficit as quantified by the NIH Stroke Scale (NIHSS) and therefore reflects critically hypoperfused tissue. A group of 20 patients with a first acute stroke and stroke MRI within 24 h of symptom onset were retrospectively analyzed. Perfusion maps were derived using four different algorithms to estimate relative mean transit time (rMTT): (1) cerebral blood flow (CBF) arterial input function (AIF)/singular voxel decomposition (SVD); (2) area peak; (3) time to peak (TTP); and (4) first moment method. Lesion volumes based on five different MTT thresholds relative to contralateral brain were compared with each other and correlated with NIHSS score. The first moment method had the highest correlation with NIHSS (r=0.79, P<0.001) followed by the AIF/SVD method, both of which did not differ significantly from each other with regard to lesion volumes. TTP and area peak derived both volumes, which correlated poorly or only moderately with NIHSS scores. Data from our pilot study suggest that the first moment and the AIF/SVD method have advantages over the other algorithms in identifying the pMRI lesion volume that best reflects clinical severity. At present there seems to be no need for extensive postprocessing and arbitrarily defined delay thresholds in pMRI as the simple qualitative approach with a first moment algorithm is equally accurate. Larger sample sizes which allow comparison between imaging and clinical outcomes are needed to refine the choice of best perfusion parameter in pMRI.