Abstract #0500
Predicting tissue outcome in acute human stroke using temporally adaptive MRI-based algorithms
Wu O, Koroshetz W, Gonzalez R, Benner T, Sorensen A, Schwamm L, Copen W
Massachusetts General Hospital
Tissue viability after stroke depends on depth and duration of the ischemic insult. Infarction risk was estimated on a voxel-wise basis by combining serial DWI and PWI and ischemia duration (i.e., time from symptom onset to MR acquisition) from stroke patients imaged <12h from onset. Models incorporating temporal information performed significantly better (p<02) in predicting final infarction at chronic timepoints (Youdens J=.46.2) compared to models that did not (J=.40.2) and performed equivalently acutely. Ischemia duration is an important factor that modulates interpretation of DWI and PWI in relation to tissue salvageability; therefore, its incorporation into MR-based algorithms improves infarction prediction.