The principle of optimal patient selection by imaging confirmation of the relevant pathology was first demonstrated by the results of the intra arterial pro-urokinase stroke study, PROACT II. Four major factors are hypothesized to predict tissue response and clinical efficacy in stroke trials: time to treatment, the salvageable tissue-at-risk, the relevance of the patient sample to the treatment and the intrinsic effectiveness of the therapeutic strategy. In using MRI as a selection criterion, the goal would be a sample selection based upon a positive imaging diagnosis of a pathology rationally linked to the drug's mechanisms of action. MRI measurements of lesion volume acutely and chronically have proven to be a marker of clinical severity. MRI is increasingly used as a selection tool and an outcome measure in stroke trials, reflecting recognition that direct pathophysiological imaging may provide a more rational approach to stroke therapeutics.