OBJECTIVE: Ischemic lesion volume is assumed to be an important predictor of poststroke neurological deficits and functional outcome. This critical review examines the methodological quality of MRI studies and the predictive value of hemispheric infarct volume for neurological deficits (at body function level) and functional outcome (at activities level).
METHODS: Using Medline, PiCarta, and Embase to identify studies, 13 of the 747 identified studies met the authors' inclusion criteria. Subsequently, studies were tested for adherence to the key methodological criteria for internal, statistical, and external validity. Each criterion was weighted binary, and studies with 6 points or more were judged to be valid for assessing the predictive value of MRI for outcome.
RESULTS: The 13 included studies had several methodological weaknesses with respect to internal validity, and none of them took lesion location into account. Only a few used outcome measures according to the International Classification of Functioning, Disability and Health and followed patients beyond 6 months. Correlation coefficients between MRI lesion volume and outcomes were higher for outcomes defined at body function level (National Institutes of Health Stroke Scale; median 0.67; range: 0.57-0.91) than for those defined at the level of activities (Barthel Index; median -0.49; range: -0.33 to -0.74).
CONCLUSIONS: Methodological shortcomings of most studies confound the prognostic value of MRI in predicting stroke outcome, and few studies have focused on functional outcome. Future studies should investigate the added value of MRI volume over clinical neurological variables in predicting functional outcome beyond 6 months poststroke.