The natural history of acute stroke is well defined. Predicting outcome in
individuals, however, remains difficult, because prognostic studies examini
ng associations between clinical signs or syndromes and outcome differ in p
atient selection, timing and choice of neurological assessments and outcome
measures. Accuracy has been disappointing. Osler in 1892 stated that the '
course of the disease... is dependent on the situation and extent of the le
sion'. Until recently, it has not been possible to examine the stroke progn
osis, using Osler's approach, with any great accuracy. The advent of diffus
ion weighted magnetic resonance imaging (DWI), which is highly sensitive to
the pathophysiological changes underlying stroke, offers this possibility
as it measures the site and extent of irreversible infarction. This review
summarises the results of syndrome or sign-based predictive studies and sho
ws how DWI may explain different outcomes in patients with identical neurol
ogical presentations, according to the 'situation and extent of the lesion'
.