Background and purpose: A mismatch between diffusion- and perfusion-weighted MRI is thought to define tissue at risk of infarction. This concept is based on the assumption that diffusion slowing of and decreases in the apparent diffusion coefficient (ADC) serve as indicator of tissue proceeding to infarction. We tested this hypothesis.
Methods: MRI (diffusion weighted, perfusion weighted, MRA, T2 weighted) was performed in 15 patients with acute stroke within 2.9+/-0.8 hours (mean+/-SD) of onset and on days 1 and 7. After intraindividual realignment of the ADC maps, the development of ADC range volumes and ADC values was determined.
Results: An increase (354%, group A1) in the total ADC-based lesion volume below a threshold of < 80% occurred in 4 patients on day 1, persisting on day 7 with a pronounced increase of ADC range volumes with low ADC values. An increase in total ADC-based lesion volume (201%, group A2) followed by a secondary drop to day 7 was found in 7 patients. A significant reduction in total ADC-based lesion volume (14%, group B) was found in 4 patients. ADC-based lesion volume increase was associated with persistent vessel occlusion in group A, whereas recanalization in group B resulted in ADC volume decrease. ADC normalization was observed independently from the degree of the initial ADC decrease on days 1 and 7 in group B.
Conclusions: In line with results from animal experiments, ADC decreases do not reliably indicate tissue infarction Even severely decreased ADC values may normalize in human stroke, and it seems likely that ADC normalization depends on the duration and severity of ischemia rather than the absolute value.