MR perfusion imaging reveals regions of hypoperfusion associated with aphasia and neglect

OBJECTIVE: To evaluate diffusion-weighted imaging (DWI) and MR perfusion imaging (MRPI) as tools for identifying regions of infarct and hypoperfusion associated with aphasia and neglect in hyperacute stroke. Secondary goal: to establish a functional correlate of a radiologically defined "ischemic penumbra." METHODS: Forty subjects underwent DWI, MRPI, and standardized tests for lexical deficits or hemispatial neglect within 24 hours of stroke onset or progression. Ten patients had repeat DWI, MRPI, and cognitive testing after 3 days (in some cases after reperfusion therapy). Pearson correlations between error rate on cognitive testing and volume of abnormality on DWI versus MRPI were determined at each time period, and regions of hypoperfusion corresponding to specific cognitive deficits were identified. RESULTS: Error rate was more strongly correlated with volume of hypoperfused tissue on MRPI (r = 0.65 to 0.93; p < 0.01 to p < 0.0003) than with volume of lesion on DWI (r = 0.54 to 0.75; p = 0.14 to p < 0.01) for dominant and nondominant hemisphere stroke, at each time point. Forty-eight percent of aphasic patients and 67% of those with hemispatial neglect had either no infarct or only small subcortical infarct on DWI, but had focal cortical hypoperfusion. Patients who had successful reperfusion therapy showed resolution of the hypoperfused territory beyond the infarction on repeat MRPI and showed resolution of corresponding deficits. CONCLUSIONS: MRPI shows regions of hypoperfused cortex associated with lexical deficits or hemispatial neglect, even when DWI shows no infarct or only small subcortical infarct. MRPI-DWI mismatch indicates regions of functionally salvageable tissue.