The clinical-DWI mismatch: a new diagnostic approach to the brain tissue at risk of infarction.

OBJECTIVE To evaluate the usefulness of a mismatch between the severity of acute clinical manifestations and the diffusion-weighted imaging (DWI) lesion in predicting early stroke outcome and infarct volume. METHODS One hundred sixty-six patients with a hemispheric ischemic stroke of <12 hours' duration were studied. The NIH Stroke Scale (NIHSS) score and the volume of DWI lesion were measured on admission and at 72 +/- 12 hours. Infarct volume was measured on T2-weighted or fluid-attenuated inversion recovery images at day 30. Early neurologic deterioration (END) was defined as an increase of > or =4 points between the two NIHSS evaluations. Thirty-eight patients received IV thrombolysis or abciximab. Clinical-DWI mismatch (CDM) was defined as NIHSS score of > or =8 and ischemic volume on DWI of < or =25 mL on admission. The adjusted influence of CDM on END, DWI lesion enlargement at 72 hours, and infarct growth at day 30 was evaluated by logistic regression analysis and generalized linear models. RESULTS CDM was found in 87 patients (52.4%). Patients with CDM had a higher risk of END than patients without CDM because NIHSS < 8 (odds ratio [OR], 9.0; 95% CI,1.9 to 42) or DWI lesion > 25 mL (OR, 2.0; 95% CI, 0.8 to 4.9). CDM was associated with an increase of 46 to 68 mL in the mean volume of DWI lesion enlargement and infarct growth in comparison with non-CDM. All the effects were even greater and significant in patients not treated with reperfusion therapies. CONCLUSIONS Acute stroke patients with an NIHSS score of > or =8 and DWI volume of < or =25 mL have a higher probability of infarct growth and early neurologic deterioration. The new concept of CDM may identify patients with tissue at risk of infarction for thrombolytic or neuroprotective drugs.