Background and purpose: Based on previous observations that infarcts encompassing the insula were linked to unfavorable clinical outcome, we hypothesized that insular damage was directly associated with worsened infarction in ischemic but potentially viable neighboring brain tissue.
Methods: Using acute diffusion- and perfusion-weighted MRI within the first 12 hours of symptom onset and a follow-up MRI on day 5 or later, we calculated the percentage of mismatch lost (PML) in 61 consecutive patients with ischemic stroke within the middle cerebral artery territory. PML denoted the percentage of mismatch tissue between diffusion-weighted imaging and mean transit time maps that eventually underwent infarction. We explored the relationship between PML and insular location using a regression model.
Results: The median PML was 17.7% (interquartile range, 3.5% to 54.2%) in insular and 2.5% (0.0% to 12.7%) in noninsular infarcts (P<0.01). The PML correlated with the volume of abnormal regions on diffusion-weighted imaging (P<0.01), mean transit time (P<0.01), cerebral blood flow maps (P<0.01), and cerebral blood volume maps (P<0.01). A linear regression model with PML as response and with acute MRI volumes, age, and the site of vascular occlusion as covariates showed that insular involvement was an independent predictor of PML (P=0.01). The regression model predicted an approximately 3.2-fold increase in PML with insular involvement.
Conclusions: Infarction of the insula is associated with increased conversion of ischemic but potentially viable neighboring tissues into infarction. The unfavorable tissue outcome in insular infarcts may not be a mere bystander effect from proximal middle cerebral artery occlusions.