Perfusion- and diffusion-weighted magnetic resonance imaging for monitoring decompressive craniectomy in animals with experimental hemispheric stroke

Object. The aim of this study was to use two types of serial magnetic resonance (MR) imaging-perfusion-weighted (PW) and diffusion-weighted (DW)-to monitor craniectomy in rats with hemispheric stroke. Methods. Focal cerebral ischemia was induced in 36 rats by using an endovascular method of occlusion of the middle cerebral artery (MCAO). Craniectomy was performed 4 or 24 hours later in 12 animals each. Twelve control animals underwent occlusion but did not receive treatment. Perfusion-weighted, DW, and T-2-weighted MR images were obtained at 4, 24, 48, 72, and 168 hours postocclusion in all animals. Relative regional cerebral blood volumes and apparent diffusion coefficients (ADCs) were calculated for the cortex and basal ganglia. Hemispheric lesion volumes (expressed as percentages of total brain volumes; %HLV) as they appeared on DW and T,-weighted MR images and on histological slices stained with 2,3,5-triphenyltetrazolium chloride were compared. Neurological performances and infarct volumes measured 7 days postocclusion were used as study end points. Both PW and DW images demonstrated ischemic tissue 4 hours after MCAO in all animals. Early treatment by performing craniectomy significantly improved cortical perfusion (p < 0.01), whereas the same procedure conveyed no benefit to the basal ganglia. Compared with findings in control animals, the DW image-derived %HLV was significantly reduced (p < 0.01) and the cortical ADCs at 4 and 24 hours postocclusion were significantly higher in animals treated early (p < 0.05). Late treatment with craniectomy did not significantly affect cerebral perfusion. The correlation between the DW imaging-derived %HLV and the histologically derived %HLV at 4 to 72 hours postocclusion was good (r = 0.74), whereas at Day 7 postocclusion the %HLV was underestimated up to 41 % on DW imaging. At 4 hours postocclusion T.-weighted imaging failed to demonstrate the ischemic lesion, whereas from 24 to 72 hours postocclusion the correlation between the T-2-weighted imaging-derived %HLV and the histologically derived %HLV was good (r > 0.81). Neurological performance was significantly improved in animals treated using craniectomy. Conclusions. Early craniectomy significantly improves cortical perfusion through leptomeningeal collateral vessels, significantly reduces infarct size, and improves neurological performance in animals with experimental acute hemispheric infarction. Both PW and DW imaging are suitable for noninvasive monitoring of the effects of decompressive craniectomy.