PURPOSE: This study was performed in order to compare unenhanced CT with diffusion-weighted, T2-weight-ed,and fast FLAIR MR imaging in the detection of hyperacute cerebral ischema induced in a dog and to deter-mine whichmodality first detected cerebral ischemia. MATERIALS AND METHODS: Experimental cerebral infarction was induced bythe occlusion of intracerebral arter-ies using embolic materials (polyvinyl-alcohol, 300 -6 00 micro) introducedthrough a microcatheter into the internal carotid artery of five dogs weighing 12 -20 kg. Serial CT and MR imageswere obtained at one hour intervals from one to five hours after occlusion, and were analyzed independently by tworadiologists. We assessed changes in attenuation, as seen on unenhanced CT and the signal intensity of the lesionon each MR image, and measured the contrast-to-noise ratio (CNR) of the lesions. RESULTS: Ischemic lesions weredetected on unenhanced CT 1 -3 hours after occlusion of cerebral arteries. In all dogs, the lesions were detectedearliest on diffusion-weighted images obtained at 1 hour. They were detect-ed on T2-weighted images at 3 -5 hoursand on fast FLAIR images of 2 -5 hours. The CNR of ischemic lesions increased gradually during the 5-hour period.It was highest on diffusion-weighted images, while on unen-hanced CT, T2-weighted, and fast FLAIR images it wassimilar. CONCLUSION: Hyperacute ischemic lesions were detected earliest on diffusion-weighted images, and earlieron unenhanced CT than on fast FLAIR or T2-weighted MR image.