PURPOSE: To study MR patterns of venous sinus occlusive disease and to relate them to the underlying pathophysiology by comparing the appearance and pathophysiologic features of venous sinus occlusive disease with those of arterial ischemic disease. METHODS: The clinical data and MR examinations of 26 patients with venous sinus occlusive disease were retrospectively reviewed with special attention to mass effect, hemorrhage, and T2-weighted image abnormalities as well as to abnormal parenchymal, venous, or arterial enhancement after intravenous gadopentetate dimeglumine administration. Follow-up studies when available were evaluated for atrophy, infarction, chronic mass effect, and hemorrhage. RESULTS: Mass effect was present in 25 of 26 patients. Eleven of the 26 had mass effect without abnormal signal on T2-weighted images. Fifteen patients had abnormal signal on T2-weighted images, but this was much less extensive than the degree of brain swelling in all cases. No patient showed abnormal parenchymal or arterial enhancement. Abnormal venous enhancement was seen in 10 of 13 patients who had contrast-enhanced studies. Intraparenchymal hemorrhage was seen in nine patients with high signal on T2-weighted images predominantly peripheral to the hematoma in eight. Three overall MR patterns were observed in acute sinus thrombosis: 1) mass effect without associated abnormal signal on T2-weighted images, 2) mass effect with associated abnormal signal on T2-weighted images and/or ventricular dilatation that may be reversible, and 3) intraparenchymal hematoma with surrounding edema. CONCLUSION: MR findings of venous sinus occlusive disease are different from those of arterial ischemia and may reflect different underlying pathophysiology. In venous sinus occlusive disease, the breakdown of the blood-brain barrier (vasogenic edema and abnormal parenchymal enhancement) does not always occur, and brain swelling can persist up to 2 years with or without abnormal signal on T2-weighted images. Abnormal signal on T2-weighted images may be reversible and does not always indicate infarction.