OBJECTIVE Primary aim is to compare the diagnostic value of contrast-enhanced 3D GRE T1-weighted sequences with unenhanced MR venography and conventional magnetic resonance imaging (MRI), in detection of dural venous sinus (DVS) and cortical venous thrombosis; secondary aim is to determine the relationship between DVS thrombosis/site and gender, age, infarction or hemorrhage. METHODS We retrospectively reviewed conventional MR images, unenhanced MR venography and immediate post-contrast 3D GRE T1-weighted MR images in 30 patients (17 male and 13 female, 21-70 years old, mean age 40.1) with clinically suspected DVS thrombosis. MR examinations had been performed with 1.5T or 3T MR Scanners. DVSs were evaluated in 10 sub-segments, including cortical veins. Each set of MR images were examined separately, blinded to the final diagnosis. Associated findings were also noted and sensitivity, specificity and accuracy of each MRI technique were calculated. RESULTS Final diagnosis of cortical venous and/or dural sinus thrombosis was established in 24 (80%) of 30 cases and 67 (22.3%) out of 300 segments. For detection of the thrombotic segment, sensitivity, specificity, and accuracy were 83.6%, 95.3%, and 92.7% by conventional MR sequences, 89.6%, 91.8%, and 91.3% by unenhanced MR venography, and 92.5%, 100%, and 98.3% by contrast-enhanced 3D GRE T1-weighted sequence, respectively. Infarction and hemorrhage were more frequent in cases with cortical venous thrombosis, while gender and age had no significant relation with DVS thrombosis or its site. Conventional MR sequences and unenhanced MR venography were helpful due to additional information they provided in some cases with isolated cortical venous thrombosis, with hyperintense thrombus material and with associated hemorrhage or infarction. CONCLUSIONS Contrast-enhanced 3D GRE T1-weighted MRI is the most accurate imaging method for the detection of DVS and/or cortical venous thrombosis. Infarction and hemorrhage were more frequent in cases with cortical venous thrombosis.