OBJECTIVE. Morphologic changes associated with papilledema may be masked by partial volume averaging effects in images obtained at a slice thickness greater than normal optic nerve thickness. We aimed to compare the diagnostic accuracy of high-resolution 3D T2-weighted imaging performed at submillimeter slice thickness with conventional T2-weighted imaging performed at 5-mm slice thickness for detection of papilledema. MATERIALS AND METHODS. Two blinded neuroradiologists evaluated conventional and high-resolution axial T2-weighted imaging across orbits from 25 patients with clinically proven papilledema and 66 control participants without papilledema. They graded optic nerve sheath distention and optic nerve head configuration, also making a binary determination for presence or absence of papilledema for each set of images. The diagnostic accuracy of each technique was assessed in terms of sensitivity, specificity, positive likelihood ratio, and interobserver agreement. These parameters were compared using the homogeneity of odds ratio and McNemar tests. RESULTS. High-resolution T2-weighted imaging was associated with higher sensitivity (83.3% vs 56.2%, p = 0.0072 for reader 1; 87.5% vs 54.2% for reader 2, p = 0.0001) but unchanged specificity. High-resolution T2-weighted imaging was significantly better than conventional T2-weighted imaging in detecting optic nerve head deformity in patients with papilledema, but there was no difference between two techniques in detection of optic nerve sheath distention. High-resolution imaging also enabled greater interobserver agreement (κ = 0.82) compared with conventional T2-weighted image (κ = 0.62). CONCLUSION. Improved visualization of the optic nerve head afforded by high-resolution T2-weighted imaging translates into better diagnostic performance of MRI in detection of papilledema, with higher sensitivity and interobserver reliability.
Keywords: MRI; idiopathic intracranial hypertension; papilledema; pseudotumor cerebri.