Magnetic Resonance Imaging Findings in Pediatric Pseudotumor Cerebri Syndrome.

BACKGROUND

Revised diagnostic criteria for pseudotumor cerebri syndrome require three of four neuroimaging findings in the absence of papilledema. We examined the sensitivity and specificity of three or more of four of these magnetic resonance imaging (MRI) findings for pseudotumor cerebri syndrome in children.

METHODS

As part of clinical care, patients in whom there was suspicion for pseudotumor cerebri syndrome underwent neurological and fundoscopic examinations, lumbar puncture, MRI, or magnetic resonance venogram. For this retrospective study, we used this information to classify 119 subjects into definite (n = 66) or probable pseudotumor cerebri syndrome (n = 12), elevated opening pressure without papilledema (n = 23), or controls who had normal opening pressure without papilledema (n = 24). A neuroradiologist, unaware of the clinical findings or original MRI report, reviewed MRIs for pituitary gland flattening, flattening of the posterior sclera, optic nerve sheath distention, and transverse venous sinus stenosis.

RESULTS

The presence of three or more MRI findings has a sensitivity of 62% (95% confidence interval: 47% to 75%) and a specificity of 95% (95% confidence interval: 77% to 100%), compared with controls. Two of three (transverse venous sinus stenosis, pituitary gland flattening, flattening of the posterior sclera) had a similar sensitivity and specificity. Transverse venous sinus stenosis alone had a slightly higher sensitivity (74%, 95% confidence interval: 60% to 85%) and specificity (100%, 95% confidence interval: 80% to 100%).

CONCLUSIONS

In children, three of four of the proposed neuroimaging criteria and transverse venous sinus stenosis alone have a moderate sensitivity and robust specificity for pseudotumor cerebri syndrome. MRIs should be reviewed for these criteria, and their presence should raise suspicion for pseudotumor cerebri syndrome in children, particularly if the presence of papilledema is uncertain.