Abstract
PURPOSE
To present our initial experience with MR cisternography, an application of fast spin-echo MR with fat suppression, and compare it with routine MR cranial studies in the evaluation of the subarachnoid cisterns and their contents.
METHODS
MR cisternography is a heavily T2-weighted fast spin-echo technique with high spatial resolution; it uses fat suppression and video reversal of the images. A small number of individual sections (two to four) are compressed into a composite image by a maximum-intensity projection algorithm, providing better depiction of anatomy in three dimensions. MR cisternography enhances the signal intensity of the cerebrospinal fluid (CSF) with suppression and subtraction of the background. A total of 41 patients were examined during a period of 6 months. MR cisternography was performed as an additional one (n = 31) or two (n = 10) sequences after conventional MR study.
RESULTS
Twenty-one cases of disease were examined by MR cisternography, including 8 neoplasms, 4 CSF fistulas, and 3 large intracranial aneurysms. MR cisternography provided information unavailable by conventional MR studies in 17 cases. These included visualization of fistulous tracks in patients with CSF rhinorrhea, origin of a large suprasellar aneurysm, an additional loculus of a posterior fossa aneurysm and its relation to surrounding structures, and proper location of three tumors (intraaxial versus extraaxial). Clear depiction of the pituitary gland separate from the cavernous sinus was noted in 60% of the cases, and a new observation of a CSF sleeve around the third nerve in the posterior cavernous sinus was made in 85% of the cases.
CONCLUSION
MR cisternography is superior to conventional MR studies in depicting anatomic structures within the subarachnoid spaces. This technique is recommended in the evaluation of cranial CSF fistulas and suprasellar and posterior fossa masses and in diagnosis of intraaxial versus extraaxial location of intracranial tumors.
Full Text
The Full Text of this article is available as a PDF (5.3 MB).