Investigating multiple sclerosis with spectroscopic imaging: harbinger of a new paradigm.

sionally the patient’s chief complaints. Stupor is an unusual complication that we encountered in one patient because of a central herniation syndrome. Although CSF pressure in SIH is usually low, this is not always the case. Mokri et al noted pressures between 70 and 130 mm in four of 26 patients (2), and we observed the same range in three patients with SIH. CSF lymphocytosis between 8 and 80 cells/mm is not uncommon, and can escalate as high as 226 cells/mm with a mild increase in CSF protein (3). These findings may lead to the erroneous diagnosis of “viral meningitis.” MR findings characteristic of SIH reflect an increase in venous volume throughout the brain and dural structures, including the spine. MR reveals diffuse dural enhancement (as if the dura were outlined with a felt-tip pen), prominent dural sinuses, an enlarged epidural venous plexus and pituitary gland, subdural collections from effusions or hematomas, and downward displacement of the chiasm, pons, and cerebellar tonsils. Although diffuse dural enhancement is almost invariably present, descent of the tonsils occurs in approximately 75% of SIH patients. Patients with elevated intracranial pressure may sometimes demonstrate descent of posterior fossa structures and chiasm on sagittal T1-weighted MR scans that appear similar to MR findings of SIH. With elevation of intracranial pressure, however, the venous dural sinuses are effaced, not enlarged, and usually diffuse enhancement of the dura is not present. In patients with dural metastases, infection, or inflammation, the enhancement of the dura is usually patchy, quite different from the appearance of SIH. Dural biopsy is not indicated in patients with clinical and radiologic features of SIH, and we have not recommended dural biopsy in over 40 cases. Localization of a spinal fistula, usually secondary to a ruptured perineural (Tarlov) cyst, is not required unless the patient has failed to respond to at least two largevolume (20–30 mL) lumbar blood patches followed by a 10-minute, 30° lowering of the head for increased blood flow to the thoracic and cervical levels. If these fail, then localization of the CSF leak using simultaneous (single dural puncture) myelography and cisternography followed by CT at 3-mm increments are indicated. Consecutive decubitus cross table views obtained by positioning the patient on the left and then right side during myelography may facilitate the leakage of contrast medium from a laterally placed perineural cyst. Radiologic localization of the leak is necessary prior to direct instillation of an epidural blood patch at the precise location of the fistula. If this fails, neurosurgical intervention is required. Epidural blood patching has been successful in the majority of our patients. Only three patients required surgical intervention, and surgery was successful in two. The spinal CT and MR manifestations of SIH may also be misleading. Enlargement of the cervical epidural venous plexus may be misinterpreted as meningioma. Leakage of CSF may occur along a nerve root lateral to the spinal canal, resulting in the easy detection of the leak by isotope cisternography and CT myelography. Alternatively, the CSF leak may collect and extend within the “gutter” of the epidural space inside the spinal canal. In this instance, chronic leakage of CSF may become walled off within the epidural fat, creating a tubular pseudocyst which on cross-sectional imaging appears as a semicircular “dog ear” shaped collection, ventral or dorsal to the thecal sac. The wall of this “pseudocyst” may prevent an epidural blood patch from contacting the site of the leak, accounting for some treatment failures. In addition, the location of an epidural collection may not always reflect the actual site of the spinal fistula because CSF may extend to within the epidural space several levels away from the actual fistula site. Detection of the site of the CSF leak is usually not necessary since most pateints are treated effectively by an epidural blood patch. Some patients, however, are refractory to this therapy. In our experience treatment is unsuccessful if the leak is too large, is positioned laterally along a nerve root sleeve, or enters a “pseudocyst” collection that is walled off from the epidural space. The accurate localization of spinal fistulae and the treatment of these patients is rewarding. Most get better, and the radiologist plays an important role in the diagnosis and therapy of these patients. WILLIAM P. DILLON, MD Senior Editor ROBERT A. FISHMAN, MD University of California, San Francisco