Cholesterol granuloma of the petrous apex: MR and CT evaluation.

We reviewed the clinical, CT, and MR findings in seven consecutive patients who had a total of nine cholesterol granulomas. Preoperative MR scans were available for five of the seven patients; two patients were studied with MR after treatment only (one had a recurrent lesion and the other was asymptomatic at the time of study). Preoperative CT scans were available for all patients, except one patient who was examined after developing a symptomatic recurrence. All lesions were detected by both imaging methods. Seven preoperative lesions (five patients) and one symptomatic recurrence (one patient) demonstrated increased signal intensity of both T1- and T2-weighted MR images. Three surgically drained lesions (three patients) showed a marked reduction in signal intensity on T1-weighted images. Pre- and postoperative lesions had different patterns of signal intensity on the chemical-shift images, which were obtained in two instances. The MR appearance of cholesterol granuloma differs from that of most other lesions that occur in the petrous apex. CT did not differentiate between pre- and postoperative lesions in all cases, while MR demonstrated a dramatic change on T1-weighted images and chemical-shift studies. Our findings indicate that MR is more specific than CT in suggesting the correct diagnosis of cholesterol granuloma and that MR appears to be the technique of choice in the follow-up of previously treated patients.

The role of MR and CT in evaluating clival chordomas and chondrosarcomas.

Sixteen chordomas and nine chondrosarcomas of the clivus were evaluated with CT and MR either before (22 cases) or after (three cases) treatment with proton beam irradiation. The ability of these imaging techniques to provide information necessary to direct patient treatment was studied. The tumor was detected and its gross margins were identified by both techniques in all instances. No reliable diagnostic features allowing differentiation between these two tumors were encountered. MR generally was superior in defining the exact position of the brainstem and optic chiasm relative to the tumor, and it frequently provided superior information about tumor extension into the nasopharynx and cavernous sinus. CT was always better than MR in demonstrating tumoral calcification and in defining the exact anatomy of bone destruction. MR was generally superior to CT in demonstrating the position of the cavernous internal carotid artery relative to the tumor and often provided superior visualization of the vertebral and basilar arteries. In cases in which bone-induced artifact obscured the interface between the neural axis and tumor in the CT image, or in which the tumor had suprasellar extension and was likely to compress the optic chiasm and tracts, MR was of great value in planning irradiation therapy. The high occurrence of clinically asymptomatic signal intensity alterations in the MR studies of previously treated patients appears to limit the differential diagnostic value of this information. Given its greater availability and lower cost, CT appears to be the technique of choice for routine follow-up of previously treated patients.

Susceptibility induced MR line broadening: applications to brain iron mapping.

Magnetic susceptibility variations due to the presence of iron in neural tissue can result in a shift of local resonance frequency, decreased T2 resulting from water diffusion through local field gradients, and line broadening due to field inhomogeneity within a voxel. In this study, modified spin echo phase contrast pulse sequences were used to map proton resonance line widths in phantoms and in vivo. In agar gels containing varying concentrations of Fe3O4 (magnetite), line broadening mechanisms permitted accurate spatial localization of iron deposits from measurement of local resonance line widths. Furthermore, line widths estimated from the data were strongly correlated with iron concentrations, indicating the potential quantitative applications of the method. The technique was applied clinically to map iron in subacute brain hemorrhages. These data suggest that resonance line widths may be a useful measure of brain iron content.

The evolution of neonatal herpes encephalitis as demonstrated by cranial ultrasound with CT correlation.

Three neonates with herpes encephalitis had serial cranial ultrasound and CT studies performed during their hospitalization. Initially, subtle changes of diffuse brain edema were present but with no ventricular compression. A second phase revealed progressive brain edema with ventricular compression. A third phase with enlarging ventricles indicates developing encephalomalacia. Occasionally cyst formation may be noted within the cerebral white matter, a phenomenon initially more readily evident by ultrasound.

The detection of intracranial calcifications by MR.

Twenty patients in whom CT had unequivocally demonstrated the presence of calcification in a diversity of lesions and who had undergone MR, performed at 0.6 T and with standard T1- and T2-weighted pulse sequences, were retrospectively studied to determine the MR signal-intensity characteristics of the calcifications and to assess the ability of MR to detect the presence of this abnormality. CT proved superior to MR in detecting and characterizing calcification. In seven of 20 cases, the apparent extent of calcification was equal by both imaging techniques, and in 13 of the 20 cases, CT showed more extensive abnormality. In five of the 20 cases, the calcifications were seen by MR as regions of profoundly reduced signal intensity, approximately equal to cortical bone, in all pulse sequences. In 12 of the 20 cases, the signal intensity was profoundly reduced in one or more, but not all, pulse sequences. T2-weighted pulse sequences were most sensitive in detecting calcification of signal void. Reviewed without knowledge of the CT findings, the MR images were interpreted as definitely indicative of the presence of calcification in three of the 20 cases. In seven of the 20 cases, the MR images raised the possibility of calcification but were less definitive than the CT findings. In 10 of the 20 cases, MR was judged indeterminate for the presence of calcification.

High-flow angiopathy: cerebral blood vessel changes in experimental chronic arteriovenous fistula.

Profound vascular damage secondary to high-flow extracranial states has been well characterized. However, changes in cerebral vasculature secondary to high-flow states have not been studied. To determine changes related to high-flow states in cerebral vessels, a rabbit model was developed in which torrential flow was created in the vertebrals, carotids, basilar, and vessels of the circle of Willis by means of a carotid-jugular shunt after ligation of the proximal carotid. The clinical, angiographic, and histologic changes noted in the animal model include: abrupt clinical deterioration after a variable interval with some animals developing ptosis, afferent vessel dilatation and the development of prominent anastomotic channels, variable cerebral vessel histopathology--related to duration and relative proximity to the shunt--affecting all three vessel layers, plump, irregular, and clumped endothelium, denuded with adherent platelets, irregular, duplicated, and thinned internal elastic membrane, frayed with invasion of the intima by mesenchymal cells, vacuolization and necrosis of the media muscle, and invasion of adventitia by foreign cells and small blood vessels. The high-flow angiopathy seen in this model may help explain vascular changes associated with high-flow cerebral vascular lesions, as well as other types of vascular damage.

Computed tomography of osteogenic sarcoma of nasal cavity and paranasal sinuses.

Computed tomography was performed in four patients with osteogenic sarcoma of the nasal cavity and paranasal sinuses. In all cases CT demonstrated an abnormal soft tissue mass with associated aggressive bony destruction. In only one case were areas of calcification and new bone formation identified, allowing a preoperative diagnosis. The CT appearance of these four patients plus an additional four cases previously reported in the literature are reviewed.

Computed tomography of lymphoproliferative disease of the orbit. Report of 50 patients.

Orbital lymphoproliferative disease is one of the most common causes of exophthalmos and palpable masses in the anterior part of the orbit. Axial and coronal computed tomography (CT) is the method of choice for localizing the lesion and determining the extent of involvement. The CT features are non-specific but are highly suggestive in the proper clinical setting. The most important CT findings consist of: round, oval or elongated lobular masses, commonly in the extraconal space, with intraconal extension in large tumors. In some lesions, linear infiltrates radiate from the bulk of the mass into the adjacent fascial compartments. Tumors in proximity to the globe, wrap around the scleral-uveal coat causing no indentation but displacement of the globe from the mass effect. Similar molding occurs along the orbital walls, especially laterally from extraconal lesions. Bone destruction or remodeling of bone is rarely seen in orbital lymphoma. No diagnostic enhancement patterns were observed following contrast infusion. The majority of lymphoid tumors were situated anteriorly and superiorly.

CT appearance of clival chordomas.

Computed tomography of 30 patients with chordomas arising from the clivus was reviewed. All examinations demonstrated bone destruction with a soft tissue mass and varying degrees of encroachment on the adjacent brain. Other CT findings included contrast enhancement, soft tissue calcification and well defined areas of low attenuation.

NMR imaging of intracranial hemorrhage.

Twenty-one intracranial hemorrhagic lesions were imaged at 0.15 and 0.6 T using inversion recovery (IR), spin echo (SE), and multiple SE (Carr-Purcell-Meiboom-Gill, CPMG) pulse sequences. Two subarachnoid hemorrhages (SAH), nine acute intraparenchymal hemorrhages (IPH), ten chronic IPH, and one subdural hematoma were studied. Acute SAH could not be identified on the T1-weighted, IR images but was clearly seen on a T2-weighted, CPMG image. Acute (7 days or less) intraparenchymal hematoma showed signal intensity on IR and CPMG images similar to white matter. The T1 and T2 times of acute intraparenchymal hematoma were also similar to white matter. Some small acute hematomas could not be distinguished from white matter on IR and CPMG images. Acute hemorrhagic tissue showed image intensities and relaxation times similar to gray matter. All acute hemorrhages were identified on CT. Chronic IPH lesions (14 days or more) showed high signal intensity, greater than white matter, on IR, SE, and CPMG images. The T1 of the chronic lesions was similar to the acute lesions but T2 was significantly longer (p less than 0.05). Available evidence suggests that the nonspecificity of acute IPH signal and relaxation times may not be restricted to our pulse sequences or magnetic field strengths.

Evaluation of postoperative osteomyelitis of the sternum using tomography and computerized tomography.

Thirty-five patients were evaluated for possible postoperative sternal osteomyelitis by tomography. Frank bone destruction and severe demineralization were the only radiologic signs that helped to distinguish patients with sternal osteomyelitis from those with infections limited to soft tissues. We report a patient in whom sternal osteomyelitis was shown by computerized tomography (CT). This observation suggests that CT may provide added information, especially if the sternum is severely demineralized.