Glial neoplasms: dynamic contrast-enhanced T2*-weighted MR imaging.

PURPOSE: To evaluate the role of T2*-weighted echo-planar perfusion imaging by using a first-pass gadopentetate dimeglumine technique to determine the association of magnetic resonance (MR) imaging-derived cerebral blood volume (CBV) maps with histopathologic grading of astrocytomas and to improve the accuracy of targeting of stereotactic biopsy. MATERIALS AND METHODS: MR imaging was performed in 29 patients by using a first-pass gadopentetate dimeglumine T2*-weighted echo-planar perfusion sequence followed by conventional imaging. The perfusion data were processed to obtain a color map of relative regional CBV. This information formed the basis for targeting the stereotactic biopsy. Relative CBV values were computed with a nondiffusible tracer model. The relative CBV of lesions was expressed as a percentage of the relative CBV of normal white matter. The maximum relative CBV of each lesion was correlated with the histopathologic grading of astrocytomas obtained from samples from stereotactic biopsy or volumetric resection. RESULTS: The maximum relative CBV in high-grade astrocytomas (n = 26) varied from 1.73 to 13.7, with a mean of 5.07 +/- 2.79 (+/- SD), and in the low-grade cohort (n = 3) varied from 0.92 to 2.19, with a mean of 1.44 +/- 0.68. This difference in relative CBV was statistically significant (P < .001; Student t test). CONCLUSION: Echo-planar perfusion imaging is useful in the preoperative assessment of tumor grade and in providing diagnostic information not available with conventional MR imaging. The areas of perfusion abnormality are invaluable in the precise targeting of the stereotactic biopsy.

MR of spinal cord ganglioglioma.

PURPOSE: Our purpose was to describe the MR imaging features in a series of spinal intramedullary gangliogliomas and to compare these findings with the MR features of intramedullary astrocytomas and ependymomas. METHODS: A retrospective analysis was performed of 76 MR examinations in 27 patients with histologically proved spinal ganglioglioma; these were then compared with imaging findings in a representative sample of histologically proved spinal cord astrocytomas and ependymomas. RESULTS: Statistically significant observations regarding spinal gangliogliomas included young age of the patients (mean, 12 years), long tumor length, presence of tumoral cyst, presence of bone erosion and scoliosis, absence of edema, presence of mixed signal intensity on T1-weighted images, and presence of patchy enhancement and cord surface enhancement. A trend (not statistically significant) was noted for holocord involvement and lack of magnetic susceptibility. CONCLUSION: Spinal ganglioglioma can be strongly suspected if MR images reflect the above criteria; however, the ultimate diagnosis still depends on radical resection and appropriate histopathologic investigation.

Spinal cord ependymomas: MR imaging features.

PURPOSE: To assess the magnetic resonance (MR) imaging characteristics of spinal intramedullary ependymomas. MATERIALS AND METHODS: MR images obtained in 25 patients (aged 12-73 years) with proved intramedullary ependymomas were retrospectively reviewed. T1- and T2-weighted images were obtained in all patients. Gadopentetate dimeglumine was intravenously administered in 23 patients; enhanced sagittal and axial T1-weighted spin-echo images were reviewed. RESULTS: All tumors had hyperintense signal on T2-weighted images. In the 23 patients who received contrast material, all tumors became enhanced; enhancement was heterogeneous in 15 patients and homogeneous in eight patients. Twenty tumors had sharply defined, enhanced borders. Nineteen tumors were centrally located in the spinal cord. A hypointense rim on T2-weighted images was noted in five patients. CONCLUSION: Intramedullary ependymomas become enhanced after administration of gadopentetate dimeglumine; the enhanced borders are usually sharply marginated. They are characteristically located centrally in an expanded spinal cord. Hemosiderin is often present at the periphery of cervical ependymomas.

Screening protocol for MR imaging of the internal auditory canal.

We have previously shown that 5 mm axial T1-weighted images following Gd-diethylenetriamine pentaacetic acid (DTPA) administration were sufficient as a screening examination of the internal auditory canal (IAC) and cerebellopontine angle cistern in 80% of cases with sensorineural hearing loss. In the remaining 20% of cases 3 mm slices were necessary for confirmation of the diagnosis. To reduce the number of cases requiring the additional examination, we have modified our protocol to include a coronal "scout" T1-weighted sequence followed by 5 mm axial sections angled through the IACs. Routine axial T2-weighted images of the posterior fossa were also obtained. One hundred eleven patients were studied with this protocol. In all but two of these the 5 mm sections unequivocally provided the diagnosis and in those two cases the coronal images confirmed the diagnosis suspected on the axial study. Three millimeter slices were not necessary in any patient. This three sequence, post-Gd-DTPA protocol, which requires only 12 min of scan time, is recommended as a screening protocol for IAC and cerebellopontine angle disease.

Diagnosis of carotid artery stenosis: comparison of 2DFT time-of-flight MR angiography with contrast angiography in 50 patients.

Fifty patients underwent 2DFT time-of-flight MR angiography and intraarterial contrast angiography for evaluation of possible carotid atherosclerotic disease. The MR angiography technique employed contiguous axial flow-sensitive (short TR/TE) slices that were reformatted and postprocessed by using a maximum-intensity projection algorithm to provide 16 angiographic views of the carotid arteries. Both studies were independently reviewed by two observers in a blinded manner. Carotid arteries were categorized as normal, mildly stenotic, moderately stenotic, severely stenotic, or occluded. For the 94 carotid arteries available for review, one observer reported a 70% agreement between the two techniques and the second observer reported a 56% agreement (p = .0001). The best correlation was in the severely stenotic category and the worst was in the occluded category. Agreement between observers was 67% for MR angiography and 72% for contrast angiography, which was similar to that between the two techniques. Although not all carotid atherosclerotic disease was visualized equally well, 2DFT time-of-flight MR angiography had a good overall correlation with the "gold standard" of intraarterial contrast angiography, supporting its use as a screening technique. While further improvements are needed, use of MR angiography as the primary diagnostic tool for many patients with suspected carotid stenosis should continue to increase.

Role of slice thickness in MR imaging of the internal auditory canal.

The relative efficacy of post-Gd-DTPA 5 and 3 mm axial T1-weighted images was compared in the detection of lesions in the internal auditory canal and cerebellopontine angle. One hundred twenty consecutive patients were prospectively evaluated with 5 mm axial T1-weighted slices. If these were negative or questionable. 3 mm axial slices were immediately obtained as the next sequence. Eighteen percent of cases were positive and in none of these was the 5 mm study normal. However, in 22 negative cases and two positive cases, a 3 mm study was necessary for confirmation. Five millimeter axial scanning post Gd-DTPA is recommended as the initial study for detection of masses in the internal auditory canal and cerebellopontine angle. Because this study requires fewer acquisitions than 3 mm sections and can be done satisfactorily on low and midfield systems, there are potential time- and cost-saving benefits to this approach.

Dural "tail" associated with meningiomas on Gd-DTPA-enhanced MR images: characteristics, differential diagnostic value, and possible implications for treatment.

Gadolinium-enhanced magnetic resonance (MR) imaging studies of 30 histologically proved cranial meningiomas revealed a linear enhanced structure or "tail" extending away from the tumor mass along the dural surface in 18 cases (60%). Contrast material-enhanced computed tomographic studies available in 10 of these 18 cases did not depict this structure. Characteristic features of this MR finding were reviewed, and criteria were defined to distinguish this tail from other enhanced structures. To assess the differential diagnostic value of this finding, gadolinium diethylenetriaminepentaacetic acid-enhanced MR studies of a control group of other extraaxial lesions as well as superficial intraaxial tumors that abut the meninges were reviewed. These cases failed to show this MR feature. Pathologic correlation was available in three meningiomas with the finding of a tail, and these demonstrated meningothelial tumor nodules in the samples taken from areas corresponding to the enhanced regions on MR images. This is in agreement with other recent pathologic studies of the dura mater surrounding meningiomas. The authors believe that the linear enhanced structure described may represent tissue containing tumoral nodules. Awareness of this MR sign may be useful in distinguishing meningiomas from other lesions and in planning total resection of the infiltrated dura mater.

Potential misinterpretation of cervical spondylosis with cord compression caused by metallic artifacts in magnetic resonance imaging of the postoperative spine.

Tiny metallic particles produced by the contact of untempered surgical instruments with a diamond drill produce magnetic susceptibility artifacts that can both limit the diagnostic quality of postoperative magnetic resonance imaging of the cervical spine and cause diagnostic error. A case of cervical pseudospondylosis with apparent cord compression due to such an artifact is presented. The source of these metallic particles is considered, and the nature of the susceptibility artifact is discussed.

Intracranial hemorrhagic lesions: evaluation with spin-echo and gradient-refocused MR imaging at 0.5 and 1.5 T.

Twenty patients with intracranial hemorrhage were examined with magnetic resonance (MR) imaging at 0.5 and 1.5 T within 2 hours on the two imagers for lesions less than 30 days old and within 24 hours for lesions older than 30 days. MR studies included T1- and T2-weighted spin-echo (SE) and T2*-weighted gradient-refocused (GR) pulse sequences at each field strength. The number of lesions identified and the characteristics (ie, signal intensity of the margin, body, and core) of each hemorrhagic lesion were assessed and compared by means of the three pulse sequences at each field strength. Lesion depiction and characterization were superior (P less than .01) at 1.5 T with T2-weighted SE sequences. Improved depiction and characterization of lesions 300 or more days old (P less than .01) accounted for this result. With the GR sequence, depiction and characterization were similar at both field strengths. The GR sequence did not provide significant additional information about hemorrhage at 1.5 T in this series, but it improved depiction and characterization of hemorrhage at 0.5 T.

Lumbar spine after surgery: examination with intravenous contrast-enhanced CT.

Computed tomography (CT) of the lumbar spine without and with intravenous administration of contrast medium was performed in 143 consecutive patients who had previously had spine surgery for disk herniation and had persistent or recurrent symptoms. Fifty-two patients underwent surgical reexploration. It was possible to make the diagnosis of normal postoperative status, disk herniation, or scarring in 31 (60%) of the 52 patients with the use of CT scans without intravenous contrast medium. CT with intravenous contrast enhancement was useful in 12 of the remaining patients (23%). Enhancement of the margins of a herniated disk occurred in 37 (71%) of the patients. There was near-homogeneous enhancement of postoperative scarring in 34 (65%) patients. Intravenous contrast medium was particularly helpful when disk herniation and scarring were both present, by delineating the margins of a herniated disk and enhancing the entire substance of the scar. In symptomatic postoperative patients, CT of the lumbar spine without intravenous contrast medium should be performed initially. If a definitive diagnosis is not established, CT with intravenous contrast enhancement should be considered.

Intramedullary spinal cord tumors: MR imaging, with emphasis on associated cysts.

Fifteen magnetic resonance (MR) imaging studies in 16 patients with intramedullary spinal cord lesions were correlated with surgical findings in all patients and with results of intraoperative spinal sonography (IOSS) in eight. Tumor extent was defined accurately in all but two patients by MR imaging; these lesions were found at surgery to have been underestimated by one to two vertebral body segments. Most tumors had lengthened T1 and T2 relaxation times. Intratumoral cysts were detected by MR imaging in one of eight cases, and rostral and caudal cysts were defined in only two of ten cases. The poor detection rate is attributed to poor spatial and contrast resolution. Preoperative MR imaging to define the extent of cord enlargement and IOSS to define cystic structures appears to be the optimal method of tumor management pending our further study.

Computed tomography of spinal chordomas.

Computed tomography of 16 spinal chordomas revealed lytic lesions, or lytic and blastic lesions, of 2 or more adjacent vertebrae and a fairly well-delineated soft tissue mass in 16 patients. Intravenous contrast enhancement revealed multiple zones of hypodensity in 5 tumors pathologically proved to contain areas of cystic degeneration. Invasion of the epidural space was noted on computed tomography in 2 vertebral and 7 sacral lesions, and on myelography in 7 vertebral lesions. Computed tomography is ideal for delineation of the soft tissue component of these lesions, detection of sacral lesions at an early stage, planning of surgical excision or irradiation treatment, and detection of recurrence.

Posttraumatic syrinx formation: experimental study.

An experimental study was performed to examine posttraumatic spinal cord cavitation in an animal model by evaluating immediate and delayed computed tomographic (CT) scans obtained after administration of intrathecal contrast material. Four cats underwent midthoracic laminectomy and spinal cord contusion using a standard 400 g-cm model. All animals were studied by CT with intrathecal contrast enhancement before and 4-5 days, 3-4 weeks, and 7-13 weeks after experimental cord contusion. Either metrizamide or iopamidol was used as the contrast agent. Two of the four cats had CT and pathologic evidence of cord cavitation at the site of injury. Another animal had uptake of contrast material into the spinal cord without pathologic evidence of cyst formation, which was believed to represent malacic change. The fourth animal had a normal-appearing cord by both CT and pathologic criteria. Animals that received metrizamide after cord contusion had generalized myoclonic seizures. This did not occur when iopamidol was administered.

Computed tomography of nontuberculous spinal infection.

The CT findings in 16 patients with nontuberculous spinal infections were reviewed. The specificity of certain CT features as well as the usefulness of intravenous contrast medium administration are discussed. The associated clinical presentations and predisposing factors are outlined. Emphasis is placed on a combined clinical, radiographic approach in facilitating an early diagnosis.

Complications of digital intravenous angiography: experience in 2488 cervicocranial examinations.

All complications were recorded from the initial 2488 cases studied with digital intravenous angiography (DIVA) at New York University Medical Center. Mechanisms of producing these reactions were categorized into procedure-related, contrast-medium-related, or disease-related. The complications included extravasation of contrast material into the arm (11 patients) and mediastinum (two), acute pulmonary edema (four), hypotension (23), thrombophlebitis (two), and grand mal seizure (one). Recommendations are made that would allow DIVA to be performed more safely.

Iopamidol and Conray 60: comparison in superselective angiography.

Iopamidol (280 mgI/ml) was compared with Conray 60 (iothalmate meglumine, 282 mgI/ml) during selective and superselective cerebral arteriography. Twelve patients were examined, and a total of 51 selective injections were compared. Iopamidol produced significantly less pain and burning than Conray 60 and caused no patient motion.

CT of lumbar spine disk herniation: correlation with surgical findings.

Computed tomography (CT) of the lumbar spine was performed with selectively positioned 5-mm-thick axial cross sections to examine each disk level from the top of the neural foramen to the pedicle of the next caudad vertebra. One hundred consecutive patients with 116 surgical disk explorations were reviewed. There was agreement between the CT and surgical findings in 89 patients (104 explorations) in determination of presence or absence of a herniated nucleus pulposus (HNP). Discrepancy occurred in 12 instances (11 patients): two because of incorrect interpretations, five in previously operated patients, three in spondylolisthesis, and two in spinal stenosis. There were 97 true-positives, eight false-negatives, seven true-negatives, and four false-positives. If nine previously operated patients are excluded from the study, then CT was accurate in detection of presence or absence of an HNP in 93% of the disk explorations.