Search

Objective: The Single-Voxel Proton Brain Exam (PROBE/SV) is an automated MR spectroscopic technique. The purpose of this study was to investigate the use of PROBE/SV as a diagnostic tool in patients with primary brain tumors and to compare our findings with the known information obtained from conventional nonautomated MR spectroscopic techniques., Subjects and Methods: We used PROBE/SV to image 10 normal adults and 46 patients with primary brain tumors: 29 glioblastoma multiformes (GBMs), five anaplastic astrocytomas, and 12 low-grade astrocytomas. All studied were performed on a 1.5-T Signa unit. Average voxel sizes were 6-8 cm3. A corresponding point-resolved spectroscopy spectrum was obtained from normal-appearing brain parenchyma in each patient for comparison with the spectra from known areas of pathology., Results: In patients with low-grade gliomas (grades 1 and 2), we observed decreased N-acetylaspartate (12 of 12) and slightly increased choline (11 of 12) when we compared these metabolites with those in the spectra of patients' normal brains. This comparison in patients with GBM yielded markedly decreased N-acetylaspartate (29 of 29) and prominently increased choline (27 of 29). In the short TE spectra, we frequently saw lipid signal in high-grade tumors, especially in GBMs (12 of 20). We identified lactate peaks in high-grade tumors (anaplastic astrocytoma and GBM, 29 of 34) and also in low-grade tumors (four of 12). The creatine signal in all gliomas was slightly less than that of healthy brain tissue. The lowest N-acetylaspartate, choline, and creatinine levels in conjunction with the highest lactate levels were usually found in necrotic portions of high-grade tumors., Conclusion: PROBE/SV is a simplified MR spectroscopy technique that reduces setup time and provides automatic on-line data processing and display. The voxel location can be selected to focus on the area of interest and to minimize voxel contamination from unwanted tissue. The results from our experimentation with PROBE/SV in patients with brain tumors generally concur with published reports of tumor spectra obtained by conventional MR spectroscopic techniques. The ease and accuracy of this new technique make it a useful clinical tool in differentiating human brain tumor grades.

Objective: Measurements of the volume of the hippocampus have been used preoperatively to determine the side of the hippocampus involved in patients with intractable temporal lobe epilepsy. However, the method used is time consuming and requires special training. The purpose of this study was to determine the validity of using measurements of the volume of the hippocampal body as a substitute for measurements of the volume of the whole hippocampus in patients with temporal lobe epilepsy., Subjects and Methods: The study group consisted of 33 patients with intractable temporal lobe epilepsy due to presumed hippocampal sclerosis and 30 control subjects. Of the 33 patients with intractable temporal lobe epilepsy, 30 had hippocampal sclerosis and three had normal findings on pathologic examination of the hippocampus. T2-weighted fast spin-echo images were used to determine hippocampal volumes. Volumes were calculated by summing the sectional areas of the entire hippocampus and of the hippocampal body. Correlation between the volume of the hippocampal body and the volume of the whole hippocampus was determined. The sensitivity and specificity of measurement of the volume of the hippocampal body for lateralizing the foci responsible for seizures in patients with presumed hippocampal sclerosis were compared with those of measurement of the volume of the whole hippocampus., Results: Significant linear relationships were noted between volumes of the hippocampal body and volumes of the whole hippocampus (p < .001). The sensitivity and specificity of measurements of volumes of the hippocampal body were identical to those of measurements of volumes of the whole hippocampus (87% and 100%, respectively)., Conclusions: Our results show that segmental MR measurements of the body of the hippocampus are as accurate as measurements of the whole hippocampus for lateralizing temporal lobe epilepsy before surgery. Because segmental measurements are less time consuming and require less experience to perform, they are considered the procedure of choice.

The limbic system refers to the part of the brain that is involved in emotional expression and in cognitive and somatomotor control systems; it most likely also has a significant role in the pathogenesis of certain dementias, neuropsychiatric disturbances, and seizure disorders. This system, as originally defined, has no clear anatomic boundaries. Limbic regions of the telencephalon include a continuous medial zone of cortical regions, including the hippocampal formation; a perihippocampal zone, including the cingulate gyrus, prefrontal region, and perirhinal region; along with subcortical areas, including the septum pellucidum and the amygdala. A distinguishing characteristic of the limbic region is that it is highly interconnected and appears to form the only major route for information transfer between the neocortex and the hypothalamus. With the multiplanar capability and resolution of MR imaging, the individual parts of the limbic system and the diseases that affect them can be studied. In this review, we discuss MR imaging of the diseases that affect this system.

Cephalocervical fibromuscular dysplasia is an uncommon vascular disease that affects less than 1% of the population. Although often an incidental finding, patients may present with transient ischemic attacks, cerebral infarctions, or subarachnoid hemorrhage. Fibromuscular dysplasia is usually diagnosed in the fourth or fifth decade of life; it occurs in women more often than in men and in whites more frequently than in blacks. The purpose of this essay is to illustrate the broad spectrum of neuroimaging findings seen on CT scans, MR images, and angiograms of patients with cephalocervical fibromuscular dysplasia. The value of MR imaging in distinguishing tubular fibromuscular dysplasia from arterial dissection and arterial hypoplasia is demonstrated.

Objective: The purpose of this study was to evaluate the usefulness of diffusion-weighted echoplanar MR imaging in the examination of high-grade brain gliomas compared with that of conventional spin-echo (SE) or fast spin-echo (FSE) MR imaging. We hypothesize that diffusion-weighted MR imaging may enable us to differentiate various tumor components on the basis of differences in the diffusion of water., Subjects and Methods: Conventional SE and FSE MR images were obtained in 10 patients with high-grade brain glioma. Diffusion-weighted echoplanar MR images were obtained with a head gradient coil capable of providing diffusion-weighted imaging along the cephalocaudal axis. Using SE and FSE MR images as a baseline, we evaluated the diffusion-weighted MR images for usefulness in distinguishing tumor components on the basis of differences in diffusion., Results: Areas of tumor that showed significant enhancement on T1-weighted SE MR images obtained after injection of contrast material were markedly hyperintense on diffusion-weighted images and had a lower apparent diffusion coefficient (ADC) than the ADCs for nonenhancing tumor and peritumoral edema. Cystic or necrotic portions of tumor showed the most signal suppression on diffusion-weighted images and were associated with the highest ADCs. On T2-weighted FSE MR images, areas of hyperintensity observed in white matter oriented parallel to the direction of the diffusion gradient could be differentiated into two patterns on the basis of findings on diffusion-weighted images: areas that showed marked signal suppression with a higher ADC, most likely representing areas of predominantly peritumoral edema, and areas that showed a lesser degree of signal suppression with similar but slightly lower ADCs than those of edema, most likely representing areas of predominantly nonenhancing tumor., Conclusion: Diffusion-weighted echoplanar MR imaging is a useful technique for examining high-grade cerebral gliomas. It enabled us to differentiate various components of the tumor (e.g., enhancing, nonenhancing, cystic, or necrotic) and to distinguish areas of predominantly nonenhancing tumor from areas of predominantly peritumoral edema when the abnormality was located in the white matter aligned in the direction of the diffusion-weighted gradient. Diffusion-weighted echoplanar MR imaging appears to be a powerful tool in the characterization of brain neoplasms.

Objective: With the increased survival of patients with severe immunosuppression, it has become more important to recognize the various forms of cerebral and craniofacial aspergillosis. Currently, only small series of patients with this infection have been described; the radiographic diagnosis of cerebral and craniofacial aspergillosis has varied and has been relatively nonspecific. The purpose of our study was to identify neuroimaging patterns in patients with cerebral and craniofacial aspergillosis. Recognition of radiographic patterns of aspergillosis may facilitate earlier radiologic diagnosis and prompt therapy., Materials and Methods: The imaging and clinical data of eight immunosuppressed patients with cerebral aspergillosis and one patient each with aspergillosis of the orbit, paranasal sinus, and calvaria were evaluated retrospectively. All patients were at risk of developing infection by virtue of poorly controlled diabetes or other types of congenital or acquired immunosuppression (e.g., steroids, chemotherapy). Patients were selected for study if the diagnosis of aspergillosis was established by means of biopsy or autopsy and CT scanning or MR imaging was available for review. CT scans and MR images were compared by two experienced neuroradiologists, who were aware of the diagnosis of aspergillosis, to see if common radiographic patterns could be identified that could be used as predictors of this type of infection., Results: Five patients with cerebral aspergillosis had multiple ring-enhancing lesions consistent with abscesses. Characteristic findings were multiple lesions, an irregular ring of contrast enhancement, and hypointensity of the ring on T2-weighted MR images. Three patients had cortical and subcortical hypodensities on CT scanning or hyperintensities on MR imaging consistent with cerebral cortical and subcortical infarction. Two of these three had superimposed hematoma formation. Three patients had craniofacial aspergillosis. One patient each had enhancing mucosal thickening of the paranasal sinus with secondary intracranial dural enhancement, abnormal enhancement of the optic nerve and sheath with infiltrating enhancing soft tissue within the intraorbital fat, and an enhancing diploetic lesion of the calvaria with underlying dural enhancement., Conclusion: Three different neuroimaging patterns of cerebral aspergillosis were identified in immunosuppressed patients. The first pattern was multiple areas of hypodensity on CT scans or hyperintensity on T2-weighted MR images involving the cortex and/or subcortical white matter consistent with multiple areas of embolic infarction. This pattern could be seen with or without superimposed hemorrhage, identified as hyperdensity on CT scans or as hyperintensity on T1-weighted MR images. The second pattern was multiple intracerebral ring-enhancing lesions consistent with abscesses. The ring was irregular and of low signal on T2-weighted MR images. The third pattern was dural enhancement associated with enhancing lesions in the adjacent paranasal sinus structure or calvaria or dural enhancement of the optic sheath with associated optic nerve and intraorbital fat enhancement. Recognition of these three patterns of aspergillosis in immunosuppressed patients may lead to more effective diagnosis and treatment planning.

Current methods of neuroimaging have revolutionized the examination of patients with dementia. Clinical examination of these patients has identified cortical and subcortical patterns of dementia. Each of these patterns can be caused by various diseases. Mixed patterns also occur and also have several possible causes. Neuroimaging can be quite useful in confirming the clinical assessment of the pattern of dementia and in further refining the diagnosis. We review the pathophysiologic basis of and the clinical and neuroimaging findings in dementia, both cortical and subcortical. The roles of neuroimaging in diagnosis, prognostic assessment, and treatment evaluation are discussed.

Herpesvirus infections of the CNS are prevalent in all segments of the population, and are an important health care issue. The family of herpesviruses consists of a large group of double-stranded DNA viruses that includes herpes simplex virus type 1 (HSV-1), herpes simplex virus type 2 (HSV-2), cytomegalovirus, Epstein-Barr virus, varicella-zoster virus, B virus, herpesvirus 6, and herpesvirus 7. In addition to producing infection when the host initially acquires the virus, an important property shared by these viruses is the ability to produce latent infection and to be reactivated.

Objective: Mesenrhombencephalitis is a serious form of brainstem inflammation predominantly involving the deep and vital portions of the brain, that is, the mesencephalon (midbrain) and rhombencephalon (pons, medulla). Mesenrhombencephalitis is difficult to diagnose on the basis of clinical and laboratory findings alone, and access to this portion of the brain for surgical biopsy carries high morbidity. We describe the MR appearance of mesenrhombencephalitis and correlate the imaging findings with clinical information., Materials and Methods: Unenhanced and contrast-enhanced MR images of nine patients with mesenrhombencephalitis were reviewed retrospectively and correlated with clinical, laboratory, and pathologic data. The patients were categorized according to the cause of the disease: three had herpes simplex, one had Listeria monocytogenes, and five had mesenrhombencephalitis of undetermined cause. The three patients with clinical and MR evidence of herpes simplex mesenrhombencephalitis (one confirmed by brain biopsy) were comatose at presentation, with cranial nerve abnormalities in two and seizures in one. One patient with L. monocytogenes (established by blood culture) had cranial nerve palsies, fever, and pain in the ear. Five additional patients had headache (three), fever (three), nausea and vomiting (four), cranial nerve palsies (three), coma (two), and hyporeflexia (one) or hyperreflexia (four). Brain biopsy performed in two patients revealed chronic inflammation of unspecified cause; in one, it was compatible with viral encephalitis., Results: MR images in three patients with herpes simplex mesenrhombencephalitis showed T2 signal hyperintensity in the midbrain (two), pons (one), medulla (one), and temporal lobes (three). Parenchymal foci of hemorrhage (methemoglobin, one patient) and leptomeningeal enhancement (one patient) were identified in the temporal lobes. T2-weighted MR images in one patient with L. monocytogenes showed signal hyperintensity in the brainstem, vermis, midbrain, and internal capsules. On T1-weighted images, low signal was present in these areas, which enhanced with paramagnetic contrast agents. In the remaining five patients, T2-weighted MR images showed patchy signal hyperintensity in the pons, medulla, and thalamus in three each and in the midbrain and temporal lobes in one each. T1-weighted MR images showed normal findings (two) or signal hypointensity in the thalamus and pons in one patient each. Areas of leptomeningeal and parenchymal enhancement were identified in one patient each. Brainstem swelling was seen in three patients, one of whom had petechial hemorrhage in the pons and hydrocephalus., Conclusion: Mesenrhombencephalitis is a serious illness that is diagnosed by a combination of imaging, clinical, laboratory, and pathologic studies. MR imaging may be crucial to the early diagnosis of this illness, and radiologists must be familiar with this uncommon entity and its MR findings in order to make timely diagnoses and facilitate treatment.

Objective: Pleomorphic xanthoastrocytoma is a rare, usually benign brain tumor. This pleomorphic supratentorial tumor involves the leptomeninges and superficial cortex in young patients with seizures. The MR imaging features of this distinct tumor have not been reported. We describe the MR imaging findings in six patients with pathologically proved pleomorphic xanthoastrocytoma., Materials and Methods: We retrospectively reviewed the MR images of six patients with pathologically proved pleomorphic xanthoastrocytoma. MR images were reviewed by two neuroradiologists. The pathologic slides of the tumors were reviewed by two neuropathologists. We also analyzed the patients' clinical features., Results: Most of the tumors were cortical based and isointense with gray matter on T1-weighted images and mildly hyperintense on T2-weighted images. All the masses enhanced with contrast material. Cystic components and gyriform and leptomeningeal enhancement were seen occasionally. The tumors occurred in the temporal lobes in four of the six patients. Three patients had seizures, and the other three had headaches., Conclusion: Pleomorphic xanthoastrocytoma is a rare, usually benign, cortical-based mass that often enhances intensely with contrast material. The most common location is in the temporal lobes. Seizures and headaches are common clinical features. Familiarity with this lesion is important in the differential diagnosis of enhancing cortical-based masses.

This essay illustrates the appearances of sections of the normal hippocampus and adjacent temporal lobe on high-resolution heavily T2-weighted fast spin-echo MR images and correlates them with histologic sections. We found that this MR examination showed the detailed anatomy of the normal hippocampus in a much shorter time than is possible with conventional spin-echo techniques. The information provided in this essay can be used as a baseline for distinguishing between normal and abnormal hippocampi in a variety of disease states.

Objective: This study evaluated the effectiveness of steady-state free-precession (SSFP) MR imaging of complex cystic masses of the brain compared with that of conventional T1- and T2-weighted spin-echo imaging. Our hypothesis is that SSFP MR images provide better characterization of these masses and facilitate more appropriate preoperative diagnoses and planning., Subject and Methods: Axial T1-weighted and SSFP MR images and specimens for pathologic examination were obtained in seven consecutive patients, 9-81 years old, with cystic mass lesions of the brain and neurologic symptoms and signs directly related to the masses. Axial contrast-enhanced T1-weighted images were obtained in six patients, surgical exploration was done in five patients, and stereotaxic biopsy was done in two. After examination of the routine spin-echo and SSFP images, the usefulness of SSFP images was determined by how well they facilitated correct preoperative diagnosis., Results: On SSFP MR images, the solid or inhomogeneous components of a cystic mass had extremely low signals in contrast to the high signal of surrounding fluid. On routine spin-echo images, however, the signals of these components were masked by the signal of the surrounding fluid. SSFP MR images helped markedly in diagnosis of hemorrhagic, epidermoid, and arachnoid cysts. In cases of enhancing brain tumors, SSFP MR images provided the same information that contrast-enhanced images did. Overall, when SSFP MR imaging was used, more information about the texture and constituents of the cystic mass was obtained, and a more useful diagnosis was made., Conclusion: Initial results show that SSFP MR imaging is a more useful technique than conventional spin-echo imaging for characterizing complex cystic masses of the brain. SSFP MR imaging (1) allows distinction of edema from tumor, (2) helps establish where biopsy has the best chance of providing tissue that will show pathologic changes, and (3) helps distinguish simple cysts from tumors, tumor-cyst, or multicompartmental cyst and may be particularly helpful in detecting the contents of hemorrhagic cysts.

Advances in MR imaging continue to improve our ability to evaluate temporal bone anatomy and disease. CT remains the procedure of choice for fine-detail imaging of bone structures such as ossicular anatomy, but it is not the ideal imaging technique for soft-tissue structures (e.g., the membranous labyrinth and neural structures). Conventional spin-echo MR techniques used to image these structures cannot yield excellent contrast and spatial resolution in clinically acceptable time frames. Conventional spin-echo T1-weighted images lack tissue contrast between fluid (e.g., CSF, endolymph, perilymph), neural tissue, otic capsule septa, and surrounding temporal bone. Conventional T2-weighted imaging of the inner ear is needed to reveal the natural contrast between fluid, neural structures, and bone; unfortunately, the use of conventional T2-weighted images is limited by time constraints when large-matrix, thin-section techniques with more than one excitation are used. Fast spin-echo imaging is a recently developed technique that can provide T2-weighted, thin-section (2-mm) high-resolution images with excellent contrast in a fraction of the time needed for conventional spin-echo techniques. This speed advantage allows us to obtain high-resolution images in clinically acceptable time frames. Images produced by this technique are a useful addition, in conjunction with routine T1- and T2-weighted spin-echo images, in the diagnosis of disorders of the inner ear.

Fat-suppression MR techniques have been useful in imaging both the head and neck and the orbit. The lumbar spine is not as complex as the head and neck and does not contain as much fat as the orbit. Nevertheless, high signal from epidural and marrow fat in the lumbar spine can make it difficult to detect and define pathologic processes. In addition, the administration of gadopentetate dimeglumine can cause some infectious and neoplastic lesions of the spine to become isointense with fat on T1-weighted images. Compared with conventional MR imaging without fat suppression, contrast-enhanced fat-suppression MR imaging can better differentiate these enhancing lesions from surrounding fat. In this essay, we illustrate the findings of fat-suppression MR imaging in a variety of diseases of the lumbar spine.

Rasmussen's encephalitis is a devastating disease of childhood causing progressive neurologic deficits and intractable seizure activity. Patients frequently have episodes of epilepsia partialis continua and, much less frequently, generalized status epilepticus. The seizures are intractable despite aggressive medical management. In advanced cases, hemispherectomy appears to be the only option to control the seizures. Permanent physical and mental impairments are inevitable. The cause of this disease is unknown, although pathologic specimens demonstrate nonspecific changes that are compatible with viral encephalitis. The progressive brain damage is typically so insidious in onset and gradual in course that it is difficult to make an accurate diagnosis on the basis of clinical evidence. We retrospectively evaluated the CT, xenon CT, positron emission tomographic, and MR neuroimaging findings in four young patients with pathologically suspected Rasmussen's encephalitis (three patients had CT scans, two had xenon CT scans, two had MR scans, and one had a positron emission tomogram). All studies showed abnormalities of the involved cerebral hemisphere: CT and MR revealed nonspecific atrophy, xenon CT showed decreased cerebral blood flow, and positron emission tomography revealed a hypometabolic state. Rasmussen's encephalitis is a diagnosis of exclusion; however, the information obtained from neuroimaging studies in combination with the clinical course should suggest this disorder.

Neurosyphilis, a sexually transmitted disease that can cause neurologic damage, has become increasingly prevalent in the AIDS era. HIV carriers can contract neurosyphilis without the presence of other concurrent opportunistic infections. Because MR findings of neurosyphilis are seldom reported, we retrospectively reviewed and evaluated contrast-enhanced MR images of six young (average age, 33 years) HIV-positive men with high serum and CSF VDRL titers indicative of neurosyphilis. All six patients tested negative for concurrent opportunistic infections. Five patients had acute or subacute strokelike symptoms involving the basal ganglia or middle cerebral arteries; one had a parietal convexity mass mimicking meningioma with headache and ataxia. Contrast-enhanced MR images showed patchy enhancement involving the basal ganglia and middle cerebral artery territories in the first five patients and the convexity mass in the sixth patient. On the basis of brain biopsy, a convexity mass was diagnosed in the patient with syphilitic gumma. The imaging findings of the remaining five patients represented ischemic infarct caused by meningovascular syphilis. After penicillin treatment, serum and CSF VDRL titers decreased, and neurologic signs and symptoms improved in all six patients. A follow-up MR study in the patient with the gumma showed that the lesion resolved almost completely. In young HIV patients with stroke symptoms or a convexity mass, neurosyphilis should be considered. Contrast-enhanced MR can reveal the extent of involvement by neurosyphilis and should be used to facilitate diagnosis and proper treatment.

Crossed cerebellar diaschisis is a condition in which cerebellar hypometabolism is ascribed to functional disconnection of the contralateral hemisphere from the cerebral cortex. Interruption of the cerebropontine-cerebellar pathway is thought to be the most likely mechanism of this remote transneuronal metabolic depression. This phenomenon can be diagnosed by positron emission tomography, as there is hypometabolism in the affected cerebellar hemisphere. The morphologic features of the affected cerebellar hemisphere in patients with the diagnosis of crossed cerebellar diaschisis have not previously been studied by MR imaging. We retrospectively reviewed 26 patients in whom the diagnosis of crossed cerebellar diaschisis was suggested by positron emission tomography; these patients also had MR studies. In all 26 patients, supratentorial diseases were documented by MR imaging. Twenty-four of the 26 patients had a pathologic diagnosis. Detailed clinical history was also obtained for all patients. MR findings of morphologic change in the affected cerebellum were correlated with the patient's clinical symptoms and supratentorial disease. Of the 26 patients, eight had cerebellar atrophy; the remaining 18 patients showed no MR abnormality of the affected cerebellum. In the 18 in whom no atrophy was seen on MR imaging, the majority of the supratentorial lesions were tumors. The eight patients in whom cerebellar atrophy was demonstrated usually had significant contralateral supratentorial hemispheric atrophy. The patients with atrophy were also generally younger (average age, 14 years) than the 18 patients without cerebellar atrophy (average age, 42 years). Our experience shows that a significant number of patients with crossed cerebellar diaschisis have morphologic changes of cerebellar atrophy shown by MR imaging. These patients usually have significant contralateral supratentorial hemispheric atrophy.

It has been reported that on MR images of the brain in neonates the entire pituitary gland is hyperintense. This is quite different from the well-established observation that portion is isointense. To better understand the development of the neonatal pituitary gland, we studied the timing and magnitude of changes in the size, shape, and MR signal of the pituitary gland in early childhood and infancy. The study included 76 infants and children ages 3 days to 4 years old. In each MR study, the signal intensity of the pituitary gland on T1-weighted coronal (n = 24) and sagittal (n = 76) images was compared with the signal intensity of the pons by using region-of-interest methods. The size and shape of the pituitary glands were recorded. Statistical comparisons were made for signal intensity and size of the pituitary gland between boys and girls and among age groups. The results showed that the MR signal of the posterior lobe remained hyperintense relative to the pons for all age groups represented in our study (through 4 years). However, the signal intensity of the anterior lobe gradually diminished and by approximately age 2 months was isointense with the pons. Thus, by approximately age 2 months the MR signal intensity of the infant's pituitary gland has changed from being entirely hyperintense (as seen in neonates) to isointense in the anterior lobe and hyperintense in the posterior lobe (as seen in adults). In the first 2 months of life the pituitary gland is bulbous in shape; later it has a flatter upper surface. We found no significant changes in the length or signal intensity of the pituitary gland; no differences among sexes in size, shape, or signal intensity were demonstrated. Findings of a bulbous pituitary gland and high MR signal of the entire pituitary are normal in neonates up to 2 months of age. After that, the gland gradually flattens and the MR signal of the anterior lobe decreases until it has characteristics similar to the adult pituitary. Lack of high signal or discovery of a small pituitary gland in a neonate should alert the radiologist to the possibility of pituitary malformation or dysfunction.

In an effort to study the blood supply of the pituitary gland, we determined the sequence in which various portions of the gland enhanced on MR images obtained after administration of gadopentetate dimeglumine in seven healthy volunteers. The pituitary gland has a unique blood supply: the pars nervosa is supplied by the inferior hypophyseal arteries; the infundibular stem is supplied by the superior hypophyseal arteries, which are more distal than the inferior hypophyseal arteries; and the pars distalis is supplied indirectly by portal systems from the infundibular stem. Therefore, the expected sequence of contrast enhancement of the pituitary gland is: the pars nervosa first, the infundibular stem second, and the pars distalis last. However, it is difficult to evaluate the temporal sequence of contrast enhancement with "slow" conventional MR techniques, such as routine enhanced images acquired minutes after contrast administration. Accordingly, we used rapid enhanced gradient-echo sequences (14 sec/image) and the region-of-interest method to determine contrast enhancement for various portions of the gland. Enhancement-time curves were then plotted for each portion and compared to determine the sequence of enhancement. Our results show that the timing of contrast enhancement of various parts of the pituitary gland as demonstrated on MR imaging correlates closely with the expected perfusion sequence based on knowledge of regional vascular anatomy. Normal enhancement-time curves of different portions of the pituitary gland were established. This technique and baseline data can be used to evaluate patients with hypopituitarism caused by ischemia.

Fat-suppression techniques are useful in MR imaging to eliminate strong signals from fatty tissues that interfere with signals from adjacent areas. Various methods of fat suppression have been devised, but when suppression of fat is used in combination with contrast enhancement employing paramagnetic agents (e.g., gadopentetate dimeglumine), the definition of normal anatomic structures is significantly improved, enhancing lesions become more conspicuous, and lesional margins are better defined in regions of the body with large amounts of fat, whose signal is suppressed. Contrast-enhanced fat-suppressed T1-weighted images provide more information than do conventional MR images. In this review, several types of fat-suppression techniques and their clinical applications in neuroradiology are described. Gadopentetate dimeglumine-enhanced, fat-suppressed T1-weighted images appear to have significant advantages over conventional T1-weighted contrast-enhanced images and should replace them in imaging regions of the body where large amounts of fat are present.

One tenth of all CNS neoplasms involve the ventricles of the brain. These mass lesions are located in areas that are difficult to reach surgically, and because they are intraventricular, spread via the CSF is common. CT and MR imaging have been useful in demonstrating these masses, but imaging characteristics are usually nonspecific. The location of the mass and the patient's age provide the most helpful information in the differential diagnosis.

MR images of pineal region tumors were analyzed in 26 patients with histologically proved tumors: seven germ-cell tumors, six astrocytomas, five teratomas, three pineoblastomas, two meningiomas, one dermoid, one epidermoid, and one metastasis. In an attempt to identify specific MR characteristics of these lesions, Gd-DTPA was administered to six patients. CSF and blood serum were assayed for alpha-fetoprotein (AFP) and human chorionic gonadotropin-beta subunit (HCG-beta) in 18 patients. MR findings were correlated with age, sex, the presence of biochemical tumor markers, and surgical outcome. We found that the most important factors in the determination of tumor type were the patient's age and the tumor markers. Increased levels of both HCG-beta and AFP were specific for patients with malignant teratomas and undifferentiated germ-cell tumors. HCG-beta alone was elevated in the patient with choriocarcinoma; only AFP was elevated in the patient with an endodermal sinus tumor. Tumor markers were not present in other patients in this series. The tumor size and the presence of fat were also helpful in determining tumor type. Hemorrhage was rare, seen only in the patient with a choriocarcinoma. Gd-DTPA did not enhance diagnostic specificity but aided in the detection of tumor seeding through CSF. We conclude that, although MR is sensitive in the detection of pineal region tumors and provides superb anatomic detail, MR signal characteristics are usually nonspecific. Correlation with the patient's age and the tumor markers significantly improves diagnostic accuracy.

Studies of twenty-five patients with loculated leptomeningeal tumor metastases diagnosed by CT and/or MR were analyzed retrospectively. Medulloblastoma was the most frequent primary tumor (8/25, 32%). Four subgroups of loculated patterns were identified. Type A included mass(es) limited to the subarachnoid space without obvious direct parenchymal infiltration; this pattern occurred in 12 patients, of whom five had associated diffuse pattern. Type B was characterized by mass(es) still predominantly in the subarachnoid space but with minor transpinal parenchymal infiltration; this pattern was found in five patients. Type C comprised subarachnoid mass(es) with marked transpinal extension mimicking parenchymal lesion; this pattern was observed in three patients. Type D consisted of subarachnoid mass(es) growing along the perineural CSF space; this pattern was noted in two patients. Additionally, two patients presented with combined A and C patterns, and one patient had a combined B and C pattern. More than half the patients (14/25, 56%) presented with a single lesion. The most frequent locations were the suprasellar cistern, ventricular walls, and lateral recesses of the fourth ventricle, Gd-DTPA-enhanced T1-weighted MR images appeared best for demonstrating the site and extent of disease. Recognition of the loculated patterns of leptomeningeal metastases, which are less common than the diffuse pattern, is important to radiologists and clinicians for correct diagnosis and proper management of patients with this disease.