| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Neuroradiology |
1 From the Departments of Radiology (J.L.D., M.B., D.G.), Pneumology (D.V., H.T.L.), and Neurology (C.B.), Hôpital Avicenne, 125 route de Stalingrad, 93009 Bobigny, France; and the Department of Internal Medicine, Groupe Hospitalier Pitié-Salpêtrière, Paris, France (C.C.A., J.C.P.). Received July 31, 1998; revision requested September 1; revision received May 5, 1999; accepted July 14. Address reprint requests to J.L.D.
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: The abnormalities detected in 24 patients (mean follow-up, 36 months) were compared before and after therapeutic periods (n = 75) that were divided into attack (high-dose), upkeep (decreased-dose), and minimal (low-dose) periods. Parenchymal lesions were classified as type 1 (enhanced with gadolinium), type 2 (demyelinating), or type 3 (lacunar) and were assessed as regressing, stable, or progressing.
RESULTS: Seven of the 24 patients had several types of lesions. Isolated type 3 lesions (six patients) were the only lesions not associated with neurologic deficit. Type 1 lesions (13 patients) regressed in 22 of 22 attack periods and progressed in nine of 27 upkeep and minimal periods. MR imaging depicted relapses in patients with multifocal CNS involvement or long-standing CNS impairment or in those who had previously received steroid therapy. Type 2 (seven patients) and type 3 (13 patients) lesions remained stable in 68 of 68 therapeutic periods. Type 1 lesions appeared in three patients with type 2 and type 3 lesions during two upkeep and three minimal periods. Findings at follow-up MR imaging contributed to the reintroduction of high-dose corticosteroid therapy in eight patients.
CONCLUSION: MR imaging can be used to differentiate between reversible and irreversible lesions in CNS sarcoidosis. MR imaging can be a useful tool for adjusting treatment to prevent irreversible CNS damage.
Index terms: Brain, MR, 10.121411, 10.121412, 10.12143 • Sarcoidosis, 10.22, 30.22 • Skull, MR, 122.121411, 122.121412, 122.12143 • Spine, MR, 30.121411, 30.121412, 30.12143 • Steroids • Vasculitis, 17.62
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
There is no consensus about how to treat neurosarcoidosis. High doses of corticosteroids are recommended for initial treatment (6), but the most appropriate ways of tapering and discontinuing treatment in the light of the response have not been codified (1). Immunosuppressive agents are occasionally used as adjuncts to corticosteroids (3,6), although there is no guarantee of a clinical response (5). The difficulty in therapeutic management arises mainly from the lack of correlation between the clinical signs and the progress of the illness (4,5).
Magnetic resonance (MR) imaging has been used to monitor CNS sarcoidosis (7–10) and has proved useful in the identification of lesions such as granulomatous inflammation and vasculopathy (11). Nevertheless, there is limited information about how MR imaging findings change in response to treatment (9,10,12,13) and about the characteristics of lesions in the white matter (14,15).
This study was, therefore, performed to document the capacity of MR imaging in the evaluation of reversible and irreversible CNS lesions in cases of neurosarcoidosis treated with variable-regimen steroid therapy. We also analyzed how MR imaging can help physicians to adapt corticosteroid treatment.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Patients
The 24 patients (15 women and nine men) were aged 19–58 years (mean age, 38 years) at the time of inclusion. Sarcoidosis was diagnosed on the basis of the compatible clinical, radiographic, and clinical laboratory findings; the evidence of noncaseous granuloma; the absence of mycobacterial and fungal infection; and the absence of exposure to airborne contaminants or medications known to cause granulomatous disorders.
All patients had manifestations at extraneurologic sites, which were thoracic (n = 4), extrathoracic (n = 6), or both (n = 14). Histopathologic evidence was obtained in 21 patients by means of a biopsy of the bronchi (n = 5), skin (n = 5), peripheral lymph nodes (n = 3), salivary glands (n = 3), other extraneurologic sites (n = 4), or brain (n = 1). The remaining three patients had systemic sarcoidosis, which was diagnosed from findings of uveitis and aseptic lymphocytic meningitis that responded to steroid therapy. All 24 patients had sarcoidosis for 0–30 years (mean, 7 years); this was estimated from the time between the diagnosis and the inclusion of the patient.
At the time of inclusion, all patients presented with MR imaging evidence of parenchymal CNS lesions that reasonably could be attributed to sarcoidosis. Patients with lesions from other causes, such as other neurologic disease, were excluded; those with negative findings at work-up for multiple sclerosis, arterial hypertension, or diabetes mellitus were included. No patient with MR imaging findings suggestive of a single meningeal involvement was included.
Of the 24 patients, 18 had clinical symptoms of CNS involvement and/or hypothalamic dysfunction (Table 1). The duration of CNS involvement was estimated as the time between the onset of manifestations and the inclusion of the patient, which was between 0 and 14 years (mean, 3 years). In 11 patients, CNS sarcoidosis was of recent onset (within 2 years or less), and seven patients had CNS sarcoidosis for more than 2 years. The remaining six patients had clinically silent brain abnormalities that recently had been discovered at MR imaging.
|
Immunosuppressive agents were added to the long-term low-dose (<15 mg daily) treatment with corticosteroids in six patients. No patients underwent cerebral radiation therapy. Twelve patients had never been given corticosteroids before initial MR imaging. Two patients had been treated previously, but they had not received treatment during the 2 years before inclusion. Ten patients had received corticosteroids within the 2 years before inclusion (treatment had lasted longer than 2 years in seven patients), but three had been without treatment for more than 3 months.
MR Imaging
Evaluation included MR imaging examinations of the brain and sella turcica in all patients. Spinal MR imaging was performed when it was indicated by the clinical manifestations. MR imaging was performed with a 0.5-T unit (Vectra; GE Medical Systems, Milwaukee, Wis) with a quadrature head or spine coil. Contrast material enhancement was provided by means of an intravenous injection of 0.1 mmol/kg gadoterate meglumine (Dotarem; Guerbet, Roissy, France).
Each brain examination included the acquisition of transverse T1-weighted spin-echo images (550/18, repetition time [TR] msec/echo time msec; two signals acquired) before and after enhancement and transverse conventional dual-echo intermediate-weighted and T2-weighted spin-echo images (2,400/40–100, one signal acquired). The sections were 7-mm thick with a 1-mm gap. The image matrix was 192 x 256. Additional images were obtained in the coronal plane, with a section thickness of 5 mm for most of the patients. The region of the hypothalamic-pituitary axis was investigated in both the sagittal and coronal planes by using a T1-weighted spoiled gradient-echo sequence (50/12, 45° flip angle) with contiguous 2-mm-thick sections before and after the administration of contrast material.
The spinal cord was imaged in the sagittal and transverse planes, before and after enhancement, with a T1-weighted spin-echo sequence (450/25, four signals acquired) and a T2-weighted fast spin-echo sequence (2,500/100, four signals acquired) with 4-mm-thick sections and a 1-mm gap.
Follow-up and Therapeutic Periods
Follow-up MR images were acquired in the most appropriate sites and planes by using the same parameters as before. Contrast material enhancement was always used whether the lesions enhanced on the initial images or not. The times for follow-up imaging were set as follows: (a) before each substantial reduction in the dose of the corticosteroids when high doses were used for a sufficient time (3 months or more), (b) each time a high dose of corticosteroids was reintroduced (regardless of the reason), and (c) routinely during the long-term low-dose maintenance periods after a good response was obtained.
A total of 162 MR imaging examinations were performed (mean number per patient, 6.7; range, two to 21). The mean interval to MR imaging follow-up was 6.5 months (range, 1–46 months; SD, 5.7). Patients were followed up for an mean of 36 months (range, 3–98 months).
The following therapeutic periods were identified according to the corticosteroid regimen and the time of the follow-up MR imaging examination: attack, upkeep, and minimal. The attack period involved high doses of 50–80 mg/d for at least 3 months to obtain regression of the observed lesions. The upkeep and minimal periods involved reduced doses of corticosteroids. During the upkeep period, the dose was slowly tapered over several months to 20–40 mg/d to prolong the improvement already achieved. The minimal period involved a low maintenance dose of 5–15 mg/d to prevent relapses. The demarcation between the successive therapeutic periods was taken to be the time at which a follow-up MR imaging examination was associated with the change in the category of steroid therapy.
A total of 75 therapeutic periods were followed at MR imaging, with 33 attack periods (mean duration, 7 months; range, 3–14 months), 30 upkeep periods (mean duration, 15 months; range, 4–53 months), and 12 minimal periods (mean duration, 16 months; range, 8–35 months).
Lesion Classification
Parenchymal lesions were characterized by their contrast enhancement, signal intensity characteristics, morphology, location, and distribution. For each MR imaging examination, lesions were classified as one of three types on the basis of the presence of contrast enhancement or isolated white matter disease. Any parenchymal enhanced lesion on T1-weighted images was classified as type 1, regardless of the corresponding abnormalities seen on T2-weighted images. There was no evidence of enhancement in the type 2 and type 3 lesions. These lesions were the most striking on images acquired with long TRs, which showed white matter disease. The lesions showed high signal intensity on both T2-weighted and intermediate–weighted images.
Type 2 and type 3 lesions were defined by their distribution and pattern on T2-weighted images. Type 2 lesions included nodular or confluent abnormalities with high signal intensity, which closely resembled those of multiple sclerosis, in the periventricular regions and in the deep white matter. Type 3 lesions included multifocal and patchy lesions of increased signal intensity that ranged from 2 to 8 mm in diameter in the subcortical white matter; these lesions had an appearance very similar to that seen in small-vessel atherosclerotic disease. A single patient could, therefore, display lesions of different types. Type 2 lesions were found in the brain and spinal cord parenchyma, whereas type 3 lesions were located in the brain. Cerebral atrophy was not investigated.
The sites of involvement were divided into the brain parenchyma, the hypothalamic-pituitary axis, and the spinal cord. Abnormal brain enhancement corresponded to a deep parenchymal and periventricular distribution, to gyral and leptomeningeal enhancement with subcortical extension, or to an involvement of the optic chiasma. We assessed the combinations of the three types of lesions and the combined involvement of several sites.
Study Design
We compared the parenchymal abnormalities observed on the MR images at a time T that indicated the end of one therapeutic period and the beginning of another. The change in MR imaging features between T and T + 1 were assessed for each type of lesion according to the category of the corresponding therapeutic period. Lesions of different types in a single patient were, therefore, assessed separately. Each consecutive therapeutic period for a given patient was also analyzed separately. Finally, we attempted to define each type of lesion by its development profile.
All MR imaging examinations were retrospectively reviewed by a senior neuroradiologist (J.L.D.) who was blinded to the clinical course and treatment of the patients. MR imaging examinations that revealed new abnormalities or changes in the preexistent abnormalities were reassessed in a consensual, nonblinded fashion (to avoid any disagreement about the clinical response to therapy when it became apparent) by clinicians who were experienced with neurosarcoidosis (D.V., C.C.A., C.B.).
Each study was evaluated for the following five markers of lesion development: complete regression, partial regression, stability, progression, and appearance. The complete regression of lesions was defined as their disappearance; partial regression was defined as a decrease in lesion size and/or number. Stable lesions were defined as lesions that showed no change between times T and T + 1. Progression was defined as an increase in lesion size and/or number. Lesion appearance indicated the appearance of additional lesions.
The predominant MR imaging pattern identified when a patient was included was also correlated with the recorded clinical data. The evolution of the MR imaging findings for a subgroup of 12 patients with type 1 lesions who were monitored for at least 12 months while they received reduced doses of corticosteroids was analyzed with reference to the duration of the neurologic disease, the start of steroid therapy, and the radiographic appearance. All of these had been evaluated at the time of the patient's inclusion. Finally, the usefulness of MR imaging features in deciding the therapeutic management of the patients was assessed.
The mean values of the data in the groups of patients (two groups or more) were compared by using the Mann-Whitney U test and the Kruskal-Wallis test.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
|
|
|
|
|
Changes in the MR Imaging Findings
Only the type 1 lesions evolved (Figs 1b, 2b, 3b). They regressed during all attack periods (22 of 22) but progressed in one-third of the reduced-dose periods (nine of 27) (Table 4). Type 1 lesions also appeared during two upkeep and three minimal periods in three patients who initially had no type 1 lesions (Fig 6). After we began tapering the corticosteroid treatment, the mean interval to follow-up MR imaging during which type 1 lesions appeared was 14.8 months (SD, 12.3). The interval during which they progressed was 13.7 months (SD, 9.1), and the interval during which they regressed or remained stable was 15.7 months (SD, 8.1). The Kruskal-Wallis test showed no significant difference (P = .8) between the three subgroups. Initial lesions classified as type 2 or type 3 did not change in any of the therapeutic periods (n = 68). New type 3 lesions appeared during only one upkeep period in a patient whose MR images showed leptomeningeal enhancement (Fig 7).
|
|
|
|
|
Contribution of MR Findings to Changes in Steroid Therapy during Follow-up
After each attack period, the regression of type 1 lesions depicted on MR images was considered to be a contributive criterion for starting the tapering of corticosteroid therapy. The clinical responses to therapy always matched the findings on the MR images. During the upkeep periods, findings on the MR images also contributed to the adjustment of steroid therapy as follows: (a) When the regression induced by the attack treatment was prolonged, the corticosteroid dose was further decreased, and (b) when type 1 lesions progressed or appeared, high doses were reintroduced.
In eight patients, MR imaging data contributed to the reintroduction of periods of high-dose treatment (n = 13) with (n = 6) or without (n = 7) clinical correlation (Table 5). Finally, of the 16 patients with type 1 lesions in whom MR imaging findings helped us to adjust the treatment, 11 had favorable long-term outcomes, two were no longer followed up at MR imaging after the last effective attack period, and three showed a progression of the type 1 lesions on the last MR images that were acquired while the patient was receiving reduced doses of corticosteroids. In the eight other patients who did not show type 1 lesions, the long-term treatment was based on extraneurologic features (n = 7) or neurologic clinical signs (n = 1).
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The development of neurosarcoidosis is primarily leptomeningeal and vascular in nature. It may result in the disruption of the leptomeningeal blood-brain barrier (9), which permits the granulomatous infiltrate to enter the brain parenchyma along the so-called perivascular or Virchow-Robin spaces (16) that accompany the penetrating arteries up to the capillaries (17). Vasculitic lesions and perivascular involvement may cause stenosis, which results in vasculopathy (9,16), and multiple granulomas may coalesce to produce intraaxial masses, often with adjacent edema (13,18,19). Findings from this study show that the MR patterns can be used to assess the specific histopathologic lesions of CNS sarcoidosis. Although these lesions are all caused by the same basic process, each could have a specific practical importance.
Type 1 lesions exhibit gadolinium enhancement that reflects damage to the blood-brain barrier and have been extensively reported. They are generally considered to suggest neurosarcoidosis (9,10). They occurred frequently in our patients and were typically distributed, with cortical enhancement. They spread perivascularly into the parenchymal extension (9,16), particularly along the basal perforating arteries (Fig 1a), where the perivascular space that separated the two sheets of leptomeninges (20) could act as efficient perivascular drainage pathways (Fig 8).
|
Type 2 lesions resemble those of multiple sclerosis (21) and have been described in neurosarcoidosis; the difficulty in diagnosing them has been pointed out (14). These lesions occurred fairly frequently in our patients. They could be related to neurosarcoidosis because they generally occurred concomitantly with type 1 lesions, which were sometimes present initially and which sometimes appeared secondarily (Fig 6). Like type 1 lesions, type 2 lesions were also associated with neurologic impairment, but this impairment was 2 years or longer.
Patients with type 2 lesions had been previously treated with corticosteroids, generally for more than 2 years. Long-term follow-up MR imaging showed that none of the type 2 lesions responded to any dose of corticosteroids. These findings suggest that they are ancient, irreversible lesions. This pattern of evolution has been reported in the literature (13). We postulate that type 2 lesions are the chronic glial sequelae of previous inflammatory lesions. The development of irreversible chronic reactions after inflammation has been previously reported on the basis of the clinical data (2). The follow-up of our patients revealed that no type 2 lesions were observed as the sequelae of type 1 lesions, probably because the type 1 lesions were treated early with corticosteroids in an attack period.
The large medullary arterioles that supply the whole depth of the white matter (22) are surrounded by a continuous pial channel (23) that correlates well with the deep periarterial draining of the inflammatory cells, which is presumed to have occurred in the type 2 lesions (Fig 8).
Type 2 lesions may arise in one of two ways. Some type 2 lesions may be directly due to granulomatous masses. Two patients with a history of thoracic sarcoidosis and pseudotumorous lesions in the deep white matter had type 2 lesions in similar locations at the time of inclusion. In one, examination of a biopsy sample taken at the time of a ventricular derivation confirmed the presence of a noninflammatory lesion. Type 2 lesions may also be caused by ischemia that ranges from ischemic gliosis to infarction due to invasion of the perivascular space of the perforating vessels by a granuloma (16). These white matter lesions are closely associated with local perivascular enhancement (15). One of our patients appeared to have an ischemic lesion in the external capsule with an area of enhancement along the course of the lateral striate arteries (Fig 1a).
Type 3 lesions, which are similar to those of microangiopathy, have also been reported in neurosarcoidosis (9,10,24,25). They have not been specifically evaluated in isolation from the other lesions seen in white matter disease, probably because such abnormalities in patients with vasculitis are difficult to separate from the lesions of multiple sclerosis (26). In lesion of the centrum semiovale, we considered the subcortical patterns rather than the periventricular patterns and the focal patterns rather than the extensive patterns as the usual patterns for type 3 lesions. The lesions appear to occur frequently, and we considered them to be important indicators of disease because more than half of the patients with a type 3 lesion also had the other types of lesions. The remaining patients with isolated type 3 lesions had systemic disease and were young. As there was no other possible cause, they have to be considered to be related to sarcoidosis and to be distinct from the age-related vascular changes in elderly individuals (27,28).
Type 3 lesions are believed to be different from type 1 and type 2 lesions because isolated type 3 lesions were not associated with any neurologic deficit and because their discovery was fortuitous. Follow-up MR imaging showed that these lesions did not respond to any treatment. We found that new type 3 lesions were detected in a patient whose MR images displayed a thin layer of enhancement that followed the contour of the brainstem, which illustrated pial involvement (Fig 7) (9,29).
These MR imaging features and the ability of neurosarcoidosis to cause multivisceral vasculitis (30) suggest that these type 3 lesions were microarteriolar ischemic complications consistent with true neurosarcoid vasculitis (31–33). They could be linked to the infiltration of the entire wall of small leptomeningeal vessels, which led to thrombotic occlusion and to the surrounding patchy areas of ischemic tissue (30,31). The small subpial blood vessels entering the cortex have no layer of pial cells (34), and the perivascular spaces around the capillaries are obliterated by the fusion of the endothelial and glial basement membranes (23). This prevents the perivascular draining of the inflammatory cells into the subcortical white matter that is supplied by the terminal twigs of the longest cortical arterioles (22) (Fig 8).
These abnormal hyperintense foci at the corticomedullary junction on T2-weighted images are nonspecific. They can be seen in a variety of noninflammatory and inflammatory disorders, including granulomatous vasculitis of the nervous system (35), also known as primary vasculitis of the central nervous system. Histopathologically, primary vasculitis of the CNS is very similar to neurosarcoid vasculitis (33). Primary vasculitis of the CNS has a special predilection for the small leptomeningeal vessels and may appear with prominent leptomeningeal enhancement and minimal parenchymal findings on MR images (36).
Similarly, we believe that these vasculitic lesions in sarcoidosis were initially those of leptomeningeal vasculopathy in which there had been no perivascular propagation of the granulomatous process into the brain (9), other than specific parenchymal sites. There could also be an association between the hyperintense white matter foci and another systemic nonspecific thrombus-inducing pathophysiologic mechanism, such as circulating antiphospholipid antibodies. The presence of antiphospholipid antibodies in disseminated sarcoidosis (37) and similar white matter abnormalities in patients with nonsystemic lupus erythematosus and with antiphospholipid antibodies (38,39) support this hypothesis. An increased frequency of focal white matter lesions has also been described in patients with inflammatory bowel disease (40).
Our experience with sequential MR imaging indicates that MR imaging can be used to predict and to evaluate the response to steroids. This is likely to be useful for the indication of the appropriate initial treatment and the subsequent adjustment in dosage. MR imaging clearly depicts the irreversible lesions for which steroid therapy, if any, should be minimal. Because it shows the changes in active lesions, MR imaging helps the physician to adapt the corticosteroid therapy and, especially, to determine the minimum dose required to prevent progression.
The use of lower doses reduces the complications associated with the long-term use of corticosteroids and maintains the possibility of a subsequent increase. Since clinical reports (4) make it clear that low doses of corticosteroids do not dependably prevent neurologic manifestations and that the earliest possible steroid treatment is recommended, we emphasize the value of follow-up MR imaging during the administration of low maintenance doses. This program could contribute to the improvement in the monitoring of patients to detect relapses, to reintroduce high doses of corticosteroids, and therefore, to prevent irreversible CNS damage.
MR imaging could be combined with clinical indicators to establish the prognosis of disease progression. The initial MR images that indicated the involvement of more than one site by type 1 and/or type 2 lesions, and/or the association of type 2 and type 3 lesions in the brain parenchyma (Fig 6) correlated with the subsequent progression or appearance of type 1 lesions in patients who had presented with CNS impairment for more than 2 years and/or who had previously received steroid therapy. These findings show that initial MR imaging could be a useful tool in recognizing these particular forms of lesion that make long-term treatment difficult.
Because inflammation is the main factor that impairs the integrity of the blood-brain barrier (41), the effect of steroids in reducing gadolinium enhancement reflects the changing pattern of the inflammatory infiltrate. The specific effect of steroids in restoring barrier integrity (42) may also play a part. In addition to the decrease in blood-brain barrier permeability, we also observed a regression of infiltrate masses on T1-weighted images and a disappearance of the surrounding edema on T2-weighted images.
Moreover, no rebound effect was observed after the tapering of corticosteroid therapy was started since the delay in the progression of inflammatory lesions mirrored the other changes on MR images. These data suggest that our use of enhanced T1-weighted images prompts the consideration of a role for steroids in the control of the major features of inflammatory cell invasion. As MR imaging is very sensitive in the depiction of clinically silent CNS lesions (6), it can also be used as an anatomic outcome variable to evaluate the effectiveness of steroid therapy on clinically silent active lesions.
Computed tomography was used in the same way to assess lesion reversibility in pulmonary sarcoidosis (43). The possibility of clinically silent lesions in neurosarcoidosis is well established (6,11). At autopsy, neurosarcoidosis was detected in 15%–27% of the cases (44); this was higher than the percentage of cases that showed clinical CNS involvement. Nevertheless, a limitation in the assessment of treatment accuracy could be the fact that the change steroid dosage for the clinically silent active lesions is based on their response at MR imaging.
The information obtained in this study about the changes of each type of lesion on MR images can be used to develop schedules for follow-up MR imaging. The effectiveness of the attack therapeutic periods appropriated for type 1 lesions appears to be confirmed at MR imaging, and follow-up MR imaging can also help in the determination of the dosages for long-term steroid therapy. Follow-up MR images should be obtained frequently if indicators of poor prognosis are initially present. In patients with type 2 lesions and without type 1 lesions, the use of follow-up MR imaging with gadolinium enhancement is justified for the detection of active inflammatory lesions, especially in those with both type 2 and type 3 lesions. Isolated type 3 lesions should not require specific MR follow-up in the absence of clinical neurologic events. Because neurosarcoidosis is a long-term disease that requires repeated assessment (4), follow-up MR imaging is appropriate because it provides important information for the treatment of patients with CNS sarcoidosis. By using this therapeutic strategy in association with clinical and laboratory parameters, we observed no deaths or uncontrolled worsening of the neurologic deficit during the follow-up period.
|
Footnotes |
|---|
Author contributions: Guarantors of integrity of entire study, J.L.D., D.V., D.G.; study concepts, D.V., J.L.D.; study design, J.L.D., D.V.; definition of intellectual content, J.L.D., D.V., M.B., D.G., J.C.P.; literature research, J.L.D., J.C.P.; clinical studies, C.C.A., C.B., D.V., H.T.L.; data acquisition, J.L.D., M.B.; data analysis, J.L.D., D.V.; statistical analysis, M.B.; manuscript preparation, J.L.D., M.B.; manuscript editing, J.L.D., D.V.; manuscript review, J.C.P., C.C.A., C.B., D.G.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. K. Demir, F. V. Aker, O. Akinci, and A. Ozgultekin Case 134: Primary Leptomeningeal Melanomatosis Radiology, June 1, 2008; 247(3): 905 - 909. [Full Text] [PDF]
|
|
|
|
 |
|
 F G Joseph and N J Scolding Sarcoidosis of the nervous system Practical Neurology, August 1, 2007; 7(4): 234 - 244. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. K. Smith, M. G. Matheus, and M. Castillo Imaging Manifestations of Neurosarcoidosis Am. J. Roentgenol., February 1, 2004; 182(2): 289 - 295. [Full Text] [PDF]
|
|
|
|
 |
|
 E. S. White and J. P. Lynch III Sarcoidosis Involving Multiple Systems : Diagnostic and Therapeutic Challenges Chest, May 1, 2001; 119(5): 1593 - 1597. [Full Text] [PDF]
|
|
| ||||||||||
). (b) Image obtained after 4 months of attack treatment shows that the hypothalamic lesions have regressed, but the punctate infundibular thickening (arrow) persists. The lesions described above progressed when the dose of corticosteroids was decreased. These later regressed when high doses were reintroduced. (Follow-up MR images obtained later are not shown.)
