| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Pediatric Imaging |
1 From the Departments of Child Neurology (M.S.v.d.K.) and Radiology (J.V.), Free University Hospital, De Boelelaan 1117, 1081 HV Amsterdam, the Netherlands; the Department of Radiology, Johns Hopkins Medical Institute, Baltimore, Md (S.N.B.); the Department of Neurogenetics, Kennedy Krieger Institute, Baltimore (S.N.); and the Department of Clinical Epidemiology and Biostatistics, University of Amsterdam, the Netherlands (A.A.M.H.). Received April 2, 1998; revision requested June 29; final revision received November 30; accepted March 16, 1999. Address reprint requests to M.S.v.d.K. (e-mail: ms.vanderknaap@azvu.nl).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: MR images of 92 patients (55 male, 37 female; mean age, 9.3 years) with a leukoencephalopathy were examined by using a scoring list of 68 items. Seven major categories were defined according to the predominant location of the white matter abnormalities. Statistical analysis was used to assess the validity of these seven categories.
RESULTS: Statistical analysis results showed that the seven categories could be well distinguished by either using the defining variables initially accepted as inclusion criteria or selecting a few other variables found to have discriminating value. The additional variables confirmed that the categories are essentially distinct and vary systematically with regard to items other than the inclusion criteria. The existence of two recently defined leukoencephalopathies was confirmed, but no consistent evidence of other new disease entities could be provided.
CONCLUSION: Establishing these seven categories helps in the interpretation of individual studies by demonstrating features that the patient has in common with other patients, and it may facilitate further research on homogeneous subgroups of patients and allow pooling of data across multiple centers.
Index terms: Brain, diseases, 18.871, 18.8721, 18.873 • Brain, MR, 18.12141, 18.121411 • Magnetic resonance (MR), in infants and children, 18.12141, 18.121411
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
A few years ago at our institution, a pattern recognition program was developed for the evaluation of magnetic resonance (MR) images of leukoencephalopathies involving a known basic defect (3). Use of this MR pattern recognition for leukoencephalopathies of unknown cause led to the definition of two recently discovered disease entities, the clinical and pathologic definitions of which followed the initial radiologic definitions (2,4–6). We decided to use this pattern recognition program in a large number of MR imaging investigations of unclassified leukoencephalopathies. The purposes of the study were to (a) provide a means of categorizing leukoencephalopathies into a few major groups according to MR imaging criteria to facilitate further studies and (b) investigate whether "new" (ie, until now unknown) disease entities can be defined within these major groups.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Inclusion Criteria
In 97 patients, the following inclusion criteria were fulfilled:
1. The MR images showed substantial white matter anomalies, either in the form of a signal intensity abnormality (diffuse or focal) or in the form of a serious delay in myelination (ie, myelination age younger than 6–8 months in children with a minimum chronologic age of 1 year).
2. The MR images showed a predominance of white matter pathologic features, although there could be additional pathologic features of the cortex and basal nuclei.
3. Comprehensive clinical and laboratory investigations had been performed, and no cause of the leukoencephalopathy had been found. In all children, at least the following tests were performed: general physical examination, neurologic and ophthalmologic examinations, neurophysiologic studies (ie, electroencephalography, evoked responses, nerve conduction velocity), routine hematology and chemistry panels, and measurements of ammonia, amino acids, organic acids (including N-acetylaspartate), lactate, pyruvate, very long chain fatty acids, phytanic acid, and lysosomal enzyme activity (to exclude metachromatic leukodystrophy, Krabbe disease, GM1 and GM2 gangliosidoses, mannosidosis, and fucosidosis). Whenever appropriate, other tests were performed; these included measurements of copper in blood and urine and of urine sulfatides, chromosome tests, serum cholesterol and cholestanol measurements, studies to exclude immune-mediated white matter disease (ie, oligoclonal bands in cerebrospinal fluid, immunoglobulin G index, autoimmune parameters), studies to exclude direct infections (especially neurotropic viruses, Mycoplasma pneumoniae, and human immunodeficiency virus), DNA repair studies, mitochondrial DNA studies, muscle biopsy, nerve biopsy, and skin biopsy. In children younger than 6 months, congenital infections were excluded (especially toxoplasmosis, rubeola, cytomegalovirus, and herpes). In all children with hypomyelination, the level of free sialic acid in urine was assessed, and in all male patients, DNA studies for Pelizaeus-Merzbacher disease were performed. The laboratory tests results either were normal or showed inconsistent abnormalities that were not suggestive of any diagnosis. In two patients who died, autopsy was performed, and a diagnosis of Alexander disease was made. Children who had periventricular white matter abnormalities on MR images and a history of substantial perinatal problems were given a diagnosis of periventricular leukomalacia and excluded from the study.
Evaluation of the MR Images
Five patients were excluded from the study because the MR images were of very poor quality and thus prevented a detailed evaluation of the white matter pathologic features. The images obtained in the remaining 92 patients (55 male, 37 female; mean age, 9.3 years; age range, 0.5–64.0 years) were reviewed by a neurologist (M.S.v.d.K.) and a neuroradiologist (S.N.B.) independently, and a consensus was reached when there was a disagreement between their interpretations. The MR images were analyzed according to a previously established scoring list (Table 1) (3). For the scoring, the cerebral white matter was divided into the following three zones: the subcortical white matter (arcuate [or U] fibers), the periventricular white matter (a rim of white matter abutting the ventricular lining), and the lobar white matter (the zone between the arcuate fibers and periventricular white matter). The first MR study was used for scoring when serial examinations were present. Follow-up MR studies were compared with the initial images for improvement, deterioration, or static disease.
|
Grouping of Patients
The patients were grouped, according to the predominant pathologic feature and the location of the white matter abnormalities, into the following seven categories: A, myelination abnormalities; B, global cerebral white matter abnormalities involving all or almost all cerebral white matter. All cerebral white matter zones and all cerebral lobes had to be involved, and only minimal sparing of arcuate fibers was accepted; C, extensive cerebral white matter abnormalities combined with involvement of the putamen. "Extensive" differed from "global" in that not all zones of the cerebral white matter were necessarily involved (the arcuate fibers could be spared) and not all cerebral lobes were necessarily affected to the same degree (only frontal predominance with relative sparing of the occipital white matter was accepted); D, predominantly periventricular cerebral white matter abnormalities; E, predominantly lobar cerebral white matter abnormalities; F, predominantly subcortical cerebral white matter abnormalities; and G, predominantly posterior fossa white matter abnormalities.
The criteria used to define each category (ie, defining variables) are listed in Table 2. Both in the case in which myelination abnormalities were predominant and in that in which posterior fossa pathologic features were predominant, additional lesions could be present in the cerebral white matter, with predominant involvement of a particular cerebral zone. A predominance of myelination abnormalities and a predominance of posterior fossa pathologic features should lead to inclusion in categories A and G only; thus, these variables were excluded for the other categories.
|
2 statistics and the Mantel-Haenszel approach (7). |
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
|
|
|
|
|
From a clinical point of view, 12 (57%) of the patients had static neurologic signs, and nine (43%) showed progressive deterioration (Table 4). The deterioration was mainly seen among the patients with signs of atrophy at MR imaging. The findings on follow-up MR images, if available, were unchanged in most patients (Table 5).
In the statistical analysis of the univariate relationships between the defining variable of category A and all the other variables, it was found that myelination abnormality as the predominant pathologic feature was strongly related to absence of lesions in other structures. Relationships with a P value of less than .0001 were found to exist with an absence of lesions in the periventricular white matter, lobar white matter, arcuate fibers, frontal white matter, parietal white matter, occipital white matter, temporal white matter, cerebellar white matter, inner rim of the corpus callosum, anterior and middle parts of the corpus callosum, posterior limb of the internal capsule, and external and extreme capsules. Confluency of the white matter abnormalities and atrophy of the cerebellar hemispheres also were important variables (P < .0001). In the forward selection procedure, the variables delayed myelination, arrested myelination, and abnormal and irregular myelination were not included because of their logical relationship with the defining variable. Myelination abnormality as the predominant pathologic feature was almost perfectly associated with absence of lesions in the frontal white matter (P < .00001). A further separation was possible on the basis of the absence of lesions in the external and extreme capsules (P = .001). Given these two variables, only one patient was incorrectly classified as belonging to category A.
In the decision tree, the second step involved the pathologic features of the posterior fossa. Category G was that of patients in whom there was a predominance of involvement of posterior fossa white matter structures. This category contained a single patient, in whom the most striking white matter abnormalities were seen in the cerebellar white matter, middle cerebellar peduncles, and pyramidal tracts in the brainstem. The posterior limb of the internal capsule also contained abnormalities. The supratentorial white matter abnormalities were minor. The white matter changes were confluent, homogeneous, and symmetric. The cerebellar vermis and hemispheres were atrophic.
Analysis of the univariate relationships between the defining variable of category G and the remaining variables revealed no relationships, probably because only one patient belonged in this category.
In the decision tree, the third category defined was category C, in which there were 11 patients. The characteristic findings in this category were extensive, homogeneous, and confluent symmetrical white matter abnormalities combined with either signal intensity abnormality (in eight [73%] patients) (Fig 4) or atrophy (in three [27%] patients) (Fig 5) of the basal nuclei. The signal intensity abnormality of the basal nuclei was always mild, and there often was evidence of some swelling (Fig 4). In all cases, the putamen, globus pallidus, and caudate nucleus were simultaneously involved. In three (27%) patients, the thalami also were abnormal. In most cases (nine [82%] patients), a fronto-occipital gradient in the extent of white matter involvement was seen with relative sparing of the occipital lobes (Figs 4, 5); in a minority of the patients (n = 2 [18%]), global involvement of the cerebral lobes was seen. Either global involvement of the cerebral zones was seen (in nine [82%] patients), or there was a periventricular predominance with some sparing of the arcuate fibers, particularly in the occipital area (in two [18%] patients). The cerebellar white matter, middle cerebellar peduncles, hilusof the dentate nucleus, and, strikingly, the medulla oblongata were frequently involved (Fig 4). The corpus callosum also was often affected. The internal capsule was frequently involved, and the external and extreme capsules were always involved. Two patients had intraparenchymal cysts in the frontal or parietal area. In seven (64%) patients, the abnormal white matter appeared to be slightly swollen.
|
|
|
The results of statistical analysis showed that category C differed from the remaining categories B, D, E, and F with respect to involvement of the medulla oblongata, caudate nucleus, and globus pallidus, and the presence of white matter swelling (P < .0001). In the forward selection procedure, category C appeared to be almost perfectly defined by caudate nucleus involvement (P < .00001). A further separation was possible on the basis of involvement of the globus pallidus (P = .00008). Given these two variables, only one patient was incorrectly classified as not belonging to category C. If involvement of not only the putamen but also the globus pallidus and caudate nucleus would be considered as a prerequisite for category C, then these variables should be excluded from the forward selection procedure. Then, involvement of the medulla oblongata (P = .00001), presence of white matter swelling (P = .0021), noninvolvement of the posterior part of the corpus callosum (P = .0096), and involvement of the anterior limb of the internal capsule (P = .0091) contributed to the separation of category C from categories B, D, E, and F. By using these variables, three patients were not classified correctly.
Following the decision tree, the final concern was the predominant involvement of a particular zone of the cerebral white matter, which defined categories B, D, E, and F. Thirty patients were in category B, which was defined by global involvement of the cerebral white matter zones and lobes. Other MR imaging findings that were characteristic of this category were a high frequency of involvement of the cerebellar white matter (23 [77%] patients) and other posterior fossa white matter structures, high frequency of involvement of the inner rim of the corpus callosum (28 [93%] patients) versus low frequency of involvement of the outer rim (two [7%] patients), high frequency of involvement of the posterior limb of the internal capsule (26 [87%] patients) versus low frequency of involvement of the anterior limb (seven [23%] patients), high frequency of involvement of the external and extreme capsules (29 [97%] patients), and sparing of the basal nuclei in most cases. The white matter changes were always confluent and symmetric.
All but one of the patients in category B had clinically progressive disease (Table 4), and in the majority of them in whom follow-up MR imaging data were available, worsening was demonstrated (Table 5).
Category D included 15 patients and was defined by a predominance of periventricular cerebral white matter abnormalities (Fig 6). As a rule, the arcuate fibers were spared. The inner rim of the corpus callosum was frequently involved (11 [73%] patients), and the outer rim was less frequently involved (seven [47%] patients). The internal capsule often contained abnormalities, and the posterior limb was involved more often (eight [53%] patients) than the anterior limb (three [20%] patients). The white matter abnormalities were confluent in nine (60%), isolated and multifocal in two (13%), and confluent as well as isolated and multifocal in four (27%) patients. The abnormalities were always symmetric. Incidental involvement of the basal nuclei was noted. In a minority of the patients, the posterior fossa structures contained abnormalities.
|
The 13 patients in category E had a predominance of lobar cerebral white matter abnormalities (Fig 7). The periventricular white matter and arcuate fibers were largely spared. The posterior fossa structures, corpus callosum, and internal capsule were rarely involved, and the external and extreme capsules were involved in three (23%) cases. The white matter abnormalities were isolated and multifocal in 12 (92%) patients (Fig 7). Symmetric abnormalities were observed in 10 (77%) cases. The basal nuclei in these patients did not contain abnormalities.
|
Two patients in category E had similar, rather peculiar imaging findings. The first MR studies, obtained in the first year of life in both children, showed two rounded lesions with a symmetric location in the centrum semiovale on both sides (Fig 8a). Clinically, the disease was rapidly progressive. Follow-up MR images obtained a few months later showed global cerebral white matter changes with little sparing of any area. (The second MR study was not scored for the study.) In both patients, the initial lesions were still demarcated from the remainder of the cerebral white matter by a thin rim of less abnormal signal intensity (Fig 8b).
|
|
The defining variable of categories B, D, E, and F—that is, predominant involvement of a particular zone of the cerebral hemispheres—appeared to be related to involvement of many structures. In the forward selection procedure, the variables periventricular involvement, lobar involvement, and arcuate fiber involvement were not included because of their logical relation to the defining variable. Differentiation between the categories could be made on the basis of the following variables: confluency of white matter abnormalities (P < .0001), low signal intensity on intermediate-weighted images (P = .00016), and subcortical cysts (P = .0034). Instead of the last variable, white matter swelling was equally useful. When these variables were used, five patients were not classified correctly.
Within category B, two subcategories—B1 and B2—could be defined. Together, these two categories included all but four of the patients in category B. The MR imaging findings in the five patients in subcategory B1 were very similar to each other and shared most of the characteristics previously described for category B, but there were some important differences. The characteristics that distinguished the five patients in category B1 from the remainder of the patients in category B were the presence of subcortical cysts, which were always in the anterior temporal area and sometimes also in the frontal or parietal area, and the presence of white matter swelling (Fig 9). The white matter abnormalities were always confluent, homogeneous, and symmetric. The basal nuclei were never involved. Within the posterior fossa, the cerebellar white matter and hilus of the dentate nucleus were always involved; other posterior fossa structures were less frequently involved. At follow-up, the MR image findings were generally unchanged (Table 5); however, clinically, the condition of most of the patients was deteriorating slowly (Table 4).
|
|
The MR imaging findings in the 21 patients in subcategory B2 also were very similar to each other and shared most of the characteristics previously described for category B, with some important differences. The white matter abnormalities were confluent and symmetric, but not homogeneous; the signal intensity was partly high and partly low on the intermediate-weighted images in all of these patients (Fig 10). Thus, part of the abnormal white matter was similar to cerebrospinal fluid in signal intensity. Another distinguishing feature of patients in category B2 was the presence of a stripelike radiating pattern that was visible within the abnormal white matter on sagittal MR images (in 17 [81%] patients) (Fig 10). This finding often appeared as a stippled pattern on the axial images. In some cases, the globus pallidus or thalamus had some signal intensity abnormality. Within the posterior fossa, brainstem abnormalities were relatively frequent, particularly those involving the central tegmental tracts (in 20 [95%] patients) (Fig 11). Cerebellar vermis atrophy was relatively frequent (in nine [43%] patients). At follow-up MR imaging, worsening was usually seen, although in one case, there was improvement of the brainstem abnormalities (Table 5). Clinically, all patients showed signs of further neurologic deterioration (Table 4).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Category A included patients in whom severely deficient myelination was the predominant finding (Figs 2, 3). Severely deficient myelination is seen in Pelizaeus-Merzbacher disease (8,9), which was excluded in all the male patients in this study. Manifestation of Pelizaeus-Merzbacher disease in female patients due to unfortunate X inactivation (10) is extremely rare and does not explain the high number of female patients in category A. Other disorders that can lead to severe myelin deficiency include Salla disease (11), Cockayne syndrome type II (12), and Tay syndrome (13), none of which were present in the patients in this study. With the exception of the proteolipid protein gene, other myelin protein genes were not evaluated. It is highly probable that mutations of other myelin protein genes also can cause a serious disturbance in myelination. In addition, early-onset neuronal degenerative disorders, such as Alpers disease, Menkes syndrome, infantile neuronal ceroid lipofuscinosis, Sandhoff disease, and Tay-Sachs disease, can lead to severely abnormal myelination (14). The progressive nature of these diseases is apparent in clinical deterioration and progressive cerebral atrophy at follow-up MR imaging. In the four patients with cerebral atrophy in our study, it is likely that the underlying cause was an unknown neuronal degenerative disorder.
Category B included patients with global cerebral white matter involvement. Among these patients, two diseases were readily recognized as recently described leukoencephalopathies. Category B1 was that of patients with typical image abnormalities that have been described as "infantile onset leukoencephalopathy with swelling and a discrepantly mild clinical course" (4), which proved to be a vacuolating leukoencephalopathy at histopathologic examination (5). The typical MR imaging findings in category B1 included diffuse cerebral white matter involvement with some swelling of the abnormal white matter and the presence of subcortical cysts invariably in the anterior temporal area and often also in the frontal and/or parietal area (Fig 9). These findings have been confirmed by several groups (15–18). It is now generally accepted that this condition represents a disease entity with a homogeneous clinical picture, an autosomal recessive mode of inheritance, and an as yet unidentified basic defect.
Category B2 represented those patients with another disease entity that has been recently described by several different groups (6,19–21). This condition is variably referred to as "CACH" (childhood cerebellar ataxia and central hypomyelination) (20) and "the disease of the vanishing white matter" (6). Findings at MR imaging and spectroscopy suggest the disappearance of cerebral white matter, which is confirmed at autopsy (6). In our study, the MR images showed a diffuse leukoencephalopathy in which increasing portions of the cerebral white matter behaved like cerebrospinal fluid, as was apparent on the intermediate-weighted spin-echo and fluid-attenuated inversion-recovery images (Fig 10). Brainstem abnormalities were frequent and typically involved the bilateral tracts of the pontine tegmentum (Fig 11). This disease also has an autosomal recessive mode of inheritance. The basic cause is not known.
Category C was a special category of patients who had extensive cerebral white matter abnormalities. In the majority of cases, these abnormalities were present with a fronto-occipital gradient and relative sparing of the occipital lobes (Figs 4, 5). In a minority of the patients, the first MR images had been obtained at the end stage of disease and showed a global distribution of the white matter abnormalities. In two-thirds of the cases, the abnormal white matter had a slightly swollen appearance. The distribution and appearance of white matter abnormalities were compatible with the diagnosis of Alexander disease (22–26). Generally, the demonstration of Rosenthal fibers in brain tissue is considered a prerequisite to a definitive diagnosis. In none of our patients did the parents grant permission to perform brain biopsy for histopathologic confirmation of the diagnosis. However, in two patients in category C, the results of autopsy performed later confirmed the MR imaging–based suggested diagnosis. In addition to the white matter changes, all patients had basal nuclei abnormalities, which were either changes in signal intensity and some swelling (Fig 4) or marked atrophy (Fig 5). To our knowledge, the involvement of the basal nuclei in Alexander disease is not described in the MR imaging literature (22–26), but from histopathologic analysis, it is known that the basal nuclei contain a relatively high volume of Rosenthal fibers (27–29). It is also known that in Alexander disease, the course of the white matter abnormalities is characterized by an increase in volume followed by cystic degeneration and a loss of volume (29,30). The evolution of the white matter abnormalities from swelling to atrophy also seems to be true for the basal nuclei (28,30). All three patients with atrophic basal nuclei at MR imaging also had white matter volume loss with enlarged lateral ventricles, and two of these patients had cystic degeneration of the frontal white matter. A striking observation in the patients was the high frequency of abnormalities involving the medulla oblongata. Prominent involvement of the brainstem, particularly the medulla oblongata, is known from histopathologic analysis (27,29,30), but it is not described with Alexander disease in the imaging literature (22–26).
Category D, that of patients showing predominantly periventricular white matter abnormalities (Fig 6), was very heterogeneous, and we could not detect one or more consistent subcategories that might represent disease entities. Autosomal dominant leukoencephalopathy with adult onset is one of the causes of a periventricular leukoencephalopathy in older patients (31,32). There is at present, however, no biochemical or DNA test to prove this diagnosis. Periventricular white matter changes with relative sparing of the arcuate fibers are seen in several leukodystrophies, such as metachromatic leukodystrophy, Krabbe disease, and X-linked adrenoleukodystrophy (33). The imaging patterns of these disorders share several characteristics, such as invariable involvement of the corpus callosum and confluency of the white matter changes. In several patients in category D, the pattern of abnormalities was indistinguishable from that of metachromatic leukodystrophy, Krabbe disease, or X-linked adrenoleukodystrophy, which were, however, excluded by using appropriate biochemical tests.
Category E mainly consisted of patients with isolated multifocal white matter abnormalities with a predominantly lobar location and relative sparing of arcuate fibers and periventricular white matter (Fig 7). In the previous MR pattern recognition study (3), isolated multifocal white matter abnormalities were found mainly among acquired, noninherited conditions, particularly inflammatory (acute disseminating encephalomyelitis, multiple sclerosis) and infectious disorders. A pattern of lobar isolated and multifocal white matter changes is seen in congenital cytomegalovirus infection (34,35). A clinical picture consisting of a static encephalopathy with deafness, combined with MR imaging findings of lobar isolated and multifocal white matter changes, which are unchanged at follow-up, is highly suggestive of congenital cytomegalovirus infection. However, after the age of 6 months this diagnosis can no longer be proved. Unfortunately, in none of the children was a computed tomographic (CT) scan of the brain obtained. The presence of calcium depositions would form another argument in favor of the diagnosis of congenital cytomegalovirus infection. The follow-up MR images in two other patients in category E showed decreasing abnormalities that were compatible with a diagnosis of acute disseminating encephalomyelitis. It is important to recognize that not all leukoencephalopathies are inherited or progressive. Some are acquired and either static or improving, as can be shown at follow-up MR imaging.
Two patients in this category required special attention (Fig 8). In both patients, the onset of disease occurred during the 1st year of life, and at MR imaging, symmetric lesions in the centrum semiovale were seen. At both clinical examination and MR imaging, the disease was rapidly progressive, with involvement of all cerebral white matter at follow-up a few months later. The most peculiar finding was that the initial lesion was still demarcated at follow-up MR imaging. The cause of disease was not known in either of the patients, but from an MR imaging point of view, they probably had the same, as yet unclassified disorder.
Category F was that of patients with white matter changes in a predominantly subcortical location. There are several known disorders that preferentially involve the arcuate fibers, such as L-2-hydroxyglutaric aciduria (36) and Kearns-Sayre syndrome (37,38). In the one patient in this category, these disorders were excluded.
Category G consisted of patients with white matter disorders in a predominantly posterior fossa location. In such cases, the differential diagnosis usually includes Refsum disease (39), cerebrotendinous xanthomatosis (40), and adrenomyeloneuropathy (41), which were excluded in the one patient in this category.
One purpose of the present study was to provide a means of subcategorizing unclassified leukoencephalopathies according to simple yet robust MR imaging criteria. The criteria used in this study to subcategorize patients, which were based on the predominant location of the white matter abnormalities, were shown to be valid and resulted in a workable subdivision with a reasonable distribution of the patients in the categories. The categories could be well distinguished either by using the defining variables initially accepted as inclusion criteria or by selecting a few other variables that were found to have discriminating value in the forward selection procedure. The facts that in most cases, the defining variables were related to many other variables and that a clear separation of the categories could also be reached by means of newly defined variables confirmed that the categories are essentially distinct and vary systematically in terms of items other than the inclusion criteria alone.
Another purpose of the study was to investigate the possibility of defining new disease entities. We could not find consistent evidence of these, but we were able to confirm the existence of the two recently defined disease entities vacuolating leukoencephalopathy with subcortical cysts (2,4,5,15–18) and the disease of the vanishing white matter (6,19–21). A striking finding was that together, these two disorders, which affected 26 patients, were responsible for 28% of all the patients with unclassified leukoencephalopathy, 32% of all the patients aged 21 years or younger, and 87% of all the patients with a global leukoencephalopathy.
Most of the children involved in this study had an inherited and progressive encephalopathy. However, it is important to recognize that not all leukoencephalopathies are progressive and not all are inherited. In this respect, it is important to note the difference between the word "leukoencephalopathy," which refers to all forms of white matter abnormality, both inherited and acquired, and the word "leukodystrophy," which refers specifically to progressive, inherited demyelinating disorders. Because MR imaging alone cannot enable differentiation between the two entities, we consistently used the broader term leukoencephalopathy. All of the patients in this study with arrested myelination in an early stage probably had an inherited condition, as suggested by the frequent instance of two affected siblings in one family, but many of them had clinical signs of a stable encephalopathy, without any deterioration up to the time this article was written. Most of the patients in category E probably had an acquired encephalopathy, with infections and immunologic conditions as the most frequent causes. The importance of this observation is that in these patients, further testing for metabolic disorders was not necessary, and the risk of recurrence in subsequent siblings was equal to that in the general population.
This study focused on findings on nonenhanced MR images. MR spectroscopy, contrast material–enhanced MR imaging, and CT of the brain were performed in only a few patients. Of course, these modalities can add valuable information. CT of the brain is superior in demonstrating calcium depositions, the presence of which can provide an important diagnostic clue, for instance, in congenital infections and several inherited disorders such as Aicardi-Goutières syndrome, Cockayne syndrome, and mitochondrial disorders (33). MR spectroscopy of the brain may add essential neurochemical information (6,19,21,33). Contrast enhancement may occur in a pattern that is more or less diagnostic for a disease, as, for instance, in X-linked adrenoleukodystrophy (33,39). However, a separate study is necessary to systematically assess the additional value of these techniques.
Leukoencephalopathies in children are considered to be rare. However, it has been shown that progressive childhood encephalopathies occur with a frequency of approximately 0.6 in 1,000 (42). White matter disorders encompass a major part of these abnormalities. It is disappointing to discover how many of these entities in patients remain unclassified. We estimate that at least half of childhood leuko-encephalopathies remain unclassified and are a diagnostic problem for neuroradiologists and pediatric neurologists. For the interpretation of MR images obtained in patients with an unclassified leukoencephalopathy, we suggest the use of a categorization system that facilitates pooling of data across multiple sites worldwide and thus may lead to new diagnoses.
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, M.S.v.d.K.; study concepts and design, M.S.v.d.K.; definition of intellectual content, M.S.v.d.K., J.V.; literature research, M.S.v.d.K.; clinical studies, S.N., M.S.v.d.K.; data acquisition and analysis, M.S.v.d.K., S.N.B.; statistical analysis, A.A.M.H.; manuscript preparation, M.S.v.d.K.; manuscript editing, S.N.; manuscript review, J.V., S.N.B.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. E. Steenweg, G. S. Salomons, Z. Yapici, G. Uziel, E. Scalais, D. I. Zafeiriou, M. L. Ruiz-Falco, V. Mejaski-Bosnjak, P. Augoustides-Savvopoulou, M. Wajner, et al. L-2-Hydroxyglutaric Aciduria: Pattern of MR Imaging Abnormalities in 56 Patients Radiology, June 1, 2009; 251(3): 856 - 865. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Schiffmann and M. S. van der Knaap Invited Article: An MRI-based approach to the diagnosis of white matter disorders Neurology, February 24, 2009; 72(8): 750 - 759. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bizzi, G. Castelli, M. Bugiani, P.B. Barker, E.H. Herskovits, U. Danesi, A. Erbetta, I. Moroni, L. Farina, and G. Uziel Classification of Childhood White Matter Disorders Using Proton MR Spectroscopic Imaging AJNR Am. J. Neuroradiol., August 1, 2008; 29(7): 1270 - 1275. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Rossi, R. Biancheri, F. Zara, C. Bruno, G. Uziel, M.S. van der Knaap, C. Minetti, and P. Tortori-Donati Hypomyelination and Congenital Cataract: Neuroimaging Features of a Novel Inherited White Matter Disorder AJNR Am. J. Neuroradiol., February 1, 2008; 29(2): 301 - 305. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. van der Knaap, T. Linnankivi, A. Paetau, A. Feigenbaum, K. Wakusawa, K. Haginoya, W. Kohler, M. Henneke, A. Dinopoulos, P. Grattan-Smith, et al. Hypomyelination with atrophy of the basal ganglia and cerebellum: Follow-up and pathology Neurology, July 10, 2007; 69(2): 166 - 171. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Bugiani, S. Al Shahwan, E. Lamantea, A. Bizzi, E. Bakhsh, I. Moroni, M. R. Balestrini, G. Uziel, and M. Zeviani GJA12 mutations in children with recessive hypomyelinating leukoencephalopathy Neurology, July 25, 2006; 67(2): 273 - 279. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M.S. van der Knaap, M. Kriek, W.C.G. Overweg-Plandsoen, K.B. Hansson, K. Madan, J.S. Starreveld, P. Schotman-Schram, F. Barkhof, and S.A.M.J. L. Oberstein Cerebral White Matter Abnormalities in 6p25 Deletion Syndrome AJNR Am. J. Neuroradiol., March 1, 2006; 27(3): 586 - 588. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Blumkin, N. Watemberg, D. Lev, G. Malinger, Y. Luckman, B. Ben-Zeev, and T. Lerman-Sagie Nonprogressive Familial Leukoencephalopathy With Porencephalic Cyst and Focal Seizures J Child Neurol, February 1, 2006; 21(2): 145 - 148. [Abstract] [PDF]
|
|
|
|
 |
|
 A. Vanderver, R. Schiffmann, M. Timmons, K. A. Kellersberger, D. Fabris, E. P. Hoffman, J. Maletkovic, and Y. Hathout Decreased Asialotransferrin in Cerebrospinal Fluid of Patients with Childhood-Onset Ataxia and Central Nervous System Hypomyelination/Vanishing White Matter Disease Clin. Chem., November 1, 2005; 51(11): 2031 - 2042. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. van der Voorn, P. J. W. Pouwels, W. Kamphorst, J. M. Powers, M. Lammens, F. Barkhof, and M. S. van der Knaap Histopathologic Correlates of Radial Stripes on MR Images in Lysosomal Storage Disorders AJNR Am. J. Neuroradiol., March 1, 2005; 26(3): 442 - 446. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. B. Barker and A. Horska Topical Review: Neuroimaging in Leukodystrophies J Child Neurol, August 1, 2004; 19(8): 559 - 570. [Abstract] [PDF]
|
|
|
|
|
|
E. M. Kaye and H. Moser Where has all the white matter gone?: Unraveling the mysteries of leukoencephalopathies Neurology, May 11, 2004; 62(9): 1464 - 1465. [Full Text] [PDF]
|
|
|
|
|
|
N. I. Wolf, M. A.A.P. Willemsen, U. F. Engelke, M. S. van der Knaap, P. J.W. Pouwels, I. Harting, J. Zschocke, E. A. Sistermans, D. Rating, and R. A. Wevers Severe hypomyelination associated with increased levels of N-acetylaspartylglutamate in CSF Neurology, May 11, 2004; 62(9): 1503 - 1508. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. van der Knaap, G. Vermeulen, F. Barkhof, A. A. M. Hart, J. G. Loeber, and J. F. L. Weel Pattern of White Matter Abnormalities at MR Imaging: Use of Polymerase Chain Reaction Testing of Guthrie Cards to Link Pattern with Congenital Cytomegalovirus Infection Radiology, February 1, 2004; 230(2): 529 - 536. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. van der Knaap, S. Naidu, P. J.W. Pouwels, S. Bonavita, R. van Coster, L. Lagae, J. Sperner, R. Surtees, R. Schiffmann, and J. Valk New Syndrome Characterized by Hypomyelination with Atrophy of the Basal Ganglia and Cerebellum AJNR Am. J. Neuroradiol., October 1, 2002; 23(9): 1466 - 1474. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. B. Kang, J. V. Hunter, and E. M. Kaye Lactic Acid Elevation in Extramitochondrial Childhood Neurodegenerative Diseases J Child Neurol, September 1, 2001; 16(9): 657 - 660. [Abstract] [PDF]
|
|
|
|
|
|
M. S. van der Knaap, S. Naidu, S. N. Breiter, S. Blaser, H. Stroink, S. Springer, J. C. Begeer, R. van Coster, P. G. Barth, N. H. Thomas, et al. Alexander Disease: Diagnosis with MR Imaging AJNR Am. J. Neuroradiol., March 1, 2001; 22(3): 541 - 552. [Abstract] [Full Text]
|
|
|
|
|
|
U. Senol, K. Karaali, I. A. Alorainy, Y. G. Patenaude, A. O’Gorman, D. N. Black, and K. Meagher-Villemure Cree Leukoencephalopathy and Other Leukoencephalopathies Involving Arcuate Fibers Dr Alorainy and colleagues respond: Radiology, January 1, 2001; 218(1): 303 - 303. [Full Text]
|
|
|
|
|
|
M. S. van der Knaap, S. Naidu, B. K. Kleinschmidt-DeMasters, W. Kamphorst, and H. C. Weinstein Autosomal dominant diffuse leukoencephalopathy with neuroaxonal spheroids Neurology, January 25, 2000; 54(2): 463 - 463. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
