Congenital Cytomegalovirus Infection of the Brain: Imaging Analysis and Embryologic Considerations

Transcription

1 Congenital Cytomegalovirus Infetion of the Brain: Imaging Analysis and Embryologi Considerations A. James Barkovih and Camilla E. Lindan PURPOSE: T o analyze the ortial gyral patterns and myelination patterns in a series of patients with ongenital ytomegalovirus infetions involving the entral nervous system, to orrelate them with known developmental events, and to develop a onsistent theory regarding their em bryogenesis. METHODS: The MR (11 patients) and CT (four patients) studies of 11 patients with ongenital ytomegalovirus infetions involving the brain were retrospetively reviewed. Analysis was made of myelination patterns, ortial gyral patterns, other areas of maldeveloped brain, and foal brain lesions. RESULTS: Lissenephaly was found in four patients. These patients had very thin erebral orties, extremely diminished volume of white matter, delayed m yelination, small erebella, and very enlarged lateral ventriles. Foal areas of dysplasti ortex, presumably polymirogyria, were found in fi ve patients. These patients had slightly thikened irregular erebral orties, slightly diminished volume of white m atter, delayed m yelination, variably small erebella, and slightly enlarged lateral ventriles. T wo patients had normal erebral orties, slightly diminished volume of white matter, delayed m yelination, normal erebella, and slightly enlarged lateral ventriles. Periventriular lesions, representing alifiation, or perhaps blood, were seen in all groups. CONCLUSIONS: We postulate that the patients with lissenephaly suffer injury before 16 or 18 weeks gestational age, whereas those with regions of polym irogyria are injured between approximately 18 and 24 weeks gestational age. Those with norm al gyral patterns are probably injured during the third trimester and may have ative infetions at birth. Moreover, we propose that the finding of erebellar hypoplasia and myelination delay in assoiation with diffuse lissenephaly or ortial dysplasia should suggest the diagnosis of ongenital ytomegalovirus infetion. Index terms: Cy tomegali inlusion disease; Infa nts, entral nervous system ; Infants, diseases; Brain, omputed tomography ; Brain, magneti resonane AJNR Am J Neuroradio/15:73-715, Apr 1994 Although it has been widely reported that patients with ongenital ytomegalovirus infetion have assoiated disorders of neuronal migration and ortial organization (1-6), little has been written onerning the spetrum of erebral ortial appearane in ongenital ytomegalovirus. We had noted a wide variation of ortial gyral anomalies in those patients with ongenital ytomegalovirus whom we had imaged with magneti resonane (MR) and omputed tomography (CT). We therefore undertook a detailed retrospetive study of the CT and MR examinations of patients with ongenital ytomegalovirus in order to doument the variation in ortial gyral patterns and to attempt to explain the various gyral patterns based on assoiated brain abnormalities and known embryologi data. Reeived Marh 9, 1993; aepted pending revision J une 2 1; revision reeived July 8. From the University of Ca liforn ia San Franiso, Department of Radiology, Neuroradiology Setion. Address repri nt requests to A. James Barkovih, MD, Assoiate Professor of Radiology, University of California San Franiso, Department of Radiology, Neuroradiology Setion, 55 Parnassus Ave, L-371, San Franiso, CA AJNR 15:73-715, A pr / 94/ Amerian Soiety of Neuroradiology 73 Patients and Methods We retrospetively reviewed MR (11 patients) and CT (four patients) of 11 patients with ongenital ytomegali inlusion disease involving the brain. Presenting signs and symptoms are listed in Table 1. Six of the infants were born prematurely (range of 33 to 36 weeks gestational age), and five were born at term. The diagnosis was established in all ases by harateristi linial and imaging findings in onjuntion with positive serologi testing for ytomegalovirus or ulture of the virus from the urine

2 -...) TABLE 1: Eleven patients with ongenital ytomegali inlusion disease involving the brain.1::> Patient Age Presentation Sequenes/ Foi of Cerebral Cortex White Matter Ventriles Hemorrhage Califiation Posterior Fossa Exams Abnormal Signal OJ )> CT, 15 Hepatospleno- MR: Ax 4 mm Normal Delayed myeli- Smoothly en- Large bilateral Intraventriular? Periventriu- Normal ;:u :A days megaly SE 3/ Dysplasti hip- nation, ysts larged, left Cb, superfiial Probable peri- lar days Sag 3 mm SE porallobes sizes from 3 x MR, 27 Spastiity 6,12 poampus anterior tern- >right retrotrigonal, ventriular < n 616/ 11 3 mm to 15 x Ax 4 mm SE 15 mm. Short 549/ 16 Tl, T2, small- CT puntate around ventriles (supraand infratentorial) and in basal ganglia, erebellum Short Tl, T2 2 I day Miroephaly MR: Sag 5 mm Agyri Delayed myeli- Marked, Peri ventriular None Subependymal Tiny erebellum (brain weight SE 617/ 2 Very thin nation smooth with short T1 masses in and brain!5 g), skin pe- Cor 5 mm SE ortex Very dimin- enlarged size, 3-12 erebrum stem small tehiae, hepato- 5/ 2 Dysplasti hip- ished vol- partially mm and erebel- (1 mm) yst megaly Ax 4 mm SE poampus ume absent lum 6/ 2 septum 3 3 days Miroephaly, MR: Sag 4 mm Agyri Delayed myeli- Marked, Puntate peri- None Subependymal Tiny erebellum lymphadenop- SE 5/2 Very thin nation smooth ventriular foi athy, hepato- Ax 5 mm SE ortex Very dimin- enlarged, foi of short splenomaegaly, 28/9 Dysplasti hip- ished vol- absent Tl, T2 horioretinitis, CT poampus ume septum, thromooyto- subependypenia, anemia, mal punpetehiae tate short Tl, T2 4 8 months Seizures, develop- MR : Sag 3 mm Bilateral ortial Delayed myeli- Small avum None None Subortial Normal mental delay, 616/12 dysplasia. nation, ysts left parietal hypotonia Cor 5 mm Left fronto- anteromedial lobe 5/ 16 temporopari- temporal Ax 4 mm etal and right lobes 3/ frontotempo- 6,12 roparietal. )>... Most severe :z posteriorly. ;:u Right parieljl,_.. tal-subor- - )> tial short -o Tl, dysplas- 2:,_.. ti hippo- ampus ::> :::

3 5 6 years Developmental MR: Ax 5 mm Bilateral fran- Bilateral asy- Smooth, Puntate, sub- None Subependymal Small erebel- )> <... :z ;;a... \Jl -...J \Jl delay, seizures, SE 2/ tooperular metrial foi slightly ependymal foi lum )> spastiity 4,8 dysplasia ex- of periven- enlarged foi of short " CT tending to triular and T 2 2:... parietal lobe subortial \ \ long T2. Di-.1::> minished periventriular white matter 6 8 months Developmental MR : Sag 5 mm Bilateral frontal Bilateral Smooth, None None None apparent Normal delay, hypo- SE 55/ 25 dysplasia ex- symmetria l slightly tonia, seizures Ax 5 mm SE tending to foi of peri- enlarged 55/ 25 parietal lobe ventriular Cor 5 mm SE Most severe and subor- 55/ 25 posteriorly tial long T2 Ax 5 mm SE Dysplasti hip- in frontal 3/ 4,8 poampus and ternporal lobes. Diminished white matter 7 2 weeks Miroephaly, MR: Sag 6 mm Diffuse ortial Very dimin- Marked, Puntate foi of None Subependymal Small erebelhepatomegaly, SE 5/ 15 dysplasia, ished vol- smoothly short T1, T2 foi lum seizures Ax 5 mm SE most severe ume enlarged around frontal 6/ 15 (smooth) in Delayed myeli- horns Ax 5 mm SE frontal lobes nation 3/ Dysplasti hip- 6,12 poampus 8 1 day Miroephaly, MR : Sag 3 mm Agyri Delayed myeli- Marked, Puntate, sub- None Subependy- Small erebelhepatospleno- SE 5/ 2 Very thin nation smoothly ependymal mal, sub- lum megaly, pete- Ax SE 5 mm ortex Very dimin- enlarged foi of short ortial, Calium dorsal hiae, at birth. 3/ Dysplasti hip- ished vol- T2 basal gan- midbrain Ultrasound at 6, 12 poampus ume glia, ere- 26 weeks CT bellar, messhowed large enephali n -< ventriles months Seizures, miro- MR: Sag 4 mm Dysplasti fran- Slightly dimin- Smooth, Few subependy- None Subependymal Hypogeneti ephaly SE 6/ 15 ta l, temporal ished vol- slightly mal puntate foi? vermis 3 rn Ax 4 mm SE ortex ex - ume enlarged foi of short Slightly small 6/ 15 tending to Subortia l T2 hemispheres )> r Ax 4 mm SE parietal lobe long Tl, T2 < 3/ Most severe ;u 6, 12 posteriorly : (/)

4 76 BARKOVICH AJNR: 15, Apri11994 UJ...J o < 1- ill [L.'!.. :;:; u ;.;:: o u Ol -E E :r: Ol o E [L ::.. < ~ > X.'! u --...!:! E :J X o-w (/) :: :;:; 2 :: ~.. Ol <.!!!.. ~.E ' vi ~..I: Ol. E (/) ~ ~ E :2 ~ I :;; :: :: :;:; :>, :; N.I: Q. a:; e u 3: u ~.I: E \D ~I\ E :: :: E :J -o E:g» 2~-EE~ o..= o n..!s! Q) Q) = o. E (/).?;>»- - "E.E ~ g. E - ~ u ~ > :: :Eo -.I: E \ :,. u ~ 1- Ol :: -~ _Q OJ ~ N....'!1 1- i.i: u C:.. w (/) -o :: :: u u X u '6 - ~ X < 1..'! of the hild. Five hildren presented in the immediate neonatal period with miroephaly (four patients), hepatosplenomegaly (three patients), hepatomegaly (two patients), skin and muous membrane petehiae (three patients), lymphadenopathy (one patient), and horioretinitis (one patient). Six patients presented later (one at age 4 months, one at age 6 months, two at age 8 months, one at age 9 months, and one at age 19 months) with seizures (infantile spasms, five patients), developmental delay (three patients), and miroephaly (three patients). Imaging studies were performed during the first 2 weeks of life in four of the five patients diagnosed in the neonatal period and at ages 15 days (CT) and 27 days (MR) in the fifth. Of the six patients diagnosed later in infany, five were imaged at the time of presentation and the sixth (patient 5) at 6 years of age. CT sans were obtained in four patients (Table 1) and onsisted of ontiguous 1-mm axial setions through the brain without the use of intravenous ontrast. The MR studies were performed at 1.5 T and inluded sagittal 3- to 6-mm (1-mm gap) spin-eho 5-65/ 11-25/ 1-2 (repetition time/ eho time/ exitations) in 1 patients, axial 4- to 5-mm (1 -mm gap) spineho 549-6/ in seven patients, axial 4- to 5-mm (2- to 2.5-mm gap) spin-eho 2-3/ 3-6,8-12 images in nine patients, and oronal 5-mm (1-mm gap) spin-eho 5-55/ images in three patients. The aquisition matrix was 128 to 192 X 256. Intravenous ontrast was not used in any of the patients. Imaging studies were assessed for erebral ortial pattern and thikness, white matter volume (assessed subjetively), state of myelination (as ompared with normal standards) (7), ventriular size and shape, erebellar size and development, and the presene of parenhymal masses, hemorrhage, or alifiation. Differentiation of hemorrhage from alifiation was problemati, as will be addressed in the Disussion. Patient 2 died at 3 days of age from ompliations of the ytomegali inlusion disease and underwent a limited autopsy. Results The erebral ortial gyral pattern was abnormal in nine patients. In three, the ortex was abnormally thin and essentially agyri (Fig 1 ), showing no suli other than primitive, vertially oriented Sylvian fissures. In the other six, the ortex showed multiple onfluent areas of thikening with irregular ortial surfae and irregularity of the gray matter-white matter juntion (Figs 2-4). The frontal ortex, extending posteriorly to inlude the perirolandi regions, was dysplasti in all six patients with loalized dysplasia. The temporal lobe was involved in four of these patients (Fig 3A). In addition, the parietal region of patient 4 had a urvilinear fous of Tl shortening (Fig 3C) at the gray matter-white matter juntion,

5 AJNR: 15, April 1994 CYTOMEGALOVIRUS 77 A B Fig. 1. Patient 2. A, Coronal spin-eho 5/ 2 image shows an agyri brain with a very thin ortex (blak arrows). Foi of short Tl (large white arrows), proved to be alium at autopsy, are present in multiple subependymal sites. The hippoampi (small white arrows) are shrunken and alified. B, The erebellum is markedly small and demonstrates some surfae alifiation (open white arrow). Large masses of alium (losed white arrows) line the lateral ventriles. believed to represent alifiation or laminar nerosis (8, 9). The hippoampal formations were abnormal in all seven patients in whom they ould be adequately visualized. In all of these, the temporal horns of the lateral ventriles were signifiantly enlarged, and the hippoampi themselves were vertial (as ompared with their normal horizontal orientations) and abnormally small (Fig 1A). The hippoampal region demonstrated a short T 1 in patient 2 (Fig 1A), verified as alifiation at autopsy. Myelination was delayed or absent in 1 patients. In the newborn patients, this was manifest on the MR studies as absene of T1 and T2 shortening in the posterior brain stem and the posterior limbs of the internal apsules, regions in whih myelination is present in healthy newborns. In patients 6, 9, 1, and 11 the hypomyelination was rather subtle, manifest as lak of T1 and T2 shortening in the middle and peripheral portions of the entra semiovale (Fig 4). The deep white matter strutures, the orpus allosum, and the internal apsules appeared normally myelinated. The white matter in patient 5 was ompletely myelinated but showed regions of T2 prolongation (Fig 2A), presumably resulting from tissue destrution, in the subortial white matter. On the aompanying CT san, these regions were hypodense (Fig 26). Volume of white matter was judged to be diminished in the erebral hemispheres in this patient and nine others (Figs 1, 2, and 5), implying that either lak of formation or destrution of white matter had ourred in these patients. Ventriular enlargement was present in all 1 patients who had diminished erebral hemispheri white matter. In all of these patients, the ventriular walls were smoothly expanded, without foal irregularities. There were areas with T1 and T2 harateristis idential to erebrospinal fluid, presumably ysts, in the anterior temporal lobes, anterior to the temporal horns of the lateral ventriles, in four patients (Figs 5F and 5G). These were separated from the ventrile by only a thin membrane. A small amount of white matter separated the yst from the erebral ortex in two patients; however, in the other two, the ysts extended all the way to the temporal ortex. Well-defined signal abnormalities of variable size, ranging from 1 to 15 mm in diameter (Figs 1 and 5) were seen in the erebral hemispheres of seven patients. In an eighth patient (patient 1), no abnormal foi were seen; however, some motion artifat was present, possibly obsuring small abnormalities. Most foi were erebral, periventriular, or subependymal in loation, but there were subortial, erebellar, and basal ganglia foi in one patient (patient 1, Fig 5), subortial, basal ganglia, erebellar, and mesenephali foi in another, and subependymal and erebellar foi in a third. In five patients, the masses exhibited short T1 and T2 relaxation times with respet to brain tissue. CT, obtained in three of these patients, showed high attenuation, ompatible with either alium or blood. A fourth patient

6 78 BARKOVICH AJNR: 15, April 1994 Fig. 2. Patient 5. A, A xial spin-eho 2/ 4 image shows dysplasti ortex involving most of both frontal lobes (open arrows) and asymmetri foi of long T2 (losed arrows) in the subortial regions bilaterally. B, A xial nonontrast CT through the same level as A shows subependymal alifiation and regions of hypoattenuation (arrows) that orrespond to the areas of long T2. A 8 Fig. 3. Patient 4. A, A xial spin-eho 3/ 12 image shows an abnormal gyral pattern in both frontal lobes, most marked in the anterior left frontal region (losed arrows). An infolding of abnormal ortex is present in the ri ght parietal region (open arrows ). B, Coronal spin-eho 6/ 15 image shows foal T1 shortening (open arrows) at the gray matter-white matter juntion in the right parietal lobe, presumably alifia tion. The left temporoparietal ortex is abnormally thikened (losed arro ws). A 8 (patient 2), studied only by MR (Fig 1), was found to have alium, but no blood, in the loation of these masses at autopsy. It is of interest that, in patient 1, the MR showed puntate, periventriular foi better than the CT (Figs 5B-5D). Presumably, these foi represented subaute hemorrhage and not alifiation, as CT is known to be muh more sensitive in the detetion of alium than MR. Supporting this assumption is the fat that patient 1 had intraventriular blood and, presumably, had an ative infetion at birth. Hemorrhage was identified definitely in only one patient and was intraventriular (Fig 5). However, we annot be ertain that some of the foi of T 1 and T2 shortening that we are asribing to alium are not, in fat, partially or wholly hemorrhagi. In fat, as alluded to above, we suspet the parenhymal foi of short T 1 in patient 1 are hemorrhages. The erebellum was judged to be small in seven patients. In two patients, the diminished erebellar size was dramati (Fig 1) and the erebellum appeared inompletely formed, whereas in the other five patients, the diminution in size was less dramati, and although small, the erebellum appeared ompletely formed. None of the affeted erebella had any foal abnormalities or any features of infartion. Disussion Congenital ytomegalovirus disease is the most ommon serious viral infetion among newborns in the United States ( 1 ). Congenital ytomeg-

7 AJNR: 15, April 1994 CYTOMEGALOVIRUS 79 Fig. 4. Patient 9. The pattern seen in this patient was the m ost ommon in this series. A, A xial spin-eho 6/ 15 image shows ventriular enlargem ent, abnorm al hypointensity of the subortial white m atter, and an abnormal gy ral pattern in the frontal lobes bilaterally. The gyral pattern is som ewhat diffiult to assess, beause the ortex and underlying white m atter are nearly isointense. 8, A xial spin-eho 3/ 12 image m ore learly shows the bilateral frontal ortial gyral abnormalities, with shallow suli, thikened ortex, and irregular gray m atter-white matter juntion. A B alovirus infetion ours in approximately 4 newborns eah year, or approximately 1% of all births. Of these, 1% have the various hematologi, neurologi, and developmental symptoms and signs that define the disease, inluding hepatosplenomegaly, miroephaly, impaired hearing, and small head size. An additional 1% to 15% of infeted infants subsequently develop neurologi or developmental abnormalities in the first year of life (11 ). In a large reent study from the ongenital ytomegalovirus disease registry (12), hepatosplenomegaly (52%) and petehiae (51%) were found to be the most ommon linial signs among affeted hildren. Severe permanent neurologi onditions were found in 55%, inluding intraranial alifiations ( 43% ), miroephaly (27 %), horioretinitis (15%), and seizures (1 % ). "Less severe neurologial abnormalities" (not speified) were found in 31 %, and hearing loss of varying degree was noted in 27%. Although the sonographi findings of ongenital ytomegalovirus (2, 13, 14) and the CT findings (2, 15, 16) are better defined, MR findings have reently been reported (2, 3, 6, 17, 18). MR findings in ongenital ytomegalovirus have inluded erebellar hypoplasia (2, 18, 19), erebral atrophy (2), ortial gyral anomalies (3, 6, 18), dilated ventriles (6, 17), large subarahnoid spaes (6), Tl and T2 prolongation of the white matter and delayed myelination (6, 17, 18), pahygyria (17), paraventriular ysts (6), and intraranial alifiation (1 %) (6, 17). We observed similar anomalies; in addition, however, our sans of our patients demonstrated hemorrhage, foal white matter injury, and hippoampal abnormalities. Moreover, gyral anomalies varied from omplete lissenephaly to loalized ortial dysplasia. There was abnormal T1 and T2 prolongation of the white matter in eight of the 1 patients. In patients 2, 3, 7, and 8, who had lissenephaly with thin erebral orties and marked diminution of erebral white matter (Fig 1), this appearane may be the result of diffuse postenephaliti gliosis, as suggested by Titelbaum et al (3). However, patients 1, 4, 9, 1, and 11 had little diminution in the amount of white matter and had areas of normal-appearing ortex (Figs 3 and 4) beneath whih the white matter had diffusely prolonged T1 and T2 relaxation times. The mildly diminished volume of white matter and the normal-appearing overlying ortex would suggest that the enephaliti proess was not loally very severe; therefore, we onur with Boesh et al (6) that the T1 and T2 prolongation most likely results, at least in part, from delayed or defiient myelination. Delayed myelination is expeted in patients with ongenital ytomegalovirus on a theoretial basis, as muh of the brain damage from the infetion is in the periventriular region, the region of the germinal matrix. The lineage of the ells being formed in the germinal zones differs depending upon the gestational age of the fetus (2-23). Neurons are formed from about 8 weeks gestational age until 16 to 2 weeks (exat timing is not learly established); their migration to the erebral ortex ontinues until about 24 to 26 weeks and is followed by a period of ortial

8 A 8 D Fig. 5. Patient 1. A, Axial CT shows blood layering in the oipital horn of the left lateral ventrile (large arrow). A small fous of hemorrhage is present in the right basal ganglia region (small arrow). 8, Axial CT at the level of the top of the lateral ventriles. A irular region of hemorrhage is present in the right hemisphere with a rim of high attenuation (arrows) surrounding it. C, Axial spin-eho 549/16 image at the same level as B shows marked T 1 shortening from hemorrhage in the right hemisphere (open arrows) orresponding to the high attenuation CT fous. A seond area of hemorrhage (losed arrow) that was not seen on the CT is present in the left parietal lobe. A few smaller foi are present (small arrows). D, Axial spin-eho 3/12 image at the same level as Band C shows that both of the larger lesions identified in C have a rim of marked T2 shortening surrounding heterogeneous signal in the enter. Multiple puntate foi of hemorrhage, manifest as regions of T2 shortening, that were not identified on the CT or T1-weighted MR are present along the margins of the ventrile and in the frontal white matter. G, Axial spin-eho 549/16 image at the level of the medulla shows three areas of hemorrhage (arrows) in the erebellum. F, Axial spin-eho 3/ 12 image 6 mm rostral to E shows multiple puntate foi of hemorrhage that were not apparent on the CT san or on the short-repetition-time/ short-eho-time MR images. The fous of hemorrhage in the lateral left erebellar hemisphere in E has a high signal intensity (white arrow) on this image. A yst in the right anterior temporal lobe (blak arrows) is faintly seen. G, Sagittal spin-eho 616/ 11 image shows the right temporal lobe yst (arrow) surrounded by low intensity white matter, anterior and inferior to the temporal horn of the lateral ventrile. 71

9 AJNR: 15, April 1994 CYTOMEGALOVIRUS 711 organization (22). Near the end of the period of neuronal prodution, astroyte generation begins (2, 21, 24). The germinal zone attains its maximum size at 26 weeks (2, 25), at whih time astroytes are the predominant ell type being formed (2). The early stages of astroyte formation and migration seem to be neessary for normal ortial organization (21 ). Oligodendroytes are produed by differentiation of astroytes produed in the first half of the third trimester (25) (van der Knaap M, Myelination and Myelin Disorders: a Magneti Resonane Study in Infants, Children, and Young Adults, Thesis, Free University of Amsterdam and University of Utreht, 1991 ), before the germinal zone beomes depleted at 32 to 34 weeks. As a result of this timing of ell prodution, an early injury to the germinal zone is likely to produe a redution in both neurons and glia, whereas an injury after 22 to 24 weeks is expeted to result in the prodution of a normal number of neurons but a redued number of glia. Moreover, an injury after 24 to 26 weeks would not be expeted to affet the migration of neurons to the erebral ortex (they have ompleted the migration). Injury late in the migration phase or in the organizational phase, if vasular in nature, is believed to result in polymirogyria (22, 25). Therefore, in order to disuss the interrelationship of the white matter and ortial anomalies in patients with ongenital ytomegalovirus, a disussion of the most likely ause of brain injury in ongenital ytomegalovirus is neessary. Some debate exists in the literature as to the mehanism of brain injury in ongenital ytomegalovirus. The earliest and most widely repeated theories stated that ytomegalovirus has a speial affinity for the immature ells within the germinal zone and that the loss of periventriular brain tissue and the abnormalities of the erebral ortex were the result of the injury to the germinal zone and the ells produed therein (4, 15, 26, 27). Marques-Dias et al (28), however, noted that no neuropathologi differenes were apparent histologially between the ortial anomalies of patients with ytomegalovirus and those with polymirogyria of other auses. Furthermore, they noted no orrelation between the loation of the periventriular injuries and the loation of the ortial anomalies. Finally, they noted no differene between the periventriular white matter injuries of patients with ytomegalovirus and those of patients with ongenital toxoplasmosis infetion or with ishemi periventriular white matter alifiations, as desribed by Banker and Larrohe (29). Moreover, animal models indiate that polymirogyria is the result of a loalized ortial injury that disrupts ortial organization (3-33). The injury may involve the developing ortex itself (3, 31, 33) or may be limited to the more superfiial pial-glial membrane, whih aids in ortial organization (32, 34). Polymirogyria has not been shown to result from germinal matrix injury. In ontradistintion to diret brain injury by the virus, Marques-Dias et al postulated that the brain injury in patients with ongenital ytomegalovirus was the result of ishemia (28). In support of their postulate, they noted that thromboti vasulitis aused by ytomegalovirus has been seen in the plaenta (35, 36) and in the fetal lung (35). Moreover, they pointed out that ytomegalovirus inlusions seem to be preferentially present in proliferating apillary endothelia in animal models (37). More reent literature supports the ausative role of ytomegalovirus in vasular injury (38), thrombosis (39), and retinal angiitis (4). However, although they found numerous ytomegalovirus inlusions in apillaries, Marques-Dias et al (28) found no evidene of thrombosis or alteration of vessel walls and did not note any randomly multifoal lesions, as would be expeted from a vasuliti proess. They found the distribution of lesions more onsistent with systemi insuffiieny of fetal irulation. Therefore, they postulated that the injuries were the result of a plaentitis and seondary hroni perfusion insuffiieny. The patterns of brain injury in some of the patients in our series are ompatible with systemi vasular insuffiieny as the underlying ause of brain damage. The dysplasti ortex tended to be symmetri and to involve the operular region, a pattern we and others (41-43) have noted in polymiorgyria seondary to fetal hypotension. Using the known timing of ell formation in the germinal zones (disussed earlier), we estimate that patients 2, 3, and 8, who had ompletely smooth (agyri) brains, extremely thin erebral orties, and markedly diminished erebral hemispheri white matter, presumably were infeted while neurons were being formed and in the early stages of migrating to the erebral ortex (probably before 16 or 18 weeks gestation). Injury at this time would result in few neurons arriving at the ortex to establish axonal ramifiations with the remainder of the brain; moreover, the destroyed germinal zone would be

10 712 BARKOVICH AJNR: 15, April 1994 unable to produe a suffiient number of oligodendroytes to myelinate the axons that had formed. We estimate that patients 4, 5, 6, 7, 9, and 1, who had symmetrial frontotemporoparietal ortial dysplasia (presumably polymirogyria) and delayed myelination with relatively little ventriular dilatation, were infeted later, at a time when neuronal migration was nearly ompleted and neuronal organization was taking plae (probably about 18 to 24 weeks) (22, 24, 44). The fat that this pattern of injury is nearly idential to the most ommon pattern of ortial dysplasia/polymirogyria in patients without a history of ytomegalovirus (42, 45, 46) also supports a vasular ause of the brain damage in ytomegalovirus disease, beause polymirogyria is presumed to be aused primarily by ishemia (22, 25, 47). Patients 1 and 11, who had normal gyral patterns, presumably were infeted during the third trimester, after ortial organization is ompleted. Patient 1, who had intraventriular and, presumably, peri ventriular hemorrhage deteted on his sans, may have had ongoing infetion (ytomegalovirus erebritis) at birth, as reported by Bray et al (48) and Bale et al (49). Beause the erebritis presumably lasts several weeks, or even months (2, 48), and myelination ontinues for years after birth (5, 51), the presene of both delayed myelination and ative infetion in a term infant is preditable. The MR appearane of delayed myelination and a thin ortex in a patient with agyria (suh as patients 2, 3, and 8) should strongly suggest the possibility of ytomegalovirus infetion, beause the most ommon form of agyria, type 1 lissenephaly, is usually assoiated with a thik layer of neurons lying entral to a "ell-sparse" zone (52, 53). Moreover, the white matter in type 1 lissenephaly is almost always normally myelinated (53). Finally, the erebellum is typially normal in type 1 lissenephaly but was abnormally small in all three of the patients with smooth brains in this series. Differentiation from miroephalia vera, a ondition aused by premature exhaustion of the germinal matrix (22, 25, 54), may be diffiult on the basis of imaging studies; however, the assoiated hepatosplenomegaly and skin petehiae of ytomegalovirus and the family history (miroephalia vera is thought to have an autosomal reessive pattern of inheritane) should allow differentiation. Differentiation of type 2 lissenephaly, as is seen in Walker-Warburg syndrome and in Fukuyama ongenital musular dystrophy, from the patterns of ortial dysplasia seen in patients 6, 7, 9, and 1, is somewhat more diffiult using imaging riteria alone. Both disorders have myelination delay, erebellar anomalies, and thikening of the erebral ortex with irregular gyral patterns (22, 55-58). Absene of the harateristi bundles of dysplasti neurons plunging into the white matter at the gray matter-white matter juntion, a typial finding in type 2 lissenephaly (22), should help to differentiate the disorders. In addition, patients with Walker-Warburg syndrome have allosal hypogenesis or agenesis, persistent primary hypertrophi vitreous, hydroephalus, and, ommonly, oipital ephaloeles that should allow differentiation. Moreover, neither Walker-Warburg syndrome nor Fukuyama ongenital musular dystrophy are reported to have subependymal or periventriular alifiations, as were seen in two of our patients. Finally, hepatosplenmegaly and skin petehiae, ommon in ytomegalovirus disease, are not reported with type 2 lissenephaly. The pattern of brain injury orrelated with age of patient presentation in that the patients with agyria all presented at birth and died by the age of 4 months, whereas those with a more pahygyri or polymirgyri pattern presented later in infany with seizures and developmental delay, presumably aused by the ortial and myelination anomalies. The timing of presentation is in agreement with our theory regarding timing of infetion in that those patients infeted earlier, when organogenesis was still in progress, would presumably have more severe multiorgan involvement and have more diffiulty adapting to the extrauterine environment. The exeption to the trend of earlier presentation of more severe brain injury was patient 1, who presented at birth despite a normal gyral pattern. However, as has been disussed, patient 1 is presumed to have had an ative infetion at birth and, therefore, would be expeted to be siker than those whose infetions were no longer ative. The observation that the erebellum was abnormally small in eight of the 11 patients in this series may be of some diagnosti signifiane. Although Dobyns (56, 59) has reported a few patients with lissenephaly and erebellar hypoplasia, most disorders of neuronal migration and organization and most destrutive brain disorders have normal-appearing erebelli on imaging studies. In hypoxi-ishemi injury, for example, sparing of the erebellum is a lassi imaging finding that is used to establish the diagnosis (6-64).

11 AJNR: 15, April 1994 CYTOMEGALOVIRUS 713 We suggest that the presene of erebellar hypoplasia on MR studies, when seen in assoiation with ortial dysplasia or lissenephaly, myelination delay, and diminished white matter, should prompt strong onsideration of a diagnosis of ongenital ytomegalovirus infetion. Seven patients had temporal lobe abnormalities, inluding hippoampal dysplasia and anterior temporal lobe ysts. The hippoampal dysplasia, manifest as vertial (as opposed to the normal horizontal) orientation of the hippoampus (Fig 1A), is similar to the pattern desribed by Baker and Barkovih (65) in developmental anomalies suh as type I lissenephaly and agenesis of the orpus allosum. Therefore, it is likely that this hippoampal appearane is a result of an arrest of hippoampal development by the infetion. It is a nonspeifi finding. The signifiane of the temporal lobe ysts (Figs 5F and 5G) is unertain; they were seen only in four patients. They are probably the result of loalized tissue destrution. Analysis of more ases is neessary to determine whether this finding is at all speifi for ytomegalovirus. Periventriular foi of signal abnormality were seen in eight of the 11 patients in this study (Figs 1 and 5). The lesions were globular in patients 1 and 2, ranging in size from 3 to 15 mm in diameter. The lesions in patient 1 had short T1 and T2 relaxation times; patient 2 was imaged only with short-repetition-time/short-eho-time spin-eho images, so only short T1 relaxation time was established. Based on the MR appearane of these lesions, we suspeted that they were foi of hemorrhage; however, autopsy of patient 2 revealed only periventriular alifiation, with no evidene of blood. The diffiulty in distinguishing hemorrhage from alium on MR has been disussed both theoretially (9) and in referene to injuries of the pediatri brain (62). Cytomegalovirus infetion seems to be another ategory in whih the differentiation of subaute blood from alium is problemati. The diffiulty was ompounded in patient 1, in whom hemorrhage was learly present within the lateral ventriles (Fig 5A), and who, as suggested earlier, probably had an ative infetion at birth. Although the MR harateristis of the lesions in this patient were idential to those in patient 1, the CT harateristis suggest that they were hemorrhagi. Parenhymal injury an learly progress postnatally in patients with ongenital ytomegalovirus (2, 48, 49); thus, the possibility annot be exluded that some of the lesions were initially hemorrhagi and evolved into alified lesions via dystrophi alifiation. Whether this transformation would be detetable by MR is unknown. Puntate periventriular lesions were present on MR images of five patients. In two, the lesions had short T 1 and short T2 relaxation times ompared with surrounding white matter, whereas in three patients, the lesions ould not be distinguished from surrounding brain on Tl-weighted images, but had short T2 ompared with surrounding brain. We suspet that most of these puntate foi represent alifiation, although some may represent residual blood or blood breakdown produts similar to the periventriular foi in patient 1 (Fig 5). Calified subependymal foi were learly shown on the CT of patient 5 but were not seen on the aompanying MR (Fig 2). Although the MR of patient 5 was relatively old, this observation one again proves the advantage of CT over MR in the detetion of small, puntate alifiation. The presene of alifiation, however, does not seem essential to make the diagnosis. In summary, we have desribed a wide spetrum of ortial anomalies, ranging from omplete agyria to diffuse polymirogyria to grossly normal, in infants and hildren with ongenital ytomegalovirus infetions. We have proposed an explanation, based on the timing of the infetion during gestation, to explain this spetrum. Moreover, we have desribed parenhymal lesions, of variable size, that may represent alifiations or hemorrhage and may be diffiult to distinguish by MR imaging alone. Finally, we have noted myelination delay and erebellar hypoplasia as onsistent abnormalities in ongenital ytomegalovirus and suggested that these findings may be valuable in differentiating ongenital infetion from various geneti and sporadi ongenital disorders. Aknowledgments We thank Erik Gaensler, MD, for the use of ases 9 and 1, and Wallae Pek, MD, for the use of ase 2. Referenes 1. Bignami A, Appiiutoli L. Miropolygyria and erebral a lifiation in ytomegali inlusion disease. Ata Neuropathol 1964;4: Perl man JM, Argyle C. Lethal ytomegalovirus infetion in preterm infants: linial, radiologial, and neuropathologial findings. Ann Neural 1992;31 :64-68

12 71 4 BARKOVICH AJNR: 15, April Titelbaum DS, Hayward JC, Zi mmerman RA. Pahygyrilike hanges: topographi appearane at MR imaging and CT and orrelation with neurologi status. Radiology 1989; 1 73: Friede R, Mikolasek J. Postenephaliti porenephaly, hydranenephaly or polymirogyria. A review. Ata Neuropatho/ 1978; 43: Crome L, Frane N. Mirogyria and ytomegali inlusion disease in infany. J Clin Patho/1 959; 12: Boesh C, lssa kainen J, Kewitz G, Kikinis R, Martin E, Boltshauser E. Magneti resonane imaging of the brai n in ongenita l ytomegalovirus infetion. Pediatr Radio/ 1989; 19: Barkovih AJ, Kjos B, Jakson Jr D, Norman D. Normal maturation of the neonatal and infant brain: MR imaging at 1.5 T. Radiology 1988; 166: Boyko, Burger P, Shelburne J, Ingram P. Non-heme mehanisms for T 1 shortening: pathologi, CT, and MR eluidation. A JNR Am J Neuroradio/ 1992; 13: Henkelman M, Watts J, Kuharzyk W. High signal intensity in MR images of alified brain tissue. Radiology 1991; 179: Alford C, Stagna S, Pass R, Britt W. Congenital and perinata l ytomegalovirus infetions. Rev Infet Des 199; 12 (suppl 7):S7 45-S Yow M. Congeintal ytomegalovirus disease: a now problem. J Infet Dis 1989; 159: Dobbins JG, Stewart JA, Demmler GJ. Surveillane of ongenital ytomegalovirus disease, Collaborating Registry Group. MMWR CD C Surveil/ Summ 1992;41: Butt W, Makey R, DeCrespigny L, et al. Intraranial lesions of ongenital ytomegalovirus infetion deteted by ultrasound sanning. Pediatris 1984;73: Dykes F, Ahmann P, Lazzara A. Cranial ultrasound in the detetion of intraranial alifiations. J Pediatr 1982; 1: Barkovih AJ. Infetions of the nervous system. In: Barkovih AJ, ed. Pediatri neuroimaging. New York : Raven, 199: Bale J, Bray P, Bell W. Neuroradiographi abnormalities in ongenital ytomegalovirus infetion. Pediatr Neural 1985; 1 : Hayward JC, T itelbaum DS, Clany RR, Zimmerman RA. Lissenephaly-pahygyria assoiated with ongenital ytomegalovirus infetion. J Child Neuro/1 991;6: Sugita K, Ando M, Makino M, Takanashi J, Fujimoto N, Niim i H. Magneti resonane imaging of the brain in ongenital rubella virus and ytomegalovirus infetions. Neuroradiology 1991 ;33: Volpe J. Viral and related intraranial infetions. In: Volpe J, ed. Neurology of the newborn. Philadelphia: Saunders, 1987: Jammes J, Gilles F. Telenephali development: matri x volume and isoortex and alloortex surfae areas. In: Gilles F, Leviton A, Dooling E, eds. The develop ing human brain. Boston: Wright, 1983: Gressens P, Rihelme C, Kadhim H, Gadisseux J-F, Evrard P. The germinati ve zone produes the most ortia l astroytes after neuronal migration in the developing mam malian brain. Bioi Neonate 1992;61 : Barkovih AJ, Gressens P, Evrard P. Formation, maturation, and disorders of brain neoortex. AJNR A m J Neuroradio/1 992; 13: Altman J, Bayer S. Horizontal ompartmentation in the germinal matries and intermediate zone of the embryoni rat erebral ortex. Exp Neuro/ 199;17: Sidman RL, Raki P. Development of the human entral nervous system. In: Haymaker W, Adams RD, eds. Histology and histopathology of the nervous system. Springfield: T homas, 1982: Evrard P, de Saint-Georges P, Kadhim H, Gadisseux J-F. Pathology of prenatal enephalopathies. In: Frenh J, ed. Child neurology and developmental disabilities. Baltimore: Brookes, 1989: Naeye R. Cytomegali inlusion disease: the fetal diso rder. Am J Clin Patho/ 1967;47: Merer R, Luse S, Guyton D. Clinial diagnosis of generalized ytomegali inlusion disease. Pediatris 1953; 11 : Marques-Dias M, Harmant-van Rijkevorsel G, Landrieu C, et al. Prenatal ytomegalovirus disease and erebral m irogyria: evidene for perfusion failure, not disturbane of histogenesis, as the major ause of fetal ytomegalovirus enephalopathy. Neuropediatris 1984;15: Banker B, Larrohe J. Peri ven triular leukomalaia of infany. A rh Neuro/1 962;7: Dvorak K, Feit J. Migration of neuroblasts through partial nerosis of the erebral ortex in newborn rats. Contribution to the problems of morphologial development and developmental period of erebral mirogyria. A ta Neuropatho/1 977;38: Dvorak K, Feit J, J urankova Z. Experimentally indued foal mirogyria and status verruosis deformis in rats-pathogenesis and interrelation, histologial and autoradiographial study. A ta Neuropathol 1978;44: Humphreys P, Rosen G, Press D, Sherman G, Galaburda A. Freezing lesions of the developing ra t brain: a model for erebral mirogyria. J Neuropathol Exp Neural 1991 ;5: Suzu ki M, Choi B. Repair and reonstrution of the ortial plate following losed ryogeni inj ury to the neonatal rat erebrum. Ata Neuropathol 1991 ;82: Rosen G, Sherman G, Rihman J, Stone L, Galaburda A. Indution of moleular etopias by punture wounds in newborn rats and mie. Dev Brain Res 1992;67: Benirshke K, Mendoza G, Bazeley P. Plaental and fetal manifestations of CMV. Virhows Arh [BJ 1974;16: Altshuler G, MAdams J. CMV inlusion disease of a 19 week fetus. Am J Obstet Gyneo/ 1971; 11 1 : Vogel F. Enhaned suseptibility of proliferating endothelium to salivary gland virus under naturally ourring and experimental onditions. A m J Patho/ 1958;34: Span AH, Grauls G, Bosman F, van BC, Bruggeman CA. Cytomegalovirus infetion indues vasular injury in the rat. A theroslerosis 1992;93: Jenkins RE, Peters BS, Pinhing AJ. Thromboemboli disease in AIDS is assoiated with ytomegalovirus disease (letter). AIDS ;5: Spaide R, Vitale A, Toth I, Oliver J. Frosted branh angiitis assoiated with ytomegalovirus retinitis. Am J Ophtha/mo/1 992; 11 3: Bankl J, Jellinger K. Zentralnervose Shaden nah fetaler Kohlenoxidvergiftung. Beitr Pathol Anat 1967;1 35: Barth PG. Disorders of neuronal m igration. Can J Neural Si 1987; 14: Hallervorden J. Ueber eine kohlenoxydvergiftung im fetalleben mit entwiklunsstorung der hirnrinde. A llg Z Psyhia tr 1949; 124: Barkovih AJ, Lyon G, Evrard P. Formation, maturation, and disorders of white matter. AJNR Am J Neuroradio/1 992;13: Barkovih AJ, Kjos B. Shizenephaly: orrelation of linial findings with MR harateristis. AJNR Am J Neuroradio/ 1992; 13: Barkovih AJ, Kjos B. Non lissenephali ortial dysplasia: orrelation of imaging findings with linia l defiits. AJNR Am J Neuroradio/ 1992; 13: Rihman DP, Stewart RM, Caviness BS. Cerebral mirogyria in a 27 week fe tus: an arhitetoni and topographi analysis. J Neuropathol Exp Neuro/ 1974;33: Bray P, Bale J, Anderson R, Kern E. Progressive neurologia l disease assoiated with hroni ytomegalovirus infetion. Ann Neural 1981 ;9: Bale J, Sato Y, Eisert D. Progressive postnatal subependymal nerosis in infant with ongenita l ytomegalovirus infetion. Pediatr Neural 1986;2:367-37

13 AJNR: 15, April 1994 CYTOMEGALOVIRUS Yakovlev P, Leours A. The m yelogeneti yles of reg ional maturation of the brain. In: Minkowski A, ed. Regional development of the brain in early life. Oxford : Blakwell, ] g67: Brody B, Kinney H, Klom an A, Gilles F. Sequene of entral nervous system m yelination in human infany. I. A n autopsy study of m yelination. J Neuropathol Exp Neural 1987;46: Aia rd i J. The agyria-pahygyria omplex: a spetrum of ortial malformations. Brain Dev 1991;13: Barkovih AJ, Koh T, Ca rrol C. The spetrum of lisse nephaly: report of ten ases analyzed by magneti resonane imaging. A nn Neural 1991 ;3: Parain D, Gadisseux J, Henoq A, T ayot J, Evrard P. Diagnosti prenatal et etude d 'une miroephalia vera a 26 semaines de gestation. In: Szliwowski H, Bormans J, eds. Pragres en neuralogie p ediatrique. Brussels: Prodim, 1985: Bordarier C, Aia rdi J, Boutieres F. Congenital hydroephalus and eye abnormalities with severe developmental brain defets: Warburg's syndrome. Ann Neural 1984; 16: Dobyns WB, Kirkpatrik JB, Hittner HM, Roberts RM, Kretzer FL. Syndromes with lissenephaly: 2. Walker-Warburg and erebral oular musular syndromes and a new syndrome with T ype 2 lissenephaly. Am J /o1ed Genet 1985;22: Takada K, Nakamura H, Takashima S. Cortial dysplasia in Fuk u yama ongenital musular dystrophy: a Golgi and angio arhitet toni analysis. Ata Neurapathol 1988; 76: Yoshioka M, Saiwai S, Kuroki S, Nigami H. MR imaging of the brain in Fukuyama-type ongenita l musular dystrophy. AJNR Am J Neuroradioll 99l ;12: Dobyns WB, Gilbert EF, Opi tz JM. Further omments on the lissene phaly syndromes. Am J /o1ed Genet 1985;22: Barkovih A J, Truwit CL. MR of perinatal asphyxia: orrelation of gestational age with pattern of damage. AJNR Am J Neuroradiol 199; 11 : Barkovih AJ. Metaboli and destrutive brai n disorders. In: Barkovih AJ, ed. Pediatri neuroimaging. New York : Raven, 199: Barkovih AJ. MR and CT evaluation of profound neonatal and infantile asphyxia. AJNR Am J Neuroradiol l992; 13: Keeney S, Adok E. MArdle C. Prospetive observa tions of I high-ri sk neonates by high field (1.5 tesla) magneti resonane imagin g of the entral nervous system: I. Intraventriular and ex traerebrallesions. Pediatris 1991;87: Keeney S, Adok E, MArd le C. Prospetive observations of 1 high-risk neonates by high field (1.5 tesla) magneti resonane imaging of the entral nervous system : II. Lesions assoiated with hypoxi-ishemi enephalopathy. Pediatris 1991 ;87: Baker L, Barkovih A J. The large temporal horn : MR analysis in developmental brain anomalies versus hydroephalus. AJNR Am J Neuraradioll 992; 13: