Neuroradiology DOI 10.1007/s00234-014-1379-2

PAEDIATRIC NEURORADIOLOGY

Vasogenic edema characterizes pediatric acute disseminated encephalomyelitis Giulio Zuccoli & Ashok Panigrahy & Gayathri Sreedher & Ariel Bailey & Ernest John Laney IV & Luca La Colla & Gulay Alper

Received: 5 March 2014 / Accepted: 7 May 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract Introduction MR imaging criteria for diagnosing acute disseminated encephalomyelitis (ADEM) have not been clearly established. Due to the wide spectrum of differential considerations, new imaging features allowing early and accurate diagnosis for ADEM are needed. We hypothesized that ADEM lesions would be characterized by vasogenic edema due to the potential reversibility of the disease. Methods Sixteen patients who met the diagnostic criteria for ADEM proposed by the International Pediatric Multiple Sclerosis Study Group (IPMSSG) and had complete MR imaging studies performed at our institution during the acute phase of the disease were identified retrospectively and evaluated by experienced pediatric neuroradiologists. Results Vasogenic edema was demonstrated on diffusionweighted imaging (DWI) and corresponding apparent diffusion coefficient (ADC) maps in 12 out of 16 patients; cytotoxic edema was identified in two patients while the other two G. Zuccoli (*) : A. Panigrahy : G. Sreedher : A. Bailey : E. J. Laney IV Department of Radiology, Section of Neuroradiology, Children’s Hospital of Pittsburgh of UPMC, 4401 Penn Avenue, Pittsburgh, PA 15244, USA e-mail: [email protected] E. J. Laney IV Department of Diagnostic Radiology, Rush University Medical Center, Chicago, IL, USA L. La Colla Department of Anesthesiology, University of Parma, Parma, Italy L. La Colla Department of Emergency Medicine, UPMC Shadyside Hospital, Pittsburgh, PA, USA G. Alper Department of Pediatric Neurology, Neuroimmunology Clinic, Children’s Hospital of Pittsburgh of UPMC, Pittsburgh, PA, USA

patients displayed no changes on DWI/ADC. ADC values for lesions and normal-appearing brain tissue were 1.39±0.45× 10−3 and 0.81±0.09×10−3 mm/s2, respectively (p=0.002). When considering a cutoff of 5 days between acute and subacute disease, no difference between ADC values in acute vs. subacute phase was depicted. However, we found a significant correlation and an inverse and significant relationship between time and ADC value. Conclusion We propose that vasogenic edema is a reliable diagnostic sign of acute neuroinflammation in ADEM. Keywords Acute disseminated encephalomyelitis . ADEM . Vasogenic edema . DWI . ADC Abbreviations ADEM Acute disseminated encephalomyelitis IPMSSG International Pediatric Multiple Sclerosis Study Group

Introduction Acute disseminated encephalomyelitis (ADEM) is classically defined as an immune-mediated inflammatory demyelinating disease of the CNS predominately involving the white matter of the brain and spinal cord. It arises spontaneously and sporadically in a predominantly pediatric age group, likely precipitated by a viral or bacterial infection or recent vaccination [1–4]. A prodromal phase of fever, malaise, headache, nausea, and vomiting may occur prior to the development of neurological signs [5]. The specific neurologic features have been reported to include unilateral or bilateral pyramidal signs, acute hemiplegia, ataxia, cranial nerve palsies, vision loss, seizures, and speech impairments [1, 6–8]. There is normally a good prognosis with complete recovery usually

Neuroradiology

Patients were identified from the Pittsburgh Pediatric Demyelinating Registry. This study was approved by the institutional review board (IRB) of the University of Pittsburgh. Among the 112 patients presenting with acute CNS demyelination between January 2003 and October 2013, 16 ADEM patients with available diffusion-weighted imaging (DWI)/apparent diffusion coefficient (ADC) images were identified. All patients included in this study met the diagnostic criteria for ADEM proposed by the IPMSSG [13]. Clinical features and neuroimaging findings were reviewed. MR imaging examinations were performed during the acute (≤5 days) or subacute (>5 days, ≤3 weeks) phase of the disease at a field strength of 1.5 T (Signa; GE Healthcare, Milwaukee, WI). Imaging sequences of the brain included T1-weighted, T2-weighted, FLAIR, proton density, gradient echo, and contrast-

enhanced T1-weighted sequences in multiple planes. DWI and ADC maps were also obtained in all patients. MR imaging findings were represented by symmetric or asymmetric signal intensity alterations on short and long TR images within the brain cortex, subcortical white mater, deep gray matter structures including basal ganglia, thalami and brainstem and other brain areas including the corpus callosum and the optic pathways, and areas of contrast enhancement after injection of a standard dose of gadolinium-based contrast material (gadobenate dimeglumine 0.5 mol/L solution, MultiHance; Bracco, Milan, Italy). Findings were recorded following consensus by two expert pediatric neuroradiologists [GZ and AP]. DWI and an ADC map were evaluated together for signal intensity changes with regard to vasogenic vs. cytotoxic edema. Lesions isointense or hyperintense on DWI and hyperintense on the ADC map were considered consistent with vasogenic edema. Lesions hyperintense on DWI and hypointense on the ADC map were considered consistent with cytotoxic edema as previously described (Fig. 1) [14]. The diffusion measurements were performed using an echoplanar imaging (EPI) sequence in axial plane, frequency direction right to left (TR 8000/TE99), section thickness of 5 mm, matrix of 128×128, and field of view of 280×280 mm. The ADC was calculated using the formula ln(S0/S1)/(b1−b0), where S0 and S1 are signal intensities at gradient settings b1 and b0. To correlate ADCs with anatomic regions of interest, a multiplanar imaging localizer system (iSite Radiology, PACS System, Philips) was used. In this, ADCs were calculated from the largest lesion and also from normal-appearing brain tissue (NABT) by manually drawing a circular area of 2–3 cm2. The NABT ADC ROIs were calculated from the white matter in those patients with the largest lesion centered in the white matter and from the gray matter in those patients with the largest lesion centered in the gray matter. The number of hours between the reported onset of neurological symptoms and the MR imaging examination was calculated by reviewing the patients’ electronic charts (Cerner Co. Millennium). The phase of the disease was further classified into

Fig. 1 Brain MR imaging demonstrates extensive vasogenic edema with areas of increased T2-FLAIR signal in the periventricular white matter (a, arrows) and thalami (a, arrowheads), corresponding to areas of

vasogenic edema demonstrating iso-to-high signal intensity signal on DWI (b, arrows and arrowheads, respectively) and increased signal on ADC images (c, arrows and arrowheads)

occurring between 1 and 6 months [9]. However, there is also some evidence that undiagnosed and untreated ADEM may possibly have lasting neurological sequels, including deficits in behavior, executive function, and attention [10, 11]. MR imaging is an intrinsic part of the clinical diagnosis proposed by the International Pediatric Multiple Sclerosis Study Group (IPMSSG), although there is no lesion pattern pathognomonic for ADEM. The MR imaging abnormalities seen in ADEM are most often characterized on T2-weighted and fluid-attenuated inversion recovery (FLAIR) sequences as large, asymmetric, and poorly demarcated areas of increased T2 signal intensity typically involving the subcortical and central white matter and the gray-white matter junction of the cerebral hemispheres, cerebellum, brainstem, and spinal cord [5]. Contrast enhancement is variable and, when present, typically involves the majority of lesions simultaneously [12]. The aim of the present study was to investigate the relationship between stage of disease and vasogenic and/or cytotoxic edema among pediatric ADEM patients.

Case series

Neuroradiology Table 1 Demographic, clinical data, and neuroradiological findings

Note that all symptoms displayed in the History column are to be intended in addition to an altered level of consciousness. Altered level of consciousness was the only presenting symptom in patient 3 G gray matter involvement, W white matter involvement, U unilateral involvement, B bilateral involvement, +, −, and N/A positive, negative, or non-available data about enhancement after contrast material administration

Patient

Age (years)

Sex

History

MRI findings

Outcome

1 2 3 4 5

16 12 7 4 6

F M M M F

Headache, vomiting Focal neurological deficits Focal neurological deficits Headache

G, W, U, − G, W, B, + G, W, B, N/A G, W, B, + G, W, B, +

Disability Disability Recovery Recovery Recovery

6 7 8 9 10 11 12 13 14 15 16

4 5 6 9 3 7 2 9 14 5 11

F M M F F M M F M M F

Ataxia Headache Headache Headache, focal neurological deficits Ataxia Seizures Ataxia Headache Focal neurological deficits Focal neurological deficits Headache

G, W, B, N/A W, B, − G, W, B, + G, W, B, − G, W, B, + G, B, − G, W, B, + G, B, + G, B, − G, B, − G, B, −

Recovery Recovery Recovery Recovery Recovery Recovery Recovery Recovery Recovery Recovery Recovery

acute (≤5 days of the onset of initial neurological symptoms) and subacute (after 5 days from the onset of initial neurological symptoms up to 3 weeks) similarly to a previous study on ADEM [15]. Shapiro-Wilk test was used to test normality of continuous variables. T test (or the corresponding non-parametric test) was used for continuous data. Frequencies were compared using the chi-square or the Fisher exact test. The correlation between ADC values and time was performed using Spearman’s rho correlation coefficient. Data were shown as mean±SD or median (25th–75th percentile), when appropriate. Statistical analysis was performed with SPSS 21.0.

Results In our study, 9 (56 %) of the 16 patients were males (Table 1). Patients’ age ranged from 2 to 16 years. Vasogenic edema was demonstrated on DWI and corresponding ADC maps of 12 patients (75 %). Cytotoxic edema was observed only in two patients (12.5 %). The same number of patients (two patients, 12.5 %) displayed no MR imaging signs on DWI/ADC map images likely accounting for pseudonormalization phenomenon (Table 2). With specific reference to lesion distribution, 15 patients (93.75 %) showed gray matter involvement including the supratentorial and infratentorial cortex (ten patients, 62.5 %) and deep gray nuclei/brainstem (13 patients, 81.25 %). Eleven patients (68.75 %) showed white matter involvement. Nine patients (56.25 %) displayed other alterations. Among these patients, there were four cases of optic tract involvement (two bilateral, one left, and one right), five cases of optic chiasm involvement, and one case in which

there was involvement of the splenium. In two patients, gadolinium-based contrast material was not injected, and 14 patients underwent contrast-enhanced MR imaging. Among these, seven patients (50 %) showed some type of enhancement. Neuromyelitis optica (NMO) was reasonably excluded in all patients presenting with optic pathway involvement Table 2 ADC values in NABT and ADEM lesions in relation to time of imaging Patient

Time to scan (hours)

Phase of disease

ADC (×10−3 mm/s2) NABT

Acute lesions

Subacute lesions – – – – 1.26 – – 1.75 – 1.27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

9 24 29 72 504 108 48 192 72 168 24 244 106 48 96

Acute Acute Acute Acute Subacute Acute Acute Subacute Acute Subacute Acute Subacute Acute Acute Acute

0.61 0.73 0.79 0.77 0.79 0.78 0.77 0.77 0.91 0.79 0.94 0.97 0.81 0.83 0.90

2.44 1.86 1.47 1.67 – 0.68 1.29 – 1.32 – No edema – 0.68 No edema 1.17

16

264

Subacute

0.76

–

1.27 – – 1.26

Time to scan is represented in hours (column 2) from onset of symptoms. Phase of disease (column 3) is considered acute when ≤5 days and subacute when ≥5 days

Neuroradiology Fig. 2 Mean±SD of ADC values in lesions vs. NABT. Displayed values are 1.40±0.56, 0.81±0.09, and 1.36±0.22 for acute, NABT, and subacute lesions, respectively. Asterisk indicates a statistically significant difference from ADC values in both acute and subacute stages

based on long-term clinical outcome and/or negative NMOIgG aquaporin antibody. ADC values for lesions and NABT are displayed in Table 2 and Fig. 1. They ranged from 0.68×10−3 to 2.44×10−3 mm/s2 (1.39±0.45×10−3 mm/s2) and from 0.61×10−3 to 0.97× 10−3 mm/s2 (0.81±0.09×10−3 mm/s2) in NABT, respectively, with a statistically significant difference with the pathological brain tissue (p=0.002). Interestingly, when considering a cutoff of 5 days (120 h) between acute and subacute disease, no statistical difference between ADC values in acute over Fig. 3 The most accurate curve estimated (R2 =0.556, p=0.002) for ADC values in lesions vs. time. Inverse refers to the inverse model, following the general equation ADC values= b0+(b1/time)

subacute phase (1.40±0.56×10−3 vs. 1.36±0.22×10−3 mm/s2, p=0.867) was found. Notwithstanding, the difference between ADC in acute stages and NABT as well as ADC in subacute stages and NABT was statistically significant (p=0.013 and p= 0.003, respectively) (Fig. 2). Despite this lack of correlation between acute/subacute stage of ADEM and ADC map, a significant correlation between ADC values and hours after symptom onset was found (Spearman’s rho correlation coefficient of −0.585, p=0.028). The most accurate curve estimated for ADC data over time (R2 =0.556, p=0.002) was that

Neuroradiology

described by an inverse model following the general equation Y=b0+(b1/t), with a constant value of 1.144 and a b1 value of 11.820 (Fig. 3). No statistical correlation between DWI/ADC data, localized MR imaging findings on pre- and post-contrast images, and clinical outcome was found.

Discussion It is evident that more specific clinical and radiological criteria are needed to distinguish ADEM from other acute demyelinating and inflammatory processes affecting the white and gray matter, especially in the pediatric population. In pediatric patients, the clinical and laboratory findings of ADEM may not be distinguishable from multiple sclerosis (MS) with CSF analysis being positive for oligoclonal bands in up to 29 % of cases of ADEM and in not more than 70 % of MS cases [16, 17]. Although the prognosis of ADEM is often favorable, severe cases can result in persisting disability. It is thus necessary that reliable indices for diagnosing and predicting outcome be developed. DWI and ADC findings are useful in differentiating various neurological disorders. Isolated case reports from the literature suggest that the course and outcome of ADEM could be assessed using MRI with DWI and ADC maps [18, 19]. These studies have found that those patients with cytotoxic edema particularly involving the brainstem experience a greater extent of tissue necrosis and poorer outcomes than those patients with primarily vasogenic edema. To date, the larger case series on DWI/ADC imaging evaluation in ADEM have been published by Balasubramanya et al. (eight patients) and Donmez et al. (eight patients) [15, 20]. Balasubramanya et al. found that ADC maps directly correlated to the stage of the disease, while Donmez et al. did not correlate interval from symptom onset to MR imaging and concluded that brainstem involvement rather than cytotoxic or vasogenic edema could represent an indicator of poor prognosis. Balasubramanya et al. found that when initial DWI/ ADC was obtained within 5 days from clinical onset (acute phase), restricted diffusion was observed, which subsequently evolved into a pattern of increased diffusion thereafter (subacute phase). Based on these findings, the authors speculated that during the acute stage of ADEM, cytotoxic edema may reflect underlying swelling of myelin sheaths, reduced vascular supply, and dense inflammatory cell infiltration followed by vasogenic edema likely underlying expansion of the extracellular space. Our demonstration of vasogenic edema in the majority of patients and the inverse correlation between ADC values and the time interval from symptom onset to MR imaging suggests that vasogenic edema could be involved in the pathogenesis of ADEM starting from the earliest phase of the disease. Furthermore, our findings are supported by the known clinical outcomes of ADEM, that is, the majority of patients have a favorable prognosis, which correlates with the

reversible nature of vasogenic edema. In the present study, cytotoxic edema was infrequently identified in pediatric ADEM patients and was not necessarily associated with a poor prognosis. Of note, the prior investigations on DWI/ ADC in ADEM were performed on smaller and mixed-age (both pediatric and adult patients) populations and the patients were not enrolled based on the recently introduced IPMSSG diagnostic criteria [15, 20]. Study limitations Our study has some limitations, including the small sample size. To confirm our findings, additional large-scale investigations are needed. ADEM may present with lesions in the brain as well as spinal cord; however, we did not evaluate the spinal cord at presentation in all patients and hence did not include the data for final analysis. Furthermore, the clinical implications of optic chiasm/tract involvement are unknown and must be further investigated in order to understand its role in this disease.

Conclusion ADC values of vasogenic edema inversely correlate with interval from symptom onset to MR imaging, supporting the hypothesis that vasogenic edema characterizes the initial pathophysiological manifestation of pediatric ADEM.

Ethical standards and patient consent We declare that all human and animal studies have been approved by the University of Pittsburgh IRB and have therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. We declare that all patients gave informed consent prior to inclusion in this study. Conflict of interest We declare that we have no conflict of interest.

References 1. Tenembaum S, Chamoles N, Fejerman N (2002) Acute disseminated encephalomyelitis: a long-term follow-up study of 84 pediatric patients. Neurology 59:1224–1231 2. Hynson JL, Kornberg AJ, Coleman LT, Shield L, Harvey AS, Kean MJ (2001) Clinical and neuroradiologic features of acute disseminated encephalomyelitis. Neurology 56:1308–1312 3. Amit R, Shapira Y, Blank A, Aker M (1986) Acute, severe, central and peripheral nervous system combined demyelination. Pediatr Neurol 2:47–50 4. Menge T, Hemmer B, Nessler S et al (2005) Acute disseminated encephalomyelitis: an update. Arch Neurol 62:1673–1680 5. Tenembaum S, Chitnis T, Ness J et al (2007) Acute disseminated encephalomyelitis. Neurology 68(Suppl 2):S23–S36

Neuroradiology 6. Leake J, Albani S, Kao A et al (2004) Acute disseminated encephalomyelitis in children: epidemiologic, clinical and laboratory features. Pediatr Infec Dis J 23:756–764 7. Murthy K, Faden H, Cohen M et al (2002) Acute disseminated encephalomyelitis in children. Pediatrics 110:21–28 8. Anlar B, Basaran C, Kose G et al (2003) Acute disseminated encephalomyelitis in children: outcome and prognosis. Neuropediatrics 34: 194–199 9. Dale R, de Sousa C, Chong W et al (2000) Acute disseminated encephalomyelitis, multiphasic disseminated encephalomyelitis and multiple sclerosis in children. Brain 123:2407–2422 10. Hahn C, Miles B, MacGregor D et al (2003) Neurocognitive outcome after acute disseminated encephalomyelitis. Pediatr Neurol 29:117– 123 11. Schwarz S, Mohr A, Knauth M et al (2001) Acute disseminated encephalomyelitis: a follow-up study of 40 adult patients. Neurology 56:1313–1318 12. Rossi A (2008) Imaging of acute disseminated encephalomyelitis. Neuroimag Clin N Am 18:149–161 13. Krupp LB, Banwell B, Tenembaum S (2007) International Pediatric MS Study Group. Consensus definitions proposed for pediatric multiple sclerosis and related disorders. Neurology 68(16 Suppl 2):S7– S12

14. Alper G, Sreedher G, Zuccoli G (2013) Isolated brain stem lesion in children: is it acute disseminated encephalomyelitis or not? AJNR Am J Neuroradiol 34:217–220 15. Balasubramanya KS, Kovoor JM, Jayakumar PN et al (2007) Diffusionweighted imaging and proton MR spectroscopy in the characterization of acute disseminated encephalomyelitis. Neuroradiology 177–83 16. Gallucci M, Caulo M, Cerone G, Masciocchi C (2001) Acquired inflammatory white matter diseases. Childs Nerv Syst 202–10 17. Dale RC, de Sousa C, Chong WK et al (2000) Acute disseminated encephalomyelitis, multiphasic disseminated encephalomyelitis and multiple sclerosis in children. Brain 12:2407–22 18. Axer H, Ragoschke-Schumm A, Böttcher J et al (2005) Initial DWI and ADC imaging may predict outcome in acute disseminated encephalomyelitis: report of two cases of brain stem encephalomyelitis. J Neurol Neurosurg Psychiatry 76:996–998 19. Kawashima S, Matsukawa N, Ueki Yet al (2009) Predicting the motor outcome of acute disseminated encephalomyelitis by apparent diffusion coefficient imaging: two case reports. J Neurol Sci 280:123–126 20. Donmez FY, Aslan H, Coskun M (2009) Evaluation of possible prognostic factors of fulminant acute disseminated encephalomyelitis (ADEM) on magnetic resonance imaging with fluid-attenuated inversion recovery (FLAIR) and diffusion-weighted imaging. Acta Radiol 50:334–9