Elsevier

Brain and Development

Volume 40, Issue 2, February 2018, Pages 116-125
Brain and Development

Original article
Use of high b value diffusion-weighted magnetic resonance imaging in acute encephalopathy/encephalitis during childhood

https://doi.org/10.1016/j.braindev.2017.07.012Get rights and content

Abstract

Aim

To determine the use of high b value diffusion-weighted imaging (DWI) in the diagnosis and assessment of acute febrile encephalopathy/encephalitis in childhood.

Subjects and methods

We enrolled 22 children, for whom we examined DWI with b = 1000 s/mm2, DWI with b = 3000 s/mm2, and apparent diffusion coefficient (ADC) map with b = 1000 during the acute phase of febrile encephalopathy/encephalitis. Clinical diagnoses included acute encephalopathy with biphasic seizures and late reduced diffusion (AESD; n = 6), clinically mild encephalopathy/encephalitis with a reversible splenial lesion (MERS; n = 6), and herpes simplex virus encephalitis (HSE; n = 3), unclassified acute encephalopathy/acute encephalitis (n = 2); acute encephalitis with refractory, repetitive partial seizures (AERRPS; n = 1); other encephalopathy (n = 1); infarction (n = 1); head injury (n = 1); or mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes (n = 1). The diagnostic quality of brain lesions was compared between b = 1000 and b = 3000 DWI images by visual inspection. In addition, we attempted a quantitative assessment using apparent diffusion coefficient (ADC) value and an index of signal intensity (SI) ratio, defined as the mean SI at the affected lesion divided by the mean SI at the pons.

Results

High intensity lesions were either visible only on b = 3000 DWI (n = 5; 2 AESD, 1 MERS, 1 HSE, and 1 unclassifiable encephalopathy) or more effectively identified on b = 3000 DWI than on b = 1000 DWI (n = 17). The outcome of the former five subjects was favorable, without motor or intellectual sequelae. The mean SI ratio of b = 3000 was significantly greater than that of b = 1000 in AESD and MERS subgroups as well as in all 22 subjects. Mean ADC values were lower in the AESD and MERS than that in the HSE subgroups.

Conclusion

We concluded that b = 3000 DWI was superior to b = 1000 DWI in detecting abnormal lesions in acute encephalopathy/encephalitis during childhood.

Introduction

Diffusion-weighted imaging (DWI) visualizes the movement of water molecules and is superior to classical magnetic resonance imaging (MRI) for the visualization of abnormal lesions, including brain infarct, tumor, and posterior reversible encephalopathy syndrome. This modality has also been applied to neurological disorders during childhood, which has facilitated the diagnosis and assessment of hypoxic–ischemic encephalopathy/acute febrile encephalopathies [1]. DWI has contributed to the establishment of acute encephalopathy with biphasic seizures and late reduced diffusion (AESD) and mild encephalopathy with reversible splenial lesion (MERS) by detecting white matter and callosal lesions, respectively [2], [3]. Additionally, DWI has proven to be useful in the diagnosis and assessment of neonatal herpes simplex encephalitis [4]. Apparent diffusion coefficient (ADC) demonstrated the degree of restricted movement of water molecules. It is not subject to the T2 shine-through effect and represents the exact water diffusivity, which helps to identify whether high intensity on DWI represents a true restriction of diffusion. In addition, ADC values decreased and increased in intracellular and extracellular edema, respectively, and are useful in the determination of lesion pathophysiology.

Clinically, MRI scanners with a static magnetic field up to 1.5 T have been typically used to date. This has limited b values to 1000–1500 s/mm2 in routine brain DWI [5]. With the recent advent of more powerful gradient hardware along with a magnetostatic intensity of 3 T, it is possible to generate greater b values and thereby obtain higher diffusion sensitivity [5]. The use of high b value DWI has been established for the diagnosis of acute cerebral infarction [6], [7]; however, studies of high b value DWI in pediatric neurological disorders have been limited [8]. In the present study, using a 3-T MR system, we compared DWI acquired at b = 3000 s/mm2 with those at b = 1000 s/mm2 in children with acute febrile encephalopathy/encephalitis to evaluate the effect of this approach on the quality of diagnosis and assessment of these conditions.

Section snippets

Patients

We reviewed the medical records and MRI of children who were admitted to the Tottori University Hospital between August 2008 and July 2015 and for whom brain high b value DWI was performed for suspicion of serious acute etiologies. Of the 49 enrolled subjects, the final diagnoses in 27 patients were febrile convulsion (n = 15), epilepsy (n = 4), and others. The remaining 22 patients (Table 1; mean, 40.8 months; age range, 20 days–126 months; 11 males and 11 females) were diagnosed with either AESD (n = 

Qualitative analysis

High intensity lesions in the 22 subjects were either visible only on b = 3000 DWI (n = 5; shaded in Table 1) (Fig. 1, Fig. 2, Fig. 3), or more effectively identified on b = 3000 DWI than b = 1000 DWI (n = 17; Table 1; Fig. 4). The five subjects with lesions only detectable on b = 3000 DWI included two cases of AESD (Fig. 1), one case of MERS, one case of HSE (Fig. 2), and one case of unclassifiable acute encephalopathy (Fig. 3). None of these subjects demonstrated atrophic changes on routine MRI, or

Discussion

With DWI, a pair of motion probing gradients (MPG) is added at the time of image acquisition. During the time of MPG application, spinning protons moved by diffusion generate phase dispersion, the degree of which determines the degree of decrease in SI, i.e. higher SI indicates more restricted diffusion. The b value expresses the effect of MPG. A higher b value, indicating a smaller amount of dispersion, represents slow diffusion of water molecules, which also reflected as phase shifting in the

References (13)

There are more references available in the full text version of this article.
1

These authors equally contributed to this article.

View full text