Supratentorial hemorrhage suggested on susceptibility-weighted magnetic resonance imaging in an infant with hydranencephaly

Introduction

Hydranencephaly is a fetal central nervous system disorder in which most of the cerebral hemispheres are damaged and largely replaced with cerebrospinal fluid and debris covered with a thin, membranous sac [2, 6]. Several theories on the pathogenesis of hydranencephaly have been proposed; these include infection, toxin, aplasia, genetic disorders, vascular origin, maternal hypoxia, and twin-to-twin transfusion syndrome [6]. Although the exact cause is still unknown, the prevailing theory is that hydranencephaly is the result of the occlusion of the cerebral arteries above the supraclinoid level, mostly affecting the major vessels of the anterior circulation because of a characteristic distribution of the preserved brain parenchyma as well as the results of animal studies [4].

Susceptibility-weighted imaging (SWI) has been introduced as a very sensitive magnetic resonance (MR) technique to detect susceptibility changes such as blood, iron, calcification, and air [8]. SWI uses a high-spatial-resolution three-dimensional, fast low-angle shot MR imaging technique with postprocessing of phase data [7]. SWI is based on the blood oxygen level-dependent contrast effect, with additional magnetic susceptibility weighting using phase data. SWI depicts hemorrhage as foci of signal loss as well as high-resolution venography. It provides new imaging contrast, which is different from conventional MR images or MR angiography. Several investigators have shown that SWI is more sensitive in detecting hemorrhage than other MR sequences or computed tomography (CT) [8]. The imaging findings with respect to the presence of hemorrhage can be of additional information to consider the etiology of hydranencephaly. We present an infant with hydranencephaly who showed signal loss, mainly at the peripheral portion of both thalami and the falx cerebri on SWI, suggesting previous hemorrhage.

Presentation of the case

A 36-year-old woman was admitted at 34 gestational weeks. Three previous pregnancies had resulted in term deliveries of normal infants. She had not attended antenatal care until the day before the delivery, when fetal ultrasonography demonstrated cerebral hemispheres replaced with fluid, suggestive of hydranencephaly. She delivered a female infant the next day. Antenatal information regarding potential infection was lacking. Apgar scores were 8 and 9 at 1 and 5 min, respectively. The general physical findings of this newborn were unremarkable. Her birth weight was 2190 g (p 50–80), and her head circumference 36.5 cm (>97^th percentile). The infant fed well and showed no circulatory or respiratory problems.

MR imaging was performed on day 2 to confirm the intracranial status using a 3-T MR unit (Achieva; Philips Healthcare, Best, The Netherlands). T1- and T2-weighted images showed an almost complete absence of the cerebral hemispheres, replaced by fluid-filled space. A small part of the posterior portion of the cerebral tissue could still be identified. Both thalami, the brain stem, and the cerebellum were preserved (Figure 1A, B). The cerebral cortex was not found peripherally from the fluid-filled sac. Part of the lateral ventricles was still visible (Figure 1C). These findings led us to the diagnosis of hydranencephaly. SWI showed foci of signal loss, mainly at the peripheral portion of both thalami and the falx cerebri, suggesting hemorrhage (Figure 1D–F). Faint signal loss at the margin of the cerebellum and brain stem was identified. Cystic dilatation of the posterior fossa was also identified. IV collagen α1 (COL4A1) sequencing was performed because the COL4A1 mutation is related to massive intracranial hemorrhage resembling hydranencephaly. However, no mutation was found. The infant is still alive at 12 months of age, has developed severe macrocephaly (69 cm at 12 months), and has not developed any milestones but drinks and feeds well.

Figure 1

(A) T2-weighted sagittal image. (B) T1-weighted coronal image. (C) T2-weighted coronal image. (D–F) Transverse susceptibility-weighted images. Images show almost complete absence of the cerebral hemispheres, replaced by fluid-filled spaces. Part of the posterior cerebrum (Figure A, arrows) and both thalami (Figure B, arrows) are shown. Part of the lateral ventricles is revealed (Figure C, arrows), which is not dilated. SWI shows foci of signal loss at the peripheral portion of both thalami (Figure D, E, arrows), the falx cerebri (Figure F, arrows), and the preserved cerebrum (Figure F, arrowheads), suggesting hemorrhage.

Discussion

In this case, SWI showed signal loss mainly at the periphery of the thalamus and falx cerebri. Signal loss on SWI is related to a susceptibility effect (i.e., disturbance of a homogeneous magnetic field), caused by various paramagnetic, ferromagnetic, or diamagnetic substances [8]. Therefore, SWI can depict blood products, including deoxyhemoglobin, methemoglobin, and hemosiderin, being paramagnetic substances. There have been a few reported cases of intracranial hemorrhage resulting in hydranencephaly [2, 3]. In these cases, massive bright echogenicity indicating hemorrhage was detected filling the supratentorial region on fetal ultrasonography, which was subsequently absorbed and replaced by an almost completely fluid-filled space. These ultrasound changes confirmed by postmortem examination. In our case, SWI did not show extensive signal loss, but showed small foci or linear signal loss at the peripheral portion of the supratentorial preserved structures. On the basis of the reported observations, the SWI findings in the present case can be a chronic status of previous hemorrhage. Owing to the distribution, the signal loss on SWI in this case is highly unlikely to be due to acute hemorrhage. Although signal loss on SWI can be ascribed to other substances such as calcium and iron, these are less likely in this infant because it is unusual that calcium deposition is seen at the peripheral portion of the thalamus, and it is known that there is almost no cerebral iron deposition at the neonatal period. In this case, small foci of signal loss were identified in the peripheral portion of the brain stem and the cerebellum. These may be due to the limited amount of hemosiderin, circulating in the cerebrospinal fluid because there were no destructive changes of the brain stem and cerebellum.

The etiology of most cases of hydranencephaly is obscure, and this case is no exception. However, an ischemic etiology can be a major cause, which is suggested by the typical vascular-like distribution of the lesion and the results of animal models with carotid ligation [4]. The reported findings regarding intracranial arteries in hydranencephaly vary. Greene et al. [3] reported autopsy findings of a patient with hydranencephaly, which revealed intact internal carotid artery and cerebral vessels, suggesting that a transient vascular occlusion had led to cerebral necrosis and subsequent reperfusion hemorrhage. On the contrary, there are reports about the interruption or hypoplasia of the anterior and middle cerebral arteries on CT angiography [4], and the interrupted internal carotid arteries and posterior cerebral arteries on MR angiography [6]. In the present case, owing to the characteristic damaged area of the brain parenchyma, a vascular origin of the hydranencephaly has been suggested. It can be speculated that hemorrhage may have occurred because of rupture of the fragile internal carotid artery after occlusion [2].

It is important to distinguish hydranencephaly from severe hydrocephalus. Children with hydrocephalus, even when the condition is serious, respond well to a cerebrospinal fluid diversion procedure and may have a subsequent acceptable cognitive and motor function if shunting is performed at an early age [1]. Pathologically, hydranencephaly is distinguished from massive hydrocephalus. In hydranencephaly, only leptomeninges are present, with occasional remnants of gliotic neural tissue. In hydrocephalus, on the other hand, a continuous cortical layer is identifiable [3]. On MR imaging, the thin rim of cerebral cortex and white matter can be identified around the markedly dilated ventricle in hydrocephalus [2]. In addition, hemorrhage in hydrocephalus is usually noted in the ventricular system. However, SWI findings suggest hemorrhage at the peripheral portion of the thalami and falx cerebri, as shown in this case, and can be an indicator to suggest destruction of the brain parenchyma and therefore can be distinguished from hydrocephalus. Although part of the signal loss at the marginal portion of the thalami can be the hemorrhage in the lateral ventricle, entirely peripheral signal loss around the thalami may result from not only intraventricular hemorrhage but also destructive parenchymal hemorrhage.

In this case, cystic dilatation of the posterior fossa with a high tentorium is noted. There was no absence or hypoplasia of the cerebellar vermis. The tentorium settles in its position, and fuses along the midline by the end of the embryonic period (i.e., 8 weeks of development following fertilization) [5]. This means that the cisterna magna was already markedly enlarged at that time. Meanwhile, the morphology of the brain stem and cerebellum suggests that the event that led to hydranencephaly might have occurred at the third trimester of pregnancy. The suggested timing of the occurrence of both supra- and infratentorial abnormalities seems to be different in this case, and the relation of these posterior fossa findings and hydranencephaly remains uncertain.

In summary, we present an infant with hydranencephaly showing supratentorial signal loss on SWI, suggestive of a preceding hemorrhage and parenchymal destructive changes.