British Journal of Radiology (2006) 79, e140-e144
© 2006 British Institute of Radiology
doi: 10.1259/bjr/44174031
Spontaneous thrombosis of congenital cerebral arteriovenous malformation complicated by subdural collection: in utero detection with disappearance in infancy
B DeCesare, MD2,
M F Omojola, MB, FRCPC1,
E F Fogarty, MD1,
J C Brown, MD1 and
C Taylon, MD3
Department of 1Radiology 2Creighton University Medical School3Department of Surgery, Creighton University Medical Center, Omaha, Nebraska, USA
Correspondence: Matthew F Omojola, MB, FRCPC, Associate Professor, Vice Chairman, Director of Neuroradiology, Department of Radiology, Creighton University Medical Center, 601 North 30th Street, Omaha, NE 68131, USA. E-mail: momojola{at}creighton.edu.
 |
Abstract
|
|---|
We report a case of congenital left temporal lobe arteriovenous malformation (AVM) detected by cranial ultrasound in utero and confirmed immediately after birth by cranial Doppler ultrasound and cranial MRI. The AVM disappeared on follow-up cranial MRI 4 months later. A small left frontal subdural collection was present on these follow-up MR images, which subsequently resolved by the 7 month MRI study. The cause of the spontaneous thrombosis of the AVM is uncertain. The frontal subdural collection may be secondary to volume loss. This case documents the perinatal presence of AVM. The baby was neurologically intact before, during and after the thrombosis of the AVM.
|
Introduction
|
|---|
Arteriovenous malformations (AVMs) were originally described in the mid 1800s, and although medicine has learned a great deal about them, much is still to be discovered. The aetiology, incidence, treatment and follow-up care have all been debated and theorized. A unifying consensus has yet to be reached with regards to the natural history of this lesion. To complicate the matter further, intracranial AVMs have been demonstrated to undergo spontaneous regression. Many hypotheses exist as to the cause of spontaneous regression. Most of these are derived from case studies [1–13]. Because of the rarity of this event, larger studies are difficult to perform, although our search included four retrospective studies [14–17] and two prospective studies [18, 19].
In review of the literature, several common themes are evident regarding intracranial AVMs: they are rare in the population and spontaneous regression/thrombosis is very rare (0.8% [14], 2–3% [18, 20]). Only 1% of AVMs are seen in patients less than 2 years old [4] and spontaneous regression in infancy is exceedingly rare. Pre-natal diagnosis has not been documented previously [4]. Abdulrauf et al [14] presented six cases of spontaneous angiographic obliteration of cerebral AVMs in their study, and their literature review revealed 24 other cases. Only three infant case reports of spontaneous thrombosis of AVM have been identified in the literature [4, 7, 12].
This report documents a case of AVM detected as a vascular brain mass during pregnancy. Neuroultrasound and MRI of the brain on the first day of life confirmed a left temporal lobe AVM that underwent spontaneous disappearance early in life with a follow-up MRI at 4 months demonstrating non-visualization of the lesion. A small subdural collection over the adjacent left frontal lobe resolved spontaneously on follow-up MRI at 7 months.
|
Case report
|
|---|
This baby boy was born to a 37-year-old Gravida 2, Caucasian mother. The mother had mitral valve prolapse and gestational diabetes controlled with diet. During a routine ultrasound at 32 weeks, a complex brain mass was noted on the left side of the fetal brain, measuring about 4 cm in diameter and causing a midline shift to the right. The brain mass did not change significantly during the third trimester. The baby was delivered by elective caesarean section at 39 weeks gestation.
The baby was born without incident. Apgar scores were 9 at 1 min and 9 at 5 min. The initial Chemstrip demonstrated hypoglycaemia. The baby weighed 3390 g and measured 52 cm. The head circumference was 36.5 cm at >95th percentile. The anterior fontanelle was soft, flat and normal, measuring 4 cm x 2 cm. On physical examination there was acrocyanosis, petechiae over his groin and back, peeling skin over the lower extremities, no signs of heart failure and no cranial bruit. The remainder of the physical examination, including the neurological examination, was normal. Vital signs remained stable throughout the hospital course. On the third day of life, phototreatment was started because of a peak total bilirubin of 15 mg dl–1, with no subsequent sequelae. Platelets were borderline low.
Cranial imaging was performed soon after birth. Cranial ultrasound with colour Doppler demonstrated a 4 cm left temporal lobe vascular lesion with mixed echogenicity. There was no significant mass effect demonstrated (Figure 1a,b
). Later, on the same day, a cranial MRI with contrast showed a well-circumscribed 5 cm x 3.5 cm ovoid lesion in the left temporal lobe with heterogeneous signal intensity on T1, T2 and fluid attenuated inversion recovery (FLAIR) weighted images. Multiple tubular signal voids were present in the mass and in the adjacent suprasellar cistern. Post-contrast T1 weighted images revealed marked contrast enhancement with contrast draining into the transverse and straight sinuses. There was mild mass effect on the brain stem. There was no evidence of extra-axial fluid collection. The lesion was clearly an AVM (Figure 1c–f). Neurosurgical opinion was to observe the baby's haemodynamic status before possibility of intervention. Hearing tests were performed, which demonstrated some deficit in the left ear. The baby was discharged home in good condition to be followed up in the clinic.
View larger version (167K):
[in this window]
[in a new window]
|
Figure 1. (a) Coronal cranial ultrasound at birth through the temporal lobes showing a 4 cm hyperechoic left temporal lobe mass. (b) Sagittal Doppler cranial ultrasound at birth through the left temporal lobe showing a left temporal lobe vascular mass. (c) Axial T1 weighted MR image at birth through the temporal lobes showing a well defined, ovoid, left temporal lobe mass of heterogeneous high signal, causing mild compression of the brain stem. (d) Post-intravenous gadolinium DTPA axial T1 weighted image through the lesion demonstrating intense enhancement and venous drainage into dilated transverse sinuses. (e) Coronal T2 weighted MR image at birth through the temporal lobes shows a round left temporal lobe mass of heterogeneous intensity with multiple tubular signal voids and large basilar and left posterior cerebral arteries. (f) Axial T2 weighted MR image at birth through the suprasellar cistern showing multiple tubular signal voids in the left temporal lobe. The left middle cerebral and left posterior cerebral arteries are large. (g) Axial T2 weighted MR image through the temporal lobes at the age of 4 months showing disappearance of the left temporal lobe arteriovenous malformation (AVM). No evidence of significant residual lesion. There is left temporal lobe volume loss. (h) Axial fluid attenuated inversion recovery (FLAIR) MR image through the level of the lateral ventricles at 4 months showing a small left frontal subdural collection. Left hemisphere atrophy is present. (i) Axial T2 weighted MR image through the suprasellar cistern at 7 months showing no evidence of a vascular malformation. There is left temporal lobe atrophy.
|
|
About 4 months after birth, a follow up cranial MRI showed complete disappearance of the left temporal lobe AVM. There was no evidence of abnormal vessels or draining veins. A small subdural collection of high signal intensity on FLAIR and T2 weighted sequences was present in the left frontal region. There was significant loss of volume in the left temporal lobe. There was no focal brain parenchymal abnormal intensity or mass effect (Figure 1g,h
). It was concluded that there was spontaneous thrombosis of the congenital AVM. Clinical examination at this time revealed no neurological abnormality.
At the age of 7 months, another follow up cranial MRI showed extra-axial cerebrospinal fluid (CSF) spaces were enlarged bilaterally, the left larger than the right. The left temporal lobe was smaller in size compared with the right. The AVM remained thrombosed (Figure 1i). The left frontal subdural collection had disappeared. Subsequent follow up MRI up to the age of 2 years showed that the AVM remained thrombosed. Throughout the boy's hospitalization and follow up, no seizure activity or abnormal neurological activity was demonstrated.
|
Discussion
|
|---|
Considering the paucity of clinical studies, many proposed mechanisms exist concerning the formation of AVMs and the cause of spontaneous regression [1, 3, 6, 10, 13, 14, 16, 19]. There has been debate about whether AVMs form pre-natally or post-natally. However, as this case demonstrates, intracranial AVMs are able to form prior to birth. In fact, most descriptions of these AVMs are that they are "congenital", thereby implying their presence at birth. While most people would likely agree with this rationalization, lack of proof with imaging studies has prevented certainty. There remains uncertainty as to the progression of AVMs once formed.
Many theories regarding spontaneous regression exist. The most consistent example includes intracranial haemorrhage with subsequent mass effect thereby decreasing blood flow to the AVM and/or occlusion of draining vein leading to haemostasis and thrombus formation [10, 13, 17]. Similar to a haemorrhagic event, but distinct, is mass effect caused by haematoma formation [14]. Haemorrhage and subsequent organization of the clot with gliosis may result in a shift of nearby pathological vessels, causing kinking and slowed circulation, predisposing to thrombus formation [6]. Dyck [1] presents such a case involving a 4-year-old girl who seemed to have an AVM regress with this mechanism. A single feeding artery in a unicompartmental AVM, and a stenosis thereof, has been theorized to play a significant role in thrombus formation [10, 11, 17]. Other possible mechanisms include clot embolisation from feeding vessels [3, 10, 11], haemodynamic changes within the lesion [7] and hypercoagulable states such as pregnancy [17]. In our patient, the mechanism for spontaneous disappearance is not clear. There was no neurological symptom to suggest a cerebrovascular event. There was no mass effect on the follow-up MRI at 4 months to suggest brain parenchymal haemorrhage or swelling. The left frontal subdural collection may represent a chronic subdural haematoma or high protein containing effusion secondary to volume loss.
The majority of cases reviewed the order of presentation for a previously asymptomatic patient with AVM included haemorrhage, seizure activity and headache [10, 14, 15, 17, 19, 20]. Ondra et al [18] prospectively followed 166 symptomatic patients and found the rate of major re-bleeding was 4.0% per year, the combined morbidity/mortality rate was 2.7% per year and the mortality rate was 1.0% per year. These annual rates remained constant, for the most part, for the entire study (mean follow up was 23.7 years). On the other hand, Brown et al [15] have shown the risk of bleeding to be 2.2% per year, but this rate increases over time. However, this study differs from Ondra's in that these patients were asymptomatic. Using the Kaplan-Meier method in Brown's study, the risk of haemorrhage was calculated to be 1.3% per year at 1 year, 1.7% per year at 5 years, 1.5% per year at 10 years and 2.2% per year at 15 years. Others have shown the risk of haemorrhage at 5 years to be 10% for large AVMs and 52% for small AVMs, thereby further distinguishing the lesions by size [19]. Spetzler et al [19] showed relationships involving perfusion pressure and size to the risk of haemorrhage from ruptured AVMs. For example, small AVMs (<3 cm) presented with haemorrhage much more often than large AVMs; the incidence being 82% vs 21%. After an acute bleed, haematoma size was inversely related to size of the AVM. The increased risk of haemorrhage apparently associated with partial embolisation of large AVMs may be due to conversion to small AVM and subsequent increased pressure in the feeding artery/arteries [19].
Various modalities may greatly aid in delineating the extent of an AVM. Angiograms, ultrasound, MRI and CT scans are the most popular and widely used imaging techniques. Anatomic detail can be specifically delineated with the use of MRI as shown in this presentation. For a choice of definitive therapeutic options, cerebral angiography is imperative. Spetzler et al [19] believe that angiography should always be repeated after angiographic proven resolution of an AVM. Recanalization of a previously asymptomatic spontaneously thrombosed AVM has been reported [8]. Although extremely rare, an instance such as this is reason for various authors to strongly urge performance of regular follow-up angiographic examinations on patients with "cured" AVMs for a minimum of 3 years; otherwise surgical removal is recommended [8, 9]. MRI was sufficient in establishing the characteristics of the AVM and follow up in our patient. Angiography was deferred until definitive therapy could be considered. Fortunately, on follow-up studies all pathological vessels disappeared. The residual left temporal lobe atrophy indicates complete extirpation of the AVM. It may be pertinent to mention that the more frequently diagnosed congenital vascular malformation during the perinatal period that may thrombose spontaneously in infancy is the aneurysm of the Great Vein of Galen [21]. This entity is usually midline in location behind the splenium of the corpus callosum and quadrigeminal plate and drains principally through the straight sinus. Both the location and the draining route are different from our present case.
Perhaps the most controversial topic involving AVMs is what treatment course should be taken and when this should be implemented. Several authors have applied various decision analyses to help determine therapeutic courses. These analyses apply the Markov Process (Quality of Life assessment) or a decision tree to assist in objectifying the decision to treat an AVM. Iansek et al [22] apply the decision tree analysis in concluding that, under no circumstances should unruptured AVMs be operated upon, and thus opting for conservative treatment. Fisher [23] states that with the Markov model, the patient's age, desires, and anticipated morbidity and mortality are better accounted for in the statistical model, thus providing the possibility of neurosurgical intervention with careful consideration. Fisher [23] believes this allows surgeons to feel more comfortable in recommending treatment to younger patients who have never had an AVM rupture. This seems reasonable for an adolescent, or possibly even a child, but an infant would not seem to apply to the same model.
With little exception, treatment strategies are very difficult to generalize regarding intracranial AVMs. Nevertheless, it seems well accepted that definitive treatment should be applied to an AVM, as opposed to only palliative treatment, since this course of action may only accelerate collateral flow and aggravate the cortical steal phenomenon [24]. A conservative approach was chosen in our patient because of stable haemodynamic state and lack of neurological findings.
|
Conclusion
|
|---|
This patient's AVM presented on a routine screening pregnancy ultrasound. Very little literature exists regarding spontaneous regression of intracranial AVMs in the paediatric population and, more specifically, infants and young children. There may be a progression from pre-natal, post-natal, childhood, to adult forms of intracranial AVMs. Previously known antecedents to spontaneous disappearance such as haematoma and seizures were not present in this infant. We discuss annual bleeding and morbidity/mortality rates to demonstrate the serious and complicated nature of these lesions. Subsequent treatment decisions may be excruciatingly difficult in infant patients and a form of decision analysis does not seem to be as easily applicable as it is with adult patients. A period of observation was decided upon in our patient and, fortunately, a short time later the lesion regressed. The lack of haemodynamic instability and neurological event strengthened the decision to take a conservative approach.
Presented as a scientific exhibit, 41st Annual Meeting, American Society of Neuroradiology, Washington, DC, 27 April–2 May 2003.
Received for publication August 11, 2005.
Revision received December 9, 2005.
Accepted for publication January 3, 2006.
|
References
|
|---|
- Dyck P. Spontaneous thrombosis of an arteriovenous malformation. Neurosurgery 1977;1:287–90.[Medline]
- Ezura M, Kagawa S. Spontaneous disappearance of a huge arteriovenous malformation: case report. Neurosurgery 1992;30:595–9.[Medline]
- Hamada J, Yonekawa Y. Spontaneous disappearance of a cerebral arteriovenous malformation: case report. Neurosurgery 1994;34:171–3.[CrossRef][Medline]
- Hanigan WC, Brady T, Medlock M, Smith EB. Spontaneous regression of giant arteriovenous fistulae during the perinatal period: case report. J Neurosurg 1990;73:954–7.[Medline]
- Krapf H, Siekmann R, Freudenstein D, Kuker W, Skalej M. Spontaneous occlusion of a cerebral arteriovenous malformation: angiography and MR imaging follow-up and review of the literature. AJNR Am J Neuroradiol 2001;22:1556–60.[Abstract/Free Full Text]
- Lakke JPWF. Regression of an arteriovenous malformation of the brain. J Neurol Sci 1970;11:489–96.[CrossRef]
- Mabe H, Furuse M. Spontaneous disappearance of a cerebral arteriovenous malformation in infancy: case report. J Neurosurg 1977;46:811–5.[Medline]
- Mizutani T, Tanaka H, Aruga T. Total recanalization of a spontaneously thrombosed arteriovenous malformation: case report. J Neurosurg 1995;82:506–8.[Medline]
- Mullan S. Reflections upon the nature and management of intracranial and intraspinal vascular malformations and fistulae. J Neurosurg 1994;80:606–16.[Medline]
- Omojola M, Fox AJ, Vinuela FV, Drake CG. Spontaneous regression of intracranial arteriovenous malformations: report of three cases. J Neurosurg 1982;57:818–22.[Medline]
- Omojola M, Fox AJ, Vinuela FV, Debrun G. Stenosis of afferent vessels of intracranial arteriovenous malformations. AJNR Am J Neuroradiol 1985;6:791–3.[Abstract]
- Pascual-Castroviejo I, Pascual-Pascual JI, Blazquez MG, et al. Spontaneous occlusion of an intracranial arteriovenous malformation. Childs Brain 1977;3:169–79.[Medline]
- Pesqualin A, Vivenza C, Rosta L, et al. Spontaneous disappearance of intracranial arteriovenous malformations. Acta Neurochir (Wien) 1985;76:50–7.[CrossRef][Medline]
- Abdulrauf SI, Malik GM, Awad IA. Spontaneous angiographic obliteration of cerebral arteriovenous malformations. Neurosurgery 1999;44:280–9.[CrossRef][Medline]
- Brown RD Jr, Wiebers DO, Forbes G, et al. The natural history of unruptured intracranial arteriovenous malformations. J Neurosurg 1988;68:352–7.[Medline]
- Enam SA, Malik GM. Association of cerebral arteriovenous malformations and spontaneous occlusion of major feeding arteries: clinical and therapeutic implications. Neurosurgery 1999;45:1105–11.[CrossRef][Medline]
- Nehls DG, Pittman HW. Spontaneous regression of arteriovenous malformations. Neurosurgery 1982;11:776–80.[Medline]
- Ondra SL, Troupp H, George ED, Schwab K. The natural history of symptomatic arteriovenous malformations of the brain: a 24-year follow-up assessment. J Neurosurg 1990;73:387–91.[Medline]
- Spetzler RF, Hargraves RW, McCormick PW, et al. Relationship of perfusion pressure and size to risk of hemorrhage from arteriovenous malformations. J Neurosurg 1992;76:918–23.[Medline]
- Wilkins R. Natural history of intracranial vascular malformations. Neurosurgery 1985;16:421–30.[Medline]
- Gerards FA, Engels MA, Barkhof F, van den Dungen FA, Vermeulen RJ, van Vugt JM. Prenatal diagnosis of aneurysm of the vein of Galen (vena magna cerebri) with conventional sonography, three dimensional sonography and magnetic resonance imaging; report of two cases. J Ultrasound Med 2003;2:1363–8.
- Iansek R, Elstein AS, Balla JI. Application of decision analysis to cerebral arteriovenous malformations. Lancet 1983;1:1132–5.[Medline]
- Fisher WS III. Decision analysis: a tool of the future: an application to unruptured arteriovenous malformations. Neurosurgery 1989;24:129–35.[Medline]
- Mullan S, Mojtahedi S, Johnson, DL, et al. Embryologic basis of some aspects of cerebral vascular fistulas and malformations. J Neurosurg 1996;85:1–8.[Medline]
Top
Abstract
Introduction
