The use of hypercapnic challenge blood oxygen level dependent (BOLD) MRI for the investigation of childhood steno-occlusive arteriopathy

BACKGROUND: In childhood, cerebral arteriopathy causes cerebral ischaemia and infarction via two related mechanisms. The first, thrombotic vaso-occlusive stroke is the more typical mechanism of stroke in childhood. The second mechanism of infarction related to arteriopathy and not seen in other forms of ischaemic stroke is that of chronic hypoperfusion of the brain. The infarcts in this case are typically located in watershed zones and can accumulate gradually over time. Moyamoya is the prototypic arteriopathy representing the hypoperfusion injury. It is characterized by chronic progressive narrowing of the distal internal carotid, proximal middle cerebral and anterior cerebral arteries; chronic low flow infarction with accumulating ‘string of pearls’ in the white matter (Fig 1). It is this chronic hypoperfusion of the brain that is the subject of my PhD thesis. Cerebrovascular reactivity is a marker of cerebrovascular reserve and has been shown to be a biomarker of ischaemic risk in adults. OBJECTIVES: The primary objectives were to, in a group of children with moyamoya: 1) Validate the use of a qualitative measure of cerebrovascular reactivity as a biomarker of ischaemic risk, namely, hypercapnic challenge BOLD MRI CVR (hBOLD CVR) via two methods: a) breath-holding and b) induced hypercapnia during general anaesthesia as reliable and repeatable for use in the paediatric population 2) Assess the utility of qualitative assessment of cerebrovascular reactivity using hBOLD CVR as a tool for the identification of the risk of ischaemia in children with arteriopathy Method Hypercapnic challenge hBOLD CVR studies were obtained in children with steno-occlusive arteriopathy prospectively enrolled in The Hospital for Sick Children Stroke Registry. Semi-quantitative methods of measuring CVR were devised and used for the purpose of analysis. hBOLD CVR studies were analysed for reliability and reproducibility of the method of analysis. Clinical and radiologic data were collected and hBOLD CVR findings described for all children enrolled. Exploratory analysis of hBOLD CVR as a potential biomarker of ischaemic risk in the paediatric population were conducted. In particular association of hBOLD CVR with clinical symptomatology; parenchymal and vascular indicators of arteriopathy; neuropsychological outcome and cortical thickness were examined. Results Forty seven children (37 Bilateral or unilateral Moyamoya arteriopathy, 6 Unilateral Non-moyamoya arteriopathy [Transient Cerebral Arteriopathy] and 4 Bilateral Non-moyamoya arteriopathy [2 PHACE(S), 1 Takayasu arteritis, 1 Sickle Cell Disease]) were enrolled and had hBOLD CVR studies. The mean age of diagnosis of arteriopathy across all groups was 8.1 years (SD 4.2) (range 7 months - 18 years). Clinical and radiographic features differed across arteriopathy groups. Most presented with acute stroke, however, among children with NF1-MM most (almost 50%) were asymptomatic and diagnosed on screening MRIs. Infarction patterns differed, with deep watershed infarction being the typical pattern in the moyamoya group in contrast to thrombotic vaso-occlusive infarction pattern in the Non-moyamoya groups. Qualitative hBOLD CVR abnormalities were concordant with moyamoya laterality, and in unilateral moyamoya demonstrated tissue level microvascular dysfunction in the contralateral unaffected hemisphere. Qualitative hBOLD CVR abnormalities demonstrated concordance with clinically important manifestations of ischaemia including stroke, transient ischaemic attacks, cortical thinning and IQ. There was a lack of concordance with indices of executive function. In addition the moderate to severe steno-occlusive arteriopathy seen in the children with Transient Cerebral Arteriopathy was not associated with abnormality of hBOLD CVR. Conclusion The thesis studies demonstrated that qualitative assessment of hBOLD CVR using breath-hold or general anaesthetic is feasible, reproducible and reliable in paediatric population. The utility of hBOLD CVR as a measure of tissue level microvascular dysfunction and thus a biomarker of ischaemic risk was demonstrated. However, larger longitudinal studies are required to characterize this further