Novel structural and functional imaging in cerebral arteriovenous malformations
Clarification of the angio-architecture and haemodynamic effects of cerebral arteriovenous malformations (AVMs) using non-invasive imaging may advance our knowledge and understanding of their natural history, resulting in improvements in the management of patients with these lesions. The aims of the work in this dissertation was to investigate the haemodynamic effects of AVMs and to determine whether newer non-invasive imaging techniques allow an accurate enough assessment of the angio-architecture of AVMs to be able to replace conventional digital subtraction angiography (DSA) in some clinical situations. The hypotheses were: 1. Non-invasive structural imaging techniques, such as CT angiography (CTA) and MR angiography (MRA), provide adequate structural and volumetric information to replace the more invasive technique of DSA.2. MR perfusion imaging is able to demonstrate the alteration of cerebral haemodynamics by AVMs. 6 Rapid frame rate DSA (RFRDSA) provides useful quantitative data on the blood flow within cerebral AVMs. In this work CT and MR angiography (CTA and MRA) were compared with conventional digital subtraction angiography (DSA). Twenty patients were examined with CTA and ten with MRA. Both techniques were able to detect most of the important angio-architectural features, but were not as accurate as DSA as decision-making tools, in particular because temporal resolution and nidus definition were poor. The use of gadolinium enhancement during MRA improved the visualisation of both nidus and draining veins. The nidal volume of ten AVMs was calculated from DSA and three magnetic resonance (MR) sequences. With biplanar DSA, an ellipsoid volume was calculated using orthogonal projections. MR images showed potential but were difficult to interpret due to varied appearances of flowing blood and, on gadolinium-enhanced MRA, the enhancement of abnormal brain. The cerebral haemodynamics of fifteen patients with AVMs were examined with contrast bolus tracking. This semi-quantitative technique was able to demonstrate consistent differences in cerebral blood flow and volume, mean transit time and time to minimum signal intensity in brain distant from the AVM. Changes in the perinidal regions were dominated by the presence of draining veins. A vascular phantom was calibrated to allow calculation of flow rates from rapid frame rate DSA. The technique for quantifying flow was assessed in five patients and compared with values measured by transcranial doppler. It was not possible to calculate accurately flow and velocity for AVM feeding vessels. These imaging modalities allowed improved appreciation of the structure and haemodynamic effects of cerebral AVMs but further development is needed before they will be of use as reliable clinical tools.