Development of analysis methods to enable clinical use of Arterial Spin Labelling perfusion MRI in neurological patients
Measuring and monitoring cerebral perfusion levels is important in a number of brain diseases with vascular components.
MRI can produce quantitative cerebral blood flow maps (CBF) using dynamic susceptibility contrast sequences and Gadolinium-based contrast agents. Some of these contrast agents have however been linked to nephrogenic systemic fibrosis and are contraindicated in patients with impaired kidney function; there is also concern about the retention of Gadolinium and its potential long-term health effects.
Arterial spin labelling (ASL) measures perfusion completely non-invasively using blood water as an endogenous contrast agent. As such, ASL is suitable for longitudinal studies without the concerns associated with Gadolinium retention. ASL-based CBF maps have been shown to be comparable to PET based metabolism and ASL is starting to be employed in selected clinical applications. However, the radiological detection of perfusion abnormalities in individual patients is complicated by within-subject variability (due to the physiological state of the subject at the time of the scan) as well as subject-to-subject variability of cerebral perfusion, both of which are well documented even in healthy subjects.
For interpretation of PET images there exist databases of healthy controls stratified by age and gender that can support radiologists in their reporting. This project aims to build a healthy control database for cerebral perfusion, similar to the ones currently available for PET. To account for the variable within-subject variance observed in ASL data, a heteroschedastic approach will be employed. The potential clinical benefit of advanced ASL sequences developed by established collaborators in Oxford and Bremen will also be investigated. These include Hadamard-encoded acquisitions allowing mapping of regional blood arrival time in an efficient manner, and vessel-encoded ASL protocols, allowing mapping of separate vascular territories. The implementation of partial volume correction algorithms for perfusion maps to be used for clinical reporting will also be explored.
Though assessing cerebral perfusion in a completely non-invasive fashion has great potential for diagnosis and monitoring vascular problems, ASL is still rarely being used in clinical settings. Whilst the main scanner manufacturers now provide basic ASL capabilities, the typical product sequences are well behind state-of-the-art protocols. On-line, on-scanner generation of parametric perfusion maps is either non-existent or too basic to be useful. This project focuses on bringing the benefits of the recent developments in brain perfusion mapping with ASL to the clinical arena. It also aims to implement a robust and easy-to-access processing pipeline that will make the reporting of ASL acquisitions as easy as for PET images. Reaching these goals will contribute to make ASL an effective and more widely used clinical tool for a number of diseases including Alzheimer’s disease.