Research Commentary

A take on functional imaging

BOLD measurement: BOLD fMRI is a powerful technique to detect brain dysfunction, using changes in tissue oxygenation. However, it remains a qualitative approach to assess normal and abnormal brain functions. BOLD is a complicated signal that assumes intact neuro-vascular and neuro-metabolic coupling. There is a complex interaction between cerebral blood flow (CBF), cerebral blood volume (CBV), oxygen metabolism, and neurotransmission, which elicits a BOLD signal change. While a change in CBF will induce a BOLD signal change, a change in CBV may not necessarily do so. It is therefore, essential to isolate the contributors to the BOLD signal. I believe that in any disease, functional (BOLD, CBF, CBV) changes precede tissue loss. Neuronal imbalances, including synaptic losses and altered neuronal activities, are likely to be responsible for altered perfusion (CBF and CBV) but do not entail tissue loss detection with current imaging methods. Continued imbalances and persistent perfusion impairments, however, will eventually cause significant, detectable neuronal cell death. Therefore, the detection of perfusion abnormalities is key to understanding the BOLD signal differences observed in disease as well as for preventing irreversible damage that follows. My motivation in measuring CBF and CBV separately from the BOLD signal is to detect the first signs of brain dysfunction.

CBF measurement:  CBF is typically measured using arterial spin labeling (ASL) approaches and serves as an excellent surrogate for cerebral metabolism. It has been shown to correlate strongly with glucose metabolism and is the prime driver of the BOLD signal.  There is growing evidence the CBF abnormalities are linked to cognitive deficits and may be  early indicators of neuropathology. I used CBF measurements extensively in my research/

CBV measurement: One approach to measure CBV without invasive contrasts is the inflow-Vascular Space Occupancy (iVASO) approach. Autoregulatory mechanisms to maintain CBF and functional hyperemia (activity dependent CBF increases) are related to relaxation of smooth muscle surrounding the arterioles, making arterial CBV (aCBV) a valuable parameter to measure. I developed a near whole-brain, Inflow-VASO (iVASO) method to measure aCBV using arterial blood water as an endogenous contrast. CBV measurement is the most under-developed technique in functional imaging, and iVASO can potentially remedy this problem.

The BOLD signal is confounded by the presence of cardiac and respiratory pulsations. BOLD signal varies with echo time (TE) depending on intra/extravascular sensitivity. Shorter echo times provide high signal-to-noise ratio (SNR) but contain mainly contributions from large vessels and high physiological noise while longer echo times contain mainly microvascular contributions with low physiological noise but provide low SNR. Fast acquisitions can often minimise physiological confounds but are limited by imaging resolution, volume coverage, and gradient switching speeds. I am also interested in developing processing pipelines to parse the composite signals and extract neuronal signals.  

ASL: Ready for prime time?

This is a recap of the ISMRM 2018 debate of whether ASL should replace contrast-based perfusion imaging (DSC/DCE) in the clinic. 

The FOR argument: ASL is non-invasive and has been extremely useful in evaluating stroke. ASL uses blood water protons as an endogenous contrast which can freely cross blood brain barrier, something the Gadolinium contrast agents in DSC/DCE cannot do (assumed to not do). More interestingly, it is very useful in differentiating gliomas (high perfusion) from metastasis (not as high). There were a lot of good examples illustrating this point. DSC/DCE are not quantitative. Gadolinium deposition in the brain is now proven. DSC/DCE are not suitable in individuals with poor kidney function, individuals with fragile veins, and in pediatrics. No such limitations exist for ASL.

The AGAINST argument:  Most clinical trials have used DSC/DCE (none with ASL). They are  readily available in the clinic and easy to implement. No special sequences are needed (low cost). Although Gadolinium deposits in the brain, there is not enough evidence about toxicity, except for the rare nephrogenic systemic fibrosis.

My take: ASL has had to prove itself worthy for clinical use for more than a decade. ASL is exceptionally good at evaluating cerebral perfusion. It is very easy to use in the clinic and most vendors are making it into a product sequence, which will be available readily on the scanners very soon. Regardless of whether Gadolinium deposition is toxic or not, a non-invasive approach has to increase patient compliance. It correlates very well with DSC/DCE measures. It also correlates very well with PET measures. It can be used for far more than severe pathologies like stroke and tumor. Especially, as a dementia researcher, I see ASL as a way to evaluate perfusion patterns and assess risk of dementia. It can be used a routine screening tool and is infinitely repeatable (a point made by the FOR debate as well) in all individuals.  There has been no clinical trial that compares both DSC/DCE and ASL simultaneously to understand the benefits of ASL. There is only a one-sided proof that DSC/DCE is useful marker in stroke. 

But, ASL needs planning and training for accurate measurements. ASL data needs to be processed and currently quantitation, although not complicated, is not available on the commercial vendor platforms. Furthermore, ASL-based perfusion in the cerebellum, brain stem, and the rest of the body still needs to undergo a lot of development.