Determining cerebrovascular reactivity from the pupil flash response

For most neurological disorders a diagnosis and intervention is only possible once significant progression has occurred and symptoms are present by which time the prognosis may be poor. It would be preferable to be able to make an early and accurate identification of those people who would be likely to develop such a disease later in life, allowing sufficiently early intervention to prevent it.
The premise is that with the identification of reliable biomarkers of conditions such as Alzheimer's disease, multiple sclerosis, Parkinson's disease and stroke, early identification of individuals predisposed to developing these conditions could be identified from regular screening within routine healthcare activities. The identification of a suitable biomarker has been the focus of much research.

One feature of these diseases which has the indications of being a good biomarker is a measure of how good the blood vessels in the brain are at responding to changes in demand or in response to an external stimulus. This effect is called the cerebrovascular reactivity or CVR. CVR is known to be impaired in the majority of brain diseases, and appears to be one of the earliest detectable symptoms that something is wrong. CVR is however quite complex and difficult to measure, requiring specialist, expensive equipment, and so it has not been widely studied in clinical trials of diseases or treatments. It has also been suggested that there are two different forms of CVR dysfunction, one due to a person's intrinsic biology and another due to their lifestyle, with each requiring different treatments.

The impact of lifestyle on CVR can be estimated from simple questions and physiological measurements but there is currently no simple means of determining the level of intrinsic CVR function. Because of this there is the potential that trials of new treatments could target only one cause but include patients with both types (intrinsic and lifestyle) and therefore be ineffective in a majority of subjects and be incorrectly deemed to be of no use. Being able to readily and cost-effectively determine which of these causes is operative would be highly beneficial to both research and, ultimately, clinical environments.

The hypothesis of this project is that intrinsic CVR dysfunction is caused by a general smooth muscle disorder; as smooth muscle is wrapped around blood vessels to control the flow. One of the few other places that smooth muscle occurs in the body is controlling the iris of the eye. It has been shown that groups known to have impaired CVR also tend to have a higher risk of developing neurological disorders, and the same groups have also been shown to have an impaired response of the pupil to a brief flash of light. As the pupil flash response (PFR) is potentially very cheap, quick and easy to assess it would make an excellent means of testing for intrinsic smooth muscle impairments as an indicator of impaired CVR.

By determining a range of simple physiological measurements for subjects, along with their responses to a lifestyle questionnaire and a measurement of the pupil flash response, the necessary data could be obtained. By then applying analytical machine learning techniques to the results, we propose that this will allow the development of a protocol to enable an accurate assessment of CVR, and its likely type, to be determined. This measurement could then form part of a risk assessment for a host of neurological disorders and enable early interventions to be implemented or discovered.

The potential impact of this research is wide ranging. In the short term, clinical trials would benefit as they would have the ability to easily and quickly stream participants according to the nature of their CVR impairments in order to investigate the relationship between CVR and cognitive decline and other neurological conditions. As the beta amyloid hypothesis has failed to produce any tangible results for dementia suffers it is time to consider alternative theories of the causes of these diseases, however this will require carefully designed studies to ensure the different forms of CVR are appropriately considered. Existing methods of measuring CVR in a way that can distinguish between the types is not feasible for large-scale trials.

If CVR is shown to be an important biomarker, or even a potential treatment target, then a pupillometry-based measure of CVR would enable the creation of a commercial device for this purpose. There are a small number of clinical pupillometers on the market, but these are designed for measuring changes over the short term to monitor brain injury patients in the emergency room and are not suitable for this purpose. The Department of Engineering Science has a world-class record of creating spin-out companies, particularly in the biomedical sector and thus is ideally placed to be able to facilitate the creation of such a device. Such a device could potentially be used in future clinical trials into a range of neurological conditions, as well as being available in clinical settings such as a GP surgery. Testing a patient's PFR could become as commonplace as measuring their blood pressure.

As it is not yet known which if any aspect of the PFR correlates sufficiently well with CVR for diagnostic purposes it is not possible to predict what the outcomes will be, however a best-case scenario would be that some useful data can be obtained from very simplistic measures of PFR in combination with some basic physiological and lifestyle data (age, height, weight, gender, etc). If a smartphone app can be created to collect this data then it creates the possibility for a huge dataset to be collected through people downloading the free app. This data could then be made freely available researchers around the world.

The evidence to support the hypothesis is mainly through anecdotal correlation, however the only way to test it is with a study of this scale. Without sufficient variation in CVR and PFR it is impossible to test for a genuine relationship between the two. However, if there is one, and it can be used to diagnose a smooth muscle disorder that would impair CVR then it would be a revolution in the field.