Cerebral Blood Flow Imaging - Towards an Efficient, Automated Assay of Ongoing Pain and its Treatment

Effective pain relief remains a significant challenge for medicine. Four out of five people report considerable (acute) pain after surgery. One in five Europeans suffer from chronic pain. Unmanageable pain impacts upon peoples' lives, their carers and families, and is a massive economic burden. We need to find new treatments for managing pain. Considerable time and effort is spent searching for new treatments, but the conventional way of working out if therapies are effective (efficacy) relies simply on asking people. Descriptions of pain vary a lot between people, meaning that sensitive, reliable ways of measuring pain have been difficult to come by. Developing a new treatment requires hundreds of patients, takes time and is expensive; up to fifteen years and a billion US dollars. Sometimes a lot of this money is spent before realizing treatments don't work or have too many side-effects. We must develop new, sensitive tools that give early signals of efficacy. These tools need to work quickly, in small numbers of people, to reduce costs and to focus earlier on the best candidates for new treatments.

This research uses a type of MRI scanning called arterial spin labeling (ASL), to measure small changes in blood circulation in the brain (cerebral blood flow-CBF). We have shown that ASL can detect CBF changes that occur when people experience pain, both after surgery and as a result of disease, such as osteoarthritis. These patterns of CBF changes are related to how the brain processes pain, are quantifiable and can be distinguished from those in people without pain, using less than 20 subjects. They tell us about how the brain represents 'being' in pain. Using mathematics known as 'machine learning' (ML), we can predict, solely from CBF patterns, about how likely it is that a person is in pain. Importantly these methods mean that we don't have to rely on what the patient is telling us.

We aim to demonstrate ASL is a useful, efficient tool for assessing new painkillers. We will examine pain experienced following wisdom teeth surgery. Many people with no other health problems, taking no other medicines, suddenly experience recurrent toothache, resolved by having their wisdom teeth extracted. Often when a wisdom tooth on one side of the mouth is removed, the other one must also be removed. This means that we can look at how reliable CBF patterns are between surgeries, or examine the effects of a painkiller, compared to a placebo. The study of pain following wisdom tooth surgery is ideal for research; it is well-controlled using common painkillers, eg, paracetamol. We know paracetamol works, but we don't know how it changes the representation of pain in the brain- 'analgesia'.

First, we will verify if ASL can reliably detect CBF changes representing pain, by comparing the effects of extracting one tooth to the other, using a statistic called the intra-class correlation co-efficient. Next, we will assess the sensitivity of ASL to detect the effects of paracetamol, compared to placebo, before and after wisdom tooth surgery. Looking at the effects of paracetamol while pain-free tells us if the drug changes CBF independently of any pain-killing quality. Assessing CBF changes following surgery will demonstrate the effects of analgesia. Using ML, we can make predictions about how likely it is the analgesia is working, compared to placebo. These techniques tell us where in the brain we know a good pain-killing drug is working, and the sensitivity of perfusion MRI to detect it. With this information, we can test new drugs and make rapid, cost-effective predictions about the likelihood that they will be effective.

We envisage the technology will be useful for assessing not just drugs, but other therapies, in both acute and chronic pain conditions. The technique may also help assessing people unable to verbalise if they are in pain, for example, very young individuals and those in vegetative states.

Despite our in depth knowledge of the basic mechanisms underlying pain, we still struggle to develop effective new analgesics. This is largely because the development of novel therapies and a clearer understanding of how we process pain are almost wholly reliant on self-reported measures that are inherently subjective. We will address these existing limitations, building upon a neuroimaging protocol recently developed in this laboratory that has demonstrated the central representation of ongoing pain, in both acute and persistent pain states, independently of patient self-report. First, we will validate the robustness of this quantitative, cerebral blood flow (CBF) assay in a new patient cohort, using these new, independent data to train our recently developed, automated machine learning algorithms. These algorithms will discriminate between individuals in pain-free, pre-surgical states, compared to the ongoing pain experienced following third molar surgery. In a second study, we will test this 'pain discrimination' algorithm in the presence and absence of a widely used, generic analgesic, to provide probabilistic discrimination between post-surgical pain response and treatment with analgesia. These methods will enable us to visualise the central effects of both pain and drug uptake as well as their interaction (i.e. analgesia). While focussed here on examining analgesic response in acute post-surgical pain, these methods should translate into the study and treatment of persistent pain. We envisage the development of shorter duration, reduced cost, analgesic trials requiring fewer patients, facilitating more rapid "go/no-go" decision-making.

Impact Summary

Who will benefit from this research and how?

Following successful completion of this project we will have developed a robust new way of assessing pain and visualizing the effect of existing and potential new classes of painkillers. We envisage that a wide range of patient groups throughout the lifespan could potentially benefit from this research and those patients suffering from chronic conditions who are not currently gaining effective pain relief will be the most immediate beneficiaries. Therefore this technology should help to enhance the quality of life, health and well being of pain sufferers. 7.8 million people have been living with moderate or severe chronic pain in the UK for more than 6 months(1). At least one person in a third of all households is thought to be in pain at any one time. Repeated surveys indicate that numbers are rising. The most common cause is osteoarthritis/arthritis (40%), with 90% of all calls to the Arthritis Care's helpline over the past 12 months related to pain. Other conditions that cause chronic pain include multiple sclerosis, cancer, diabetes, infections and injuries and back pain is the commonest site for chronic pain, with 1.6m UK adults developing chronic back pain each year. At any one time, a quarter of people over 40 are suffering from knee pain, and in around half of them the pain is disabling. However despite the number of people affected by chronic pain, it is not currently seen as a health priority and despite the national prevalence of chronic pain running at 13%, healthcare professional training on chronic pain is less than 1% of the medical school curriculum. To summarize, the impact of pain on individuals overall is immense with over half also developing symptoms of depression and 19% of people with chronic pain eventually losing their jobs. It has been estimated that the cost to the economy is £12.3 billion per year for back pain alone (2,3).

We also envisage that the commercial, private sector will benefit from this research as both pharmaceutical and biotechnology companies will use this technology to improve their understanding of the neurobiology of pain whilst also allowing us to visualize the brain penetration, optimal dose and distribution of the putative new medicine. We firmly believe that this methodology will speed up clinical trials and add value to the interpretation of the analgesic effect whilst also allowing scope for development of secondary indications based on the spatial fingerprint of the drug's effect. This approach may also lead to more rapid "no go" decision-making where the site and mechanism of action is no better than existing, generic medications. Due to the immense current pressure to develop less "me too" compounds with unprecedented mechanism of action, the need for a robust, rapid assay in order to triage these more pioneering approaches has never been greater. Combining such methods with our recently developed nationwide network of clinical research facilities equipped with MR imaging equipment will ensure that pivotal, early stage clinical trials will be retained within the UK. The ability of such technology to stratify potential responders and non-responders may contribute to a more personalized medicine approach for such individuals.

Given sufficient support to take this technology forward in a future MRC Developmental Clinical Study (DCS), we firmly believe that we will be in a position to appraise new medicines alongside gold standard analgesics for a suite of on-going pain conditions within 3 years from completion of our first MRC DPFS project.

(1) Breivik H, Collett B, et al. (2006). "Survey of chronic pain in Europe: prevalence, impact on daily life, and treatment." Eur J Pain.
(2) Smith et al. (1999) Chronic Pain in Primary Care. Family Practice 16 (5) 475-482.
(3) Breaking through the Barrier', Chief Medical Officer 2008 Annual Report, March 2009.