A Novel Artificial Intelligence Powered Neuroimaging Biomarker for Chronic Pain.

Lead Research Organisation: University of Sheffield
Department Name: Oncology and Metabolism

Abstract

SCALE OF THE PROBLEM: Chronic pain is a major unmet global health challenge. One in ten adults over the age of 30 suffers from chronic nerve pain (neuropathic pain, NeuP) that arises from injury to sensory nervous system. Worryingly, this is expected to increase due to aging, rising cases of diabetes and improved cancer survival. NeuP can have a negative impact on patients, their families and wider society. Patients typically report constant burning, aching or 'electric-shock' type pains in their feet, legs and hands. As the pain is felt every day, patients may have difficulty doing simple daily activities such as walking to the shop or socialising with friends. This results in a poor quality of life and depression, with one in six of people rating their quality of life as 'worse than death'. Unfortunately, current medications provide only partial benefit in around half of all patients, with many enduring inadequate pain relief and unwanted side effects. Over the past 25 years, there has been a lack of new drugs that are more effective than the ones currently in use for treating NeuP. One possible reason is that there are many different sub-types of NeuP even when caused by the same condition e.g. diabetes. Treatment response is very individual , with no good way to predict who will benefit. This often results in negative outcomes in drug trials especially when conducted on a diverse group of patients. Working with patients and industry partner AstraZeneca, our goal is to develop new biomarkers for NeuP to improve the success of future drug development programmes.

PATIENT INVOLVEMENT: Our patients with diabetes and NeuP provided input to develop a study question that 'matters most to patients'. We discussed the studies that have been conducted, how this proposal takes the next step towards validating a biomarker which would translate to improvements in clinical care.

SOLUTION: In a 'proof-of-concept' study funded by NIHR Efficacy Mechanisms Evaluation (129921), we used artificial intelligence (AI) to develop a biomarker using brain magnetic resonance imaging that accurately predicts treatment response. However, these studies have been performed in a single centre and exclusively in patients with NeuP caused by diabetes. We now want to conduct a large, multicentre study to confirm the effectiveness of our model and ensure the findings are generalisable.

APPROACH: Test our model on one of the largest NeuP neuroimaging datasets in the country comprising people with NeuP from multiple causes (diabetes and post-chemotherapy) and from different centres (Oxford and Dundee). To achieve this, we will collaborate with the PAINSTORM Advance Pain Discovery Platform [https://tinyurl.com/yeuwt8y8] - a world leading, inter-disciplinary group of clinicians (Profs Bennett, Tesfaye, Colvin and Steele) and scientists (Profs Wild, Lu, Dr Sergedahl), the pharmaceutical industry (AstraZeneca) and people living with NeuP whose focus is to study a large group of people with NeuP using several innovative technologies including brain imaging. This established partnership with AstraZeneca will bring in the ideas and needs of industry to ensure our biomarker is 'fit-for-purpose' not only for the industry but also for clinicians and patients.

EXPECTED OUTCOMES AND LEGACY: At the conclusion of this proposal we will have an effective, objective biomarker for NeuP that can be widely used. We will have established an open access online platform to maximize future collaborations. AstraZeneca are conducting a study in NeuP caused by diabetes and will be well positioned to perform 'real-world' feasibility testing of the biomarker. If successful, the biomarker will be incorporated in future clinical trials of NeuP medications.

Publications

10 25 50
 
Description What were the most significant achievements from the award?
We have completed research activities related to the first work package. This includes delivery of an optimised machine learning model, cross-site harmonisation completed and successful piloting of external validation set-up and procedures.

To what extent were the award objectives met? If you can, briefly explain why any key objectives were not met.
Objectives and milestones of Work Package 1 have been met.

How might the findings be taken forward and by whom?
The next step (Work Package 2) will involve analysis of datasets from Oxford University. This will provide multi-site, single condition generalisability of the neuroimaging based AI model for neuropathic pain.
Exploitation Route The award is still active
Sectors Pharmaceuticals and Medical Biotechnology

 
Description Industry perspective of neuropathic pain biomarker development - Partnership with AstraZeneca 
Organisation AstraZeneca
Department Research and Development AstraZeneca
Country United Kingdom 
Sector Private 
PI Contribution Meeting organised at AZ head offices to present outcomes of Work Package 1 including a presentation from industry leading experts on the role of biomarkers and use of AI in drug development. This will inform the development of AI neuroimaging biomarker for neuropathic pain. Exploring new avenues for further collaboration
Collaborator Contribution Hosting of the meeting Organisation of industry experts presenting at this meeting Exploring future testing of the biomarker in a 'real-world' setting
Impact Summary of meeting minutes will be provided
Start Year 2023
 
Description Patient group workshop (PPI/E) 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Patients, carers and/or patient groups
Results and Impact PPI/E panel helpduring the early stages of developing the study approach and concept
We reviewed the progress made to date and future work planned
Year(s) Of Engagement Activity 2023