###Next-generation inflammation imaging combining computational modelling and machine learning
Project ID: 2228bd1073 (You will need this ID for your application)
Research Theme: Healthcare Technologies
UCL Lead department: Division of Medicine
Lead Supervisor: Timothy Bray
Project Summary:
This project aims to develop a new approach to imaging inflammation, harnessing recently advances in computational modelling and machine learning to address the key gaps in our ability to diagnose and phenotype inflammatory diseases.
Inflammation – the body’s response to harm and danger signals, which can occur inappropriately to cause pathology – is a fundamental process in human disease and causes a massive global burden of ill-health across all organ systems and in patients of all ages. Inflammatory diseases affect hundreds of millions of individuals globally, with a cost to health services and national economies of many billion pounds. Unfortunately, in clinical practice, our ability to identify and treat these diseases effectively is limited by weaknesses in existing imaging methods, which offer only indirect assessments of inflammation and are interpreted in a qualitative fashion.
To address this unmet need, this project will develop a set of interlinked computational methods, with deep learning as the core technology, aiming to provide detailed, quantitative imaging assessments of inflammation. These tools will focus on (i) the generation of biologically-relevant spatial maps describing key pathophysiological processes in inflammation (ii) the extraction of biologically-meaningful summary metrics, describing the burden and type of disease in each patient and (iii) precise measurement of changes in disease over time.
These tools will allow us to quantitatively phenotype inflammation in individual patients, and thus enable therapy to be tailored to individual patients for maximum benefit. The techniques will be evaluated in a unique cohort of patients with axial spondyloarthritis (axSpA), forming the first step towards development of a tool capable of comprehensive inflammation assessment in clinical care. If successful, the methods developed will be rapidly introduced into the large (n=800) multisite UK IDEAL study in axSpA, offering a test-bed for translation into clinical care and, ultimately, introduction across the NHS.