Lead: Professor Thomas Nichols
Deep phenotyping is the precise and comprehensive categorisation or analysis of a person’s observable traits. We have captured a lot of data that have untapped potential, and we want to use them to augment existing methods, deliver more personalised medicine and ultimately produce better outcomes for patients.
For example, ankylosing spondylitis, a long-term condition in which the spine and other areas of the body become inflamed, is a disease where it takes a long time to predict prognosis, which is difficult for patients.
Currently, pictures are taken every six months, and these are interpreted by individuals, which can often be subjective. If those same images can be read using machine learning, it may result in quicker diagnosis – and identify those patients whose condition is likely to progress quicker.
Our researchers are building a regional platform that contains very rich data that will be able to support such research across a range of diseases.
Working with the Thames Valley Cancer Alliance, GE Healthcare and Roche Diagnostics, we are introducing research multi-disciplinary team to augment existing care with the integrated analysis of data from cancer imaging, blood biomarkers and digital pathology. This will improve cancer diagnosis and outcomes.
We are collaborating with the regional Inherited Cardiac Conditions service to incorporate data from cardiac ultrasound and routine digital echocardiograms to deliver a more complete picture of the patient’s health, so reducing risk and mortality.
Similarly, with BRC’s Imaging Theme, we are working with Oxford Health BRC to address the under-utilisation of image data in mental health care. This will involve using brain imaging and sensor data, and making full use of the Wellcome Centre for Integrative Neuroimaging’s FSL software library.
We are also continuing to develop the use of AI in care settings and patient populations throughout the country through our National Consortium of Intelligent Medical Imaging (NCIMI) network. This work will accelerate translational research and improve diagnosis and treatment.