Health Data Platform - Bringing together data to support patient outcomes

A Newcastle BRC-funded project has demonstrated the value of high-quality datasets linking patient data.

Background

Brain diseases such as dementia affect many older people. These can be challenging for doctors to diagnose, especially in the early stages when treatment would be most effective. One reason for this is that doctors can’t examine the brain directly. Brain scanners help, but are too expensive for routine screening. These scanners can also only produce coarse images, which may not reveal early-stage disease. However, there is one part of the central nervous system which can be examined directly; the retina at the back of the eye. Advanced eye imaging technology can now scan the retina quickly and cheaply, and in microscopic detail.

The Study

The study, Health Data Platform, laid the foundations for investigating whether eye scans could reveal signs of brain diseases like Parkinson’s and Alzheimer’s. If so, this could lead to better screening techniques, which enable earlier detection of these diseases.

Retinal scans contain vast amounts of information but so far, it has proved difficult to recognise and extract the important features using standard methods. Fortunately, new computer techniques are incredibly good at spotting patterns in large amounts of data.

Principal Investigators for the study, Dr Will Innes and Professors Anya Hurlbert and Jenny Read led an interdisciplinary team spanning ophthalmology, neuroscience and computing science. This team carried out the project to establish the data platform that makes it possible to train computers to spot subtle signs of disease, which are hard for humans to detect.

This proof of principle project demonstrated the value of the vast healthcare data repository within the Newcastle Hospitals Trust. To do this, they created the platform and applied advanced Machine Learning computer algorithms to obtain deeper insights from eye imaging and clinical data already collected.

Dr Innes comments:

It is incredibly exciting to think that we can use digital methods to improve the diagnosis of progressive conditions. The management of Parkinson’s and Alzheimer’s can be a lot more efficient if the illness is detected at the earliest stage. We believe there is a way to do this through effective manipulation of data that already exists, obtained as part of routine eye care.

Of course, it is essential that patients provide consent, and that data is completely secure while it is being analysed. The important aim of this project was to set up ways for the NHS, Newcastle University and patients to work together to ensure this happens.

Outcomes

With funding from the NIHR Newcastle BRC, this study has provided the springboard for further funded projects. The team (PI Professor Anya Hurlbert, Co-I Dr Will Innes, and six co-investigators spanning disciplines of computer science, neuroscience, neurology and ophthalmology) were successful in a subsequent application for funding from the NIHR Artificial Intelligence in Health and Care Award (AI Award). This will provide financial support of almost £150,000 for OCTAHEDRON, a project that will create larger datasets of annotated OCT scans (3D optical coherence tomography eye scans) within Newcastle Hospitals and Sunderland NHS Trusts, to develop diagnostic tools using AI technologies based on retinal imaging scans. See full story here 

Next steps

OCTAHEDRON will continue until April 2022 and will support Professor Hurlbert and the interdisciplinary team to achieve the long-term goal of positioning Newcastle as a leader in medical informatics, and to enable Newcastle Hospitals to extract value from its store of raw digital healthcare information.