The benefits of AI in our hospitals

A year on from the London Medical Imaging & AI Centre for Value-Based Healthcare opening, we look at some highlights from the past twelve months.

ai in our hospitals

King's Health Partners is part of a new centre for medical imaging and AI funded by UK Research and Innovation. The Centre brings together academic, NHS and industry partners, led by King’s College London, to train algorithms from NHS medical images to create new healthcare tools.

We are always looking to the future and using the very latest technologies, including AI, in our plans to join up care, research and education at King’s Health Partners. Data is one of our most important assets and we have unique strengths in using informatics to improve care, public health and the efficiency of our health system.

Cutting patient reporting times by 72%

In early 2019, research from King's College London, supported by the National Institute for Health Research (NIHR) Guy’s and St Thomas’ Biomedical Research Centre, showed that AI can dramatically reduce the time it takes for abnormal chest X-rays to receive expert radiologist opinion.

Using a dataset of 500,000 anonymised adult chest X-rays from patients at Guy’s and St Thomas’ NHS Foundation Trust, the researchers trained an algorithm to recognise abnormalities in chest X-rays.

When the team tested the AI system in simulation, it found it could cut the average reporting time by almost three quarters, from 11 days to less than three. That’s a reduction of over 70%. 

Growing partnerships

In May 2019, the AI Centre joined forces with its technology partner, NVIDA, to build the UK’s first AI platform (NVIDIA® DGX-2TM).

The NVIDIA® DGX-2TM AI system classifies stroke and neurological impairments, determines the underlying causes of cancers, as well as recommending the best treatment for patients. With the ability to develop algorithms using on site data from individual hospitals, this was an NHS first.

Professor Sebastien Ourselin, Head of the School of Biomedical Engineering & Imaging Sciences (IoPPN) at King’s College London, said:

The NVIDIA DGX-2 AI system’s large memory and massive computing power make it possible for us to tackle training of large, 3D datasets in minutes instead of days while keeping the data secure on the premises of the hospital.

King’s College London also use with their own imaging technologies, such as NiftyNet, as well as those from other partners, including Kheiron Medical, Mirada and Scan Computers.

By December 2019, the AI Centre broadened its wealth of partnerships by working with Owkin, a company that develops AI technologies to advance medical research, to deliver federated learning in the healthcare and life sciences sector.

Federated learning removes the need to host data on one central server. It happens by exposing AI algorithms to large amounts of data, located at different sites and therefore giving it the experience it needs in broader, high-quality datasets, to store data non-centrally. This approach is more privacy-compliant, because it means avoiding sharing personally identifiable data, which is particularly relevant within health organisations like the NHS.

Commenting on industry partnership, Professor Sebastien Ourselin said:

We are very pleased to welcome Owkin into our consortium of partners. Owkin are thought leaders in the new field of federated learning and will make an important contribution to the AI Centre by providing the software layer that allows models to be built, orchestrated, secured and traced as they travel between our hospital and university partners.

Late last year, the informatics tool CogStack, developed by AI Centre partners the Centre for Translational Informatics, part of the South London and Maudsley Biomedical Research Centre and King’s College Hospital NHS Foundation Trust, featured in an NHSX report on AI, and more recently, Matt Hancock's keynote address to the Healthtech Alliance.

The Secretary of State for Health and Social Care said:

The Cogstack AI can perform manual coding and data collection tasks in a tenth of the time that it takes a human analyst.
It’s a clear example of the latest AI helping us fix the basics, because once you’ve coded up and digitised your patient records, you can start to solve fundamental problems, like how to share those records across different parts of the NHS.

The future of informatics

Looking forward, the AI Centre continues to use informatics to improve care and efficiency in clinical practice.

Professor Reza Razavi, Director of Research at King’s Health Partners and the AI Centre Director, said:

“The Centre will provide a fantastic opportunity to transform patient pathways by using advanced imaging and AI and help make the products that will substantially improve the experience for our patients and their clinical outcomes. It will also allow us to better utilise the resources within the NHS.”

Stay up to date on how King’s Health Partners are using informatics to join up care, research and education, by visiting our website.