The five hospitals are using Artificial Intelligence (AI) to pioneer new ways of delivering exceptional levels of patient care.
The Royal Marsden: Using AI to plan and deliver radiotherapy
Accurately identifying tumours on radiotherapy planning scans is essential for successful treatment. Machine learning (ML) is a form of artificial intelligence that creates computer algorithms to emulate human performance. This project, led by Dr Alexandra Taylor, Consultant Clinical Oncologist at The Royal Marsden, aims to develop a machine learning model to automatically plan radiotherapy scans for cervical cancer patients and investigate new approaches to account for the variation that occurs between individual clinicians and imaging systems. This could improve the treatment’s quality and improve patient outcomes, as well as build confidence in its accuracy.
Dr Taylor said, “Artificial intelligence, particularly with machine learning approaches is going to significantly change how radiotherapy treatments are planned and delivered. Thanks to generous Royal Marsden Cancer Charity funding from the Lady Garden Foundation, this project will ensure The Royal Marsden can lead the way in developing innovative new radiotherapy approaches for patients with cervical cancer. From generating radiotherapy plans and structures in minutes not hours, to developing highly personalised and targeted treatments, the potential impact on patients is huge.”
Moorfields Private Eye Hospital: Using AI to detect early changes in the retina to diagnose Parkinson’s disease
Moorfields Eye Hospital and UCL Institute of Ophthalmology (IoO) researchers have developed an artificial intelligence (AI) system that has the potential to identify medical conditions up to seven years earlier than clinical presentation. The study, published in Neurology®, the medical journal of the American Academy of Neurology, identified markers of Parkinson’s in eye scans with the help of AI. The use of data from eye scans has previously revealed signs of other neurodegenerative conditions, including Alzheimer’s, multiple sclerosis and, most recently, schizophrenia, in an emerging and exciting field of research referred to as “oculonomics”.
Louisa Wickham, Moorfields’ medical director, explained that “Increasing imaging across a wider population will have a huge impact on public health in the future, and will eventually lead to predictive analysis. OCT scans are more scalable, non-invasive, lower cost and quicker than brain scans for this purpose.”
Royal National Orthopaedic Hospital: Using AI to assess bone health
At the Royal National Orthopaedic Hospital (RNOH), we are spearheading a transformative effort in bone health assessment using AI. This initiative blends advanced technology with clinical expertise to redefine screening, diagnosis, treatment, and prediction of outcomes for metabolic bone disease.
A unique collaboration with a company spun out from RNOH has led to the development of OsteoSight, software that uses AI to measure Bone Mineral Density (BMD) from conventional digital X-rays. Osteoporosis, which can lead to debilitating fragility fractures, impacts three and a half million people in the UK; however, fewer than 25% of people with the disease are diagnosed due to the expense of deploying a dedicated screening program. By applying OsteoSight opportunistically to all X-rays that are already being taken for any reason, we aim to increase levels of diagnosis cost-effectively, and prevent these fractures occurring. This could prevent 25,000 fragility fractures a year.
OsteoSight uses machine learning and classical image analysis to automatically identify suitable images and estimate BMD. This is challenging due to factors like tissue type, shape, volume, attenuation, metal implants, and labelling errors. The model has been trained on our unique dataset, ensuring robustness across various patient populations and imaging machines. All images are de-identified.
Drs Richard Keen and Jude Bubbear will be leading a prospective cross-sectional study to validate the accuracy and efficacy of OsteoSight in a real-world setting, inviting outpatients at RNOH to have their X-rays ‘enhanced’ to assess BMD, and comparing these results with the current gold-standard in diagnosing osteoporosis. We are optimistic that a positive result from this trial will lead to the deployment of this innovation into the NHS, and will seek to identify downstream opportunities, such as targeted physiotherapy at MSK Community Hubs, to increase its impact.
The OsteoSight project is just one of several innovative research programs using AI at RNOH.
Great Ormond Street Hospital for Children: Harnessing the power of data to build and test AI models
The Data Research Innovation and Virtual Environments (DRIVE) unit at GOSH has focused on development of a secure technical environment and the digital tools themselves to carry out proof-of-concept testing of AI tools. Researchers at GOSH are also interested in developing AI and some are using GOSH data to develop AI tools.
Digital infrastructure developed by DRIVE underpins our ability to test and utilise AI. The Digital Research Environment (DRE) at GOSH DRIVE allows Electronic Health Record (EHR) data to be extracted into an anonymous and secure environment. This can facilitate building of AI models. AI tools developed outside of GOSH could be brought in on this platform for further development and testing.
Patient data held in the EHR is private and secure. There are strict regulations and processes required before it can be used for secondary purposes in research and innovation. It is very important to GOSH and our academic and industry collaborators that data is kept safe, and patient’s anonymity is protected.