London Medical Imaging & Artificial Intelligence Centre for Value-Based Healthcare
"The London Medical Imaging & Artificial Intelligence Centre for Value-Based Healthcare will improve NHS patient care and health outcomes, reduce healthcare costs and support the growth of companies, supporting the economy. It will do this by applying artificial intelligence technologies to medical imaging (for example MRI scans, CT scans PET scans and ultrasound). Artificial intelligence will enable faster and earlier diagnosis, automation of reporting, improved patient screening for disease, and identification of the best treatment for each person.
We will create a powerful, dynamic Centre bringing together industry, small and medium sized enterprises (SMEs), world-leading researchers, healthcare professionals covering many areas of practice, and experts in data science/governance. The Centre is a collaboration between three excellent universities (King's, Imperial and QMUL), four leading NHS Trusts (Guy's & St Thomas', King's College Hospital, South London & the Maudsley and Barts Health), multinational industry (Siemens, NVIDIA, IBM, GSK), 10 UK-based SMEs and the Health Innovation Network. The Centre will have a physical hub embedded in St Thomas' Hospital, in the heart of one of the UK's top-performing NHS Trusts with specialist services and one of the UK's largest critical care units. It will be underpinned by the existing Wellcome/EPSRC Centre for Medical Engineering, a flagship investment in medical imaging.
The Centre will deliver well-governed, controlled access to high-quality NHS imaging and patient data for academic researchers, SMEs and industry partners. This will be done while preserving patient privacy as the first requirement. We will add dedicated expertise in health economics & statistics, care pathway design and clinical implementation to create an environment where products can be created and tested. Recognising this is a new development in healthcare, patients, the wider public and policy makers will all have opportunities to input and shape priorities for the Centre.
The Centre will drive progress through a selection of 12 exemplar projects, specifically chosen, with public input, to illustrate the breadth of opportunity-- covering early life (fetal diagnosis) to old age (dementia), various organ systems including heart, brain and lungs, and diseases such as heart failure, headache, congenital conditions and cancer.
In addition to the direct benefits of the Centre, this activity will act as a beacon to attract multinational companies, venture capital investment and AI talent from across the world, creating jobs, broader economic benefit and contributing to the UK's prosperity."
Cloud-based high-performance computing for real-time medical 3D roadmapping
GRD Development of Prototype
Hospitals worldwide are moving their storage of medical imaging data to the cloud. As well
as connecting to large storage, the cloud can connect small computers to remote massive
computing resources, effectively turning them into supercomputers. Cydar Ltd, a spin-out
from Kings College London and Guy’s and St Thomas’ NHS Trust, is developing a novel
class of clinical applications that use this potential of the cloud to deliver high-performance
computing into operating rooms to improve the outcomes for patients undergoing imageguided
surgery.
Cydar’s first product will be a ‘satnav for surgeons’ for use in X-ray guided keyhole surgical
procedures. CT scans are a common diagnostic test that contain 3D information and show the
soft tissues, but they are not useful for live guidance due to high radiation. Instead, real-time Xray
guidance is used in many types of keyhole surgery. While X-rays are good at showing
bones and the position of surgical instruments, they do not show soft tissues well and they
produce flat (2D) images that superimpose all the 3D anatomical features.
Cydar has developed software algorithms and high-performance computing technology that
detects anatomical information present in live X-ray images and matches it to the preoperative
CT scan in order to determine the patient’s exact position. This match allows a 3D
image of the relevant soft tissues (taken from their CT scan) to be accurately overlaid- giving
the surgeon a 3D ‘roadmap’. The core technology has been tested and validated in a proof-ofconcept
system in 130 operations.
The project will develop the current proof-of-concept technology into a cloud highperformance
computing prototype, in order to boost performance and test the user experience
in hospitals. This prototype stage will determine performance needs, inform