Launch of the London Medical Imaging & Artificial Intelligence Centre for Value-Based Healthcare

Thursday 28 February saw the official opening event for the London Medical Imaging & Artificial Intelligence Centre for Value-Based Healthcare, established as part of the UK Government’s Industrial Strategy Challenge Fund.

The new Centre will train sophisticated artificial intelligence (AI) algorithms from NHS medical images and patient data to provide tools for clinicians to speed up and improve diagnosis and care across a number of patient pathways including dementia, heart failure and cancer.

It brings together an ambitious consortium including two other universities (Imperial and Queen Mary’s University London), King’s Health Partners NHS Foundation Trust partners, Guy’s and St Thomas’, King’s College Hospital and South London and Maudsley, with Bart’s Health, multinational industry (Siemens, NVIDIA, IBM, GSK), 10 UK-based SME’s and the Health Innovation Network.

The event at County Hall saw presentations showcasing both the technical infrastructure, importance of the partnerships and the potential clinical impact. Speakers included:

  • Professor Reza Rezavi, Centre Director: Vision and Aims of the Centre
  • Professor Sebastien Ourselin, Head of the School of Biomedical Engineering & Imaging Sciences: Building our MedTech Hub at St Thomas’ Hospital
  • Dr M Jorge Cardoso, Chief Technical Officer:  A Federated Model for healthcare data access
  • Dr Gerry Carr-White: The role of AI in clinical cardiology: Heart Failure
  • Professor Daniel Rueckert: AI-enabled fetal screening
  • Professor Julia Schnabel: AI for early detection & prognosis of lung cancer
  • Dr James Teo: Stroke prevention with AI
  • Dr Mark Michalski, Executive Director of MGH & BWH Center for Clinical Data Science: The US experience of AI enabled healthcare

The Centre has a focus on transformation and value-based healthcare, and how advanced imaging and AI technologies can be used to improve the patient journey. From earlier diagnosis if there is a problem and reassurance if not, to moving quickly to a treatment which is tailored to the patient and will result in the best possible outcome. By optimising triage and targeting resources, these technologies will also allow the NHS to reduce wasted effort that is not supporting patient care, and deliver significant financial savings.

Importantly the Centre will also ensure that the technologies developed become products that can be used across the NHS and also exported internationally. To achieve this the Centre will co-locate researchers, clinicians and industry partners at its hub in St Thomas’ Hospital. Siemens Healthineers is making a substantial £6.6M investment into the UK by making the Centre their European Stratifies Medicine Hub. This along with other major investments from NVIDIA, IBM and GSK will help to leverage UK research strengths and clinical knowledge into becoming a leading industrial player in AI and healthcare.

Recipients of the Centre for Medical Engineering Pump-Priming Scheme Awards Announced

The Wellcome/ESPRC Centre for Medical Engineering (CME) has selected seven successful projects for the 2018 Pump-Priming Scheme Awards. This internal competition awards a share of up to half a million pounds aimed at highly innovative work in clinical challenges that would not easily be supported by conventional grant funding.

The projects awarded all address critical questions within the Centre’s Research Themes (Physics and Engineering, Computer Science and Chemistry) and Clinical Challenges (Cardiovascular, Neurology and Cancer), with a focus on feasibility.

Recipients of the Pump-priming Scheme awards and their focus areas are as follows:

  • Jana Hutter & Pablo Lamata (Project): ‘Detecting risks created during pregnancy (the placenta – heart axis). Pre-eclampsia affects about 1 in 25 pregnancies. It can cause severe problems for the mother such as fits, bleeding, heart problems later in life, and even death. The baby can be affected by a poorly functioning placenta, which leads to inadequate nutrition and thus to low birthweight with associated risks or even death. We aim to reduce the impact of this condition by a better understanding of the health of the mother’s heart and placenta, and how they affect each other. By working with cutting-edge imaging and computational technologies we hope to define strategies of effective surveillance and therapy planning.
  • Teresa Matias Correia (Fellowship): ‘Fully quantitative high-resolution free-breathing first-pass perfusion cardiac MRI for assessing myocardial ischemia. Quantitative first-pass perfusion cardiac magnetic resonance imaging (FPP-CMR) is an emerging safe and non-invasive test for detecting coronary heart disease, the most common type of heart disease. However, FPP-CMR requires ultra-fast scans whilst patients hold their breath, leading to a trade-off between spatial resolution and heart coverage. This project aims to develop an automatic image reconstruction strategy based on mathematical models of cardiac blood flow, for accurately quantifying myocardial perfusion from accelerated acquisitions (to collect more information) performed without patients having to hold their breath. This method will generate images with improved resolution and heart coverage and, hence, greater diagnostic accuracy
  • Colm McGinnity (Fellowship): ‘Quantification of brain stiffness via magnetic resonance elastography in refractory focal epilepsies. Epilepsy is a very common brain condition that affects roughly 500,000 people in the UK. Despite medication, around one in three people with epilepsy continue to have seizures. If doctors can find out where exactly in the brain the seizures begin, this area can sometimes be removed by surgery. This project will test a new type of brain scan to measure brain stiffness, in hope that this new type of scan can help with finding the area where seizures start.
  • Kavitha Sunassee (Project): ‘Assessment of novel particle platform technology (for tracking cancer cells and nuclear delivery of Auger Emitting Radionuclides for therapy). A viral structural protein when mixed with small single-stranded pieces of DNA (ODNs) spontaneously self-assembls into spherical particles in the test-tube. If a fluorochrome-tag is attached to the ODN, these particles can be seen using a fluorescent-microscope. When these particles are fed to cells in culture, they are gobbled up and they remain inside the cells for >8 days. These internalised particles can be triggered to disrupt, such that particle components disassemble and flood the nucleus (home of DNA). Attaching radioactivity to particle components would allow us to track tumour cells as they divide (tracking cell spread) but also deliver radioactivity specifically to the nucleus to kill cancer cells.
  • Rui Teixeira (Fellowship): ‘Establishing MRI as reproducible and motion robust measurement tool. Currently images acquired using MRI tailor signals to specific brain tissue properties but are not easily comparable across scanners. Quantitative MRI seeks to change this, however it is fraught with lack of reproducibility, vulnerability to motion and unacceptably long acquisition times. This proposal aims to optimise a detailed and reliable MRI acquisition method that is able to measure 3 brain tissue properties at high spatial resolution, all within a 15min examination, even if the subject moves. Achieving this will provide more detailed and reproducible clinical information in a shorter period of time, thereby revolutionising the current state-of-the-art.
  • Marta Varela (Fellowship): ‘Predicting the outcome of atrial fibrillation patients using CINE-MRI and convolutional neural networks. Before  undergoing  treatment,  atrial  fibrillation  (AF)  patients  often  have  a  dynamic  MRI  scan to assess how patients’ atria deform across the cardiac cycle. Atrial deformations are known to be subtly  different  in patients  with more  advanced  forms of  the  disease  compared  to  those  who respond well to treatment. Although dynamic MRIs are inspected visually by cardiologists, Artificial Intelligence  (AI)programs are  better  at classifying atrialde formations  in  these  images,  greatly improving their clinical value. This project will create, for the first time, AI programs that analyse atrial deformations from dynamic MRI scans and make personalized treatment predictions for AF patients
  • Tobias Wood (Fellowship): ‘Silent Quantitative Magnetization Transfer MR Imaging. Currently MR scans are very noisy and take a long time. This project will use a new, almost silent, very fast imaging method so that more quantitative data can be acquired in the same time and with increased patient comfort. Our project’s novel approach will allow the team to map chemical information across brain tissue, such as myelination using a method called Magnetization Transfer (MT) and the acidity level using a method called Chemical Exchange Saturation Transfer (CEST).

The CME, funded by Wellcome and EPSRC (Engineering Physical Sciences Research Council) focuses on the science and engineering of medical imaging and hosts the largest research group in this area in Europe.

Future CME funding rounds will be announced in March 2019.

In Conversation With: Professor Sandy Wells

This month the School of Biomedical Engineering and Imaging Sciences has been delighted to host Professor William M. Wells III (aka Sandy Wells), Professor of Radiology at Harvard Medical School and Brigham and Women’s Hospital, Member of the Affiliated Faculty of the Harvard-MIT Division of Health Sciences and Technology, and research scientist at the MIT Computer Science and Artificial Intelligence Laboratory.  Professor Wells and his groups have made seminal contributions in segmentation of MRI and in multi-modality registration. He was also involved in early work in intra-operative MRI and the development of the 3D Slicer software package.

Could you start by giving us an overview of your career?

I’ve been working in medical image analysis for 25 years, being one of the early adopters from the field of computer vision in 1993. Once I finished my PhD I secured a postdoctoral position at Brigham and Women’s Hospital where the field really took off. Although I think it was partly being in the right place at the right time, this allowed me to work through an exciting and transformative period in the field.

What current clinical challenge are you working on?

My primary appointment is in the MRI division of the Radiology department at the Brigham where our research focus is more on intervention than diagnostics – we do a lot of work in image-guidance, with emphasis on neurosurgery and prostate procedures.

In fact, we were among the first sites to have MR machines adopted for real-time use in surgery and we maintained that theme ever since.  As interventional MRI has proven too costly for widespread adoption we have been working on alternative methods for bringing pre-operative MRI to surgical procedures.

We’re now working with industry standard guidance systems produced by companies like BrainLab, but one of the current challenges we’re facing is improving image registration and its ability to compensate for tissue deformation. With image registration we use pre-surgical scans to identify and plan interventions and image registration is the tool we use to ensure the guidance systems link up with the data. However, when tissue is removed during procedures this changes the structure which can affect the quality of image registration.

I’ve been very focused on this challenge for a number of years, and we’re currently exploring the use of intra-operative ultrasound for estimating “brain shift” in neurosurgery.

What brings you to King’s?

Whilst a month doesn’t sound too long we are working to establish more long-term collaborations between King’s and the Brigham, hopefully supported by some joint grants. We have a long-standing history with many of the academics here and I hope we can work together to help solve registration problems on your large-scale projects such as image-guided neurosurgery.

How closely do you work with Brigham and Women’s Hospital and how important is this for your research?

The Surgical Planning Laboratory is in prime clinical space in the main thoroughfares of the radiology department, so I would say very closely. It’s been very important for our work over the years and the reason our lab has been able to stay relevant. In simple terms – we’re working on problems that the clinicians care about.

That said, I do think it’s crucially important to make that relationship even stronger and more widespread across MIT and Harvard to improve technical and clinical feedback.

I think your School has such an opportunity with its location in St Thomas’ Hospital. The senior level integration between technical and medical specialisms is really conducive to progress in medical AI, ensuring that the big data sets from the clinic can be exploited by the technical talent – I’m very interested to see how that works out for your research here.

What do you envisage the next ‘big’ technical developments will be in healthcare?  

We’ve just spoken about AI but there has also been a lot of progress in the basic technologies supporting robotics, these advances are making it much more feasible to incorporate these advances into surgical devices.

However, the two are certainly interlinked and any big breakthroughs in minimally invasive surgery will be combining AI and robotics to achieve unprecedented results. The flow of information we now have to train AI with will allow us to incorporate the genetics of patients, their tumours, and many other complex biomarkers which will personalize surgery in a way that will change its nature. We’re already beginning to use mass spectrometry in neurosurgery at the Brigham.  Biomarker feedback at the molecular level combined with the distilled knowledge from millions of patient’s data will enable new levels of precision medicine.  I think we will move beyond the current image-guided therapy to a more general information-guided surgery, ultimately with AI inference systems working alongside surgeons.

Any last thoughts for our readers? 

Whilst it’s no doubt exciting to be working through such an experimental time in AI & machine learning, we are moving towards the stage where we need to underpin this with robust intellectual foundations in academia. I’m beginning to work with Dr. Jorge Cardoso in developing a curriculum for a masters degree specialisation in Artificial Intelligence and Clinical Data Science, that will teach core foundation modules in areas such as probability theory. This is something we must begin investing in to ensure that we most effectively train the next generation of specialists in this area.

New £10m Centre in medical imaging and AI

Greg Clark, UK Secretary of State for Business, Energy and Industrial Strategy (BEIS), confirmed today (Tuesday 6 November) that UK Research and Innovation will invest £10million in The London Medical Imaging & Artificial Intelligence Centre for Value-Based Healthcare as part of the Industrial Strategy Challenge Fund.

He said: “AI has the potential to revolutionise healthcare and improve lives for the better. That’s why our modern Industrial Strategy puts pioneering technologies at the heart of our plans to build a Britain fit for the future. The innovation at these new centres will help diagnose disease earlier to give people more options when it comes to their treatment, and make reporting more efficient, freeing up time for our much-admired NHS staff time to spend on direct patient care.”

Led by King’s College London, the new Centre will train sophisticated artificial intelligence (AI) algorithms from NHS medical images and patient data, to provide tools for clinicians to speed up and improve diagnosis and care across a number of patient pathways including dementia, heart failure and cancer.

In order to achieve these results, the Centre brings together an ambitious consortium including two other universities (Imperial and Queen Mary’s University London), King’s Health Partners NHS Foundation Trust partners, Guy’s and St Thomas’, King’s College Hospital and South London and Maudsley, with Bart’s Health, multinational industry (Siemens, NVIDIA, IBM, GSK), 10 UK-based SME’s and the Health Innovation Network.

The Centre will have a focus on transformation and value-based healthcare, and how advanced imaging and AI technologies can be used to improve the patient journey. From earlier diagnosis if there is a problem and reassurance if not, to moving quickly to a treatment which is tailored to the patient and will result in the best possible outcome. By optimising triage and targeting resources, these technologies will also allow the NHS to reduce wasted effort that is not supporting patient care, and deliver significant financial savings.

Centre Director Professor Reza Razavi from King’s College London said: “The Centre will provide a fantastic opportunity to transform 12 different 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. This builds on a great on-going partnership between our researchers, clinicians and industry colleagues that will help put the UK at the forefront of developing and applying new technologies to improve healthcare.”

Importantly the Centre will also ensure that the technologies developed become products that can be used across the NHS and also exported internationally. To achieve this, the Centre, based at St Thomas’ Hospital, will co-locate researchers and clinicians from King’s College London, two other leading universities, Guy’s and St Thomas’ and three other leading NHS Trusts with staff from industry including 10 UK small and medium enterprises (SMEs). Siemens Healthineers is making a substantial £6.6M investment into the UK by making the Centre their European Stratifies Medicine Hub. This along with other major investments from NVIDIA, IBM and GSK will help to leverage UK research strengths and clinical knowledge into becoming a leading industrial player in AI and healthcare.

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

“This Centre marks a significant chapter in the future of AI-enabled NHS hospitals. We are bringing together a critical mass of industry and university partners which will allow us to share and analyse data on a scale that has not previously been possible for the NHS. The infrastructure is an essential part of building new AI tools which will benefit both patients and the whole healthcare system.”

Planned to open in early 2019, the Centre is one of a number of initiatives that will feed into longer-term plans to make the St Thomas’ Hospital campus a major MedTech Hub for the UK. The work is supported by the existing Wellcome / EPSRC Centre for Medical Engineering, the School of Biomedical Engineering & Imaging Sciences and TOHETI.

Prof Sir Robert Lechler, Executive Director of King’s Health Partners, said: “We welcome the government’s commitment to funding transformational research and innovations that will lead to better patient care. The new Centre gives us the opportunity to deliver improved outcomes for patients with a focus on driving quality and sustainability by using a value-based healthcare approach.

“As an Academic Health Sciences Centre, our purpose is to bring cutting-edge innovation into patient care. Partnership working enables us to deliver this ambition by bringing together advanced technologies such as imaging and AI, with our researchers, clinicians and industry partners, to accelerate and improve diagnosis and care.”

UK Research and Innovation Chief Executive Professor Sir Mark Walport said: “Early diagnosis of illness can greatly increase the chances of successful treatment and save lives.

“The centres announced today bring together the teams that will develop artificial intelligence tools that can analyse medical images varying from x-rays to microscopic sections from tissue biopsies. Artificial intelligence has the potential to revolutionise the speed and accuracy of medical diagnosis.”

Read the full announcement on the Department for Business, Energy and Industrial Strategy website

 

 

In Conversation With: Prof Terry Peters

This September we are delighted to welcome Terry Peters to the School on a six-month sabbatical. Terry is a Professor of Biomedical Engineering, Medical Biophysics and Medical Imaging and a Scientist at the Robarts Research Institute at Western University in London Canada. Until last month he was also the  Director of the Biomedical Imaging Research Centre (BIRC) at Western, and brings over 40 years of expertise in medical imaging to the School.

Read our interview to discover how the field has developed during his renowned career, what he will be working on at King’s, and what the next research breakthroughs in medical imaging might look like:

What made you choose the path of medical imaging as a career?

I was awarded my PhD in Electrical Engineering in 1974 from the University of Canterbury (Christchurch, New Zealand) producing the world’s first thesis on CT scanning, but it wasn’t until I went on to work for Christchurch Hospital that I got my real introduction to how medical imaging could make a difference. We convinced the hospital administration  to invest ($20,000 NZD) in the construction of a CT scanner, which went on to be used clinically for radiotherapy treatment planning. At the time, this was very novel, cutting-edge technology, and preceded the first commercial CT at the hospital.

Because of my early exposure to CT scanning, the NZ government asked me to tour the US, Canada, and Europe to characterize and compare CT scanners, finally making recommendations for their introduction to New Zealand. During one of these tours, a turning point for my career came in the form of exposure to MRI technology at the International CT Conference (Stanford, 1975). A seminal paper on MRI was presented by Paul Lauterbur, who later shared the Nobel Prize in Physiology or Medicine in 2003 with Sir Peter Mansfield, which opened up my own research to innovative new ideas. At the same meeting,  I met another New Zealander, Dr Chris Thompson, who was working at the Montreal Neurological Institute (MNI), and who invited me to spend 2 years there.

Two years became 19, and while at the MNI, I became  an active member of the Neurology & Neurosurgery, Biomedical Engineering and Medical Physics Departments  at McGill University where I explored the use of CT scanning in the brain in more depth, including 3D imaging, working  with Neurosurgeons to develop image-guided neurosurgical planning.  I was also responsible for introducing the first 0.5T “high-field” (yes, ½ Tesla magnets were considered high-field in 1984!) cryo-magnet MRI in Canada.

In 1997, the Imaging group at the Robarts Research Institute (RRI) at Western University asked me if I was interested in starting an image-guided surgery program in (the other) London in Ontario. This was where I established my lab, initially focussing on image-guided neurosurgery, then moving into other IGT disciplines such as cardiac and abdominal. This research group became the VASST (Virtual Augmentation and Simulation for Surgery and Therapy) Lab.

What have been the major research breakthroughs in medical imaging that you have seen?

A number of new technologies have disrupted the clinical practice since my first days as a medical physicist. When I first became excited about the fact that I could reconstruct images using CT principles, my radiology colleagues never thought this could replace X-Ray imaging.  The same thing happened with MRI – by this point CT was well established in clinical practice and it was difficult to convince teams to accept yet another imaging modality. (Fortunately, they eventually changed their minds!)

The rate of tech expansion has always amazed me. MRI is the result of an incredible interaction between so many disciplines. These almost magical machines would never have materialized without such an enormous multi-disciplinary collaboration between chemistry, physics, electrical engineering, computer science, radiofrequency coil design, and cryogenics. Since that point we’ve been able to progress from acquiring single slices to dynamic volumes, capturing thousands of images in minutes.

What do you predict as the next big changes we will see in the sector?

I’ve always found that new breakthroughs are often things you can’t predict – the whole evolution of CT and MRI is indicative of this. We certainly didn’t predict we would be doing live functional MRI when things were starting. It just didn’t seem possible.

Our work at Robarts, and here at the BMEIS School, is focused on minimally invasive surgery with dedicated, intelligent instruments, coupled with imaging to help us approach surgical targets more precisely and safely. I think we have tremendous opportunities to use developments in computing, robotics and artificial intelligence, along with imaging, to help us achieve these goals – and it’s absolutely essential that we work together.

A shift in the field is the way we’re using the latest technologies to develop more economic solutions to clinical practice. For example, Robarts has Canada’s only 7T scanner, but it would be impossible to expect to scan every patient who requires an MRI with this facility. However, with the advantage of the higher resolution and sensitivity offered by the 7T, we can identify characteristics of the brain which aren’t currently visible on lower field scanners, and this in turn informs us about how to improve protocols to extract similar information with the more widely available 3T machines.

Another example is the opportunity to use low footprint and affordable ultrasound, (whose quality has improved dramatically over the past couple of decades), combined with augmented reality visualization techniques, to develop low cost solutions for minimally-invasive therapies for widespread deployment in less developed countries.

What will you be focusing on during your sabbatical with the School?

I’m here to learn as much as I can. We’ve already started writing a joint grant between King’s and Western University concerned with using High Field imaging to better diagnose epilepsy by more accurately locating the signals we get from electrodes placed in the brain.

I also want to learn more about what’s going on in Deep Learning and if there’s a way we can use this approach to predict the histology of affected brain areas from MRI in epilepsy patients. At MICCAI, which took place just last week, there’s been a definite shift towards Deep Learning in medical image computing, and I’m particularly interested in how that can help the computer assisted image-guidance community. I’m planning to work closely with Prof Julia Schnabel, Chair in Computational Imaging, and her team to learn more about her work using Deep Learning to analyse MRI scans.

At Robarts, we also work a lot with cardiac interventions. I want to find out what techniques I can take back to Canada to understand heart valve pathologies using MRI imaging to help us design better therapies and treatments.

My lab also has an ongoing collaboration with King’s College Hospital at  Denmark Hill on mitral valve repair, where we are creating patient-specific valve models to help understand the impact of new therapies.

On the other hand, I also hope I can bring some expertise from Robarts’ to the table. My colleagues already have 10 years’ experience with 7T MRI, developing new pulse sequences and building customized coils in their radiofrequency labs. This expertise will be of huge benefit when setting up the new 7T facilities at St Thomas’.