27 ‘phase 1 and 2’ AI projects have joined the NHS AI Lab through its AI in Health and Care Award programme.
Phase 1 of the programme intends to support early stage concepts and prototypes with up to £150,000 in funding over 6-12 months. Phase 2 is aimed to develop and evaluate prototypes and generate early clinical safety and efficacy data.
The programme has 4 phases, in which 42 suppliers and projects have joined. 5 AI projects have joined phase 3 to support testing and to develop evidence to support further adoption. 10 suppliers have joined the final phase to support wider adoption.
The NHS AI Lab tests and helps scale AI products and is backed by £250m in funding from the Department for Health and Social Care. The Award is run by the Accelerated Access Collaborative in partnership with NHSX and the National Institute for Health Research. It will make £140 million available over three years to accelerate the testing and evaluation of AI technologies which meet the strategic aims of the NHS Long Term Plan.
Phase 1 projects include:
- Stewardship of Antimicrobials using Real-Time Artificial Intelligence (SamurAI) – University College London
The SamurAI system will use AI to combine historical data for patients prescribed with antibiotics with the findings of specialists in infection as they review prescriptions. The system will learn when to stop or change the use of antibiotics to ensure they are only used when really necessary.
- Deep learning for effective triaging of skin disease in the NHS – University of Dundee
This project is developing an AI (deep learning) system to distinguish common benign skin lesions from common skin cancers with state-of-the-art accuracy. This research will develop the system with representative image data from NHS clinics.
- A fully automated ultrasound tool to screen for fetal growth restriction (FGR) in the first trimester – University of Oxford
This project will further develop fully automated ultrasound tools (the OxNNet toolkit) that can provide reliable measurements of placental size and shape in the first trimester, as well as estimating the blood flow within it. This can form the basis of a population-based screening test for fetal growth restriction (FGR). By identifying women at high risk of FGR early, we can increase monitoring and deliver the baby before it is stillborn and, in the future, test new treatments that could help prevent FGR from developing.
- Personalised Preoperative (Neoadjuvant) Chemotherapy (NACT) to optimise curative treatment in breast cancer – University of Nottingham
This project will identify features from MRI scans in breast cancer patients, in combination with routine clinical data and advanced computational modelling, to predict the response to NACT. This will help clinicians to make better decisions for each individual patient and minimise unnecessary treatment.
- OCTAHEDRON : Optical Coherence Tomography Automated Heuristics for Early Diagnosis via Retina in Ophthalmology and Neurology – Newcastle University
The OCTAHEDRON project aims to use machine learning to detect early signs of neurodegenerative disease – such as Parkinson’s – in OCT scans of the retina. NHS doctors’ input and analysis of thousands of OCT scans will teach the computer system how to recognise changes in the retina that could help to detect neurological disorders sooner when treatment can be most effective.
- Improving diagnostic yields of the Faecal Immunochemical Test using artificial intelligence and machine learning – Advanced Expert Systems Limited
This project will develop a computer model to identify potential cases of bowel cancer or polyps using results from FIT and other patient data, to enhance the NHS bowel cancer screening programme. The aim is for this system to help identify high-risk patients who can be prioritised for colonoscopy, reducing unnecessary, costly procedures.
- Development of AI techniques to predict eye cancer using big longitudinal data – University of Liverpool
This project will further develop a novel, fully-automatic AI-powered diagnostic tool to support the accurate diagnosis and monitoring of choroidal naevi (patches of pigment at the back of the eye) and to predict the risk of ocular melanoma, the major form of eye cancer. The aim is to help to streamline the management of patients and reduce cost for the NHS by assessing and monitoring in the community for low-risk lesions, and follow up conditions with high risk factors in secondary care.
- An artificial intelligence macular disease treatment decision tool for patients with wet age-related macular degeneration, diabetic macular oedema, and retinal vein occlusion – Macusoft Ltd
This project will complete development of computer software that will be able to tell whether the patient’s eye is stable or not without them needing to see eye doctors for regular check-ups. The software will allow more patients to be seen and treated, by creating efficiency and capacity for busy eye clinics
- Project Rhapsody: Investigating the clinical feasibility of using AI-based deep audio and language processing techniques to diagnose neurological and psychiatric diseases – Novoic Ltd
This AI technology offers a novel way to analyse complex speech/language patterns from free speech to detect common neurological and psychiatric diseases. A key differentiator is the use of proprietary speech representation methods to build generalisability and robustness into the tool–this research will help establish the clinical feasibility of this first-of-its-kind modality.
- AI-enabled point-of-care technology for radiotherapy planning peer review – Mirada Medical Ltd
This project will determine the clinical and technical feasibility of using AI for review of RT treatment plans for cancer patients. This will be tested using anonymised clinical images to evaluate how effective the approach is, with the ultimate aim of conducting effective, faster checks to improve treatment.
- Prognosis of epilepsy using at-home EEG monitoring – Neuronostics Limited
This project will develop a smartphone-based app that can receive and segment EEG recordings from wireless headsets to assist with assessing how well epilepsy treatment is working. The project will deliver a prototype device and a roadmap for product development.
- Autonomous cardiac MR acquisition – Barts Health NHS Trust
This project aims to use AI to fully automate cardiac MRI scans. Autoscanning and autoanalysis will add precision and help to predict clinical outcomes better than current care, as well as speeding up scanning, reducing waiting times, saving money and freeing up scarce resources.
- Wearable technology to enable remote precision and predictive medicine for respiratory patients – Senti Tech Limited
This project is developing a device which enables remote chest examination for respiratory patients through sensors embedded into a jacket. The research will use AI to develop ways to predict deterioration in patients with long-term respiratory conditions, personalise treatment and enable patients to make more informed healthcare choices.
- An artificial intelligence algorithm for diagnosing attention deficit hyperactivity disorder (ADHD) in adults – University of Huddersfield
This project will develop a first-of-its-kind AI solution for diagnosing ADHD in adults. This will shorten the time people will need to wait for a diagnosis because a broader range of health professionals will be able to complete the diagnostic assessments quicker. The AI will use clinical data to guide health professionals about who requires extra assessment and who doesn’t.
- Woubot: An AI predictive system to produce personalised care recommendations for chronic lower limb wounds – Nine Health Global Ltd
This project will create a suite of automated software tools for community and wound clinics, with a user-friendly mobile application designed by doctors and nurses for their own use within the NHS. The app will generate a personalised care pathway for each patient, and use image and other automated software to monitor progress and outcomes.
Phase 2 projects
- FORE AI – Odin Vision Limited
Odin Vision’s AI technology assists doctors to detect and characterise polyps. Better early detection and instant diagnosis have the potential to improve patient outcomes, reduce costs and improve the patient experience. The FORE-AI project will evaluate this AI technology across multiple hospitals to analyse the benefits for patients and the potential cost savings.
- Clinical validation of the AI Clinician decision support system for sepsis treatment – Imperial College London
This project will test a method to automatically and continuously recommend to clinicians the correct dose of medications for treating sepsis in individual patients, personalising treatment and potentially improving survival.
- Developing Lifelight: A contactless vital signs monitor for CVD screening – Xim Limited
Lifelight is software technology that completely contactlessly measures blood pressure, heart rate, respiratory rate and oxygen levels in the blood using the camera on any smartphone. This project will collect data to allow Lifelight to measure blood pressure more accurately so it can diagnose high blood pressure more effectively.
- Digitally adapted, hyper-local real-time bed forecasting to manage flow for NHS wards – University College London
This project aims to improve a model that predicts future demand for hospital beds, allowing local teams to adjust staffing levels or reschedule operations in line with future demand. The model will be tested with clinical and operational teams to make sure it is reliable, easy to use and safe.
- Interactively trained ‘human-in-the-loop’ deep learning approach to improve cardiac CT and MRI assessment for accurate therapy response and mortality prediction – University of Sheffield
This project will develop an interactive deep learning method to measure heart health in large groups of patients, using MRI/CT scans. Existing detection algorithms will be used on these scans and the data will then be edited by experienced consultants to improve the measurements. Ultimately this could provide better predictions of responses to treatment and survival in patients with heart disease.
- Artificial Intelligence to improve cardiometabolic risk evaluation using CT (ACRE-CT) – Caristo Diagnostics Limited
Caristo Diagnostics’ FatHealth technology is using standard CT scans combined with AI techniques to detect fat tissue inflammation, which can indicate a higher risk of developing diabetes or dying from heart disease. The project will analyse 20,000 CT scans to train the AI algorithm and help develop accurate risk predictions.
- Dem Dx triage support platform for ophthalmology referrals – Dem Dx Limited
This project will develop and test a new technology to gather and process information about patients’ eye symptoms, to help healthcare staff make accurate and safe triage decisions and deal with the most common eye problems. This could help free up specialist time for urgent and complicated cases that need faster treatment.
- BioEP: From prototype to clinical evaluation – Neuronostics Limited
BioEP is a computer biomarker of epilepsy that is designed to augment electroencephalograms (EEGs) to address delays in diagnosis of the condition. The project will further develop a prototype of a diagnostic decision support tool that can provide a risk score showing how easy it is for seizures to occur.
- SMARTT critical care pathways (Safe, Machine Assisted, Real Time Transfer): an artificial intelligence based decision support tool to enable safer and more timely critical care transfer – University Hospitals Bristol and Weston NHS Foundation Trust
This research is developing an AI tool that will help to decide which patients are well enough to leave intensive care, helping to free up bed spaces and provide better care. The tool uses data from monitors, ventilators and blood tests and will help clinicians make more accurate and timely decisions. It will be tested in a real intensive care unit as part of this project.
- Prediction and prevention of asthma attacks in children – BreatheOx Limited
Albus Health have developed a small table-top device that can automatically monitor a range of symptoms and metrics without patients having to do or wear anything, helping to predict preventable asthma attacks in children. Alongside Birmingham Children’s Hospital, Imperial College London, Asthma UK and Oxford AHSN, this project will test the system within existing NHS infrastructure to generate real world evidence of clinical benefit and economic value.
- Autonomous telemedicine – cataract surgery follow-up at two NHS trusts – Ufonia Limited
The project is using Ufonia’s natural-language AI assistant delivered via a normal telephone call, to follow up with patients after cataract surgery at two large NHS hospitals. This study will evaluate Ufonia’s clinical decisions with those of expert clinicians and assess how acceptable the system is for patients and clinicians.
- Natural language processing for real-time data capture in electronic health records to improve clinical care and operational efficiency – University College London Hospitals NHS Foundation Trust
This project is developing a natural language processing system to support the conversion of clinician’s text in electronic health care records into a structured format that can be processed by computers to help support clinical decision making, planning and research. The system works at the point of care, during data entry, and clinicians are given the opportunity to validate the suggestions before they are added to the patient record. The system will be tested in a simulation environment and then tested at University College London Hospitals and Great Ormond Street Hospital.