Caristo Diagnostics intend to analyse 20,000 CT scans to train its AI algorithm and develop rapid and reliable AI analysis of CT scans to deliver early risk prediction of someone developing diabetes. The earlier identification of patients at risk of diabetes will assist primary care in managing these patients better and will reduce the risk of further complications and disease progression, which will reduce the disease burden on the whole of the NHS.
FatHealth detects fat tissue inflammation, which is a reliable indicator of diabetes-related cardiometabolic risk, using new artificial intelligence techniques applied to computed tomography (‘CT’) scans completed as part of care for other health concerns.
The Oxford AHSN carried out a feasibility study to evaluate the technology’s utility in the diabetes care patient pathway. This included feedback from real world clinical evidence generation. More clinical studies are needed prior to widespread adoption.
What is the challenge?
Diabetes can be a chronic and lifelong condition, affecting quality of life for patients and significantly impacting NHS resources. More than 4.9 million people in the UK have diabetes and an estimated 13.6 million people are now at an increased risk of type 2 diabetes.1 Diabetes is associated with reduced life expectancy owing to a greater risk of heart disease, stroke, peripheral neuropathy, renal disease, blindness and amputation.
Obesity, diabetes and cardiometabolic disease are global health economic burdens.
The NHS currently spends at least £10 billion a year on diabetes, which is around 10% of its entire annual budget, and almost 80% of the diabetes spending is on treating complications.2 Earlier identification of a patient with increased risk of diabetes picked up during a “routine” CT scan could result in earlier intervention, including medication and lifestyle changes, and prevent progression of the disease or, in the case of pre-diabetes, could help to reverse the condition entirely. This could reduce the economic burden of diabetes-related morbidity and complications such as stroke and heart attacks.
What did we do?
The Oxford AHSN conducted a feasibility study to investigate the potential clinical utility of FatHealth as a biomarker to identify individuals with diabetes or pre-diabetes (impaired glucose regulation) and identify individuals at risk of future diabetes and its cardiovascular complications (stroke, heart attacks). An initial literature review was performed to explore the evidence base surrounding the management of cardiometabolic risk using the current care pathway in the NHS and schematic diagrams of the current and proposed pathways were developed. Stakeholders were identified through both literature review and local recommendations and were interviewed across primary and secondary care settings. The study aimed to identify where the technology would best placed in the clinical pathway for maximum potential benefit to both patients and the NHS, identify its clinical utility, and explore the possible barriers to adoption of the technology.
This project was funded by the National Institute for Health Research (Artificial Intelligence to improve Cardiometabolic Risk Evaluation using CT (ACRE-CT) AI_AWARD02013) and NHS Transformation Directorate.
What has been achieved?
The Oxford AHSN feasibility study evaluated the technology’s utility in the diabetes care pathway. Stakeholders who were interviewed agreed that earlier identification of at-risk patients could lead to earlier intervention/management of the disease and therefore have the potential to improve patient outcomes and reduce the cost burden to the NHS.
The partnership between Oxford AHSN and Caristo Diagnostics has helped the company to develop their offering to more easily identify patients that are at risk of developing diabetes. Caristo Diagnostics have used the feedback provided from the semi-structured interviews with clinicians to develop their offering to draw it more in line with clinical needs and its possible place in the clinical care pathway. The study has also helped with data collection for real world clinical evidence generation. This feedback is crucial for the technology to be developed in a way that would increase the likelihood of future NHS adoption and takeup.
What people said
“The Oxford AHSN has been working closely with Caristo Diagnostics on our AI Award projects and is helping us in our NHS Adoption Strategy. They have recently completed a feasibility study for our product and have highlighted how our product is perceived in the industry. They have also directed us about the industry expectations from the product helping us plan the implementation effectively.”
Yogesh Sohan, Project Manager, Caristo Diagnostics
Further clinical studies are required to allow the product to gain regulatory approvals, and to generate the evidence required for adoption by the NHS. Many stakeholders interviewed were interested in the technology and were keen in principle to implement the technology.
Lead Health Economist & Methodologist