Developing an AI algorithm for cardiac amyloidosis
Research type
Research Study
Full title
Developing an artificial intelligence algorithm for analysing echocardiograms and Cardiovascular magnetic resonance imaging to improve clinical management of patients with amyloidosis and related disorders
IRAS ID
284514
Contact name
Marianna Fontana
Contact email
Sponsor organisation
Royal Free London NHS Foundation Trust
Duration of Study in the UK
5 years, 0 months, 1 days
Research summary
Amyloidosis is an uncommon condition where proteins fold abnormally, accumulate in different parts of the body, and can affect their structure and function. When amyloidosis effects the heart, it can result in progressive heart failure, and unfortunately, these patients often don't have any symptoms until a late stage, and even then, the symptoms can be very non-specific, making the diagnosis very difficult.
Echocardiography uses ultrasound to visualise the heart, and is the most widely used method to image the heart for diagnosing heart and monitoring heart diseases, including amyloidosis. Even when these patients have an echocardiogram, their condition is sometimes not recognised. Cardiac MRI is another imaging modality, that is less widely available and is used less frequently. Early detection and more accurate prediction of disease severity and outcomes is particularly important because this can lead to changing treatment, including discovered drugs.
The National Amyloidosis Centre at the Royal Free Hospital is going to partner with Ultromics, a spinout company from the University of Oxford. Ultromics has previously applied artificial intelligence (AI) to echocardiography images acquired for the assessment of coronary artery disease in order to improve its diagnosis and to predict potential adverse outcomes. This was performed by taking thousands of measurements from a single echocardiogram, enabling more detailed assessments with no variability compared to those taken in routine clinical practice by human operators. Ultromics was able to integrate these measurements to make an accurate diagnosis and predict outcomes.
The aims of this project are to apply AI to echocardiograms and CMR in patients with suspected amyloidosis to improve:
1. The diagnosis of amyloidosis and related disorders
2. The understanding of how amyloid deposition causes heart dysfunction
3. Physicians’ ability to estimate prognosis, stage disease and enable better stratification for clinical trials
4. The assessment of response to treatment.REC name
North of Scotland Research Ethics Committee 2
REC reference
21/NS/0012
Date of REC Opinion
25 Jan 2021
REC opinion
Favourable Opinion