Prof Alain Zemkoho (opens in a new tab)
Principal Investigator
University of Southampton
AI-assisted ICD screening
A fuller picture. More informed ICD decisions.
Deep-ICD turns 24-hour heart recordings into clear, visual insights to help clinicians assess suitability for subcutaneous and extravascular ICDs, revealing changes that state-of-the-art screening practices may miss.
ICD means implantable cardioverter-defibrillator: a device that treats potentially life-threatening heart rhythms.
Why a fuller picture matters
Heart signals change. Screening should capture that.
A short recording captures a moment in time. Deep-ICD is designed to help clinicians explore changes across the day and identify patterns relevant to device suitability and the risk of inappropriate shocks.
See changes over time
Explore the recording across 24 hours, including periods that may warrant closer review.
Make insights easier to review
Bring lengthy recordings into clear visual summaries, with detail available for further examination.
Support personalised assessment
Use the patient's individual recording to inform clinical discussion and screening decisions.
See what Deep-ICD reveals
From a day of recordings to a clearer clinical picture.
Each view leads to the next, from the full recording down to the ECG behind a single prediction.
01 · Overview
The 24-hour timeline
After submitting a recording for analysis, the clinician sees an overview of the whole day across each recorded vector.
02 · Zoom in
A closer look at a chosen time
Choosing a time on the overview opens a second timeline, showing that period in more detail.
03 · ECG sample
The ECG behind a prediction
Selecting a specific example on the second timeline shows the 10-second ECG sample it came from.
04 · Report
Predicted T:R or P:R ratio distribution
Clinicians can generate a PDF report, including how the predicted T:R or P:R ratios are distributed for each vector.
Step 1 of 4: The 24-hour timeline
Example output from a test recording. Deep-ICD is research software, not a tested medical device, and is not for diagnostic use.
The screening approach
Record. Explore. Inform.
01 · Record
Capture the 24-hour picture
A Holter recording captures how the patient's heart signals vary across the day.
02 · Explore
Review clear visual insights
Deep-ICD analyses the recording and presents patterns for closer review.
03 · Inform
Support clinical judgement
Clinicians interpret the findings, using them to inform the shared decision-making process with the patient.
Want to explore Deep-ICD in practice?
If you are a clinician interested in testing the platform or contributing to the research, we'd love to hear from you.
Research and publications
The research behind Deep-ICD
Explore the methods, clinical studies, and ongoing work behind prolonged ICD screening.
2024 · Clinical study · Indian Pacing and Electrophysiology Journal
Using artificial intelligence and deep learning to optimise the selection of adult congenital heart disease patients in S-ICD screening
ElRefai M, Abouelasaad M, Conibear I, et al.
2023 · Methods · Annals of Operations Research
Deep learning and hyperparameter optimization for assessing one’s eligibility for a subcutaneous implantable cardioverter-defibrillator
Dunn AJ, Coniglio S, ElRefai M, et al.
2023 · Clinical study · Annals of Noninvasive Electrocardiology
Correlation analysis of deep learning methods in S-ICD screening
ElRefai M, Abouelasaad M, Wiles BM, et al.
2023 · Clinical study · Annals of Noninvasive Electrocardiology
Role of deep learning methods in screening for subcutaneous implantable cardioverter defibrillator in heart failure
ElRefai M, Abouelasaad M, Wiles BM, et al.
The people behind the project
Clinical experience meets mathematical innovation.
Deep-ICD brings together cardiologists, mathematicians, AI experts, and software engineers at the University of Southampton and University Hospital Southampton.
Prof John Paisey (opens in a new tab)
Clinical Lead
University Hospital Southampton
Prof Paul Roberts (opens in a new tab)
Clinical Co-Lead
University Hospital Southampton
Dr Anthony Dunn (opens in a new tab)
Research Collaborator
University of Southampton
Dr Mohamed ElRefai (opens in a new tab)
Clinical Research Collaborator
University Hospital Southampton
Samuel Ward (opens in a new tab)
Doctoral Researcher
University of Southampton
Dr Lin-Thiri Toon (opens in a new tab)
Clinical Researcher
University Hospital Southampton
Bartosz Morawski (opens in a new tab)
Doctoral Researcher
University of Southampton
Dr Alex Touze
Clinical Researcher
University Hospital Southampton
Dr Sam Mangham (opens in a new tab)
Research Software Engineer
University of Southampton
Work with us
Interested in testing Deep-ICD?
Are you a clinician interested in exploring Deep-ICD with your patients or contributing to the research? Register your interest to learn more and request access to the platform.
Our partners
