AI in cardiovascular risk prediction: from patient history through biomarkers to imaging
How machine-learning tools estimate cardiovascular risk from the history, from blood and molecular biomarkers and from imaging, and how to judge whether they work.
- Duration
- Four live sessions of 120 minutes over two weeks
- Format
- Live online (Zoom Webinar) with worked case examples, maximum 30 participants per language cohort
- Who it is for
- Cardiologists, internists, general practitioners, cardiovascular nurses and researchers who use, buy or evaluate risk-prediction tools.
Introduction from the faculty
A short faculty introduction will be added here.
A structured programme on the tools that now estimate cardiovascular risk: from classical scores built on the patient history, through blood, genetic and multi-omic biomarkers, to models that read the ECG, the echocardiogram, coronary CT and retinal images. Each layer is examined for what it adds, how it was validated, where it fails and what it means for the patient in front of you.
Learning outcomes
- Explain how classical risk scores and machine-learning models differ in inputs, assumptions and outputs
- Judge a prediction model on discrimination, calibration, external validation and clinical usefulness
- Place biomarkers, polygenic scores and multi-omic panels against what the history already tells you
- Interpret AI-enabled ECG, echocardiography, coronary CT and retinal analyses, including their failure modes
- Apply European and UK regulatory and governance requirements when adopting a tool in practice
Curriculum
01Risk from the history
- From Framingham to SCORE2 and PREVENT
- Machine learning on electronic health records
- Lifetime versus ten-year risk
- Risk-enhancing factors and how models miss them
02Biomarkers and molecular risk
- High-sensitivity troponin, natriuretic peptides and lipoprotein(a)
- Polygenic risk scores: promise and limits
- Proteomic and metabolomic panels
- Incremental value: reclassification and decision curves
03Imaging read by machines
- Coronary calcium and AI plaque analysis on CT
- AI-enabled ECG and echocardiography
- Retinal imaging as a cardiovascular window
- Bias, drift and hidden confounders in imaging models
04From model to clinic
- Calibration, fairness and subgroup performance
- EU AI Act, medical-device rules and local governance
- Communicating risk to patients
- Where the evidence ends: open questions
Sessions
Delivered via Zoom Webinar with EU data-residency settings.
English session
Register interest
Dates to be announced
Price announced with dates
Polish session
Register interest
Dates to be announced
Price announced with dates