0
Could these things be detected from an EKG before this product?
Yes — heart failure and valve disease were detectable before, but only with echocardiograms, physical exams, or biomarkers; no ECG-based test could flag them at scale until this AI.
Why it matters: The breakthrough isn't that the conditions can be found at all — it's that a cheap, ubiquitous, 10-second test can now surface them in patients who'd otherwise go unnoticed.
- The standard route before this AI: a doctor suspects a problem from symptoms (breathlessness, fatigue) or an abnormal physical exam, then refers the patient for an echocardiogram — an ultrasound that directly images heart structure and pumping function.
- ECGs alone couldn't reliably reveal heart failure or valve disease — they only record electrical activity, which is why the research team says the AI extracts information 'the human eye can't see' from a test that was 'not designed to detect either condition'.
- Blood tests like BNP (B-type natriuretic peptide) can suggest heart failure, but they can't identify which valve is affected, and they aren't part of routine screening.
- The persistent failure this AI addresses isn't detection in principle — it's triage: patients can wait months for an echocardiogram, and the AI lets high-risk cases skip the queue.
Room for disagreement
- Some cardiologists argue the pre-AI toolkit — echocardiography and biomarkers — already caught these conditions adequately in symptomatic patients, and the real gap is only in asymptomatic screening, which the AI is still unproven at.
- Coverage of the accuracy figures varies: the researchers report 77–90% detection depending on the cohort, but the results are conference-stage, not peer-reviewed, and the 590-patient NHS trial is still pending regulatory approval — some commentators caution the real-world performance may be lower.
AI-generated with web search — may be wrong. Check the linked sources.
