AI medical trial: AI echocardiography to detect cardiac amyloidosis — a clinical trial (ClinicalTrials.gov)
A prospective study testing whether AI applied to echocardiography images can pick up patients with findings suggestive of cardiac amyloidosis (especially the ATTR type) — a condition easily missed. The AI flags are checked against confirmatory methods: bone-tracer SPECT, cardiac MRI, and blood tests.
Trial overview (primary data)
- StatusNot yet recruiting
- ConditionsCardiac Amyloidosis, Heart Failure With Preserved Ejection Fraction (HFPEF), Left Ventricular Hypertrophy
- InterventionsDIAGNOSTIC_TEST: AI-based echocardiogram
- SponsorGermans Trias i Pujol Hospital
- Target enrollment200 participants
- Period2026-07-01 〜 2028-10-01
Key points
- Tests how accurately AI applied to echocardiography identifies patients suggestive of cardiac amyloidosis (especially ATTR)
- A prospective study that judges the AI against confirmatory methods: PYP-SPECT, cardiac MRI, and blood tests as ground truth
- Layers AI onto the first-line echocardiogram to widen early detection of an easily missed disease
- Also aims to clarify the true prevalence of ATTR cardiac amyloidosis among referred patients with red-flag features
- The AI call is not taken as truth; scintigraphy, cardiac MRI and blood tests serve as the reference standard.
1Cardiac amyloidosis is hard to catch
Cardiac amyloidosis is a disease in which misfolded protein is deposited in the heart muscle, causing mainly heart-failure symptoms. It often produces heart failure even though the left ventricle's pumping function (ejection fraction) is preserved, and it shows certain features on echocardiography such as thickened (hypertrophied) walls.
Because it can resemble other heart conditions, identifying it with currently available echocardiographic tools remains challenging. Early diagnosis matters greatly, since it can connect patients to therapies that have improved survival.
2Reading echocardiograms with an algorithm
This study (NCT07529860) is a prospective study testing how accurately an AI algorithm applied to echocardiographic images (the software Us2.ai) can pick up patients with findings suggestive of cardiac amyloidosis, especially the transthyretin (ATTR) type.
Crucially, it does not treat the AI's output as the answer: accuracy is judged against methods that actually establish the diagnosis — bone-tracer SPECT (using 99mTc-pyrophosphate, PYP), cardiac MRI, and blood tests, with tissue confirmation if needed — used as ground truth.
Echocardiography is the first-choice test for patients with breathlessness or suspected heart failure, and for those with risk factors such as hypertension and diabetes. Layering AI onto these routine scans could widen the chance of catching the disease earlier.
3Not treating the AI's call as truth
What matters in the design is that the AI's determination is not taken as the answer. Methods that actually establish the diagnosis are placed as the standard.
- 1AI reads the echocardiogramThe software identifies patients with findings suggesting cardiac amyloidosis
- 2Its call is not the truthBy design it is not taken as the reference
- 3Diagnostic methods are the standardBone scintigraphy, cardiac MRI, blood tests, and tissue confirmation if needed
- 4Accuracy and prevalence togetherTrue prevalence among those with red-flag findings is also established
Cardiac amyloidosis often causes heart failure with preserved ejection fraction and is easily confused with other cardiac disease. Echocardiography is the first test chosen for suspected breathlessness or heart failure, so laying an AI reading over it could widen the chance of picking the disease up from an ordinary echo.
Why it matters
Combining an existing AI tool with the first-line echocardiogram and validating it against confirmatory methods is one example of how AI may be positioned for the early detection of an easily missed disease. The study also aims to clarify true prevalence among patients with red-flag features.
FAQ
What is cardiac amyloidosis?
Is the diagnosis made by the AI alone?
Sources (primary)
Source: ClinicalTrials.gov (U.S. NIH/NLM, public domain). This site does not provide medical advice. Verify the latest and exact details with the official source. This site is not endorsed or certified by the NIH/NLM.
- ClinicalTrials.gov (study record, original)
- NCT ID: NCT07529860