Catch Cardiac Amyloidosis Before It's Too Late
Cardiac amyloidosis is severely underdiagnosed. Us2.ai combines an FDA-cleared deep learning algorithm with automated guideline-based measurements to help clinicians identify it earlier, when treatment is most effective.
A Common Disease Hiding in Plain Sight
Cardiac amyloidosis was once considered rare. We now know it is massively underdiagnosed, found in up to 1 in 5 patients in common cardiac populations.
Studies find ATTR cardiac amyloidosis in 13% of elderly patients hospitalized with heart failure with preserved ejection fraction.
ATTR has been identified in up to 16% of patients undergoing TAVR for severe aortic stenosis, a critical dual-pathology population.
Untreated AL amyloidosis with cardiac involvement has a median survival of just 6 months, but early-stage patients can survive 9+ years.
International guidelines now recommend screening for cardiac amyloidosis in patients over 65 with unexplained increased wall thickness1, a population that routinely undergoes echocardiography. Non-invasive diagnosis of ATTR is possible via bone scintigraphy, but only if the disease is suspected in the first place.
Deep Learning on a Standard Four-Chamber Clip
Us2.ai's deep learning model analyzes a standard apical four-chamber clip and flags studies suggestive of cardiac amyloidosis for a qualified clinician to review. No manual scoring. FDA cleared and CE marked.
Pattern Recognition from a Single View
The model reads the moving four-chamber image and reports one of three results: suggestive of cardiac amyloidosis, equivocal, or not suggestive. It does not mark where in the image the result comes from, and it does not tell amyloidosis types apart, which is why the measured red flags below sit beside it.
FDA Cleared for Amyloidosis Detection
This specific algorithm has received FDA clearance (Us2.ca, K250151) and CE marking for cardiac amyloidosis detection. Its FDA validation was a retrospective, multi-center study at sites in the United States and Japan, on images from GE and Philips ultrasound systems. Us2.ai is the first platform to combine deep learning pattern recognition with guideline-based echocardiographic measurements for automated detection.
Built Into the Automated Report
Once the module is enabled, there is no extra workflow. Us2.v2 checks the images first and passes suitable four-chamber clips to Us2.ca, and the result appears among the report's main findings, where the clinician can clear or change it before signing.
Every Red-Flag Parameter, Measured and Shown
Beyond the AI flag, Us2.ai measures the parameters in the internationally recommended diagnostic criteria, so a clinician can check the flag against them before signing.
Increased Wall Thickness
LV wall thickness ≥ 12 mm in a non-dilated ventricle is the hallmark finding that triggers further workup. Us2.ai measures this automatically on every study, and it is the threshold that decides whether the Us2.ca result applies.
Reduced Global Longitudinal Strain
GLS worse than −15% with a characteristic apical sparing pattern is highly suggestive. Us2.ai computes GLS and the apex-to-base strain ratio automatically.
Diastolic Dysfunction
Grade 2 or worse diastolic dysfunction with E/e' > 11 is a core diagnostic criterion. Us2.ai performs complete diastolic assessment including tissue Doppler velocities.
RV Dysfunction
TAPSE ≤ 19 mm indicates right ventricular involvement and carries prognostic significance. Automatically measured and reported by Us2.ai.
A validated echocardiographic scoring system combines relative wall thickness, E/e', TAPSE, GLS, and apical sparing ratio to identify cardiac amyloidosis with high accuracy. Us2.ai delivers all five parameters in every automated report, providing the data clinicians need to apply this scoring at scale. The automated detection workflow for cardiac amyloidosis and hypertrophic cardiomyopathy is patented (US 12,400,762 B2).
Independently Validated in 11,000+ Patients
The AI-SCREEN-CA study is the largest independent validation of AI-based cardiac amyloidosis screening from archived echocardiograms, presented as a late-breaking presentation at the American Heart Association Scientific Sessions 2025.
AI-SCREEN-CA: AI-Assisted Screening for Cardiac Amyloidosis
AI screen-positive patients had a cardiac amyloidosis prevalence over five times higher than screen-negative patients (16.7% vs. 3.0%), demonstrating the algorithm's ability to identify high-risk individuals from routine echocardiograms.
Why Early Detection Changes Outcomes
Effective therapies now exist for cardiac amyloidosis, but they work best when started early. Echo-based screening is the key to prompt diagnosis.
Disease-Modifying Therapy
Clinical trials have demonstrated that approved disease-modifying therapies can reduce mortality and cardiovascular hospitalizations in ATTR cardiomyopathy. Treatment is most effective in early-stage disease, making timely detection critical.
Non-Invasive Diagnosis Now Possible
ATTR cardiomyopathy can be diagnosed without biopsy using bone scintigraphy, but only if suspicion is raised first. Echocardiography is the front-line imaging modality that triggers the diagnostic pathway.
Calls for AI Screening
The ESC Working Group on Myocardial Diseases2 explicitly identifies AI-based imaging tools as a key area of investigation to improve cardiac amyloidosis diagnosis at scale.
The Survival Impact of Early Detection
References
- Garcia-Pavia P, et al. Diagnosis and treatment of cardiac amyloidosis: a position statement of the ESC Working Group on Myocardial and Pericardial Diseases. European Journal of Heart Failure. 2021;23(4):512–526.
- Rapezzi C, et al. ESC Working Group on Myocardial and Pericardial Diseases position statement on the role of AI in cardiac amyloidosis. European Heart Journal. 2023.
Selected AI Echo Algorithm in Alnylam and Viz.ai's ATTR-CM Care Pathway
Us2.ai has been selected as the echocardiography AI algorithm powering a strategic collaboration to build an AI-enabled ATTR-CM care pathway, designed to help clinicians identify patients earlier in the course of disease and guide appropriate diagnostic evaluation and referral.
AI-Enabled ATTR-CM Care Pathway: Earlier Detection and Coordinated Care Across Health Systems
“By embedding AI-driven detection into everyday clinical workflows and pairing it with coordinated care pathways, this collaboration focuses on closing the gap between the first clinical signal and meaningful clinical action.”
Frequently Asked Questions
Common questions about cardiac amyloidosis and AI echocardiography.
How common is cardiac amyloidosis?
How does echocardiography detect cardiac amyloidosis?
Can AI help detect cardiac amyloidosis?
Why does early detection of cardiac amyloidosis matter?
What Cardiologists Say About AI Amyloidosis Detection
Cardiac amyloidosis remains a frequently under-diagnosed condition with significant prognostic implications. Us2.ai's AI-based pattern recognition model, applied to a standard apical four-chamber view, provides a validated and scalable solution to support early detection and enhance clinical decision-making within routine echocardiographic practice.
Us2.ai's platform already provides powerful automated measurements that support the detection of cardiac amyloidosis. But this new deep learning approach moves beyond standard measurement workflows by enabling fast, fully automated detection of amyloidosis from a single echo view. By simplifying image requirements while maintaining diagnostic accuracy, it has the potential to expand the role of echocardiography as an accessible, practical screening tool for earlier identification of patients with this life threatening yet treatable disease.
A routine echo is where suspicion of cardiac amyloidosis starts.
Raise Suspicion Earlier, on the Echo You Already Run
Us2.ca flags possible cardiac amyloidosis in adults with a wall thickness of 12 mm or more, in the same automated report as the guideline-based measurements a clinician checks it against. The clinician makes the diagnosis.