AI-Enhanced Echocardiography for Cardiac Amyloidosis Detection – July 9, 2025.
AI-Enhanced Echocardiography for Cardiac Amyloidosis Detection – July 9, 2025.
The findings were published in full European Heart Journal.
1. Innovation Overview:
• Mayo Clinic and Ultromics developed the first AI model designed specifically for screening cardiac amyloidosis using echocardiography.
• Validated in a clinical trial with >2,600 patients across a multi-ethnic international population.
2. Diagnostic Accuracy:
• Sensitivity: 85% (correctly identifying those with the disease)
• Specificity: 93% (correctly identifying those without the disease)
• Effective using a single echocardiography videoclip
• Outperformed traditional clinical scoring systems:
• TTR cardiac amyloidosis score
• Increased wall thickness score
3. Amyloidosis Subtypes:
• Successfully detected:
• AL amyloidosis: 84% sensitivity
• ATTR wild-type (wtATTR): 85%
• ATTR hereditary (hATTR): 86%
4. Comparison with Traditional Methods:
• Conventional diagnosis relies on:
• ECG + Echo
• Red-flag feature recognition → MRI, Tc-PYP scan, or biopsy
• AI model improves early detection and supports non-invasive triaging.
5. Clinical Significance:
• Cardiac amyloidosis is often underdiagnosed due to nonspecific symptoms and imaging features.
• Early detection is essential since new drug therapies (e.g., tafamidis) can slow but not reverse disease progression.
• Estimated 15% of HFpEF patients may have underlying cardiac amyloidosis.
6. Integration & Use:
• AI model is FDA-cleared
• Interfaces directly with PACS echo systems
• Currently used in multiple U.S. centers, including Mayo Clinic
7. Expert Perspective:
• Dr. Patricia A. Pellikka:
• Lead investigator and Mayo Clinic cardiologist
• Emphasized the model’s breakthrough value in identifying amyloid early
• Expressed optimism about incorporating the tool into clinical practice
8. Background & Future:
• Builds on Mayo and Ultromics’ prior AI work (e.g., FDA-cleared HFpEF detection tool in 2022)
• Part of a broader trend toward AI-guided echo diagnostics for complex cardiomyopathies
