AI-echo to diagnose Cardiac Amyloidosis: a multi-centre international development and validation study
University College London
(Presenter)
Cardiac Amyloidosis (CA) is a progressive and ultimately fatal cardiomyopathy. The majority of cases are caused by transthyretin amyloidosis (ATTR) and light-chain amyloidosis (AL). Despite advancements in cardiac imaging, CA remains significantly underrecognized and underdiagnosed, presenting a challenge in patient care.
Presented at the Late Breaking Science session at EuroEcho–Imaging 2024, this multi-centre international study introduces and validates Us2.ai’s deep learning diagnostic (black box, pattern recognition) algorithm, developed based on apical 4-chamber DICOM images to detect CA using echocardiography.
The study demonstrated the algorithm’s diagnostic accuracy and its performance in combination with a fully AI-automated multi-parametric echocardiographic score (current clinical guidelines recommendation).
These results provide strong scientific evidence for Us2.ai’s unique approach to improve CA detection and diagnosis with greater precision and automation.
Methods
Results
Conclusion
This breakthrough study marks a pivotal step forward for Us2.ai, showcasing our technology’s potential to improve the early detection of Cardiac Amyloidosis. By harnessing the power of AI, this study offers hope for more accurate diagnoses and the promise of better patient outcomes, reinforcing Us2.ai’s commitment to advancing healthcare through innovative, precision-driven solutions.
Comments from session discussant
Professor Eugenio Picano, MD, PhD, session discussant at EuroEcho–Imaging 2024, commended the study for its design, methodology and scientific approach, stating, “The study’s design enhances the generalisability of its findings by incorporating diverse ethnic groups and institutions from various geographical regions.” This global scope adds significant value to the performance findings of Us2.ai’s CA algorithm, ensuring its applicability to a wide range of patient populations.
Professor Picano further highlighted the clinical relevance of this methodology, noting, “The use of the multi-parametric echo score as the standard comparator significantly increases the clinical appeal of the findings, as it is based on parameters recommended by established clinical guidelines.“
