Fully Automated Aortic Stenosis Assessment
Us2.ai is the first AI system to deliver fully automated aortic stenosis severity grading, from image acquisition to final classification, with zero human input and published accuracy matching expert sonographers.
The Most Common Valvular Heart Disease
Aortic stenosis affects over 12% of people aged 75 and older. As populations age and TAVR referrals surge, the demand for accurate AS grading is outpacing the specialist workforce.
Aortic stenosis is the most common valvular heart disease in the elderly. Prevalence increases steeply with age, making it a growing public health challenge.
Severe aortic stenosis carries a 2-year mortality of 25.6% when managed conservatively. Timely identification and intervention are critical to survival.
Up to 50% of AS cases show discordance between Vmax, mean gradient, and AVA grading criteria, making consistent severity classification a persistent clinical challenge.
TAVR has transformed AS management, creating unprecedented demand for echocardiographic assessment: pre-procedural screening and post-procedural surveillance. Every TAVR candidate requires precise AS grading, and every TAVR patient needs serial follow-up. Automating this assessment is essential to keep pace.
Complete AS Grading, Zero Human Input
Us2.ai is the first system to perform fully automated aortic stenosis assessment, from view identification through Doppler tracing to final severity classification, without any manual interaction.
Peak Aortic Velocity (Vmax)
The primary hemodynamic parameter for AS grading. Us2.ai automatically identifies continuous-wave Doppler of the aortic valve, traces the velocity envelope, and reports Vmax with r = 0.97 correlation to expert measurements.
Mean Pressure Gradient
Derived from the full CW Doppler envelope. Us2.ai computes mean gradient automatically. Alongside Vmax, it defines mild, moderate, and severe AS. Published correlation: r = 0.94.
Aortic Valve Area (AVA)
Calculated via the continuity equation using LVOT diameter, LVOT VTI, and AV VTI, all measured automatically. AVA is critical for identifying low-flow, low-gradient severe AS. Published correlation: r = 0.88.
LV Function & Flow Assessment
LVEF and stroke volume index are essential to distinguish classical severe AS from paradoxical low-flow states. Us2.ai reports both automatically, enabling complete hemodynamic context for every patient.
Beyond individual measurements, Us2.ai integrates Vmax, mean gradient, AVA, and LVEF to output a final AS severity grade: mild, moderate, or severe.
Validated in a Peer-Reviewed Study
Us2.ai's AS assessment has been validated in a multi-center study published in the Journal of the American Society of Echocardiography, the field's leading journal.
First Fully Automated System
Us2.ai is the first published AI system to perform end-to-end automated AS assessment requiring zero human input, from raw DICOM images to severity classification, with no manual measurements or corrections needed.
Expert-Level Accuracy
Across all key parameters, Us2.ai achieved correlations of r = 0.76–0.97 against expert sonographer measurements. For the primary AS marker (Vmax), the correlation of r = 0.97 matches inter-expert variability.
Scalable Without Specialists
The growing TAVR population demands exponentially more echo assessments. Fully automated AS grading enables high-volume screening and surveillance without requiring additional specialist sonographers for every study.
Us2.ai Accuracy: AI vs. Expert Measurements
Peer-reviewed. Fully automated. Expert-level accuracy.
Automate AS Grading on Every Echo
Us2.ai delivers Vmax, mean gradient, AVA, and severity classification, automatically on every scan, with published accuracy matching expert sonographers.