Assess Pulmonary Hypertension Probability
Us2.ai measures every guideline-recommended right-heart parameter on each scan and reports the echocardiographic probability of pulmonary hypertension for the clinician to review. In a Stanford study in CHEST (2026), its peak TRV separated PAH patients from healthy controls with AUC 0.98, against 0.99 for core lab readers.
A Deadly Disease That Hides in the Right Heart
Pulmonary hypertension is frequently diagnosed late because the right ventricle is difficult to assess manually. Delayed diagnosis means patients miss the window for effective therapy.
Patients with pulmonary arterial hypertension experience years of diagnostic delay on average. By the time PH is recognized, many patients have advanced right heart dysfunction.
PAH disproportionately affects women, with a mean age of 48 at diagnosis. The disease is progressive and fatal without treatment, making early detection essential.
PH spans five distinct clinical groups with different etiologies and treatments. Echocardiography is the front-line screening tool that triggers the diagnostic pathway for all of them.
The RV's asymmetric shape, prominent trabeculations, and abnormal septal motion in PH make manual echocardiographic measurements time-consuming and prone to variability. This is why right heart analysis has lagged behind the left heart in AI development.
Every Guideline Parameter, Measured
International guidelines include echocardiographic metrics in both PH diagnosis and risk stratification. Us2.ai measures all of them accurately, from view classification through measurement, on every scan.
Peak Tricuspid Regurgitation Velocity
Peak TRV is the primary echocardiographic marker for estimating pulmonary artery systolic pressure and screening for PH. Us2.ai automatically identifies CW Doppler TR signals and traces the velocity envelope with ρ = 0.90 correlation to expert core lab measurements.
TAPSE
Tricuspid annular plane systolic excursion is a key measure of RV longitudinal function and a guideline-recommended risk stratification parameter. Us2.ai measures TAPSE automatically with ρ = 0.78 correlation and less than 2% bias versus expert readers.
Right Atrial Area
RA area reflects right heart filling pressures and is included in international guidelines' risk assessment. Us2.ai segments the right atrium automatically with ρ = 0.86 correlation. This parameter is particularly difficult and time-consuming to measure manually.
RV Size & Function
RV basal diameter, end-diastolic area, end-systolic area, and fractional area change provide comprehensive assessment of RV remodeling and systolic function. Us2.ai delivers all four measurements automatically from the RV-focused apical view.
Us2.ai performs the complete pipeline (view classification, view selection, segmentation and measurement) without manual interaction. The Stanford study tested whether those automated reads hold up against core lab readers and in a clinical referral cohort.
Independently Validated and Peer-Reviewed
Us2.ai's right heart analysis was independently evaluated by Stanford University in a study published in CHEST, a leading pulmonology and critical care journal.
Near-Expert Discrimination
For distinguishing PAH from healthy controls, Us2.ai achieved AUC 0.98 for peak TRV, compared to AUC 0.99 for expert core lab readers (221 PAH patients, 213 controls). In the harder referral cohort, with milder PH, the AI read reached AUC 0.75 against 0.79 for clinical reads, a difference that was not statistically significant.
Low Bias on Critical Parameters
For peak TRV, TAPSE, and RA area (the three most important PH echo markers) Us2.ai demonstrated less than 2% relative bias compared to expert core lab measurements. RV fractional area change was the exception, with a significant 11.5% bias.
Independent Stanford Evaluation
Data acquisition and analysis were performed independently by Stanford University researchers. The study included 221 PAH patients and 213 healthy controls, the largest AI validation cohort for right heart echocardiography in PH to date.
Us2.ai Accuracy: AI vs. Expert Core Lab
Frequently Asked Questions
Common questions about pulmonary hypertension and AI echocardiography.
How is pulmonary hypertension detected on echocardiography?
Echocardiography is the primary screening tool for pulmonary hypertension. The key parameter is peak tricuspid regurgitation velocity (TRV), which estimates pulmonary artery systolic pressure. Additional guideline-recommended markers include TAPSE, right atrial area, RV basal diameter, and RV fractional area change.
Can AI accurately assess the right heart for pulmonary hypertension?
Yes. A Stanford University study published in CHEST (2026) validated Us2.ai's fully automated right heart analysis in pulmonary hypertension patients. Against healthy controls, the AI's peak TRV separated PAH patients with AUC 0.98 (0.99 for core lab readers), and bias was under 2% for peak TRV, TAPSE and RA area. In a referral cohort with milder PH, AUC was 0.75 against 0.79 for clinical reads.
Why is right heart assessment difficult?
The right ventricle has an asymmetric shape, prominent trabeculations, and abnormal septal motion in PH, making manual measurements time-consuming and prone to variability. This complexity is why AI-based right heart analysis has lagged behind left heart applications, and why automated assessment is particularly valuable.
How does Us2.ai assess the right heart for pulmonary hypertension?
It runs the whole right heart workflow (view classification, view selection, segmentation and measurement) with no manual interaction, and reports the echocardiographic probability of pulmonary hypertension as a guideline-based conclusion for clinician review. The Stanford study (Celestin, Haddad et al., CHEST 2026) tested those automated reads against core lab readers and in a clinical referral cohort.
Stanford-validated. Published in CHEST. AUC 0.98 for PAH.
Complete PH Assessment on Every Echo
Us2.ai measures peak TRV, TAPSE, RA area, and complete RV function on every scan, and reports the guideline-based probability of pulmonary hypertension for the clinician to review.