CorVista Health today announced a poster presentation showcasing preliminary results of its machine learning-based algorithm for non-invasively estimating pulmonary capillary wedge pressure (PCWP) elevation – a key indicator of heart failure with preserved ejection fraction.
The poster titled ‘Point-of-Care Testing for Pulmonary Capillary Wedge Pressure Elevation Using Machine Learning on Non-invasive Signals’ was presented at the American College of Cardiology’s Annual Scientific Session (ACC.25) in Chicago on March 30, 2025.
Currently, PCWP is measured through an invasive right heart catheterization (iRHC) procedure that requires a cardiologist, pulmonologist or intensivist. The CorVista capture device collects non-invasive signals for 3.5 minutes in patients at rest, which could potentially provide an alternative to iRHC in certain patient populations. The data demonstrate how machine learning signal analysis of a symptomatic population may provide a robust, non-invasive alternative to measuring PCWP elevation, potentially making early detection of HFpEF and PH subtypes more accessible.
Key findings presented in poster:
- The machine learning-based test achieved 0.92 ROC-AUC, 90% sensitivity, and 76% specificity when tested in 283 patients, showing strong diagnostic performance consistently across multiple subgroups defined by sex, BMI and age – both confirming and ruling out disease with a high degree of certainty.
- Univariate analysis identified 247 features within the collected signal features that differentiate subjects with PCWP > 18mmHg vs PCWP ≤ 18mmHg when measured via iRHC.
- Model has a positive likelihood ratio of 3.8 and negative likelihood ratio of 0.13, indicating that the test has a strong impact on post-test probability in both test-positive and test-negative cases.
- Model resulted in a diagnostic odd ratio of 29, indicating that a positive result was 29 times more likely to be observed in a patient with heart failure than a patient without.
“Heart failure with preserved ejection fraction can be challenging to diagnose in part because its symptoms overlap with many other conditions and although echocardiography is an excellent modality for estimating left ventricular ejection fraction (EF), its sensitivity for diagnosing diastolic dysfunction is rather abysmal,” said Charles R. Bridges, M.D., Sc.D., Executive Vice President and Chief Scientific Officer, CorVista Health. “The new data provide preliminary results for the feasibility of our non-invasive machine learning-based test to detect elevated PCWP with a high level of accuracy. HFpEF represents over 50 percent of all cases of heart failure and has the highest morbidity and mortality. The underdiagnosis of HFpEF is arguably the largest unmet need in cardiovascular medicine today, representing approximately $25 billion in direct costs to the US healthcare system annually. Moreover, PH is also vastly underdiagnosed, and it is critically important to differentiate subtypes of pulmonary hypertension which have distinct often lifesaving guideline-directed treatments.”
The exploration of a PCWP add-on to the CorVista System represents a major advancement in cardiology, with potential implications for early diagnosis, cost reduction, and broader access to cardiovascular care. The CorVista System Add-On for elevated PCWP discussed here is currently an investigational device limited by federal law (or United States) law to investigational use and is not available for commercial distribution. The system is designed to be deployed anywhere with an internet connection – addressing the unmet need in rural and underserved populations with limited access to traditional standard of care.