First-Of-Its-Kind AI-Powered Probability Scoring System Assesses Heart Failure with Preserved Ejection Fraction

By HospiMedica International staff writers
Posted on 31 Mar 2025

Image: The AI-powered probability scoring has been introduced on the EchoGo Heart Failure platform for enhancing HFpEF diagnosis (Photo courtesy of Ultromics)

Heart failure with preserved ejection fraction (HFpEF) is one of the most difficult types of heart failure to diagnose due to the intricate interaction between various clinical and echocardiographic factors. Accurate identification of HFpEF necessitates the careful integration of these variables, which can sometimes be inconsistent or unclear in certain patients. As a result, HFpEF is often either undiagnosed or misdiagnosed, leading to delays in providing the proper treatment. However, a new innovation represents a significant breakthrough in using deep learning to improve the detection of this often-overlooked condition.

Ultromics (Oxford, UK) has introduced a new advancement to its EchoGo Heart Failure platform: a pioneering artificial intelligence (AI)-driven probability scoring system for evaluating HFpEF. EchoGo Heart Failure remains the only AI platform capable of identifying HFpEF from a standard echocardiogram. The newly added probability scoring system enhances this unique capability by generating a continuous score that reflects the likelihood of disease, offering a more detailed diagnostic assessment of HFpEF, reducing uncertainty, and supporting better clinical decision-making. Additionally, the platform continues to assist in detecting other serious heart conditions, such as cardiac amyloidosis, which can help clinicians identify underlying causes of heart failure earlier.

The updated EchoGo Heart Failure platform has undergone rigorous validation in a study that focused on real-world, complex cases, including patients with multiple overlapping comorbidities such as hypertension and diabetes—situations where traditional diagnostic models often struggle to distinguish between cases. The results showed that the AI-powered probability scoring system added significant value, effectively aiding in the diagnosis of HFpEF, predicting the likelihood of disease, and improving patient outcomes. The platform demonstrated greater sensitivity in identifying HFpEF and improved the ability to differentiate between HFpEF and complicated cases, where the condition was suspected but not confirmed, compared to standard clinical methods.

The new probability scoring system is also linked to better patient outcomes: those identified by EchoGo Heart Failure as high-risk had double the likelihood of heart failure-related hospitalizations and mortality, emphasizing AI’s potential to enable timely interventions that can enhance patient management. By incorporating all available clinical data, EchoGo Heart Failure improved the correct management of HFpEF patients by 33% compared to relying solely on the H2FPEF score. When compared to standard clinical practices, such as H2FPEF and HFA-PEFF, EchoGo Heart Failure produced fewer indeterminate results, addressing a common challenge in traditional diagnostic models and enabling confident diagnoses in over 80% of patients.

"The introduction of the probability scoring feature into the EchoGo Heart Failure platform represents a pivotal step forward in heart failure detection and patient risk stratification," said Ross Upton, PhD, Founder, CEO, and Chief Scientific Officer of Ultromics. "By delivering a precise probability score, EchoGo Heart Failure enables clinicians to assess a patient's likelihood of disease, allowing for more confident decision-making".

Related Links:
Ultromics