Aortic Stenosis
Conditions
Keywords
artificial intelligence, electrocardiogram, deep learning, cardiac disease, aortic stenosis
Brief summary
AoS-SEDAI study is an observational, multicenter, retrospective and prospective clinical study. This study aims to assess Willem Artificial Intelligence (AI) ability to distinguish between aortic stenosis (AS) and non-AS patients from 12-lead electrocardiogram (ECG) data.
Detailed description
Aortic Stenosis (AS) is a common and progressive valvular heart disease, especially in older adults, yet significantly underdiagnosed. Many individuals with significant AS may remain asymptomatic for extended periods or experience vague symptoms, delaying diagnosis for several years. In some cases, sudden cardiac events or decompensation may be the first indication of advanced AS, particularly in those who have not undergone regular cardiovascular evaluation. The primary method for diagnosing AS is echocardiography, which allows visualization of valve anatomy and assessment of transvalvular gradients. However, reliance on symptom reporting or late-stage signs, that trigger a provider to order an echo, can result in missed opportunities for earlier detection. Additionally, electrocardiographic changes-such as left ventricular hypertrophy or strain patterns-can often be detected before structural abnormalities are visible on imaging studies. This delay between electrical changes evolution and later structural findings creates a valuable opportunity to intervene sooner, for example with a valve replacement. This diagnostic latency highlights a critical window where early identification through Artificial Intelligence (AI) analysis of electrocardiograms (ECGs) and timely referral to cardiology can significantly alter disease trajectory and improve outcomes, especially in primary care and community health settings. AoS-SEDAI study is an observational, retrospective and prospective, multicenter clinical study. Even though controls will be distinguished from AS patients for ground truth and performance evaluation, this is a single-arm study since there are no differences in study interventions.
Interventions
DEVICEWillem AI ECG assessment
There is no study intervention. The Willem AI platform will assess all study electrocardiograms (ECGs) for the identification of aortic stenosis. Regardless of retrospective or prospective enrollment, Willem output will not be provided to the healthcare professional user for clinical evaluation, and therefore routine practice will not be impacted nor altered.
Sponsors
Idoven 1903 S.L.
Study design
Observational model
CASE_CONTROL
Time perspective
OTHER
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
Yes
Inclusion criteria
* Age ≥ 18 years; * All available 12-lead ECG with a 10 seconds minimum length on raw data digital format will be included * Available clinical data corresponding to each ECG to confirm patient demographics and Aortic Stenosis diagnosis * Available transthoracic echocardiogram (TTE) within +/- 90 days of each ECG recording No
Exclusion criteria
are defined for this study.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Willem performance to detect severe Aortic Stenosis | ECG will be performed at baseline, and ECG analysis by Willem AI will be performed retrospectively throughout the trial, and finalized upon recruitment completion. | Performance of Willem AI platform to distinguish between severe Aortic Stenosis (AS) patients with confirmed diagnosis and non-AS patients, by means of the following performance metrics: Area Under the Receiver Operating Characteristic curve (AUROC), diagnostic accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). The Standard Of Care (SOC) investigator diagnosis will be used as the ground truth. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Willem performance to detect moderate Aortic Stenosis | ECG will be performed at baseline, and ECG analysis by Willem AI will be performed retrospectively throughout the trial, and finalized upon recruitment completion. | Performance of Willem AI platform to distinguish between moderate Aortic Stenosis (AS) patients with confirmed diagnosis and non-AS patients, by means of the following performance metrics: Area Under the Receiver Operating Characteristic curve (AUROC), diagnostic accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). The Standard Of Care (SOC) investigator diagnosis will be used as the ground truth. |
Countries
Germany
Contacts
CONTACTManuel Marina-Breysse, MD, PhD
Outcome results
None listed