Smartphone Based Digital Screening for Aortic Valve Stenosis

Sensitivity and specificity will be calculated by comparing the classification produced by the smartphone-based algorithm with the diagnosis obtained from transthoracic echocardiography, which serves as the clinical reference standard. Aortic stenosis severity will be defined according to established guideline criteria, with moderate-to-severe aortic stenosis classified as an aortic valve area (AVA) of ≤ 1.5 cm². Smartphone recordings will be obtained during a single study visit using built-in microphones and motion sensors to capture heart sounds and chest wall vibrations. Echocardiographic measurements, performed by certified clinical personnel, will provide the comparator classification. The reported outcome will reflect how accurately the smartphone algorithm identifies participants with moderate-to-severe aortic stenosis at this time point.

Signal quality will be quantified by calculating the signal-to-noise ratio (SNR) of heart sound and vibration recordings captured using built-in smartphone microphones and motion sensors during the study visit. Higher SNR values indicate clearer cardiac signals with less background noise. The reported outcome reflects the feasibility and technical performance of the smartphone recording pipeline.

Agreement between the severity classification produced by the smartphone-based algorithm and the clinical reference standard (echocardiographic grading of aortic stenosis) will be quantified using Cohen's kappa coefficient. Echocardiographic classification will follow guideline-based severity thresholds. The reported value reflects the degree of concordance between both methods beyond chance.

The AUROC will be calculated to assess the discriminatory ability of the smartphone-based algorithm to distinguish between participants with and without moderate-to-severe aortic stenosis, as defined by an aortic valve area (AVA) ≤ 1.5 cm² on echocardiography. Higher AUROC values indicate better diagnostic performance.

Major adverse cardiac and cerebrovascular events (MACCE)-including all-cause mortality, cardiovascular mortality, non-fatal myocardial infarction, non-fatal stroke, and hospitalization for heart failure-will be recorded during follow-up. The clinical event data will be analyzed in relation to cardiac signal characteristics extracted from the baseline smartphone recordings (e.g., murmur intensity, dominant frequency patterns, signal-to-noise ratio). This outcome explores whether smartphone-derived cardiac features are associated with subsequent adverse clinical events.