Our scientific evidence covers echo automation and EKG-based predictions.
A deep learning model using temporal shift module architecture, trained on 22,527 echo studies and validated across 115 patients (1,955 segments), achieved expert-comparable performance for RWMA detection and significant CAD prediction — supporting AI as an automated screening tool for wall motion analysis.
Single-center retrospective study across 145 patients with ICM, HCM, and cardiac amyloidosis demonstrating that AI-Echo measurements match expert manual echocardiography and show comparable or superior agreement with CMR for volumetric and functional parameters.
Multicenter study across 205,916 ECG–echo pairs and four validation cohorts (hospital, primary care, health screening) demonstrating robust simultaneous detection of LVH, LAE, and LVE — outperforming conventional rule-based ECG criteria across all settings (all p<0.001).