PRESENT-SHD©


An Ensemble Deep Learning Algorithm for Structural Heart Disease Screening Using Electrocardiographic Images

About


The PRESENT-SHD© tool can be used for the automated detection and prediction of multiple structural heart diseases (SHDs) using an ensemble deep learning approach. Users can upload an image of a 12-lead ECG to obtain an SHD probability below.

A detailed description of the development and validation of this tool are described in the work by Dhingra LS and Khera R, et al.

This tool is for demonstration only and is not intended for clinical use. For questions or feedback regarding the tool, please contact us at cards@yale.edu.

Instructions


  • Click the “Access the tool here!” button below. This will open a new window with the PRESENT-SHD web application.

  • Drag and drop the ECG image file or click the “Upload ECG Image” button to select the image file.

  • Input the person’s age and sex, and click the “Proceed to Image Cropping” button.

  • Ensure that only the 12-lead ECG is included for model inference by using the crop and rotate functions.

  • Use the checkboxes to confirm that only the 12-lead ECG is visible in the image, without patient identifiers. Proceed to click the “Submit for PRESENT-SHD Inference” button.

  • Note: The web application can take up to 60 seconds to output the PRESENT-SHD model probability.