3D-Visualization-of-Myocardial-Scar-Using-CMR-Automatic-Segmentation

Under Development
This is a graduation project peformed by team of four students in Healthcare Engineering Department at Cairo University in Egypt in collaboration with Magdi Yacoub Foundation, Aswan Heart and Research Centre under the supervision of advisors from both Cairo University and Aswan Heart Centre.

Project Summary

Create automated platform to visualize accurate location of myocardium scars on 3D model of left ventricle and automatically extract clinically relevant parameters. Segmentation of myocardium scars and left ventricle will be achieved using Deep Learning. Data and mentoring needed to accomplish such project is provided by Aswan Heart Centre and Research, Magdi Yacoub Foundation. This application will potentially help closing gap between Radiology and Electrophysiology departments by offering means to validate relation between voltage maps used in Electrophysiology and contours of heart acquired in Radiology.

Project Output

Provide automated platform to detect, visualize and quantify accurate location of myocardium scars on 3D model of left ventricle from cardiac MR images as input

Illustrated Description

• Physicians in Electrophysiology department identify scars by using of an epicardial needle and measuring voltage signal of tissue to determine whether it is fibrous and signals are finally constructed as 3D electro-anatomic map as shown in previous figure which is time consuming.

• Physicians in Electrophysiology department identify scars by using of an epicardial needle and measuring voltage signal of tissue to determine whether it is fibrous and signals are finally constructed as 3D electro-anatomic map as shown in previous figure which is time consuming.

• Our solution is to combine two different types of images to benefit from various information extracted from each type. First Type is cine images, which have high resolution and large number of slices covering heart. Second type is LGE images, where scars are clearly identifiable. Then, we construct 3D model visualizing scars of heart accurately, which can later be used in Electrophysiology to make heart mapping operations more time efficient.

Funding Opportunities

This project has applied for funding from:

• Faculty of Engineering, Cairo University
• Information Technology Industry Development Agency
• National Telecommunications Regulatory Authority
• Academy of Scientific and Research Technology