Comparison of DeepStrain and Feature Tracking for Cardiac MRI Strain Analysis

Area of Science:

• Cardiovascular Imaging
• Medical Image Analysis
• Artificial Intelligence in Medicine

Background:

• Myocardial feature tracking (FT) analyzes myocardial deformation from cine bSSFP images.
• Current FT methods are time-consuming, limiting clinical application.

Purpose of the Study:

• To compare left-ventricular global radial strain (GRS) and global circumferential strain (GCS) using automated DeepStrain versus manual FT.
• Evaluate DeepStrain's accuracy and efficiency against established methods.

Main Methods:

• Retrospective analysis of cardiac MRI data from healthy subjects and patients.
• Comparison of DeepStrain automated analysis with manual FT (Circle, cvi42).
• Analysis performed on bSSFP sequences from Philips and Siemens scanners at 1.5T and 3T.

Main Results:

• High correlation and agreement for GRS and GCS between DeepStrain and manual FT across different scanners.
• DeepStrain analysis time was significantly reduced (7-fold) compared to manual FT (34.7 seconds vs. 4.1 minutes).

Conclusions:

• Automated DeepStrain provides accurate myocardial strain measurements comparable to manual FT.
• DeepStrain offers substantial time savings, facilitating clinical adoption of comprehensive myocardial deformation analysis.