EVCMR: A tool for the quantitative evaluation and visualization of cardiac MRI data

Area of Science:

• Cardiovascular Imaging
• Medical Image Analysis
• Computational Biology

Background:

• Quantitative evaluation of diseased myocardium in cardiac MRI is crucial for cardiovascular disease diagnosis and prognosis.
• Efficient post-processing and analysis of cardiac images require advanced user interfaces.

Purpose of the Study:

• To develop a custom user interface tool for quantitative evaluation of the left ventricle (LV) and myocardium from cardiac MRI.
• To integrate modules for cine, perfusion, late gadolinium enhancement (LGE), and T1 mapping data analyses.
• To implement 3D visualization for enhanced analysis.

Main Methods:

• Developed Python modules for cine, perfusion, LGE, and T1 mapping analyses.
• Implemented a 3D visualization module using the PyQtGraph library.
• Utilized U-net segmentation and manual contour correction for myocardial segmentation mask generation.
• Integrated C++ implementation with pybind11 for efficient T1 map calculation.

Main Results:

• U-net segmentation achieved a mean Dice score of 0.87 (±0.02) for diastolic myocardial segmentation.
• LV mass measurements showed excellent agreement with manual segmentation (ICC=0.97).
• Significant computational gain was observed in T1 map calculation.
• 3D visualization facilitated rapid user interaction with myocardial and scar transmurality data.

Conclusions:

• The custom tool provides efficient and comprehensive analysis of LV and myocardium from multi-parametric cardiac MRI data.
• It aids in generating labeled data for deep learning model training through effective segmentation.
• The tool has significant potential for clinical application in cardiovascular disease assessment.