CMRSegTools

CMRSegTools segments and quantifies myocardial infarct lesions from Cardiac Magnetic Resonance (CMR) images to support quantitative analysis of reperfusion injury and evaluation of cardioprotective strategies.

Key Features:

• Advanced segmentation methods: Provides access to state-of-the-art segmentation methods and adjustable parameters for myocardial infarct analysis.
• Algorithm integration: Supports integration of new algorithms and implementations using ITK and VTK libraries in C++.
• Segmentation capabilities: Offers user-assisted left ventricle segmentation, semi-automated and fully automated lesion segmentation, and statistical analysis and visualization based on the American Heart Association 17-segment model.
• Machine learning dataset generation: Facilitates creation of labeled datasets for supervised machine learning including lesion identification, microvascular obstruction, and remote regions of interest (ROI).
• Standardized results: Produces standardized outputs to enable comparative studies across clinical cohorts.
• Comparative assessment: Enables comparative evaluation of different lesion segmentation methods within a single integrated environment.

Scientific Applications:

• Clinical trials of cardioprotective therapies: Quantifies myocardial infarct size and characteristics in CMR datasets to assess treatment effects on reperfusion injury.
• Machine learning model development: Generates labeled cardiac MR datasets for training and validating supervised algorithms for lesion detection and classification.
• Method comparison and validation: Supports comparative assessment and validation of segmentation algorithms and parameters across cohorts.

Methodology:

Implemented in C++ using ITK and VTK, the software applies image processing techniques and statistical analysis including the American Heart Association 17-segment model.