Unsupervised inline analysis of cardiac perfusion MRI.

Area of Science:

• Medical Imaging
• Cardiovascular Imaging
• Magnetic Resonance Imaging

Background:

• Automated analysis of cardiac perfusion MRI faces significant technical challenges.
• Existing methods often require manual intervention, limiting efficiency and scalability.

Purpose of the Study:

• To present a fully unsupervised workflow for automated cardiac perfusion MRI analysis.
• To address the technical challenges hindering efficient image processing and parameter mapping.

Main Methods:

• Key-frame detection and consecutive motion compensation.
• Surface coil inhomogeneity correction using proton density images.
• Robust, pixel-wise perfusion parameter map generation.

Main Results:

• The entire processing chain was implemented on clinical MR systems for inline analysis.
• Validation on 260 perfusion time series demonstrated the approach's feasibility.
• Achieved unsupervised, automated analysis of cardiac perfusion MRI data.

Conclusions:

• The proposed unsupervised workflow effectively addresses technical challenges in cardiac perfusion MRI.
• This automated approach enables feasible inline analysis on clinical systems.
• Facilitates robust generation of pixel-wise perfusion parameter maps.