Robust motion correction for cardiac T1 and ECV mapping using a T1 relaxation model approach

Area of Science:

• Cardiovascular Imaging
• Medical Physics
• Image Processing

Background:

• Quantitative myocardial tissue characterization using T1 and ECV mapping in cardiac MRI is crucial for diagnosing diffuse myocardial diseases.
• Accurate pixel-by-pixel calculation of these maps requires precise spatial correspondence between MRI images.
• Motion artifacts from cardiac, respiratory, or patient movement necessitate robust retrospective motion correction methods.

Purpose of the Study:

• To develop and validate a novel robust non-rigid registration framework for retrospective motion correction in cardiac MRI T1 and ECV mapping.
• To improve the accuracy of T1 and ECV quantification by addressing motion-induced spatial misalignments.
• To enable more reliable diagnosis of diffuse myocardial diseases through enhanced image analysis.

Main Methods:

• A new non-rigid registration framework combining data-driven initialization with a model-based approach utilizing T1 relaxation properties.
• Registration of native and contrast-enhanced T1-weighted images using T1 model-fitting information to generate motion-free ECV maps.
• Validation on three diverse datasets (MOLLI and STONE protocols) acquired under breath-hold and free-breathing conditions.

Main Results:

• Significant improvements in motion correction for single T1-weighted sequences: average Dice coefficient increased from 72.6% to 82.3% and mean boundary error decreased from 2.91mm to 1.62mm (P < 0.05).
• Enhanced accuracy in registration between native and enhanced T1 sequences: average Dice coefficient improved from 63.4% to 79.2% and mean boundary error reduced from 3.26mm to 1.77mm (P < 0.05).
• Substantial reduction in T1 and ECV mapping errors: native T1 SD error decreased from 67.32ms to 58.11ms, enhanced T1 SD error from 30.15ms to 22.74ms, and ECV SD error from 10.08% to 5.42% (P < 0.05).

Conclusions:

• The proposed non-rigid registration framework effectively corrects motion artifacts in cardiac MRI T1 and ECV mapping.
• The method enhances spatial accuracy and reduces quantification errors, leading to more reliable myocardial tissue characterization.
• This advancement supports improved diagnostic capabilities for diffuse myocardial diseases using quantitative cardiac MRI techniques.