Image registration in myocardial perfusion MRI

Area of Science:

• Cardiovascular Imaging
• Medical Image Analysis
• Radiology

Background:

• Physiological and patient motion significantly degrade the quality of first-pass myocardial perfusion imaging.
• Motion artifacts reduce the sensitivity and reliability of detecting myocardial perfusion abnormalities.

Purpose of the Study:

• To develop and evaluate a robust method for correcting myocardial misregistration in first-pass perfusion imaging.
• To enhance the accuracy of quantitative myocardial perfusion assessment.

Main Methods:

• A normalized mutual information approach was employed for multi-slice registration of cardiac MRI data.
• The method was tested on 15 clinical datasets, assessing its robustness.

Main Results:

• The proposed registration procedure demonstrated considerable robustness across clinical datasets.
• Improved time courses were observed for regions of interest in both normal and infarcted myocardium.
• The correction facilitated more reliable quantification of myocardial perfusion parameters.

Conclusions:

• Normalized mutual information-based registration effectively corrects motion-induced misregistration in first-pass myocardial perfusion studies.
• This approach enhances the diagnostic potential of myocardial perfusion imaging for improved patient care.