Image registration in myocardial perfusion MRI

K Wong; E Wu; M Ng; Y Wu; H Tse; C Lau; G Lo; E Yang

doi:10.1109/IEMBS.2005.1616444

Image registration in myocardial perfusion MRI

Conf Proc IEEE Eng Med Biol Soc. 2005:2006:453-4. doi: 10.1109/IEMBS.2005.1616444.

Authors

K Wong¹, E Wu, M Ng, Y Wu, H Tse, C Lau, G Lo, E Yang

Affiliation

¹ HCNR Center of Bioinformatics and Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA.

PMID: 17282213
DOI: 10.1109/IEMBS.2005.1616444

Abstract

In clinical first-pass myocardial perfusion studies, physiological and patient motions are inevitable. Such motions impair the sensitivity and reliability in assessing myocardial perfusion abnormalities. The current study aims to correct the misregistration of myocardium during first-pass perfusion imaging by using a normalized mutual information approach. Multi-slice registrations were performed in 15 sets of clinical data sets. The proposed registration procedure was found to be fairly robust. With such registration, ROI time courses from normal myocardium and infarcted myocardium were improved, offering potential to quantify myocardial perfusion parameters in a more reliable manner.