Melanie Freed1,2, Jacco A. de Zwart3,
Jennifer T. Loud4, Riham H. El Khouli5, Mark H. Greene4,
Brandon D. Gallas1, Kyle J. Myers1, Jeff H. Duyn3,
David A. Bluemke5, Aldo Badano1
1CDRH/OSEL/DIAM, FDA, Silver Spring,
MD, United States; 2Department of Bioengineering, University of
Maryland, College Park, MD, United States; 3NINDS/LFMI/Advanced
MRI Section, National Institutes of Health, Bethesda, MD, United States; 4NCI/Clinical
Genetics Branch, National Institutes of Health, Rockville, MD, United States;
5Department of Radiology and Imaging Sciences, National Institutes
of Health, Bethesda, MD, United States
We
are developing a physical, tissue-mimicking phantom to be used for
task-based, quantitative assessment of breast MRI protocols. Here we present initial results of the
phantom characterization and comparison to human data. Measured T1 and T2
relaxation values of the adipose- and glandular-mimicking phantom components
agree with human values from the literature.
The structure of human and phantom images is compared using the covariance
kernel and found to match within patient variation.