J. Kevin DeMarco1,
Hideki Ota2, David Zhu1, William Kerwin3,
Chun Yuan3
1Michigan
State University, East Lansing, MI, United States; 2Diagnostic
Radiology, Tohoku University Hospital, Sendai, Miyagi, Japan; 3University
of Washington, Seattle, WA, United States
The purpose of this study was to test the feasibility of automated segmentation of two non-contrast MR sequences to predict the size of necrotic-core identified by manual segmentation of multi-contrast carotid plaque MRI. Forty-eight arteries from 26 consecutive patients were analyzed. Automated segmentation identified all 16 carotid plaques harboring moderate/large necrotic-core with 6 additional false positive studies and excluded 54% of the carotid arteries as not having significant necrotic core. The addition of 3D IR-FSPGR to clinical MRA evaluation of carotid stenosis could help identify patients potentially harboring moderate/large necrotic-core who might be candidates for new aggressive lipid-lowering therapies.
Keywords