Abstract
1672
Objectives Cardiac and respiratory motions make PET detection of coronary plaque extremely challenging. Fat surrounding coronary arteries makes it highly advantageous to use MRI to track fat and plaque motion. We assess the feasibility of coronary plaque imaging using fat MR-based motion correction in simultaneous PET-MR.
Methods Cardiac MR of a CAD subject was acquired using a turbo field echo sequence with fat-water separation and respiratory/cardiac gating. Two 2.5x2.5x3mm plaques were digitally inserted within coronary vessels in the fat volume. Respiratory and cardiac motion fields corresponding to 6 and 8 phases, respectively, were generated using the XCAT torso phantom and used to obtain patient water/fat MR volumes in all 48 gates. The original MR volume was transformed into tissue activity maps (myocardium:blood=2:1) and inserted into XCAT for PET Monte Carlo simulations using GATE and the WB Siemens mMR geometry. Sixteen noise realizations of lesion-absent (1:1) as well as 3 plaque-to-blood activity ratios (PBR) of 10, 15 and 25:1 were simulated for all 48 cardio-respiratory phases. Cardio-respiratory vessel motion fields were obtained using non-rigid registration of transformed fat MR volumes and incorporated into the PET system matrix for iterative reconstruction along with motion-dependent attenuation maps. Reconstruction was performed with (MC) and w/o motion correction (nMC). Channelized-Hotelling observer (CHO) SNR was used to objectively quantify plaque detectability.
Results Fat MR-based MC yielded PBR>1.2, which is the marginal lesion visual detectability threshold in our study. Compared to nMC, MC significantly improved plaque-to-blood contrast by 76-79%, 84-98% and 87-159% for PBR of 10, 15 and 25, respectively (p<1e-5). Likewise, MC significantly improved plaque detection SNR by 152-177% for all PBRs (p<1e-5).
Conclusions Fat MR-based motion correction greatly improves plaque contrast and detectability and has great potential for coronary plaque imaging using simultaneous PET-MR.