Cardiac T1 mapping: A comparison of methodologies for quantifying cardiac T1 values

Results For the phantom validation study, T1 values from MLLSR, IR SSFP, CINE IR and IR FSE for all flip angles and heart rates are shown in Figure 1 (A-I). The mean correlation coefficient between three methods versus IR FSE was 0.9952, 0.9855, and 0.9797 (HR=60bpm); 0.9955, 0.9473, 0.9684 (HR=75bpm); and 0.9955, 0.8666, 0.8757. With T1 values beyond 600ms and heartbeats higher than 75bpm, only MLLSR showed consistent and reliable correlation above 99%. A volunteer pre-contrast study is shown in Figure 1 (J-L). With Cine-IR and IR-SSFP each TI image is acquired at a different cardiac phase, so cardiac motion limits the ability to do pixel-wise T1 mapping, as can be seen from Figures 1K and 1L. Approaches where only TI images corresponding to diastole have been attempted with some success, though these are still limited by the time window over which signal recovery is evaluated and by residual motion-induced blurring. Whereas the phantom results (A-I) demonstrate fundamental limitations of the techniques for accurately estimating T1 values, the in vivo data illustrates the sensitivity of each technique both to the inherent error in the technique for estimating long T1 values in addition to sensitivity to cardiac motion.