Efficient phase-cycling strategy for high-resolution 3D gradient-echo quantitative parameter mapping

Magnetization-prepared (MP) gradient-echo (GRE) sequences suffer from signal contaminations from T-1 recovery during the readout train, which can be eliminated by paired RF phase cycling (PC) at the cost of doubling the scan time. The objective of this study was to develop and validate a novel unpaired PC strategy to eliminate the time penalty for high-resolution quantitative parameter mapping in 3D MP-GRE sequences. Based on the observation that the contaminating T-1 recovery signal along the GRE readout train is independent of magnetization preparation, its impact can be eliminated using a novel curve-fitting approach with complex-valued data without needing paired PC acquisitions. Four new unpaired PC schemes were compared with two traditional paired PC schemes in both phantom and in vivo human knee studies at 3 T using a MP angle-modulated partitioned k-space spoiled gradient-echo snapshots (MAPSS) T-1 rho mapping sequence. In the phantom study, all methods resulted in consistent T-1 rho measurements (Delta T-1 rho < 0.5%) at the center slice when B-0/B-1 values were uniform. Results were not consistent when off-center slices with nonideal B-0/B-1 were included. Two unpaired PC schemes had comparable or significantly improved quantitative accuracy and scan-rescan reproducibility compared with the paired PC schemes. There was no significant T-1 rho quantitative variability increase or spatial fidelity loss using the new unpaired PC schemes. Unpaired PC schemes also had different T-1 rho spectral responses at different B-0 frequency offsets, which can potentially be exploited to reduce sensitivity to B-0 field in homogeneities. The human knee study results were consistent with the phantom study findings. In conclusion, an unpaired PC strategy potentially allows more accurate quantitative parameter mapping with halved scan time compared with the paired PC approach to eliminate signal contaminations from T-1 recovery. It therefore offers additional flexibility in SNR optimization, spatial resolution improvement, and choice of imaging sampling points to obtain more accurate quantitative parameter mapping.