Answers are grounded in this paper and cite their sources.
PURPOSE: Quantitative imaging on mid-field (0.55 T) MRI systems can suffer from poor precision due to low SNR. This work presents an MRF pulse sequence optimized with Cramer-Rao lower bound (CRLB) and combined with a multi-echo deep image prior (DIP) reconstruction for co-registered T1/T2/T2*/PDFF mapping in the liver at 0.55 T within a single breath-hold. METHODS: A 20 s 2D multi-echo rosette MRF breath-hold acquisition was implemented on a 0.55 T wide-bore system. The sequence design was optimized via pseudorandom search using a CRLB-based objective for T1/T2 encoding and compared with reference designs. Data were reconstructed using (i) an iterative low-rank (LR) reconstruction and (ii) a multi-echo DIP jointly modeling TR- and echo-domain dynamics including B0 effects. Water/fat separation preceded dictionary matching for T1/T2; curve fitting was used to calculate T2* and PDFF was derived from M0 with noise-bias correction. The method was evaluated in numerical simulations, NIST and fat-fraction phantom scans, and in vivo acquisitions in 10 healthy subjects. A linear mixed-effects model was used to evaluate the effects of sequence design and reconstruction method on mapping accuracy and precision. Phantom experiments showed agreement with reference values. In vivo, both sequence optimization and DIP reconstruction significantly reduced the standard deviation of all four parameters in liver ROIs. CONCLUSION: CRLB-guided acquisition and multi-echo DIP reconstruction enable accurate and precise liver MRF T1/T2/T2*/PDFF mapping at 0.55 T in a single breath-hold.
No takes yet. Share an insight, caveat, or question.
Griesler et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: