Instantaneous Abdominal T₂* Mapping via Single-Shot MOLED Under Free-Breathing: A Preliminary Study of Hepatic...

To propose a gradient-echo multiple overlapping-echo detachment (GRE-MOLED) method for rapid abdominal T₂* mapping, and to systematically validate its efficacy in non-invasive monitoring of dynamic hepatic glycometabolism.

The GRE-MOLED sequence was optimized to achieve ultrafast abdominal T₂* mapping under free-breathing without respiratory gating, the scan time per slice was only 60 ms. A deep neural network architecture was developed and trained on synthetic data generated through Bloch equation simulations to achieve direct end-to-end reconstruction of T₂* maps from GRE-MOLED images. The proposed method was evaluated through numerical abdomen experiments, water phantom experiments, test-retest experiments in healthy volunteers, and dynamic glycometabolic experiments.

Quantitative analyses revealed excellent agreement between the results of our method and reference method. Numerical abdomen experiment achieved mean structural similarity index (SSIM) of 0.9818 and voxel-wise R² = 0.9163. Water phantom experiments demonstrated high linear correlation (R² = 0.9992) with a Bland-Altman bias of 0.2686 ms (95% limits of agreement: -1.1048 to 1.6420 ms). Test-retest analysis showed intra-subject coefficient of variation (CV) below 7% across multiple organs. Dynamic glycometabolism mapping successfully captured postprandial hepatic T₂* fluctuations with temporal resolution outperforming conventional breath-holding methods.

The GRE-MOLED method enables robust abdominal T₂* mapping under free-breathing, demonstrating obvious potential for non-invasive assessment of hepatic glycometabolic dynamics. Validation across numerical simulations, water phantom and in vivo experiments confirm technical reliability and high reproducibility.