Fast water/fat <i>T</i> ₂ and PDFF mapping via multiple overlapping-echo detachment acquisition and deep learning reconstruction

<i>Objective.</i>Rapid and accurate quantitative assessment of muscle tissue characteristics is valuable for the diagnosis and monitoring of neuromuscular diseases (NMDs). Quantitative magnetic resonance imaging (MRI) enables non-invasive assessment of muscle pathology by using water<i>T</i>₂values to assess muscle damage and proton density fat fraction (PDFF) to quantify fat infiltration. However, conventional methods for simultaneous water-fat separation and<i>T</i>₂quantification often require long acquisition times. This study aims to develop an ultrafast method for simultaneous water-fat separation and<i>T</i>₂quantification.<i>Approach.</i>A novel water-fat separation framework that combines chemical shift encoding with the multiple overlapping-echo detachment sequence (CSE-MOLED) was proposed. Synthetic training data and deep learning-based reconstruction were employed to address challenges in water-fat separation, including the complex multi-peak spectral characteristic of fat and the non-idealities in MRI acquisition. All experiments, including phantom and<i>in vivo</i>scans, were performed on a 3T MRI scanner. The<i>in vivo</i>experiments focused on the human thigh and involved five healthy volunteers, one subject with muscle atrophy, and one with muscle damage. The CSE-MOLED sequence was acquired with a spatial resolution of 1.72 mm × 1.72 mm × 5 mm. For reference, mDixon-TSE (Turbo Spin Echo) was performed to generate PDFF, water<i>T</i>₂, and fat<i>T</i>₂maps.<i>Main results.</i>In numerical experiments (<i>T</i>₂range: water 19-160 ms, fat 15-160 ms; PDFF range: 14%-100%), the<i>R</i>²values were all 0.999 for water<i>T</i>₂, fat<i>T</i>₂, and PDFF, with average percentage errors of 3.16%, 0.98%, and 0.78%, respectively. In phantom experiments (<i>T</i>₂range: water 35-125 ms, fat 36-47 ms; PDFF range: 14-36%), the<i>R</i>²values were 0.995, 0.733, and 0.996 for water<i>T</i>₂, fat<i>T</i>₂, and PDFF, with average percentage errors of 5.56%, 2.93%, and 1.77%, respectively. High repeatability (coefficient of variation <2.0%) was achieved in both phantom and<i>in vivo</i>experiments. In patient scans, CSE-MOLED successfully distinguished between fat infiltration and muscle damage.<i>Significance</i>. CSE-MOLED simultaneously obtains<i>T</i>₂and proton density maps for both water and fat, along with<i>T</i>₂-corrected PDFF map, in 162 ms per slice of acquisition time, offering the potential to enhance the diagnostic accuracy of NMDs without increasing the clinical scanning burden.