Abstract Purpose Respiratory motion during data acquisition may compromise image quality and quantitative accuracy of MRI, particularly in patients with limited breath-holding capability. This study aimed to compare hepatic proton density fat fraction (PDFF), T2*, and T2 estimates of the liver obtained from breath-hold (BH) acquisitions with those from corresponding free-breathing respiratory-gated (RG) MRI acquisitions. Methods In this IRB-approved study, healthy controls and patients with liver disease underwent quantitative liver MRI. Hepatic PDFF and T2* were estimated using a confounder-corrected chemical shift-encoded (mDIXON Quant) sequence, and T2 was estimated using a multi-echo Gradient and Spin Echo (mGRASE) sequence with fat suppression, each performed as both BH and RG acquisitions with identical acquisition parameters. Three observers independently drew regions of interest (ROIs) on PDFF, T2*, and T2 maps and agreement between mean values was compared using intraclass correlation coefficients (ICC) and Bland-Altman analysis. Results Nineteen participants (15.4 ± 5.5 years; 9 males) were evaluated. Scan times were 7 s and 14 s (BH) and 21 ± 6 s and 28 ± 4 s (RG) for mDIXON Quant and mGRASE respectively. Estimated values from BH acquisitions were: PDFF – median 5% (range 1–43%), T2* – mean 27 ms (range 9–41 ms), and T2 – mean 48 ms (range 34–63 ms). RG and BH acquisitions yielded comparable values with biases of −0.33% −1.42 to 0.76% for PDFF, −1.58 ms −9.50 to 6.33 ms for T2*, and 0.16 ms −2.43 to 2.12 ms for T2, and strong correlations (ICC = 0.99 0.99, 0.99, 0.85 0.66, 0.94, and 0.99 0.97, 0.99 for PDFF, T2*, and T2). Conclusion RG acquisitions provide hepatic PDFF, T2*, and T2 estimates comparable to BH acquisitions without substantial time penalty and are a viable alternative for patients with limited breath-holding ability.
Ramaniharan et al. (Wed,) studied this question.