Purpose: To evaluate the feasibility, accuracy, and reproducibility of a fast breath-hold magnetic resonance (MR) spectroscopic method for T2-corrected hepatic lipid measurement in phantoms and in humans.
Materials and methods: All experiments were institutional review board approved and HIPAA compliant; informed consent was obtained from all subjects. The 15-second breath-hold high-speed T2-corrected multiecho (HISTO) MR spectroscopic technique was developed to acquire multiple echoes in a single acquisition, which enables the quantification of water and lipid T2, and subsequently to provide a corrected measure of hepatic lipid fraction. The accuracy of T2-corrected MR spectroscopy was evaluated in eight lipid phantoms doped with iron to simulate variable T2 effects. The mean absolute error of the HISTO technique with the known lipid amounts, as well as with uncorrected MR spectroscopic measures, was evaluated. The HISTO sequence was performed in 25 male subjects (mean age, 23.0 years +/- 19.2 [standard deviation]) to evaluate measurement bias with conventional, uncorrected MR spectroscopy. Three additional male subjects (mean age, 30.0 years +/- 1.0) were examined to assess reproducibility by using analysis of variance testing within subject and between separate imaging sessions.
Results: The absolute error in quantifying lipid fraction by using iron-doped lipid phantoms was less than 11% for the HISTO technique, compared with more than 50% for uncorrected MR spectroscopy. In the 25 human subjects, hepatic lipid measured by using HISTO differed significantly from that by using uncorrected MR spectroscopic methods by 5.1% +/- 2.6. Analysis of variance of three separate imaging sessions with the HISTO technique indicated no significant variance (P = .13) in three subjects.
Conclusion: HISTO is an accurate, reproducible MR spectroscopic sequence for quantifying hepatic lipid noninvasively. Evidence has shown this method to be feasible in vivo for clinical use.