Premium
Noise properties of proton density fat fraction estimated using chemical shift–encoded MRI
Author(s) -
Roberts Nathan T.,
Hernando Diego,
Holmes James H.,
Wiens Curtis N.,
Reeder Scott B.
Publication year - 2018
Publication title -
magnetic resonance in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.696
H-Index - 225
eISSN - 1522-2594
pISSN - 0740-3194
DOI - 10.1002/mrm.27065
Subject(s) - noise (video) , estimator , statistics , imaging phantom , signal to noise ratio (imaging) , probability density function , magnetic resonance imaging , mathematics , nuclear magnetic resonance , computer science , algorithm , physics , artificial intelligence , medicine , radiology , optics , image (mathematics)
Purpose The purpose of this work is to characterize the noise distribution of proton density fat fraction (PDFF) measured using chemical shift–encoded MRI, and to provide alternative strategies to reduce bias in PDFF estimation. Theory We derived the probability density function for PDFF estimated using chemical shift–encoded MRI, and found it to exhibit an asymmetric noise distribution that contributes to signal‐to‐noise‐ratio dependent bias. Methods To study PDFF noise bias, we performed (at 1.5 T) numerical simulations, phantom acquisitions, and a retrospective in vivo experiment. In each experiment, we compared the performance of three statistics (mean, median, and maximum likelihood estimator) in estimating the PDFF in a region of interest. Results We demonstrated the presence of the asymmetric noise distribution in simulations, phantoms, and in vivo. In each experiment we demonstrated that both the median and proposed maximum likelihood estimator statistics outperformed the mean statistic in mitigating noise‐related bias for low signal‐to‐noise‐ratio acquisitions. Conclusions Characterization of the noise distribution of PDFF estimated using chemical shift–encoded MRI enabled new strategies based on median and maximum likelihood estimator statistics to mitigate noise‐related bias for accurate PDFF measurement from a region of interest. Such strategies are important for quantitative chemical shift–encoded MRI applications that typically operate in low signal‐to‐noise‐ratio regimes. Magn Reson Med 80:685–695, 2018. © 2018 International Society for Magnetic Resonance in Medicine.