BJR © 2018 The Authors.

Published by the British Institute of Radiology

https://​doi.​org/​10.​1259/​bjr.​20170959
Received: Revised: Accepted:
12 December 2017 29 March 2018 19 April 2018

Cite this article as:

Zhang YN, Fowler KJ, Hamilton G, Cui JY, Sy EZ, Balanay M, et al. Liver fat imaging—a clinical overview of ultrasound, CT, and MR
imaging. Br J Radiol 2018; 91: 20170959.

The role of imaging in obesity special feature: Review

Article
Liver fat imaging—a clinical overview of ultrasound, CT,
and MR imaging
1
Yingzhen N Zhang, MD, 2Kathryn J Fowler, MD, 1Gavin Hamilton, PhD, 1Jennifer Y Cui, BS,
1
Ethan Z Sy, BS, 1Michelle Balanay, BA, 1Jonathan C Hooker, BS, 1Nikolaus Szeverenyi, PhD and
1
Claude B Sirlin, MD
1
Department of Radiology, Liver Imaging Group, University of California San Diego, San Diego, CA, USA
2
Department of Radiology, Washington University School of Medicine, Washington University, St. Louis, MO, USA

Address correspondence to: Dr Claude B Sirlin

E-mail: ​csirlin@​ucsd.​edu

Abstract
Hepatic steatosis is a frequently encountered imaging finding that may indicate chronic liver disease, the most common
of which is non-alcoholic fatty liver disease. Non-alcoholic fatty liver disease is implicated in the development of systemic
diseases and its progressive phenotype, non-alcoholic steatohepatitis, leads to increased liver-specific morbidity and
mortality. With the rising obesity epidemic and advent of novel therapeutics aimed at altering metabolism, there is a
growing need to quantify and monitor liver steatosis. Imaging methods for assessing steatosis range from simple and qual-
itative to complex and highly accurate metrics. Ultrasound may be appropriate in some clinical instances as a screening
modality to identify the presence of abnormal liver morphology. However, it lacks sufficient specificity and sensitivity to
constitute a diagnostic modality for instigating and monitoring therapy. Newer ultrasound techniques such as quantita-
tive ultrasound show promise in turning qualitative assessment of steatosis on conventional ultrasound into quantitative
measurements. Conventional unenhanced CT is capable of detecting and quantifying moderate to severe steatosis but is
inaccurate at diagnosing mild steatosis and involves the use of radiation. Newer CT techniques, like dual energy CT, show
potential in expanding the role of CT in quantifying steatosis. MRI proton-density fat fraction is currently the most accu-
rate and precise imaging biomarker to quantify liver steatosis. As such, proton-density fat fraction is the most appropriate
noninvasive end point for steatosis reduction in clinical trials and therapy response assessment.

Introduction with steatosis severity.11,12 A smaller subset of NAFLD

Fatty liver, or hepatic steatosis, refers to the abnormal accu- patients with advanced stage fibrosis or NASH also have
mulation of triglycerides (TG) within hepatocytes.1 While increased liver-related morbidity and mortality consequent
potentially an inconsequential or self-limited finding, to the higher risk of progression to cirrhosis and cirrhosis-
hepatic steatosis is also associated with chronic liver disease, related liver transplantation.2,6,13–16 In addition to histo-
the most common of which is non-alcoholic fatty liver logical fibrosis and NASH, changes in liver steatosis may
disease (NAFLD). NAFLD comprises two main pheno- also impact NAFLD progression. Recent studies found that
types: nonalcoholic fatty liver (NAFL), which is thought steatosis severity correlates with the risk of fibrotic progres-
to have little histological progression over time, and non- sion in NAFLD and regression in NASH, and that a reduc-
alcoholic steatohepatitis (NASH), which is thought to be tion in steatosis severity is associated with improvement in
the more progressive form with higher risk of evolution to NASH.17,18
cirrhosis and its complications.2 An estimated 6–26% of all
NAFLD patients have NASH.3–5 The systemic and hepatic diseases associated with NASH
and progressive NAFLD warrant accurate detection and
Patients with NAFLD have increased overall mortality staging of these conditions. In particular, distinguishing
than the general population, with cardiovascular compli- between physiological vs pathological fat accumulation
cations being the leading cause of death followed by and longitudinal monitoring for treatment response are
metabolic and liver-related causes.6–10 The increased desired. Liver biopsy is the clinical reference standard
cardiovascular risk amongst NAFLD patients correlates for the assessment of NAFLD and histological evaluation
 BJR Zhang et al

includes features not currently detectable on imaging, such as Ultrasound

specific patterns of inflammation and hepatocyte injury seen Normal liver parenchyma is the same as or slightly more echo-
in NASH.2,19 Histological steatosis is graded on a semi-quan- genic (“brighter”) than the adjacent kidney and spleen.25 Ultra-
titative scale based on the number of hepatocytes containing sound beam scattering by lipid droplets in steatosis causes more
microscopically discernible cytoplasmic fat droplets: 0 (<5% echo signals to return to the transducer, creating the appearance
hepatocytes), 1 (5–33% hepatocytes), 2 (33–66% hepatocytes), of a “bright” or hyperechoic liver.26 Fat also attenuates the beam
and 3 (>66% hepatocytes).19,20 However, liver biopsy is observer which decreases beam penetration into tissue. This attenuation
dependent and invasive, conveying non-negligible risk of signifi- leads to poor visualisation of structures within the steatotic liver
cant morbidity and mortality.21,22 The relatively small core size of parenchyma—such as intrahepatic vessels, bile ducts and in some
biopsy also introduces sampling errors, especially as steatosis is
cases liver lesions26—and of structures deep to the liver, such as
known to be heterogeneous.23,24 These shortcomings make liver
the diaphragm. Thus, the presence of steatosis can be inferred
biopsy a suboptimal tool for screening, monitoring, and research.
if the liver is too bright and/or if liver structures are blurry or
poorly visualised.
As a non-invasive alternative to liver biopsy, imaging is increas-
ingly utilised in the diagnosis and management of NAFLD.
In addition to diagnosing steatosis, ultrasound can be used
Imaging and related non-imaging techniques can accurately
assess the important disease markers of liver steatosis and to grade the severity of steatosis by scoring the degree of liver
advanced liver fibrosis. Non-invasive techniques for staging liver brightening and/or blurring of vessels and diaphragm. See
fibrosis is beyond the scope of this review. Conventional tech- Figure 1 for examples. To avoid inaccurate assessment due to
niques for evaluating steatosis include ultrasound, CT and MR parameters of acquisition (frequency, gain, etc.), liver bright-
spectroscopy and MRI. This article outlines the performance ness is assessed by comparing to an internal standard such as the
and clinical utility of each modality. Future directions including kidney or the spleen.26 Ultrasound performs best at qualifying
quantitative approaches will also be discussed. liver steatosis when there is no other background liver disease;
however, it remains relatively insensitive to detection of mild
steatosis. See Table 1 for a summary of studies on ultrasound
diagnostic accuracy. The sensitivity and specificity of ultra-
sound at detecting moderate to severe steatosis, using histology

Figure 1. Ultrasound, CT and MR at steatosis—examples. B-mode ultrasound transverse images of the liver (first row), axial unen-
hanced CT images of the liver at the level of the spleen (second row), and axial MRI PDFF images of the liver (third row) are
shown for four patients. Steatosis grade was determined at liver biopsy with direct histological visualisation for the number of cells
with intracellular fat vacuoles: none (0% hepatocytes), mild (0–33% hepatocytes), moderate (33–66% hepatocytes) and severe
(>66% hepatocytes). As steatosis grade increases from left to right in each row, the following patterns are seen: on ultrasound,
increased liver parenchyma echogenicity and decreased definition of intrahepatic structures such as vessel walls; on unenhanced
CT, liver density on CT in HU decreases though spleen density in HU is variable; on MR, PDFF values increase. HU, Hounsfield
unit; PDFF, proton-density fat fraction.

2 of 12 birpublications.org/bjr Br J Radiol;91:20170959
 Review article: Liver fat imaging—a clinical overview of ultrasound, CT and MRI BJR

Table 1. Studies on diagnostic accuracy of ultrasound on steatosis

Author, year Reference

Design N Indication Sensitivity Specificity
(reference) standard
Palmentieri et al27 Prospective 235 Suspected liver Liver biopsy Steatosis ≥5%: 0.64 Steatosis ≥5%: 0.97
disease Steatosis ≥30%: 0.93 Steatosis ≥30%: 0.93
Lee et al28 Prospective 161 Potential liver Liver biopsy Steatosis ≥5%: 0.62 Steatosis ≥5%: 0.81
donors Steatosis ≥30%: 0.82 Steatosis ≥30%: 0.98
van Werven et al29 Prospective 46 Liver resection Liver biopsy Steatosis >5%: 0.65 Steatosis >5%: 0.77
30
Hernaez et al Meta-analysis: 49 studies 4720 Suspected/known Liver biopsy Steatosis >5%: 0.65 Steatosis >5%: 0.81
from 1967 to 2010 liver disease Steatosis ≥20–30%: Steatosis ≥20–30%:
0.91 0.99
Bril et al31 Prospective 146 High BMI with Liver biopsy Steatosis >12.5%: 0.85 Steatosis >12.5%:
suspected NAFLD and MRS 0.70
BMI, body mass index; MRS, magnetic resonance spectroscopy; N, sample size; NAFLD, non-alcoholic fatty liver disease.

as reference standard, are 80–89 and 87–90%, respectively.27–31 a grossly abnormal ultrasound indicates the presence of disease
The sensitivity and specificity drops to 65 and 81%, respectively, including steatosis, but a diagnostic test such as CT, MRI or
when all grades of steatosis are considered.27 biopsy may be considered as a next step to differentiate between
disease states (fibrosis and fat) and to accurately quantify severity.
Advantages and limitations For clinical trials, ultrasound is not considered a reliable tool to
Advantages of ultrasound include safety, wide availability, and accurately assess liver steatosis. Ultrasound lacks sufficient preci-
little associated patient discomfort.32,33 The relative cost of sion for longitudinal measurements. As quantitative ultrasound
abdominal ultrasound is low compared to CT or MR. Unlike CT techniques emerge, they may play a more central role in clinical
and MRI, liver iron has little effect on the ultrasound beam.34 and research usage. However, there is insufficient evidence to
make a recommendation regarding their use at this time.
Ultrasound has several disadvantages for steatosis detection and
grading. In patients with large body habitus, steatosis may be over- Future directions
estimated due to beam attenuation by overlying fat rather than liver Quantitative ultrasound is a technique designed to address the
fat. Echogenicity of the liver may be confounded by fibrosis, inflam- subjectivity, operator- and machine-dependency, and diagnostic
mation, and other features of chronic liver disease.27 Fibrosis and and grading inaccuracy of conventional ultrasound on steatosis.
fat can superficially resemble each other by causing coarsening of With quantitative ultrasound, two quantitative parameters—
the echotexture and increased echogenicity of the liver.34 In prin- attenuation coefficient, which is analogous to obscuration of liver
ciple, fat causes more vessel wall blurring and beam attenuation structures, and backscatter coefficient, which is analogous to
than fibrosis, but qualitative assessment of such differences is prone echogenicity—are estimated. Calibrated tissue phantoms are used
to misclassification. Thus, in the setting of chronic liver disease, it to address machine and operator variability. Importantly, quanti-
may be difficult to ascertain the extent that hyperechogenicity is tative ultrasound can be implemented on any clinical ultrasound
attributable to steatosis, fibrosis, or both.35 system and is acquired during conventional ultrasound, adding
negligibly to examination time. Preliminary studies have shown
Ultrasound yields relatively imprecise qualitative classifications that quantitative ultrasound can diagnose and grade hepatic
of mild, moderate, and severe steatosis. Additionally, conven- steatosis more accurately than conventional ultrasound38,39 and
tional ultrasound is operator- and reader-dependent, resulting in has superior intra- and inter-observer agreement.39,40
variable results and reproducibility. Among patients with known
or suspected steatosis, the intra- and inter-reader correlation Other possibilities for improvement include examining the effect
ranged from 0.5 to 0.6 and 0.4–0.5, respectively.34,36 In addition, of variable operator expertise and acquisition parameters on the
liver steatosis can be diffuse, focal or mixed; but ultrasound may accuracy of ultrasound evaluation of steatosis. Development
not visualise the entire liver due to shadowing from ribs, gas, of methods to improve agreement between readers, such as a
and other patient factors.34,37 Finally, ultrasound measurements training atlas, is another area under investigation. Automated
are indirect indices of fat and hence, the values depend on cali- tools for detecting steatosis and improved technology allowing
bration and acquisition parameters. This can lead to variations better penetration for patients with higher body mass index
between manufacturers, machines and operators that confound would also potentially improve results.
interpretation of results.
Computed Tomography
Recommendations for clinical care and clinical Similar to ultrasound, attenuation is a relevant factor in deter-
trials mining final image brightness on CT (Figure 1). CT images are
Despite its many limitations, it may be reasonable to use ultra- generated by X-ray photons traversing tissues and exposing a
sound in the appropriate clinical setting as an initial screen for detector opposite the beam. The denser the tissue, the more atten-
steatosis. However, providers and patients should recognise that uated the X-ray is and the brighter the corresponding image pixel.

3 of 12 birpublications.org/bjr Br J Radiol;91:20170959
 BJR Zhang et al

Table 2. Studies on diagnostic performance of non-contrast conventional CT

Author, year Reference

Design N Indication Sensitivity Specificity
(reference) standard
Lee et al28 Prospective 161 Potential liver donors Liver biopsy Steatosis ≥5%: 0.50 Steatosis ≥5%: 0.77
Steatosis ≥30%: 0.73 Steatosis ≥30%: 0.91
van Werven et al29 Prospective 46 Liver resection Liver biopsy Steatosis >5%: 0.74 Steatosis >5%: 0.70
49
Park et al Prospective 154 Potential liver donors Liver biopsy Steatosis ≥30%: 0.91 Steatosis ≥30%: 0.97
47
Bohte et al Meta-analysis: 1721 Potential liver donors/ Liver biopsy Steatosis >0%: 0.46 Steatosis >0%: 0.94
12 studies from NAFLD/liver resection Steatosis >10%: 0.57 Steatosis >10%: 0.88
2001 to 2009 Steatosis >25%: 0.72 Steatosis >25%: 0.72
Saadeh et al34 Prospective 25 Biopsy-proven NAFLD Liver biopsy Steatosis >33%: 0.93 N/A
N, sample size; N/A, not applicable; NAFLD, non-alcoholic fatty liver disease.

CT scanners are calibrated to yield pixel value measurements rela- Advantages and limitations
tive to water using a unit of measurement known as the Hounsfield The main advantages of CT for assessing steatosis are relatively
Unit (HU). Water, by definition, is 0 HU, and air is defined as −1000 fast acquisition, ease of performance, straightforward analysis,
HU. At unenhanced CT, normal liver parenchyma is about 60 HU and quantitative results.
and hyperattenuates (appears brighter) relative to the spleen.41,42
With increased steatosis, the liver tissue becomes hypoattenuating Like ultrasound, however, CT cannot accurately diagnose mild
(darker) relative to the adjacent fat-free spleen.43,44 In cases of steatosis. CT uses tissue density as an indirect index of steatosis
severe steatosis, the normally hypoattenuating intrahepatic vessels and thus relies on calibration which is known to vary between
may appear bright relative to the steatotic liver and mimic the effect scanners, manufacturers, and reconstruction algorithms.48,51,52
of contrast enhancement.45,46 Like hyperechogenicity of the liver on ultrasound, X-ray beam
attenuation is not specific for steatosis. Liver density is influ-
Over the years, various criteria for diagnosing steatosis at unen- enced by the presence of materials such as iron, glycogen and
hanced CT have been proposed. Because materials other than less well-understood factors including haematocrit, copper and
fat, such as iron, can also affect the attenuation of the X-ray other metallic ions; all of these can alter X-ray beam attenua-
beam,34 and because the calibration of HU varies by scanner and tion.45,53 Likewise, the spleen is an imperfect reference standard
manufacturer, some investigators have advocated comparing the as it can be affected by haemosiderosis and haemochromatosis
liver to the spleen which serves as an internal standard.41 Abso- in a small minority of patients.45 The use of ionising radiation is
lute liver HU less than 40 or liver-minus-spleen difference of less also a drawback. Finally, the vast majority of CT examinations
than −10 HU have been used to diagnose steatosis with reported performed for clinical care are performed following intravenous
sensitivity and specificity ranging 46–72% and 88–95%.29,47 contrast injection.45 Quantification of steatosis on conventional
Retrospective evaluations of steatosis at unenhanced CT have post-contrast images involves specific contrast protocols and
established absolute liver HU less than 48 as highly specific for imaging delay which limits its utility as a standard metric. While
moderate to severe steatosis.48 See Table 2 for a summary of available as research protocols, methods designed to subtract
studies on the diagnostic accuracy of CT. Like ultrasound, the iodine from contrast-enhanced studies have not been fully vali-
diagnostic performance of CT decreases with lesser severity dated or widely utilised in clinical settings.54,55
of steatosis. In mild steatosis with fat fraction of 10–20%, CT
diagnostic sensitivity ranges from 52 to 62%.47 As a quantita-
Recommendations for clinical care and clinical
tive method, liver HU at unenhanced CT has demonstrated an
inverse linear relationship with MR spectroscopic proton-density trials
fat fraction (PDFF), an MR-related biomarker currently accepted Due to exposure to ionising radiation and low sensitivity for mild
as the noninvasive reference standard for steatosis quantification steatosis, we would not recommend CT as a primary modality for
(see more under section “Magnetic Resonance”).50 Contrast- measuring liver steatosis. Ultrasound and MRI are better alterna-
enhanced CT utilising a liver-minus-spleen difference of less than tives. If CT is done for other purposes, then we recommend that
or equal to 19 HU has been found to diagnose moderate to severe radiologists assess for steatosis using conservative thresholds.
steatosis with modest sensitivity and high specificity on portal In addition, the widespread availability and quantitative metric
venous phase post-contrast.51 However, contrast-enhanced CT is of CT make it potentially useful for identifying patients with
generally not used for clinical assessment of steatosis due to the steatosis in retrospective studies.
overlap in HU between normal and abnormal liver tissues and
to the HU dependence on scan delay and contrast protocol.52,53 To diagnose steatosis at unenhanced CT, no one method is clearly
superior. Measuring absolute liver attenuation with a threshold
of equal to or less than 40 HU raises specificity but lowers sensi-
tivity. On the other hand, measuring the liver to spleen attenu-
ation ratio may increase the number of false positive cases. Fat
sparing on CT would be the one unequivocal finding for the

4 of 12 birpublications.org/bjr Br J Radiol;91:20170959
 Review article: Liver fat imaging—a clinical overview of ultrasound, CT and MRI BJR

presence of steatosis.37 If there is fat-sparing around the gall- including fat, to decompose images specific for different material
bladder fossa or other such characteristic fat-sparing patterns, composition (Figure 2). Thus, “fat maps” may be recreated from
then there is at least some degree of steatosis present in at least a study done at different energy levels.56 Likewise, the character-
some parts of the liver.37 istic attenuation profile of iodinated contrast may be utilised to
“subtract” contrast enhancement from studies and create virtual
Future directions unenhanced images.50
Recent advances in CT technology, such as dual energy CT
(DECT), show promise in separating the fat component from Although it remains an investigational technique, DECT has
water and in reconstructing virtual unenhanced CT images. shown promising results in some studies. Hyodo et al demon-
This technique exploits the observation that different tissues strated that using a method of decomposition CT, it was possible
have characteristic attenuation profiles across a range of photon to quantify fat fraction by volume.56 This is potentially useful for
energies.50 Most tissues exhibit decreased attenuation as the clinical care in which patients may have images acquired only
incident photon energies increase.8 In contrast, fat preferentially after contrast.56 DECT-based algorithm can be applied to both
attenuates high-energy photon over low-energy photons for the contrast-enhanced and non-contrast-enhanced images, though
photon range used in conventional CT imaging.8 This is thought DECT has yet to demonstrate clinical utility in the analysis of
to be due to the different mechanisms of X-ray beam interfer- single-energy unenhanced CT. Further validation of DECT
ence at different energy levels. At lower photon energies, the is required prior to routine use in the clinical setting at this
photoelectric effect predominates. Since fat has more hydrogen time.50,56
atoms (lower effective atomic number) than other soft tissues
and the magnitude of the photoelectric effect correlates with
the effective atomic number, fat causes less attenuation at lower Magnetic Resonance Imaging
photon energies.53 At higher photon energies, the Compton MRI is considered the most sensitive and specific technique for
effect, which scatters the incident beam, predominates. Since assessing steatosis. Unlike ultrasound and CT, which measure
the Compton effect correlates with electron density and fat is steatosis by proxy, MRI measures the signal intensity (bright-
more electron-dense relative to other tissue types, fat prefer- ness) of protons at different resonance frequencies.57 Water reso-
entially attenuates high-energy photons.53 This results in the nates at a single frequency, while TG in steatosis exhibits more
observation that as the tube voltage used to acquire the projec- complex behaviour (Figure 3).57,58 MRI exploits the difference
tion data increases, fat becomes denser on CT images even as in proton resonance frequencies of water and TG by acquiring
the surrounding soft tissue becomes less dense.50,53,55 DECT images at echo times at which water and TG are nominally in
uses the characteristic attenuation profiles of different tissues, and out of phase.57,59

Figure 2. Conventional unenhanced CT and DECT at steatosis– adapted from Kramer et al.50 Conventional unenhanced CT
acquired at 120 kVp (first row) and DECT acquired by rapidly switching tube voltages between 80 and 140 kVp, then post-pro-
cessed into fat-density images (second row) are shown for three patients with varying degrees of steatosis. Patients A, B, and C
have 0, 10, and 40% liver fat fraction, respectively, as determined by MRS PDFF (not shown). As liver fat fraction increases across
the rows, liver attenuation at conventional unenhanced CT visibly decreases and liver fat density on DECT visibly increases. 50
Reprinted with permission from the American Journal of Roentgenology. DECT, dual energy CT; PDFF, proton-density fat frac-
tion; MRS, MR spectroscopy.

5 of 12 birpublications.org/bjr Br J Radiol;91:20170959
 BJR Zhang et al

Figure 3. Typical liver MR spectrum showing water peak at 4.7 ppm (chemical shift measured in ppm) and multiple fat peaks
(Peaks 1–6). There is one main fat peak (Peak 5). There are also Peak 4 and Peak 6, which partially overlap with the main fat peak.
Peak 1 and Peak 2 overlap with the single water peak. Different correction techniques exist in advanced MR to address the problem
of teasing apart contributions from individual peaks. ppm, parts per million.

Hepatic steatosis assessment on MRI has evolved from early Yokoo et al recently performed a meta-analysis with histology as
methods that only gave qualitative estimates (i.e. dual echo the reference standard. MRS demonstrated superior diagnostic
chemical shift imaging) to more advanced and fully quantita- accuracy compared to other noninvasive methods for detecting
tive methods, meeting the ultimate goal of accurate and precise mild steatosis (histological grade <5–<10%) with sensitivity and
steatosis measurement. Identification and incorporation of the specificity ranging 77–95 and 81–97%, respectively.62
main confounders were essential steps. These confounders on
MRI are R2* decay, the complexity of the fat spectrum and T1 Prior studies in children and adults with known or suspected
weighting due to the differing T1 values of fat and water.57,60 NAFLD have shown that MRI-PDFF have high intra- and
Confounder corrected chemical shift-encoded MRI corrects for interexam repeatability across scanners and magnets.70–73
these three confounders by collecting images at multiple echo MRI-PDFF correlates strongly with both biochemically deter-
times with low flip angle to minimise T1 weighting and by incor- mined TG concentration and MR spectroscopy.34,72,74 See
porating the multi-peak structure of fat in the analysis algo- Table 3 for a summary of studies on the diagnostic accuracy
rithm.57,61,62 The end result is the PDFF, which is the (liver fat of MRI-PDFF. Using contemporaneous histology as the refer-
signal)/(total signal). ence standard, MRI-PDFF accurately classifies dichotomised
steatosis grades cross-sectionally, and change in PDFF accu-
The liver fat signal generated through PDFF imaging is due rately classifies change in steatosis longitudinally.18,39,47,64,75–80
almost entirely to protons in TG, which make up virtually all of MRI-PDFF demonstrates high interexam precision at all
the pathological fat accumulation in hepatic steatosis.63 While anatomic levels of the liver and accuracy in detecting as low
liver fat may contain trace contributions from other lipids, these as 1.6% change in fat fraction over time.69,72,81,82 Most MRI
are not detected on MR as they have ultrashort T2s due to being vendors offer FDA approved packages that can generate PDFF
bound in components of normal tissue, i.e. cell walls. Water also maps making it relatively accessible clinically.
has an invisible ultrashort T2 fraction, as it is bound to protein
such as collagen. Thus, PDFF signal can be defined as the frac- Advantages and limitations
tion of proton signal from mobile fat normalised by the total These advanced MRI techniques have many advantages over
proton signal from all mobile proton species. ultrasound or CT. Primarily, MRI measures the PDFF which is
a fundamental property of tissue and requires no internal cali-
Over the years, three classes of advanced technique have emerged: bration or reference standard. Advanced sequences can address
magnitude data-based MRI, complex data-based MRI, and MRS. biological confounders such as iron overload by simultaneously
Performed properly, they agree so closely with each other they measuring and correcting for R2*. The images needed for PDFF
can essentially be assumed to be equivalent (Figure 4).64–69 measurements can be acquired very quickly (the entire liver may
This review will not address MRS, as it is often used as a refer- be imaged in a single or two breath-holds). As such, volumetric
ence standard and is a biochemical rather than imaging-based assessment of steatosis is possible. This is not possible with ultra-
technique. For details on the diagnostic performance of MRS, sound or MRS.

6 of 12 birpublications.org/bjr Br J Radiol;91:20170959
 Review article: Liver fat imaging—a clinical overview of ultrasound, CT and MRI BJR

Figure 4. Complex data-based (c-MRI) and magnitude data-based (m-MRI) MR, acquisitions from 64-year-old female patient with
mild histological grade steatosis in Figure 1. Startinginary source images with the output PDFF map to the right. The echo times at
which the c-MRIs are acquired are listed on the far left. Further to the right are the magnitude source echoes for the m-MRI with
the corresponding PDFF map on the far right. The echo times at which the m-MRIs are acquired are adjacent to the magnitude
source echoes. The average PDFF measurement derived from c-MRI is 7.5% while that derived from m-MRI on the same patient is
7.4%. PDFF, proton-density fat fraction.

Despite its superior diagnostic performance to ultrasound and patient in the course of steatosis or other liver diseases, PDFF may
CT, MRI does have several drawbacks. Areas of motion and not accurately reflect the actual degree of steatosis in the patient.
parallel image artefact negatively impact measurement accuracy,
and so these regions need to be identified and avoided when Besides limitations related to fat fraction analysis, MRI can also
placing ROIs. Magnitude data-based MRI does not readily differ- be limited by patient factors, operators and institutions. While
entiate fat fraction greater than 50% from fat fraction less than commercially available on new MR platforms, the software pack-
50%.59 Though rare, human liver fat fraction does occasionally ages capable of processing PDFF maps may not be readily avail-
exceed 50%. Another issue is the limited ability of current MRI able due to budgetary and hardware constraints in some imaging
techniques for R2* correction. In cases of extreme iron overload, centres. MRI suitability may be limited by patient factors including
the signal loss may be so fast that it is not feasible to measure the claustrophobia, implanted devices, and discomfort. There is also
oscillation. a higher relative charge for MRI compared to ultrasound and CT.
This is an area for future research and development to help offset
Finally, we have limited knowledge about proton pool variability the cost.
between patients and within patients over time. One of the
implicit assumptions in PDFF is that the pool of protons invis-
ible on MR—e.g. those bound to cholesterol crystals and to water
in collagen—remains the same across and within patients. To the
extent that the invisible pool changes between or within the same

Table 3. Studies on diagnostic accuracy of MR with liver biopsy as reference standard

Author, year PDFF threshold

Design N Indication Sensitivity Specificity
(reference) (%)
Bohte et al47 Meta-analysis: 11 569 Potential liver donor/ N/A Grade > 0%: 0.82 Grade > 0%: 0.90
studies from 2001 NAFLD/liver resection Grade > 10%: 0.90 Grade > 10%: 0.95
to 2009 Grade > 25%: 0.97 Grade > 25%: 0.76
Idilman et al75 Retrospective 70 Biopsy-proven NAFLD Grade ≥ 2: 15% Grade ≥ 2: 0.93 Grade ≥ 2: 0.85
76
Tang et al Prospective 89 NAFLD Grade ≥ 1: 6.4 Grade ≥ 1: 0.86 Grade ≥ 1: 0.83
Grade ≥ 2: 17.4 Grade ≥ 2: 0.64 Grade ≥ 2: 0.96
Grade = 3: 22.1 Grade = 3: 0.71 Grade = 3: 0.92
Paige et al39 Prospective 61 Biopsy-proven NAFLD Grade ≥ 2: 13.45 Grade >= 2: 0.85 Grade >= 2: 0.96
Grade = 3: 16.83 Grade = 3: 1.00 Grade = 3: 0.81
Middleton et al., Multicenter RCT 110 Paediatric NAFLD Grade ≥ 2: 17.5 Grade ≥ 2: 0.74 0.90a
201777 clinical trial Grade = 3: 23.3 Grade = 3: 0.60
Middleton et al., Multicenter RCT 113 NASH clinical trial Grade ≥ 2: 16.3 Grade ≥ 2: 0.83 0.90a
201777 Grade = 3: 21.7 Grade = 3: 0.84
N, sample size; NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; PDFF, proton-density fat fraction; RCT, randomised
controlled trial.
a
PDFF threshold was chosen for a target specificity of 0.90.

7 of 12 birpublications.org/bjr Br J Radiol;91:20170959
 BJR Zhang et al

Recommendations for clinical care and clinical Future directions

trials In the last decade, PDFF has grown from an experimental method
Clinical care recommendations depend largely on availability being tested at a few research centres to a validated clinical stan-
and patient tolerance. If a PDFF technique is available, it should dard for steatosis assessment. The availability and adoption of
be the method of choice in all patients for whom assessment of PDFF technique has become widespread in the last few years, and
liver steatosis is clinically requested. The currently available soft- will continue to increase in response to the worldwide steatosis
ware packages can generate parametric maps in less than 30 s, epidemic. Future directions would involve addressing the various
making them easy to incorporate into routine clinical practice. known limitations of PDFF. Techniques are being investigated that
Importantly, the PDFF technique can be done after gadolinium can reliably measure PDFF in the setting of extreme iron over-
injection because these techniques are robust in the setting of the load. Improvement in our knowledge of the biological and clinical
T1 and T2* shortening effects of gadolinium. significance of PDFF values and their longitudinal change would
also make PDFF a better diagnostic tool. For instance, there is yet
For clinical trials, PDFF is the most accurate and precise imaging to be consensus on the cutoff value(s) that differentiate normal
measure of liver steatosis. Hence, PDFF should be used as an from abnormal and the clinical significance of the range of PDFF
end point for clinical trials for inclusion criteria and in settings values. Investigators are also trying to automate PDFF analysis and
where quantifying steatosis is relevant. Since commercial variants to improve the accuracy and precision of PDFF in the low-fat frac-
of PDFF technique may not be available at all sites, trials may be tion range.
done in partnership with radiology coordinating centres that can
standardise the appropriate PDFF technique across all sites partic- Summary
ipating in a trial. Hepatic steatosis can be seen in a wide variety of chronic liver
diseases, the most common of which is NAFLD. Studies suggest
A factor to consider when assessing longitudinal data for either that steatosis severity as well as steatosis change over time influence
clinical care or clinical trials is the fine margin of error in diag- disease progression in NAFLD and its high-risk subtype, NASH.
nosing mild steatosis. While more accurate than US and CT in With the increasing global prevalence of NAFLD and the recent
quantifying steatosis, MRI does have error rates of ± 1.5% as surge of clinical trials aimed at disease-altering therapy, there is an
an order of magnitude.62,67,82–84 This margin of error is largely ever more important need for safe and accurate quantification of
independent of the actual fat fraction, which poses a problem steatosis. Liver biopsy is currently the reference standard for disease
for screening steatosis in children and healthy adults since a 1% assessment in NAFLD and NASH. However, it is observer depen-
margin of error could be consequential for a diagnosis threshold dent and invasive, and it carries non-negligible risks. Imaging tech-
of 5%. Our limited knowledge of steatosis further confounds niques for assessing steatosis range from qualitative tools available
margin of error analysis. Using metabolic indices as a reference, a at the bedside to highly accurate and precise metrics. Table 4 lists
PDFF cutoff of 3% fat fraction has been suggested for diagnosing the strengths, weaknesses, and clinical care recommendations
steatosis; using histology as the standard, however, a cutoff of 6% for these techniques. Ultrasound is a safe and widely available
fat fraction has been suggested.2,85 Further research is needed to technique that may serve in certain clinical scenarios as an initial
refine our understanding of what constitutes a normal amount of screen. Its main drawbacks are machine and operator depen-
fat in the liver. dencies, qualitative assessment, and inaccuracy at detecting mild

Table 4. Comparison of ultrasound, CT and MRI-PDFF for clinical care and clinical trials in hepatic steatosis

Advantages Disadvantages Recommendations

Ultrasound Safe Indirect measurement Clinical care:
Widely available Qualitative initial screen
Low cost Operator and calibration dependent Clinical trials:
Inaccurate for mild steatosis do not recommend
Inaccurate steatosis grading
Confounders: obesity, fibrosis
Imprecise localization
CT Fast acquisition Indirect measurement Clinical care:
Easy to perform Variable calibration retrospective with conservative
Straightforward analysis Inaccurate for mild steatosis thresholds
Quantitative Confounders: iron, glycogen Clinical trials:
Ionising radiation do not recommend
Requires standard acquisition if contrast-enhanced
MRI-PDFF Direct measurement Relatively limited access Clinical care:
Precise fat quantification Claustrophobia study of choice (if available)
Highly sensitive and specific Implantable devices Clinical trials:
Corrects for confounders study of choice (if available)
Fast acquisition
PDFF, proton-density fat fraction.

8 of 12 birpublications.org/bjr Br J Radiol;91:20170959
 Review article: Liver fat imaging—a clinical overview of ultrasound, CT and MRI BJR

steatosis. Recent innovations in quantitative ultrasound promise to availability and utilization of PDFF have grown rapidly in recent
address some of these deficiencies. Like ultrasound, conventional years, with continued progress being made in technical refinement
unenhanced CT is accessible, easy to perform, and can be highly and validation. Where available, PDFF may serve as the noninva-
specific for moderate to severe steatosis. Quantitative evaluation of sive endpoint for steatosis reduction in clinical trials and therapy
steatosis is an additional benefit of CT. However, CT is inaccurate response assessment.
in the mild steatosis range and involves the use of radiation. Newer
CT techniques such as DECT could potentially expand the utility Funding
of this modality at steatosis quantification. MRI PDFF is currently Dr Sirlin reports research grants from GE and Siemens.
the most accurate quantitative imaging biomarker of steatosis. The

References:
1. Fabbrini E, Sullivan S, Klein S, Ekstedt with NAFLD and elevated liver enzymes. et al. Liver fibrosis, but no other histologic
M, Hagström H, Nasr P. Obesity and Hepatology 2006; 44: 865–73. doi: https://​doi.​ features, is associated with long-term
nonalcoholic fatty liver disease: biochemical, org/​10.​1002/​hep.​21327 outcomes of patients with nonalcoholic fatty
metabolic, and clinical implications. 8. Graffy PM, Pickhardt PJ. Quantification liver disease. Gastroenterology 2015; 149:
Hepatology 2010; 51: 679–89. doi: https://​doi.​ of hepatic and visceral fat by CT and MR 389–97. doi: https://​doi.​org/​10.​1053/​j.​gastro.​
org/​10.​1002/​hep.​23280 imaging: relevance to the obesity epidemic, 2015.​04.​043
2. Chalasani N, Younossi Z, Lavine JE, Diehl metabolic syndrome and NAFLD. Br J Radiol 15. Fassio E, Alvarez E, Domínguez N, Landeira
AM, Brunt EM, Cusi K, et al. The diagnosis 2016; 89: 20151024. doi: https://​doi.​org/​10.​ G, Longo C. Natural history of nonalcoholic
and management of non-alcoholic fatty liver 1259/​bjr.​20151024 steatohepatitis: a longitudinal study of repeat
disease: Practice guideline by the American 9. Pickhardt PJ, Hahn L, Muñoz del Rio A, liver biopsies. Hepatology 2004; 40: 820–6.
Association for the Study of Liver Diseases, Park SH, Reeder SB, Said A. Natural history doi: https://​doi.​org/​10.​1002/​hep.​20410
American College of Gastroenterology, of hepatic steatosis: observed outcomes 16. Wong VW, Wong GL, Choi PC, Chan AW,
and the American Gastroenterological for subsequent liver and cardiovascular Li MK, Chan HY, et al. Disease progression
Association. Am J Gastroenterol 2012; 107: complications. AJR Am J Roentgenol 2014; of non-alcoholic fatty liver disease: a
811–26. doi: https://​doi.​org/​10.​1038/​ajg.​ 202: 752–8. doi: https://​doi.​org/​10.​2214/​AJR.​ prospective study with paired liver biopsies
2012.​128 13.​11367 at 3 years. Gut 2010; 59: 969–74. doi: https://​
3. Younossi ZM, Koenig AB, Abdelatif D, Fazel Y, 10. Hahn L, Reeder SB, Muñoz del Rio A, doi.​org/​10.​1136/​gut.​2009.​205088
Henry L, Wymer M. Global epidemiology of Pickhardt PJ. Longitudinal changes in liver 17. Ajmera V, Park CC, Caussy C, Singh S,
nonalcoholic fatty liver disease-meta-analytic fat content in asymptomatic adults: hepatic Hernandez C, Bettencourt R, et al. Magnetic
assessment of prevalence, incidence, and attenuation on unenhanced CT as an resonance imaging proton density fat
outcomes. Hepatology 2016; 64: 73–84. doi: imaging biomarker for steatosis. AJR Am J fraction associates with progression of
https://​doi.​org/​10.​1002/​hep.​28431 Roentgenol 2015; 205: 1167–72. doi: https://​ fibrosis in patients with nonalcoholic fatty
4. Williams CD, Stengel J, Asike MI, doi.​org/​10.​2214/​AJR.​15.​14724 liver disease. Gastroenterology 2018; In Press.
Torres DM, Shaw J, Contreras M, et al. 11. Brouha SS, Nguyen P, Bettencourt R, Sirlin doi: https://​doi.​org/​10.​1053/​j.​gastro.​2018.​04.​
Prevalence of nonalcoholic fatty liver disease CB, Loomba R. Increased severity of liver fat 014
and nonalcoholic steatohepatitis among a content and liver fibrosis in non-alcoholic 18. Patel J, Bettencourt R, Cui J, Salotti J,
largely middle-aged population utilizing fatty liver disease correlate with epicardial Hooker J, Bhatt A, et al. Association of
ultrasound and liver biopsy: a prospective fat volume in type 2 diabetes: a prospective noninvasive quantitative decline in liver fat
study. Gastroenterology 2011; 140: 124–31. study. Eur Radiol 2018; 28: 1345–55. doi: content on MRI with histologic response
doi: https://​doi.​org/​10.​1053/​j.​gastro.​2010.​ https://​doi.​org/​10.​1007/​s00330-​017-​5075-6 in nonalcoholic steatohepatitis. Therap Adv
09.​038 12. Granér M, Nyman K, Siren R, Pentikäinen Gastroenterol 2016; 9: 692–701. doi: https://​
5. Kim H, Lee K, Lee KW, Yi NJ, Lee HW, MO, Lundbom J, Hakkarainen A, et al. doi.​org/​10.​1177/​1756283X16656735
Hong G, et al. Histologically proven non- Ectopic fat depots and left ventricular 19. Kleiner DE, Brunt EM, Van Natta M, Behling
alcoholic fatty liver disease and clinically function in nondiabetic men with C, Contos MJ, Cummings OW, et al. Design
related factors in recipients after liver nonalcoholic fatty liver disease. Circ and validation of a histological scoring
transplantation. Clin Transplant 2014; 28: Cardiovasc Imaging 2015; 8: e001979. doi: system for nonalcoholic fatty liver disease.
521–9. doi: https://​doi.​org/​10.​1111/​ctr.​12343 https://​doi.​org/​10.​1161/​CIRCIMAGING.​114.​ Hepatology 2005; 41: 1313–21. doi: https://​
6. Ekstedt M, Hagström H, Nasr P, Fredrikson 001979 doi.​org/​10.​1002/​hep.​20701
M, Stål P, Kechagias S, et al. Fibrosis stage is 13. Gramlich T, Kleiner DE, McCullough AJ, 20. Brunt EM, Janney CG, Di Bisceglie AM,
the strongest predictor for disease-specific Matteoni CA, Boparai N, Younossi ZM. Neuschwander-Tetri BA, Bacon BR.
mortality in NAFLD after up to 33 years of Pathologic features associated with fibrosis in Nonalcoholic steatohepatitis: a proposal for
follow-up. Hepatology 2015; 61: 1547–54. nonalcoholic fatty liver disease. Hum Pathol grading and staging the histological lesions.
doi: https://​doi.​org/​10.​1002/​hep.​27368 2004; 35: 196–9. doi: https://​doi.​org/​10.​1016/​ Am J Gastroenterol 1999; 94: 2467–74. doi:
7. Ekstedt M, Franzén LE, Mathiesen UL, j.​humpath.​2003.​09.​018 https://​doi.​org/​10.​1111/​j.​1572-​0241.​1999.​
Thorelius L, Holmqvist M, Bodemar G, 14. Angulo P, Kleiner DE, Dam-Larsen S, Adams 01377.x
et al. Long-term follow-up of patients LA, Bjornsson ES, Charatcharoenwitthaya P,

9 of 12 birpublications.org/bjr Br J Radiol;91:20170959
 BJR Zhang et al

21. El-Badry AM, Breitenstein S, Jochum W, value of liver ultrasound for the diagnosis of 64: 218–31. doi: https://​doi.​org/​10.​1109/​
Washington K, Paradis V, Rubbia-Brandt nonalcoholic fatty liver disease in overweight TUFFC.​2016.​2588979
L, et al. Assessment of hepatic steatosis by and obese patients. Liver Int 2015; 35: 41. Piekarski J, Goldberg HI, Royal SA, Axel
expert pathologists. Ann Surg 2009; 250: 2139–46. doi: https://​doi.​org/​10.​1111/​liv.​ L, Moss AA. Difference between liver and
691–7. doi: https://​doi.​org/​10.​1097/​SLA.​ 12840 spleen CT numbers in the normal adult:
0b013e3181bcd6dd 32. Debongnie JC, Pauls C, Fievez M, Wibin its usefulness in predicting the presence of
22. Ratziu V, Charlotte F, Heurtier A, Gombert E. Prospective evaluation of the diagnostic diffuse liver disease. Radiology 1980; 137:
S, Giral P, Bruckert E, et al. Sampling accuracy of liver ultrasonography. Gut 1981; 727–9. doi: https://​doi.​org/​10.​1148/​radiology.​
variability of liver biopsy in nonalcoholic 22: 130–5. doi: https://​doi.​org/​10.​1136/​gut.​ 137.​3.​6934563
fatty liver disease. Gastroenterology 2005; 22.​2.​130 42. Boyce CJ, Pickhardt PJ, Kim DH, Taylor AJ,
128: 1898–906. doi: https://​doi.​org/​10.​1053/​j.​ 33. Mathiesen UL, Franzén LE, Aselius H, Winter TC, Bruce RJ, et al. Hepatic steatosis
gastro.​2005.​03.​084 Resjö M, Jacobsson L, Foberg U, et al. (fatty liver disease) in asymptomatic adults
23. Maharaj B, Maharaj RJ, Leary WP, Cooppan Increased liver echogenicity at ultrasound identified by unenhanced low-dose CT.
RM, Naran AD, Pirie D, et al. Sampling examination reflects degree of steatosis but AJR Am J Roentgenol 2010; 194: 623–8. doi:
variability and its influence on the diagnostic not of fibrosis in asymptomatic patients https://​doi.​org/​10.​2214/​AJR.​09.​2590
yield of percutaneous needle biopsy of the with mild/moderate abnormalities of liver 43. Wells MM, Li Z, Addeman B, McKenzie CA,
liver. Lancet 1986; 1: 523–5. doi: https://​doi.​ transaminases. Dig Liver Dis 2002; 34: Mujoomdar A, Beaton M, et al. Computed
org/​10.​1016/​S0140-​6736(86)90883-4 516–22. doi: https://​doi.​org/​10.​1016/​S1590-​ tomography measurement of hepatic
24. Arun J, Jhala N, Lazenby AJ, Clements 8658(02)80111-6 steatosis: prevalence of hepatic steatosis in
R, Abrams GA. Influence of liver biopsy 34. Saadeh S, Younossi ZM, Remer EM, a canadian population. Can J Gastroenterol
heterogeneity and diagnosis of nonalcoholic Gramlich T, Ong JP, Hurley M, et al. Hepatol 2016; 2016: 1–7. doi: https://​doi.​org/​
steatohepatitis in subjects undergoing gastric The utility of radiological imaging 10.​1155/​2016/​4930987
bypass. Obes Surg 2007; 17: 155–61. doi: in nonalcoholic fatty liver disease. 44. Park YS, Park SH, Lee SS, Kim DY, Shin YM,
https://​doi.​org/​10.​1007/​s11695-​007-​9041-2 Gastroenterology 2002; 123: 745–50. doi: Lee W, et al. Biopsy-proven nonsteatotic
25. Zwiebel WJ. Sonographic diagnosis of diffuse https://​doi.​org/​10.​1053/​gast.​2002.​35354 liver in adults: estimation of reference range
liver disease. Semin Ultrasound CT MR 1995; 35. Saverymuttu SH, Joseph AE, Maxwell JD. for difference in attenuation between the
16: 8–15. doi: https://​doi.​org/​10.​1016/​0887-​ Ultrasound scanning in the detection of liver and the spleen at nonenhanced CT.
2171(95)90011-X hepatic fibrosis and steatosis. Br Med J 1986; Radiology 2011; 258: 760–6. doi: https://​doi.​
26. Charatcharoenwitthaya P, Lindor KD. Role of 292: 13–15. doi: https://​doi.​org/​10.​1136/​bmj.​ org/​10.​1148/​radiol.​10101233
radiologic modalities in the management of 292.​6512.​13 45. Johnston RJ, Stamm ER, Lewin JM, Hendrick
non-alcoholic steatohepatitis. Clin Liver Dis 36. Strauss S, Gavish E, Gottlieb P, Katsnelson L. RE, Archer PG. Diagnosis of fatty infiltration
2007; 11: 37–54. doi: https://​doi.​org/​10.​1016/​ Interobserver and intraobserver variability in of the liver on contrast enhanced CT:
j.​cld.​2007.​02.​014 the sonographic assessment of fatty liver. AJR limitations of liver-minus-spleen attenuation
27. Palmentieri B, de Sio I, La Mura V, Masarone Am J Roentgenol 2007; 189: W320–W323. difference measurements. Abdom Imaging
M, Vecchione R, Bruno S, et al. The role doi: https://​doi.​org/​10.​2214/​AJR.​07.​2123 1998; 23: 409–15. doi: https://​doi.​org/​10.​
of bright liver echo pattern on ultrasound 37. Hamer OW, Aguirre DA, Casola G, Lavine 1007/​s002619900370
B-mode examination in the diagnosis of liver JE, Woenckhaus M, Sirlin CB. Fatty liver: 46. Jacobs JE, Birnbaum BA, Shapiro MA,
steatosis. Dig Liver Dis 2006; 38: 485–9. doi: imaging patterns and pitfalls. Radiographics Langlotz CP, Slosman F, Rubesin SE, et al.
https://​doi.​org/​10.​1016/​j.​dld.​2006.​03.​021 2006; 26: 1637–53. doi: https://​doi.​org/​10.​ Diagnostic criteria for fatty infiltration of
28. Lee SS, Park SH, Kim HJ, Kim SY, Kim MY, 1148/​rg.​266065004 the liver on contrast-enhanced helical CT.
Kim DY, et al. Non-invasive assessment of 38. Xu L, Lu W, Li P, Shen F, Mi YQ, Fan JG. AJR Am J Roentgenol 1998; 171: 659–64. doi:
hepatic steatosis: prospective comparison of A comparison of hepatic steatosis index, https://​doi.​org/​10.​2214/​ajr.​171.​3.​9725292
the accuracy of imaging examinations. controlled attenuation parameter and 47. Bohte AE, van Werven JR, Bipat S, Stoker
J Hepatol 2010; 52: 579–85. doi: https://​doi.​ ultrasound as noninvasive diagnostic tools J. The diagnostic accuracy of US, CT, MRI
org/​10.​1016/​j.​jhep.​2010.​01.​008 for steatosis in chronic hepatitis B. Dig Liver and 1H-MRS for the evaluation of hepatic
29. van Werven JR, Marsman HA, Nederveen Dis 2017; 49: 910–7. doi: https://​doi.​org/​10.​ steatosis compared with liver biopsy: a meta-
AJ, Smits NJ, ten Kate FJ, van Gulik TM, et al. 1016/​j.​dld.​2017.​03.​013 analysis. Eur Radiol 2011; 21: 87–97. doi:
Assessment of hepatic steatosis in patients 39. Paige JS, Bernstein GS, Heba E, Costa https://​doi.​org/​10.​1007/​s00330-​010-​1905-5
undergoing liver resection: comparison of EAC, Fereirra M, Wolfson T, et al. A 48. Pickhardt PJ, Park SH, Hahn L, Lee SG, Bae
US, CT, T1-weighted dual-echo MR imaging, pilot comparative study of quantitative KT, Yu ES. Specificity of unenhanced CT for
and point-resolved 1H MR spectroscopy. ultrasound, conventional ultrasound, and non-invasive diagnosis of hepatic steatosis:
Radiology . 2010; 256: 159–68. doi: https://​ MRI for predicting histology-determined implications for the investigation of the
doi.​org/​10.​1148/​radiol.​10091790 steatosis grade in adult nonalcoholic fatty natural history of incidental steatosis. Eur
30. Hernaez R, Lazo M, Bonekamp S, Kamel liver disease. AJR Am J Roentgenol 2017; 208: Radiol 2012; 22: 1075–82. doi: https://​doi.​
I, Brancati FL, Guallar E, et al. Diagnostic W168–W177. doi: https://​doi.​org/​10.​2214/​ org/​10.​1007/​s00330-​011-​2349-2
accuracy and reliability of ultrasonography AJR.​16.​16726 49. Park SH, Kim PN, Kim KW, Lee SW, Yoon
for the detection of fatty liver: a meta- 40. Han A, Andre MP, Erdman JW, Loomba R, SE, Park SW, et al. Macrovesicular hepatic
analysis. Hepatology 2011; 54: 1082–90. doi: Sirlin CB, O'Brien WD. Repeatability and steatosis in living liver donors: use of CT
https://​doi.​org/​10.​1002/​hep.​24452 reproducibility of a clinically based QUS for quantitative and qualitative assessment.
31. Bril F, Ortiz-Lopez C, Lomonaco R, Orsak B, phantom study and methodologies. IEEE Radiology 2006; 239: 105–12. doi: https://​doi.​
Freckleton M, Chintapalli K, et al. Clinical Trans Ultrason Ferroelectr Freq Control 2017; org/​10.​1148/​radiol.​2391050361

10 of 12 birpublications.org/bjr Br J Radiol;91:20170959
 Review article: Liver fat imaging—a clinical overview of ultrasound, CT and MRI BJR

50. Kramer H, Pickhardt PJ, Kliewer MA, spectrum. NMR Biomed 2011; 24: 784–90. doi: https://​doi.​org/​10.​1007/​s00330-​015-​
Hernando D, Chen GH, Zagzebski JA, doi: https://​doi.​org/​10.​1002/​nbm.​1622 3724-1
et al. Accuracy of liver fat quantification 59. Reeder SB, Cruite I, Hamilton G, Sirlin 68. Haufe WM, Wolfson T, Hooker CA, Hooker
with advanced CT, MRI, and ultrasound CB. Quantitative assessment of liver fat JC, Covarrubias Y, Schlein AN, et al.
techniques: prospective comparison with with magnetic resonance imaging and Accuracy of PDFF estimation by magnitude-
MR spectroscopy. AJR Am J Roentgenol 2017; spectroscopy. J Magn Reson Imaging 2011; based and complex-based MRI in children
208: 92–100. doi: https://​doi.​org/​10.​2214/​ 34: 729–49. doi: https://​doi.​org/​10.​1002/​jmri.​ with MR spectroscopy as a reference. J Magn
AJR.​16.​16565 22580 Reson Imaging 2017; 46: 1641–7. doi: https://​
51. Kim DY, Park SH, Lee SS, Kim HJ, Kim SY, 60. Hong CW, Fazeli Dehkordy S, Hooker JC, doi.​org/​10.​1002/​jmri.​25699
Kim MY, et al. Contrast-enhanced computed Hamilton G, Sirlin CB. Fat quantification in 69. Satkunasingham J, Nik HH, Fischer S,
tomography for the diagnosis of fatty liver: the abdomen. Top Magn Reson Imaging 2017; Menezes R, Selzner N, Cattral M, et al. Can
prospective study with same-day biopsy used 26: 221–7. doi: https://​doi.​org/​10.​1097/​RMR.​ negligible hepatic steatosis determined
as the reference standard. Eur Radiol 2010; 0000000000000141 by magnetic resonance imaging-proton
20: 359–66. doi: https://​doi.​org/​10.​1007/​ 61. Meisamy S, Hines CD, Hamilton G, density fat fraction obviate the need for
s00330-​009-​1560-x Sirlin CB, McKenzie CA, Yu H, et al. liver biopsy in potential liver donors? Liver
52. Kodama Y, Ng CS, Wu TT, Ayers GD, Curley Quantification of hepatic steatosis with T1- Transplantation 2018; 24: 470–7. doi: https://​
SA, Abdalla EK, et al. Comparison of CT independent, T2-corrected MR imaging with doi.​org/​10.​1002/​lt.​24965
methods for determining the fat content of spectral modeling of fat: blinded comparison 70. Artz NS, Haufe WM, Hooker CA,
the liver. AJR Am J Roentgenol 2007; 188: with MR spectroscopy. Radiology 2011; 258: Hamilton G, Wolfson T, Campos GM,
1307–12. doi: https://​doi.​org/​10.​2214/​AJR.​ 767–75. doi: https://​doi.​org/​10.​1148/​radiol.​ et al. Reproducibility of MR-based liver fat
06.​0992 10100708 quantification across field strength: same-day
53. Fischer MA, Gnannt R, Raptis D, Reiner 62. Yokoo T, Bydder M, Hamilton G, comparison between 1.5T and 3T in obese
CS, Clavien P-A, Schmidt B, et al. Middleton MS, Gamst AC, Wolfson T, et al. subjects. J Magn Reson Imaging 2015; 42:
Quantification of liver fat in the presence Nonalcoholic fatty liver disease: diagnostic 811–7. doi: https://​doi.​org/​10.​1002/​jmri.​
of iron and iodine. Invest Radiol 2011; 46: and fat-grading accuracy of low-flip-angle 24842
351–8. doi: https://​doi.​org/​10.​1097/​RLI.​ multiecho gradient-recalled-echo MR 71. Kang GH, Cruite I, Shiehmorteza M,
0b013e31820e1486 imaging at 1.5 T. Radiology 2009; 251: Wolfson T, Gamst AC, Hamilton G, et al.
54. Zheng D, Tian W, Zheng Z, Gu J, Guo Z, 67–76. doi: https://​doi.​org/​10.​1148/​radiol.​ Reproducibility of MRI-determined proton
He X. Accuracy of computed tomography 2511080666 density fat fraction across two different MR
for detecting hepatic steatosis in donors for 63. Choi SS, Diehl AM. Hepatic triglyceride scanner platforms. J Magn Reson Imaging
liver transplantation: a meta-analysis. Clin synthesis and nonalcoholic fatty liver 2011; 34: 928–34. doi: https://​doi.​org/​10.​
Transplant 2017; 31: e13013. doi: https://​doi.​ disease. Curr Opin Lipidol 2008; 19: 1002/​jmri.​22701
org/​10.​1111/​ctr.​13013 295–300. doi: https://​doi.​org/​10.​1097/​MOL.​ 72. Bannas P, Kramer H, Hernando D, Agni
55. Artz NS, Hines CD, Brunner ST, Agni 0b013e3282ff5e55 R, Cunningham AM, Mandal R, et al.
RM, Kühn JP, Roldan-Alzate A, et al. 64. Zand KA, Shah A, Heba E, Wolfson T, Quantitative magnetic resonance imaging of
Quantification of hepatic steatosis with Hamilton G, Lam J, et al. Accuracy of hepatic steatosis: validation in ex vivo human
dual-energy computed tomography: multiecho magnitude-based MRI (M-MRI) livers. Hepatology 2015; 62: 1444–55. doi:
comparison with tissue reference standards for estimation of hepatic proton density fat https://​doi.​org/​10.​1002/​hep.​28012
and quantitative magnetic resonance imaging fraction (PDFF) in children. J Magn Reson 73. Tyagi A, Yeganeh O, Levin Y, Hooker JC,
in the ob/ob mouse. Invest Radiol 2012; 47: Imaging 2015; 42: 1223–32. doi: https://​doi.​ Hamilton GC, Wolfson T, et al. Intra- and
603–10. doi: https://​doi.​org/​10.​1097/​RLI.​ org/​10.​1002/​jmri.​24888 inter-examination repeatability of magnetic
0b013e318261fad0 65. Yokoo T, Serai SD, Pirasteh A, Bashir MR, resonance spectroscopy, magnitude-based
56. Hyodo T, Hori M, Lamb P, Sasaki K, Hamilton G, Hernando D, et al. Linearity, MRI, and complex-based MRI for
Wakayama T, Chiba Y, et al. Multimaterial bias, and precision of hepatic proton density estimation of hepatic proton density fat
decomposition algorithm for the fat fraction measurements by using MR fraction in overweight and obese children
quantification of liver fat content by using imaging: a meta-analysis. Radiology 2018; and adults. Abdom Imaging 2015; 40: 3070–7.
fast-kilovolt-peak switching dual-energy CT: 286: 486–98. doi: https://​doi.​org/​10.​1148/​ doi: https://​doi.​org/​10.​1007/​s00261-​015-​
experimental validation. Radiology 2017; radiol.​2017170550 0542-5
282: 381–9. doi: https://​doi.​org/​10.​1148/​ 66. Achmad E, Yokoo T, Hamilton G, Heba ER, 74. Lee SW, Park SH, Kim KW, Choi EK,
radiol.​2016160129 Hooker JC, Changchien C, et al. Feasibility Shin YM, Kim PN, et al. Unenhanced CT
57. Hamilton G, Middleton MS, Heba ER, Sirlin of and agreement between MR imaging and for assessment of macrovesicular hepatic
CB. Imaging techniques for the assessment spectroscopic estimation of hepatic proton steatosis in living liver donors: comparison of
of ectopic fat in liver and skeletal muscle. In: density fat fraction in children with known visual grading with liver attenuation index.
Translational research methods for diabetes, or suspected nonalcoholic fatty liver disease. Radiology 2007; 244: 479–85. doi: https://​doi.​
obesity and cardiometabolic drug development Abdom Imaging 2015; 40: 3084–90. doi: org/​10.​1148/​radiol.​2442061177
[Internet]. London: Springer London; 2015. https://​doi.​org/​10.​1007/​s00261-​015-​0506-9 75. Idilman IS, Aniktar H, Idilman R, Kabacam
pp. 99–119. http://​link.​springer.​com/​10.​ 67. Rehm JL, Wolfgram PM, Hernando D, G, Savas B, Elhan A, et al. Hepatic steatosis:
1007/​978-​1-​4471-​4920-​0_​4. Eickhoff JC, Allen DB, Reeder SB. Proton quantification by proton density fat fraction
58. Hamilton G, Yokoo T, Bydder M, Cruite density fat-fraction is an accurate biomarker with MR imaging versus liver biopsy.
I, Schroeder ME, Sirlin CB, et al. In vivo of hepatic steatosis in adolescent girls and Radiology 2013; 267: 767–75. doi: https://​doi.​
characterization of the liver fat ¹H MR young women. Eur Radiol 2015; 25: 2921–30. org/​10.​1148/​radiol.​13121360

11 of 12 birpublications.org/bjr Br J Radiol;91:20170959
 BJR Zhang et al

76. Tang A, Desai A, Hamilton G, Wolfson treatment of nonalcoholic steatohepatitis: 1265–71. doi: https://​doi.​org/​10.​1002/​jmri.​
T, Gamst A, Lam J, et al. Accuracy of MR assessment by novel magnetic resonance 24284
imaging-estimated proton density fat imaging and magnetic resonance 83. Hong CW, Mamidipalli A, Hooker JC,
fraction for classification of dichotomized elastography in a randomized trial Hamilton G, Wolfson T, Chen DH, et al. MRI
histologic steatosis grades in nonalcoholic (MOZART trial). Hepatology 2015; 61: proton density fat fraction is robust across
fatty liver disease. Radiology 2015; 274: 1239–50. doi: https://​doi.​org/​10.​1002/​hep.​ the biologically plausible range of triglyceride
416–25. doi: https://​doi.​org/​10.​1148/​radiol.​ 27647 spectra in adults with nonalcoholic
14140754 80. Cui J, Philo L, Nguyen P, Hofflich H, steatohepatitis. J Magn Reson Imaging 2018;
77. Middleton MS, Heba ER, Hooker CA, Hernandez C, Bettencourt R, et al. 47: 995–1002. doi: https://​doi.​org/​10.​1002/​
Bashir MR, Fowler KJ, Sandrasegaran Sitagliptin vs. placebo for non-alcoholic jmri.​25845
K, et al. Agreement between magnetic fatty liver disease: a randomized 84. Heba ER, Desai A, Zand KA, Hamilton
resonance imaging proton density fat controlled trial. J Hepatol 2016; 65: 369–76. G, Wolfson T, Schlein AN, et al. Accuracy
fraction measurements and pathologist- doi: https://​doi.​org/​10.​1016/​j.​jhep.​2016.​ and the effect of possible subject-based
assigned steatosis grades of liver 04.​021 confounders of magnitude-based MRI
biopsies from adults with nonalcoholic 81. Lin SC, Heba E, Bettencourt R, Lin GY, for estimating hepatic proton density fat
steatohepatitis. Gastroenterology 2017; Valasek MA, Lunde O, et al. Assessment fraction in adults, using MR spectroscopy
153: 753–61. doi: https://​doi.​org/​10.​1053/​j.​ of treatment response in non-alcoholic as reference. J Magn Reson Imaging 2016;
gastro.​2017.​06.​005 steatohepatitis using advanced magnetic 43: 398–406. doi: https://​doi.​org/​10.​1002/​
78. Noureddin M, Lam J, Peterson MR, resonance imaging. Aliment Pharmacol Ther jmri.​25006
Middleton M, Hamilton G, Le TA, et al. 2017; 45: 844–54. doi: https://​doi.​org/​10.​ 85. Nasr P, Forsgren MF, Ignatova S, Dahlström
Utility of magnetic resonance imaging versus 1111/​apt.​13951 N, Cedersund G, Leinhard OD, et al. Using
histology for quantifying changes in liver 82. Negrete LM, Middleton MS, Clark L, a 3% proton density fat fraction as a cut-off
fat in nonalcoholic fatty liver disease trials. Wolfson T, Gamst AC, Lam J, et al. Inter- value increases sensitivity of detection of
Hepatology 2013; 58: 1930–40. doi: https://​ examination precision of magnitude-based hepatic steatosis, based on results from
doi.​org/​10.​1002/​hep.​26455 MRI for estimation of segmental hepatic histopathology analysis. Gastroenterology
79. Loomba R, Sirlin CB, Ang B, Bettencourt proton density fat fraction in obese 2017; 153: 53–5. doi: https://​doi.​org/​10.​1053/​
R, Jain R, Salotti J, et al. Ezetimibe for the subjects. J Magn Reson Imaging 2014; 39: j.​gastro.​2017.​03.​005

12 of 12 birpublications.org/bjr Br J Radiol;91:20170959