Clinical Application of Quantitative MR Imaging in Nonalcoholic Fatty Liver Disease

Abstract

Viral hepatitis was previously the most common cause of chronic liver disease. However, in recent years, nonalcoholic fatty liver disease (NAFLD) cases have been increasing, especially in developed countries. NAFLD is histologically characterized by fat, fibrosis, and inflammation in the liver, eventually leading to cirrhosis and hepatocellular carcinoma. Although biopsy is the gold standard for the assessment of the liver parenchyma, quantitative evaluation methods, such as ultrasound, CT, and MRI, have been reported to have good diagnostic performances. The quantification of liver fat, fibrosis, and inflammation is expected to be clinically useful in terms of the prognosis, early intervention, and treatment response for the management of NAFLD. The aim of this review was to discuss the basics and prospects of MRI-based tissue quantifications of the liver, mainly focusing on proton density fat fraction for the quantification of fat deposition, MR elastography for the quantification of fibrosis, and multifrequency MR elastography for the evaluation of inflammation.

Introduction

There are many causes of diffuse liver disease, such as viral hepatitis, autoimmune hepatitis, hemosiderosis, alcohol, and fatty liver.¹ Chronic and persistent inflammatory changes in the liver cause repetitive destruction and regeneration of the liver parenchyma, which eventually lead to fibrosis, cirrhosis, and the development of hepatocellular carcinoma.^2–4 In the past, hepatitis viruses, such as hepatitis viruses B and C, were the leading causes of chronic liver diseases. However, with the development of treatment methods, the number of patients with chronic liver diseases caused by hepatitis viruses has been decreasing in developed countries, such as Western countries and Japan.^5,6 The proportion of non-B and non-C chronic liver diseases have been increasing because of the globally increasing prevalence of nonalcoholic fatty liver (NAFL) disease (NAFLD), which one in four adults suffer from.^7,8 NAFLD encompasses a histological spectrum ranging from NAFL, which is characterized by minor fat deposition without inflammation, to nonalcoholic steatohepatitis (NASH), which manifests as inflammation and ballooning, with or without fibrosis.⁹ It is estimated that 25%–44% of patients with NAFL will develop NASH, 5%–38% of whom will progress to cirrhosis, and that within eight years, 2.4%–12.8% of patients with NAFLD-related cirrhosis will develop hepatocellular carcinoma.^9–13 It is of utmost importance to prevent the progression to NASH and to early detect the high-risk groups. Regarding the diagnosis of NAFLD, when fatty liver or abnormal liver function is found, the hepatitis virus or autoimmune diseases should be first ruled out. By examining the history of alcoholism, alcoholic liver disease can also be excluded and NAFLD can be considered. Scoring systems, such as the fibrosis-4 (FIB-4) index and the NAFLD fibrosis score, are used to differentiate between NAFL and NASH, but they are not absolutely accurate.^14,15 Thus, liver biopsy is used to confirm the diagnosis of NASH.¹⁶

The pathology of NASH is characterized by inflammation of the hepatic parenchyma, ballooning of hepatocytes, and macrovesicular fatty liver; as the disease progresses, fibrosis develops around the central veins.¹⁷ The NAFL Activity Score system scores the degree of fat deposition, lobular inflammation, and ballooning to diagnose NASH and, along with the stage of fibrosis, determines the degree of disease progression.¹⁸ Although liver biopsy is the gold standard for NASH diagnosis, it is not practical to perform liver biopsy in all patients with NAFLD.¹⁶ Liver biopsy can yield only 1/20000 to 1/50000 of the total liver volume, and there is a risk of sampling error. It is highly invasive and causes serious complications in approximately 0.5% of patients, and there is variability in the diagnosis even among experienced pathologists.^19–22

To compensate for the disadvantages of liver biopsy, noninvasive imaging methods, such as ultrasound (US), CT, and MRI, have been developed. In this review, we discuss the basics and prospects of MRI-based tissue quantifications of the liver, mainly focusing on the quantitative assessment of fat deposition, fibrosis, and inflammation.

Quantification of Steatosis

Pathophysiology

Fatty liver is characterized by the excessive accumulation of triglycerides in the cells of the liver. This is one of the pathological features of NAFLD, but this can also be caused by other viral liver diseases, alcoholic liver disease, lipodystrophy, glycogenosis, and fatty drugs.^23,24 Normally, fat in the liver serves as energy (triglycerides and cholesterol esters) storage when there is excess energy, releasing them during extended fasting.²⁵ When lipid homeostasis becomes imbalanced due to abnormal conditions, such as metabolic diseases or obesity, excess triglycerides accumulate in the hepatocytes. Some cases of NAFLD progress to NASH, which can lead to cirrhosis and liver cancer; thus, an early diagnosis is of great clinical significance.^2,4,24

Chemical-shift-based proton density fat-fraction

The Dixon method is widely known as a method for fat quantification (Fig. 1). In the usual two-point Dixon method, in-phase and opposed-phase images are taken using the chemical-shift difference between water and fat protons, and the fat fraction can be calculated from the obtained signal intensity (F = [I – O]/2, W = [I + O]/2, Fat fraction = F/W + F; F, W, I, O: signal intensity of fat, water, in-phase, and opposed-phase, respectively). Some reports found that fat quantification obtained by two-point Dixon method also correlates well with histopathological fat deposition.^26,27 However, this method has several problems. Triglycerides, which are the fat in the body, have six peaks on the MR spectroscopy (MRS), but the two-point Dixon method assumes only the fat signal of beta-carboxyl methylene, which peaks at 1.3 ppm.^28,29 It accounts for only approximately 70% of the total body fat, thus underestimating the fat fraction. In addition, diffuse liver disease causes intrahepatic iron deposition, which leads to the underestimation of the signal due to T2* shortening and overestimation of the signal due to the effect of the T1 value, resulting in inaccurate fat quantification.^30,31 A fat fraction map using the Dixon method with multi-echo can overcome these weaknesses. Increasing the number of echo samples allows for multi-peak fat quantification and T2* value correction, and reducing the flip angle minimizes the effect of the T1 value.^32,33 The fat quantification image adjusted for these confounding factors is called proton density fat fraction (PDFF) map (Figs. 2 and 3).

Fig. 1.

In the two-point Dixon method, in-phase and opposed-phase images can be obtained using the chemical-shift difference between water and fat protons. This is an example of a patient with nonalcoholic steatohepatitis. The signal intensity of the liver parenchyma decreases from the in-phase (a) to the opposed-phase image (b).

Fig. 2.

In the multi-echo Dixon method, five images, a water image (a), fat image (b), T2* map (c), R2* map (d), and PDFF map (e) can be obtained with a single scan. PDFF (%) is obtained by setting the ROI on the PDFF map. PDFF, proton density fat fraction.

Fig. 3.

PDFF maps of three patients. The entire liver was evaluated in only a single 16s breath-hold. ROIs were placed in the posterior segment. PDFF values of the first patient showed minimal fat deposition in the liver (normal, < 6%) (a). PDFF values of the second and third patients revealed medium and high degrees of liver fat deposition (10.5% and 15.1%), respectively (b and c). PDFF, proton density fat fraction.

PDFF measures the proportion of proton components of fat molecules and is not identical to the pathologically defined lipid droplets in hepatocytes. It has been reported that PDFF threshold of 6.4% is consistent with the fatty liver criteria in pathology.^34,35 Tang et al. proposed the PDFF cut-off to classify steatosis grades, with high-specificity cut-offs of about ≥ 6.4%, ≥ 17.4%, and ≥ 22.1% corresponding to steatosis grades of ≥ 1, ≥ 2, and 3, respectively.³⁴ PDFF has been proven to be highly accurate and reproducible in phantom experiments and clinical cases, and is optimal as an imaging biomarker since it yields minimal error in MR models and magnetic field strength.^36,37 In addition, while fat deposition in the liver is often heterogenous, the ability to evaluate the entire liver provides clinical value superior to that of biopsy.^38,39 This imaging method also simultaneously produces the R2* map, which is useful for the assessment of iron deposition.^40,41 Iron is an important micronutrient in the oxygen transport system of hemoglobin, but excessive iron deposition in the liver can cause oxidative stress that leads to mitochondrial dysfunction and deoxyribonucleic acid (DNA) damage, resulting in hepatocellular injury, liver fibrosis, and cirrhosis.^42,43 Meanwhile, hereditary hemochromatosis, long-term blood transfusion, and hematological diseases that cause hemolytic anemia (e.g., thalassemia and sickle cell disease) are common causes of iron deposition and diffuse liver diseases, such as NAFLD and alcoholic and viral hepatitis.⁴⁴ It is widely known that the R2* value correlates well with the amount of liver iron.^40,41 The cut-off value for R2* values in iron deposition was reported to be 60 s^–1 by 1.5 T MRI and 126 S^–1 by 3.0 T MRI.⁴⁰ Bashir et al. reported that in patients with NAFLD, the most influential factor on hepatic R2* values was hepatic fat deposition as measured by PDFF.⁴⁵ This study showed that R2* values may not accurately reflect hepatic iron deposition in cases of excessive fat deposition in the liver.

Limitations of the PDFF

Although PDFF is an established imaging technique for the quantification of liver fat, it requires a long breath-hold because of 3D Cartesian sampling. In patients with difficulty in breath-holding, coherent aliasing artifacts make fat quantification inaccurate. Imaging with 3D stack of radial sampling under free breathing conditions can significantly reduce motion artifacts compared with conventional methods. It has been reported to be as accurate as conventional PDFF and MRS in phantoms and healthy volunteers and to improve the image quality of PDFF in infants with difficulty in breath-holding.^46,47 Another countermeasure for the long breath-holding time can be compressed sensing (CS).^48,49 This minimizes compressible image data using incoherent undersampling and an iterative reconstruction algorithm to shorten scan time and minimize motion effects.⁵⁰ In a qualitative and quantitative study comparing PDFF with four levels of CS factors and conventional PDFF, it was reported that the same accuracy was achieved even when the imaging time was shortened to one-fourth of the original time.⁴⁹

Another weakness of PDFF is that T2* correction by the fitting model becomes inaccurate due to premature T2* shortening in cases of severe iron deposition. Although ultrashort TE is expected to be used to acquire signals in a shorter time than T2* shortening, some studies have suggested that there is a short T2* signal in the liver, and that the PDFF bias is larger in ultrashort TE than in short TE, which had to be corrected by excluding short echoes below 1 ms.^51,52 On the other hand, the good accuracy of PDFF has been reported in cases with severe T2* shortening, such as before and after superparamagnetic iron oxide administration and in patients with hemochromatosis.^53,54 The R2* values in diffuse liver diseases, such as NAFLD, alcoholic liver disease, and viral hepatitis, are much lower than those in the aforementioned severe iron deposition cases, and the effect of T2* shortening may not be significant.

Other imaging techniques

MRS is a measure for hepatic fat quantification by MRI, and the diagnostic performance of PDFF has been reported to be almost equivalent to that of MRS.^31,55 However, MRS is limited in scan coverage and usually evaluates only a single voxel. MRS is considered inferior to PDFF because fat deposition in the liver is often heterogeneous, overall evaluation is important, and reproducibility is necessary for the evaluation of treatment effects.^29,37 For these reasons, MRS has been replaced by PDFF.

In US examination, Fibroscan (approved by the Food and Drug Administration in 2013; EchosensTM, Paris, France), equipped with the controlled attenuation parameter (CAP) function, can measure the US attenuation of the generated shear waves and estimate the fat deposition. However, in comparative studies between PDFF and CAP, PDFF was superior in diagnosing mild, moderate, and severe fatty liver.^56,57 The considered reason was that PDFF directly quantifies the fat molecule proton ratio, while CAP is an indirect semi-quantitative evaluation that utilizes the correlation between US attenuation and the degree of liver fat.

Dual-energy CT (DECT) is a technique that uses two different energies for imaging. DECT can determine the concentration of constituent materials in a mixture with a process known as material decomposition.⁵⁸ Fat, iron, and healthy liver tissue have adequately different mass attenuation coefficients with one another, which allow the direct quantification of liver fat and iron with material decomposition algorithms. It was reported that, in grading hepatic steatosis, DECT had comparable diagnostic performance to MRS.⁵⁹

Quantification of Fibrosis

Pathophysiology

Liver fibrosis in chronic liver diseases is caused by the repeated inflammation, destruction, and regeneration of tissues over many years due to chronic liver diseases, resulting in the activation of stellate cells, proliferation of fibroblasts and myofibroblasts, and excessive deposition of collagen as an extracellular matrix.⁶⁰ When fibrosis progresses and the regenerative capacity decreases, this leads to cirrhosis. Since hepatic fibrosis is the most significant prognostic factor in chronic liver diseases, early diagnosis and treatment of the causative disease leads to a better patient outcome.^61,62

MR elastography

MR elastography (MRE) applies vibrations from outside of the body, transmits propagate elastic waves through the tissue, and estimates the tissue stiffness by measuring the wavelength (Fig. 4).^63,64 With the assumption that the liver is an elastic material, the elastic modulus, an index of stiffness, can be expressed using the following formula:⁶³

Fig. 4.

MR elastograms and wave images of four patients. The modulus of elasticity increased with the stage of fibrosis. The stiffness values of the liver were as follows: 2.43 kPa in the patient with F1 (biopsy specimen scored using New Inuyama Classification) (a and b), 2.97 kPa in F2 (c and d), 5.65 kPa in F3 (e and f), and 9.24 kPa in F4 (g and h).

v² = μ/ρ (v: elastic wave velocity, μ: shear modulus, ρ: material density)

The density of the liver is almost 1 g/cm³,⁶⁵ and the elastic wave velocity is the multiplication of the frequency f and the wavelength λ. The elastic modulus could be obtained by vibrating the liver at a fixed vibration frequency and measuring the wavelength. The principle of MRE is that the rotational phases of protons are aligned to a homogeneous material, but by shifting the rotational phase, the physical vibrations transmitting in the liver can be visualized on phase images. The phase images are processed by an algorithm to obtain an elastogram. When comparing the wave images taken in normal and cirrhotic livers, the wavelength is longer and the elastic waves transmit faster in the latter. MRE could be easily added to conventional liver MRI protocols to evaluate the stiffness of the entire liver. It is not affected by ascites, obesity, fatty liver, or intravenous contrast administration, resulting in high reproducibility.^66,67 Multiple studies comparing histological fibrosis scores and elastic moduli measured using MRE in patients with chronic liver disease have reported a good correlation, with elastic moduli increasing as fibrosis progressed.^68–70 According to a multicenter study, the cut-off value of MRE for liver fibrosis in patients with various etiology of chronic liver diseases was reported as follows, all of which showed high diagnostic performance:

F1: 3.45 kPa (AUC: 0.84 [95% CI: 0.76–0.92]), F2: 3.66 kPa (AUC: 0.88 [95% CI: 0.84–0.91]), F3: 4.11 kPa (AUC: 0.93 [95% CI: 0.90–0.95]), F4: 4.71 kPa (AUC: 0.92 [95% CI: 0.90–0.94])⁷¹

Consensus profile by quantitative imaging biomarker alliance suggested that ROIs should be drawn manually in the largest possible area of liver parenchyma in which coherent shear waves are visible while excluding major blood vessels seen on the MRE magnitude images.⁷² In addition, various methods of ROIs placement have been reported, with the placement of a small round ROI in each segment (four segments or eight segments), the largest possible ROI in the entire liver or right lobe, a small ROI in the anterior segment, and such others.^73–75 Lee et al. reported that placing a round ROI in each segment and a large ROI in the entire liver was significantly more reliable than placing a small ROI in the anterior segment.⁷³ Since liver fibrosis has been reported to be highly variable among the segments, Toguchi et al. tested a method of averaging liver stiffness obtained from small ROIs placed in the right lobe of several slices.⁷⁶ They speculated that measuring liver stiffness in the same segment as the biopsy site would provide a better correlation with histological fibrosis. The study showed a significantly better correlation with the stage of histological fibrosis compared to a large freehand ROI placed in the right lobe.

Limitations of MRE

Conventional 2D MRE are sensitized to detect motion in a single plane and to work on the principle that shear waves transmit through the axial plane of acquisition. This hypothesis does not apply for regions of the liver close to the dome or below or far from the passive driver. In such regions, the waves may be transmitting at an oblique angle to the axial direction, and the wavelengths of shear wave may be overestimated. The 3D MRE is sensitized to detect motion in all three planes and can correct for wave obliquity.⁷⁷ The 3D MRE, usually spin-echo based, could evaluate the entire liver with a few breath-holds. This technique corrects for wave skewing and has a higher diagnostic performance than 2D MRE.⁷⁸

Technical failures account for less than 5% of examinations and are most commonly caused by a reduced SNR due to iron overload in the liver.⁷¹ The pulse sequence typically used in MRE is the gradient-echo technique, which requires a long TE and is susceptible to T2* shortening due to iron deposition. In addition, it has been reported that this phenomenon is more likely to occur in 3T MRI scanners.⁷⁹ Spin echo-echo planar imaging (SE-EPI) was recently developed as a pulse sequence that can be used in 3T MRE.⁸⁰ SE-EPI has advantages, such as low sensitivity to the attenuation of the transverse relaxation signal and wide coverage since multiple slices can be acquired with a single breath-hold. It is expected to reduce the overall technical failure rate.^79,80

Other imaging techniques

The US transient elastography can also quantify liver stiffness by measuring the velocity of shear waves generated by the probe. In comparison between MRE and US elastography, MRE was slightly superior in terms of the diagnostic performance in all fibrosis stages.⁸¹ Another advantage is that MRE is not dependent on radiographers and has greater reproducibility.^75,82–84

Other imaging techniques using MRI that show promise for the evaluation of hepatic fibrosis include diffusion weighted imaging (DWI), intravoxel incoherent motion (IVIM), T1 mapping, and T1ρ mapping. DWI is a technique for imaging molecular motion and diffusivity. The pathological structure of liver tissue changes due to fibrosis, and diffusivity is greatly affected by the degree of fibrosis.^85,86 Although a large number of studies have shown that liver fibrosis and apparent diffusion coefficient (ADC) values are inversely correlated, a recent meta-analysis reported that DWI also showed a good diagnostic performance for histological fibrosis staging but was inferior to MRE.^85,87 Since the liver is a highly vascularized organ, perfusion has a significant impact on diffusivity. IVIM, a biexponential model of DWI introduced by Le Bihan et al., can evaluate capillary microcirculation and water diffusion separately by analyzing the signal attenuation of multi-b-value DWI.⁸⁸ However, there are some reports that MRE has a higher diagnostic performance than IVIM.^89,90 On the other hand, diffusion-weighted-based virtual elastography has recently been developed and reported to have comparable diagnostic performance to MRE.^91,92 T1 and T1ρ are sensitive to low-frequency exchange interactions between water molecules and the large, slow-tumbling macromolecules, such as collagen, and can be used biomarkers of liver cirrhosis. Although it is not currently an established imaging method, some reports have shown that T1 and T1ρ values strongly correlated with US elastography, and further studies on this topic are expected.^93,94 The liver stiffness measurement obtained with standard clinical MRE sequences represents the apparent elasticity of the liver parenchyma and cannot be distinguished from that due to inflammation, congestion, or increased vascular pressure. A more accurate quantification of liver fibrosis is expected by separating fibrosis and inflammation, as described in the next chapter.

Quantification of Inflammation

Pathophysiology

Fatty deposits in the liver are the cause of NASH, but the presence of fatty deposits does not necessarily lead to NASH. The two-hit theory and multiple parallel hit theory are known as hypotheses of the process from NAFL to NASH.^95,96 In the two-hit theory proposed by Day et al., background factors, such as overnutrition and metabolic syndrome-induced fat deposition in the liver, lead to the development of NAFL.⁹⁵ This is followed by secondary stresses, such as oxidative stress, inflammatory cytokines, and endotoxins derived from intestinal bacteria, leading to inflammation of the liver and progression to NASH. Recently, interactions with extrahepatic organs, such as adipose tissue and the intestinal tract, have been reported to be a factor in the development of NASH.⁹⁷ It is important to accurately assess persistent inflammation in the liver to detect the early development of NASH.

Multifrequency MR elastography

MRE measures the elastic modulus of the liver to estimate fibrosis. It was also found via MRE that inflammation also increases the elastic modulus. In the presence of inflammation, the elastic modulus shows a high value due to increased intra-tissue pressure caused by increased regional blood flow, congestive edema, and inflammatory cell infiltration.⁹⁸ Ichikawa et al. reported that, in patients with chronic liver disease, the presence of inflammation increases the elastic modulus at all fibrosis stages.⁹⁹ The elastic modulus of cases with low fibrosis stage and high activity grade was similar to those with high fibrosis and low activity, indicating that the degree of inflammation had a significant effect on the elastic modulus. Liver stiffness on MRE is calculated by assuming that the liver tissue is a homogeneous isotoropic elastic body. However, liver tissue is generally composed of a viscoelastic body, and elastography, by using shear wave propagation, is needed to be determined as viscoelasticity. In the Voigt model, the elastic modulus can be expressed as a complex modulus and is calculated from the elastic and viscous moduli, while conventional single frequency MRE is not capable of evaluating the elastic and viscous tissues separately.¹⁰⁰ In reality, biological soft tissues are viscoelastic, with a mixture of elastic and viscous materials. The stiffness values on MRE by using inversion algorithm are composed of storage modulus (elastic property) and shear loss modulus (viscous property). The velocity and attenuation of shear waves in the viscous tissues are greater at higher vibration frequencies due to the diffusion of elastic waves in the viscous media, while elasticity in the elastic tissues is largely independent of frequency. Thus, higher vibration frequency on MRE can emphasize the viscous property compared with the elastic property.^101,102 According to this nature, multifrequency MRE with multiple vibration frequencies may be able to evaluate liver inflammation separately from fibrosis (Fig. 5). In a study using an animal model, complex modulus and damping ratio (viscosity to elasticity) obtained from multifrequency MRE correlated with fibrosis and inflammation, respectively.¹⁰³ Sofue et al. measured the elastic moduli of 30 patients with various chronic liver diseases, including NASH, using MRE at vibration frequencies of 60 Hz and 80 Hz.¹⁰⁴ They found that the elastic moduli imaged at 80 Hz were always higher than those at 60 Hz and that the difference, ΔG, correlated strongly with the histological activity grade. The modulus at 80 Hz correlated more strongly with the activity grade than the modulus at 60 Hz, and the correlation became even stronger when ΔG was taken into account. Thus, by changing the vibration frequency, the effect of viscosity on the elastic modulus can be captured, which is expected to be useful for diagnosing inflammation. Multiparametric 3D MRE can also provide basic viscoelastic modeling of tissues and separate the complex shear moduli into elastic and viscous components. A recent study found that among several parameters obtained using 3D MRE in patients with chronic hepatitis B and C, the damping ratio correlated well with the activity score in mild liver fibrosis patients, suggesting the possibility of the sensitive detection of early liver inflammation.¹⁰⁵ However, when liver fibrosis progresses, the damping ratio decreases and loses its ability to measure early necroinflammation. Currently, the MRE technics, including 3D MRE, SE-EPI, and multifrequency MRE, described in this paper, are not clinically available.

Fig. 5.

MR elastograms at 60-Hz and 80-Hz vibration frequencies of two patients. The first patient, a 67-year-old man with chronic hepatitis C, fibrosis stage F3, and necroinflammation grade A1, had stiffness values in MR elastograms at 60 Hz and 80 Hz of 7.82 kPa and 8.23 kPa, respectively (a and b). The ΔG value was calculated as 0.41 kPa. The second patient, an 82-year-old man with chronic hepatitis C, fibrosis stage F3, and necroinflammation grade A2, had stiffness values at 60 Hz and 80 Hz of 5.40 kPa and 6.78 kPa, respectively (c and d). The ΔG value was calculated as 1.38 kPa.

Other imaging techniques

The random movement of water molecules could be assessed with DWI. ADC, a marker of diffusion, has been reported to be decreased in inflammatory liver tissue, but the results have not been sufficient to quantify inflammation of the liver.^106,107 ADC is affected by perfusion in addition to diffusion effects, and the two effects may be offset or confounded. IVIM is a more advanced method of analyzing diffusion signals and can separate diffusion effects from perfusion effects. Perfusion-related parameters, such as perfusion rate and pseudo-diffusion coefficient, may reflect microvascular changes associated with inflammation. The diagnostic performance of these parameters in noninvasively detecting inflammation is limited and has been reported to have little or no correlation with inflammation.^108,109 Some reports used T1 mapping and T2 mapping to assess inflammation and have found some correlation, while others have reported that radiomic models based on T1 values corrected for T2* or T2 weighted sequences have shown promising results.^110–112

Conclusion

In the diagnosis and treatment of NAFLD, it is important to evaluate the presence of fatty deposits, fibrosis, and inflammation in the liver, and to compare and evaluate these factors cross-sectionally during the follow-up and therapeutic monitoring period. Chemical-shift-based PDFF can reliably quantify steatosis of the liver, and MRE can accurately estimate liver fibrosis. In addition, recent papers support adoption of MRI to estimate inflammation of the liver by using multifrequency MRE and other techniques. Quantitative MRI can be expected to play a pivotal role in screening, diagnosis, and treatment monitoring for the assessment of NAFLD by evaluating pathological change in the liver comprehensively and noninvasively.