AJR:190, April 2008 993 [7–9] (Fig. 1). The proportion of fat in the liver is not thought to alter the risk of develop- ing nonalcoholic steatohepatitis. Most cases of nonalcoholic steatohepatitis present de novo. The mortality rate from liver failure in nonalcoholic steatohepatitis is low, estimated to be 2–3% [8, 9]. Biopsy studies show that progression of nonalcoholic steatohepatitis (inflammation) to fibrosis occurs in approxi- mately 40% of patients but most of these pa- tients do not show clinical or biochemical deterioration [10]. The reported progression of nonalcoholic steatohepatitis to cirrhosis is 3–10% [8, 11, 12]. Thus, in comparison with alcoholic steatohepatitis, nonalcoholic ste- atohepatitis is a relatively benign disease. The 10-year survival rate of the former has been shown to be about 15%, compared with 60% for nonalcoholic steatohepatitis [10]. It is not clear why some patients with he- patic steatosis progress to nonalcoholic steato- hepatitis and most do not. Hepatic steatosis develops when the supply of fatty acids to the liver exceeds the requirements of triglyceride, phospholipid, and cholesterol synthesis, and mitochondrial oxidation. A “two hit” theory has been proposed [13] (Fig. 2). Clinical Diagnosis and Management The original definition of nonalcoholic fat- ty liver disease required liver biopsy findings as outlined and the absence of significant al- cohol intake or hepatic C seropositivity. Dis- agreement exists as to the threshold of alcohol intake that would be expected to cause alco- holic liver disease. Currently, nonalcoholic steatohepatitis is diagnosed if the alcohol in- take is less than 20g (≈ 2 drinks) per day. Nonalcoholic Fatty Liver Disease Chandana G. Lall 1 Alex M. Aisen Navin Bansal Kumaresan Sandrasegaran Lall CG, Aisen AM, Bansal N, Sandrasegaran K 1 All authors: Department of Radiology, Indiana University School of Medicine, 550 N University Blvd., Ste. UH 0279, Indianapolis, IN 46202. Address correspondence to K. Sandrasegaran ([email protected]). Hepatobiliary Imaging • Review AJR 2008; 190:993–1002 0361–803X/08/1904–993 © American Roentgen Ray Society N onalcoholic fatty liver disease is a group of disease entities that are typified by macrovesicular fatty change in the liver, unrelated to significant alcohol intake. The spectrum in- cludes hepatic steatosis, nonalcoholic steato- hepatitis, and chronic fibrosis and cirrhosis. Epidemiology Autopsy studies suggest that approximately 20% and 3% of American adults have hepatic steatosis and nonalcoholic steatohepatitis, re- spectively [1]. In the United States, 7–10% of patients undergoing liver biopsy have nonalco- holic steatohepatitis [2], compared with only 1.5% of liver biopsy patients in Japan [3]. Obe- sity is the biggest risk factor for nonalcoholic fatty liver disease. It is estimated that 70–80% of obese individuals have hepatic steatosis and 15–30% have nonalcoholic steatohepatitis [1, 4]. In 1999, 35% of adults were overweight (as defined by a body mass index of 25–30 kg/m 2 ) and 30% were obese (body mass index of > 30 kg/m 2 ) [5]. Obesity shows an epidemic rise in prevalence, and it is estimated that 50% of adults will be obese by 2025 [6]. Given the prevalence of nonalcoholic steatohepatitis in obese individuals, it is probable that more than 25 million Americans will have nonalcoholic steatohepatitis in the next 20 years. As a result, nonalcoholic steatohepatitis is expected to be- come the most common cause of chronic liver disease, surpassing hepatitis C. Natural History of Nonalcoholic Fatty Liver Disease The progression of hepatic steatosis to non- alcoholic steatohepatitis is thought to be low Keywords: CT, fatty liver, MRI, nonalcoholic fatty liver disease, nonalcoholic steatohepatitis, sonography DOI:10.2214/AJR.07.2052 Received February 14, 2007; accepted after revision May 22, 2007. CME This article is available for CME credit. See www.arrs.org for more information. OBJECTIVE. The inflammatory subtype of nonalcoholic fatty liver disease, nonalco- holic steatohepatitis, is becoming one of the most important causes of chronic liver disease. In this article, we discuss the epidemiology, pathogenesis, and clinical and radiologic diagno- sis of the subtypes of nonalcoholic fatty liver disease. CONCLUSION. We discuss the current and evolving imaging tests in the evaluation of hepatic fatty content, inflammation, and fibrosis. Lall et al. Nonalcoholic Fatty Liver Disease Hepatobiliary Imaging Review Downloaded from www.ajronline.org by 27.70.129.20 on 03/23/23 from IP address 27.70.129.20. Copyright ARRS. For personal use only; all rights reserved
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Nonalcoholic Fatty Liver DiseaseAJR:190, April 2008 993
[7–9] (Fig. 1). The proportion of fat in the liver is not thought to alter the risk of develop- ing nonalcoholic steatohepatitis. Most cases of nonalcoholic steatohepatitis present de novo. The mortality rate from liver failure in nonalcoholic steatohepatitis is low, estimated to be 2–3% [8, 9]. Biopsy studies show that progression of nonalcoholic steatohepatitis (inflammation) to fibrosis occurs in approxi- mately 40% of patients but most of these pa- tients do not show clinical or biochemical deterioration [10]. The reported progression of nonalcoholic steatohepatitis to cirrhosis is 3–10% [8, 11, 12]. Thus, in comparison with alcoholic steatohepatitis, nonalcoholic ste- atohepatitis is a relatively benign disease. The 10-year survival rate of the former has been shown to be about 15%, compared with 60% for nonalcoholic steatohepatitis [10].
It is not clear why some patients with he- patic steatosis progress to nonalcoholic steato- hepatitis and most do not. Hepatic steatosis develops when the supply of fatty acids to the liver exceeds the requirements of triglyceride, phospholipid, and cholesterol synthesis, and mitochondrial oxidation. A “two hit” theory has been proposed [13] (Fig. 2).
Clinical Diagnosis and Management The original definition of nonalcoholic fat-
ty liver disease required liver biopsy findings as outlined and the absence of significant al- cohol intake or hepatic C seropositivity. Dis- agreement exists as to the threshold of alcohol intake that would be expected to cause alco- holic liver disease. Currently, nonalcoholic steatohepatitis is diagnosed if the alcohol in- take is less than 20g (≈ 2 drinks) per day.
Nonalcoholic Fatty Liver Disease
Chandana G. Lall1 Alex M. Aisen Navin Bansal Kumaresan Sandrasegaran
Lall CG, Aisen AM, Bansal N, Sandrasegaran K
1All authors: Department of Radiology, Indiana University School of Medicine, 550 N University Blvd., Ste. UH 0279, Indianapolis, IN 46202. Address correspondence to K. Sandrasegaran ([email protected]).
Hepatobi l iar y Imaging • Review
AJR 2008; 190:993–1002
© American Roentgen Ray Society
N onalcoholic fatty liver disease is a group of disease entities that are typified by macrovesicular fatty change in the liver, unrelated to
significant alcohol intake. The spectrum in- cludes hepatic steatosis, nonalcoholic steato- hepatitis, and chronic fibrosis and cirrhosis.
Epidemiology Autopsy studies suggest that approximately
20% and 3% of American adults have hepatic steatosis and nonalcoholic steatohepatitis, re- spectively [1]. In the United States, 7–10% of patients undergoing liver biopsy have nonalco- holic steatohepatitis [2], compared with only 1.5% of liver biopsy patients in Japan [3]. Obe- sity is the biggest risk factor for nonalcoholic fatty liver disease. It is estimated that 70–80% of obese individuals have hepatic steatosis and 15–30% have nonalcoholic steatohepatitis [1, 4]. In 1999, 35% of adults were overweight (as defined by a body mass index of 25–30 kg/m2) and 30% were obese (body mass index of > 30 kg/m2) [5]. Obesity shows an epidemic rise in prevalence, and it is estimated that 50% of adults will be obese by 2025 [6]. Given the prevalence of nonalcoholic steatohepatitis in obese individuals, it is probable that more than 25 million Americans will have nonalcoholic steatohepatitis in the next 20 years. As a result, nonalcoholic steatohepatitis is expected to be- come the most common cause of chronic liver disease, surpassing hepatitis C.
Natural History of Nonalcoholic Fatty Liver Disease
The progression of hepatic steatosis to non- alcoholic steatohepatitis is thought to be low
Keywords: CT, fatty liver, MRI, nonalcoholic fatty liver disease, nonalcoholic steatohepatitis, sonography
DOI:10.2214/AJR.07.2052
Received February 14, 2007; accepted after revision May 22, 2007.
CME This article is available for CME credit. See www.arrs.org for more information.
OBJECTIVE. The inflammatory subtype of nonalcoholic fatty liver disease, nonalco- holic steatohepatitis, is becoming one of the most important causes of chronic liver disease. In this article, we discuss the epidemiology, pathogenesis, and clinical and radiologic diagno- sis of the subtypes of nonalcoholic fatty liver disease.
CONCLUSION. We discuss the current and evolving imaging tests in the evaluation of hepatic fatty content, inflammation, and fibrosis.
Lall et al. Nonalcoholic Fatty Liver Disease
Hepatobiliary Imaging Review
Lall et al.
Liver biopsy finding of macrovesicular fat (single large vacuole) in the cytoplasm of hep- atocytes displacing the nucleus peripherally is the hallmark of hepatic steatosis caused by alcohol, diabetes, and obesity [14]. In contra- distinction, microvesicular steatosis is char- acterized by multiple small fatty inclusion bodies with a predominantly central nucleus and is associated with abnormalities of mito- chondrial fatty acid oxidation, such as in acute fatty liver of pregnancy and Reye’s syndrome [15]. In nonalcoholic steatohepatitis, addi- tional histologic features are seen, including Mallory’s bodies, cytoplasmic balloon degen- eration, perisinusoidal (zone III) fibrosis, and neutrophilic infiltrate [14] (Fig. 3). However, many atypical findings may be found, such as the presence of lymphocytic infiltration and periportal fibrosis. It is not clear if these are different clinical entities that have been
grouped as nonalcoholic steatohepatitis. Note that the histologic appearance of nonalcoholic steatohepatitis is identical to that of alcoholic liver disease, and the distinction between the two conditions has traditionally been made on the basis of the amount of alcohol intake.
Increasing evidence indicates that nonalco- holic fatty liver disease is the hepatic compo- nent of a systemic metabolic syndrome that includes obesity, insulin resistance, hyperlipi- demia, and hypertension. Thus, it is possible to have patients who have both nonalcoholic fatty liver disease (in view of their metabolic status) and alcoholic liver disease, or nonal- coholic fatty liver disease, and chronic liver disease due to hepatitis C [16].
The general goals of treating nonalcoholic steatohepatitis are to correct risk factors with exercise and appropriate diet and to avoid drugs such as alcohol, tamoxifen, and steroids that
exacerbate liver disease. Pharmacologic thera- pies that have been proposed include those that increase insulin sensitivity, such as glitazones and biguanides, and antihyperlipidemic drugs, such as gemfibrozil [17]. Bariatric surgery has also been shown to reduce the severity of he- patic steatosis and nonalcoholic steatohepatitis [18, 19]. Patients with nonalcoholic steatohepa- titis who progress to end-stage liver disease become candidates for liver transplantation.
Imaging Tests in Nonalcoholic Fatty Liver Disease Assessment of Hepatic Fat Content
Sonographic findings of fatty liver change include increased echogenicity of liver, blur- ring of vascular margins, and increased acous- tic attenuation. Hyperechogenicity is due to increased acoustic interfaces resulting from intracellular accumulation of lipid vesicles. Severe fatty liver, containing more than 30% fat by weight, is detected by sonography with a sensitivity and specificity of 67–84% and 77–100%, respectively [20, 21]. Hepatic ste- atosis may appear heterogeneous on sonog- raphy, with some areas of liver parenchyma spared by the steatosis. Sonography is poor in detecting smaller amounts of fat in liver. In addition, the degree of fatty change in the liver can only be subjectively classified as mild or severe on sonography. In general, severe ste- atosis is diagnosed when the hepatic paren- chymal echogenicity obscures visualization of the walls of the hepatic and portal veins.
Unenhanced CT can show fatty change by a reduced attenuation of liver density. Fatty change is suspected when the density of liver is more than 10 H below that of the spleen. When IV contrast material was used, a differ- ence in Hounsfield density of at least 20 units between liver and spleen (liver lower than spleen), at 80–100 seconds after the start of contrast injection, was found to have sensitiv- ity and specificity of 86–87% in diagnosing fatty liver [22]. That study used 150 mL of iothalamate meglumine (Conray 60, Mall- inckrodt) injected at a rate of 2 mL/s. How- ever, another study reported lower sensitivity and specificity values of 50–75% [23]. Dif- ferences between these studies may be due to confounding factors such as iron, copper, or fibrous tissue that alter the Hounsfield densi- ty of liver, and differences in the rate of con- trast injection and the timing of the scanning. In addition, CT is not sensitive in detecting mild or moderate elevations of hepatic lipid content (5–30%) [21]. Dual-energy CT has been reported to accurately determine the
Fig. 1—Natural history of nonalcoholic fatty liver disease. Small proportion of patients with fatty liver develop nonalcoholic steatohepatitis. Less than 10% of nonalcoholic steatohepatitis patients develop cirrhosis. Current research is aimed at detecting early stages of fibrosis that are potentially reversible. aProbable percentage of patients with hepatic steatosis progressing to nonalcoholic steatohepatitis. bDisease progression (%) of all nonalcoholic steatohepatitis patients.
Fig. 2—Pathogenesis of nonalcoholic steatohepatitis. Currently popular “two-hit” theory. First hit is insulin resistance, which leads to hepatic steatosis. Fatty liver is less able to cope with oxidative stress, which is second hit, leading to chronic liver inflammation. Purported factors causing liver damage include free radical formation, cytokine release, iron overload, and altered mitochondrial energy production [71].
Nonalcoholic steatohepatitis
Stable disease
Hepatic steatosis
Iron overload Impaired mitochondrial
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 70
.1 29
.2 0
on 0
3/ 23
/2 3
fr om
I P
ad dr
es s
27 .7
0. 12
9. 20
. C op
yr ig
ht A
R R
S. F
or p
er so
AJR:190, April 2008 995
Nonalcoholic Fatty Liver Disease
fatty content of liver in animals [24], although this finding has not been reproduced in hu- mans [25, 26].
Water and fat protons have slightly differ- ent precessional frequencies in a magnetic field. As a result, MRI has the potential for diagnosing fatty infiltration both qualitatively and quantitatively. The most commonly used quantitative method is the acquisition of the so-called in- and out-of-phase imaging, or chemical shift imaging, in which different TE parameters are used in gradient-echo images
so that the signal from fat protons is added or subtracted, respectively, from the signal from protons in water. With 1.5-T magnets, the fat and water protons are in phase or out of phase when the TE is an even or odd multiple, re- spectively, of 2.2 milliseconds. Reduction of signal on out-of phase T1-weighted images has been shown to be an accurate predictor of hepatic fat content (Fig. 4), with correlation (r) of 0.86–0.91 compared with the histologic assessment of liver fat [27–29]. MRI is supe- rior to sonography in assessing liver fat [30].
Chemical shift imaging has been shown to be useful in assessing reduction in fatty liver af- ter insulin sensitizing therapy [31].
With 3-T MRI, the order in which in-phase and out-of phase echoes are acquired varies depending on the scanner vendor. In our prac- tice, TEs of 2.5 and 6.2 milliseconds, respec- tively, are used for in-phase and out-of-phase imaging with the Magnetom Trio (Siemens Medical Solutions). With other 3-T MRI scan- ners, the out-of-phase echo may be acquired first. Susceptibility artifact from hepatic iron becomes more pronounced as the TE increas- es. Therefore, depending on the 3-T scanner, the presence of iron may variably affect the signal intensity of fatty liver. This confound- ing factor is less problematic when T2-weight- ed fast spin-echo sequences, which contain multiple refocusing pulses, are used. A study of cirrhotic and noncirrhotic patients per- formed at 1.5-T MRI found that the signal drop off on T2-weighted fast spin echo with- out and with fat saturation better correlated with fat content determined histologically than in- and out-of-phase T1-weighted gradi- ent-recalled images (r = 0.76 vs r = 0.25; p < 0.01 in cirrhotic patients; r = 0.92 vs r = 0.69; p < 0.01 in noncirrhotic patients) [32].
It is possible to use specialized MR pulse sequences to quantitatively assess the degree of fat in the liver with greater accuracy than is possible with CT. The original approach was based on the Dixon technique, a spin- echo method analogous to in-phase and out- of-phase imaging, in which a pair of spin- echo images is acquired. In one of these, the timing of the 180° focusing pulse is altered so that both in-phase and out-of-phase images
Fig. 3—Photomicrograph shows histology of nonalcoholic steatohepatitis in 62-year-old woman. Mallory’s hyaline bodies (pink filamentous structures, black arrowhead) are cytoplasmic inclusions in hepatocytes consisting of abnormal keratin, hyaline, and other proteins. They are usually found in hepatocytes that are ballooned (black arrow) and are morphologic hallmarks of alcoholic and nonalcoholic steatohepatitis. Mallory’s bodies are not cause but rather consequence of cellular injury. Usually hepatocytes with Mallory’s bodies do not contain large fat vacuoles, although microvesicular fat may be seen. In this frame, other hepatocytes are present, containing macrovesicular fat globules (white arrow), which occupy almost all cytoplasm, displacing nucleus (white arrowhead) to periphery. (H and E, × 400) (Courtesy of Romil Saxena, Department of Pathology, Indiana Universtiy School of Medicine)
A B Fig. 4—Chemical shift MRI detection of hepatic fat in 56-year-old man with nonalcoholic steatohepatitis. A and B, Liver appears diffusely hypointense on out-of-phase gradient-echo sequence (TR/TE, 130/2.2; flip angle, 70°) (B) compared with in-phase sequence (130/4.9; flip angle, 70°) (A), indicating presence of fat.
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 70
.1 29
.2 0
on 0
3/ 23
/2 3
fr om
I P
ad dr
es s
27 .7
0. 12
9. 20
. C op
yr ig
ht A
R R
S. F
or p
er so
Lall et al.
are obtained [33, 34]. These two images can then be mathematically added and subtracted to produce lipid-only and water-only images. With suitable calibration, quantification of fat is possible. Quantitative chemical shift imaging, based on an additional dimension of Fourier transformation to achieve separa- tion water and fat signals, can also be used to
measure lipid content [35]. Many additional approaches are described in the literature, including recent descriptions of quantitative gradient-echo methods that have been used to measure the degree of hepatic steatosis [36]. Many of these techniques are easy to perform and may be used as part of routine imaging assessment of the liver [32].
Proton (1H) MR spectroscopy has been found to have good correlation with hepatic lipid content (Fig. 5), as determined by liver biopsy, with correlation (r) values of 0.91–0.98 [37–39]. This technique is considered sensitive to small variation (as little as 0.5% change) in hepatic lipid content and has a potential use in assessing therapy of steatosis [39].
g
e d c a
A B Fig. 5—Proton (1H) hepatic MR spectroscopy. A and B, Single-voxel spectrum in 39-year-old-woman with hepatic steatosis (A) and 29-year-old asymptomatic male volunteer (B). In healthy volunteer, only resonances of water (f) and methylene (b), found in hepatic triglycerides and fatty acids, are discernible. Normal liver contains less than 5% fat by weight. In patient with hepatic steatosis, amplitude of methylene resonance (b) is much higher. Several other lipid resonances are now visible. Note chemical shifts of various resonances in 1H MR spectroscopy at 1.5 T: a, terminal methyl (CH3): 0.8 ppm; b, methylene (CH2)n: 1.2 ppm; c, CH2-C = C: 1.9 ppm; d, C = C-CH2-C = C: 2.6 ppm; e, CH2-O-COR: 4.15 ppm; f, water (H2O): 4.7 ppm; g, CH = CH and CH = O: 5.2 ppm.
b
f
α−ATP
β−ATP
A C Fig. 6—Phosphorus (31P) hepatic MR spectroscopy in healthy patient at 1.5 T. (Reprinted with permission from Cortez-Pinto H, Chatham J, Chacko VP, Arnold C, Rashid A, Diehl AM. Alterations in liver ATP homeostasis in human nonalcoholic steatohepatitis: a pilot study. JAMA 1999; 282:1659–1664 [42]. Copyright American Medical Association © 1999. All rights reserved.) A, There are six main resonances. PME = phosphomonoesters, Pi = inorganic phosphate, PDE = phosphodiesters; γ-ATP, α-ATP, and β-ATP = γ, α, and β phosphates of adenosine triphosphate (ATP). B, 15 minutes after fructose infusion, ATP resonances are reduced in amplitude but Pi is maintained. C, 60 minutes after infusion, ATP resonances recover. In patients with nonalcoholic steatohepatitis this recovery is impaired (not shown).
D20D15D10D5 ppm
Minimum
05
Pi
10
B
Recovery
0510
Detection of Inflammation in Nonalcoholic Steatohepatitis
Sonography, CT, and MRI are insensitive in differentiating hepatic steatosis from non- alcoholic steatohepatitis [21]. A small CT study of patients with nonalcoholic fatty liver disease found that those with nonalcoholic steatohepatitis had increased liver size and increased caudate lobe-to-right lobe size ra- tio, compared with those with steatosis only [40]. The caudate-to-right lobe size ratio was statistically higher in steatohepatitis (mean, 0.43; range, 0.31–0.55) compared with ste- atosis only (mean, 0.36; range, 0.22–0.47). However, measurements showed consider- able overlap in both categories, and it is un- likely that these measurements will be useful in individual patients.
In the future, MR spectroscopy may offer a role in the diagnosis of nonalcoholic steatohep- atitis. MR spectroscopic studies on hydrogen, phosphorus, and sodium have been reported. The latter two chemical species show promise.
The typical phosphorus (31P) hepatic MR spectroscopy at 1.5 and 3 T shows six dominant
resonances, denoting phosphomon oe sters, inorganic phosphate (Pi), phosphodiesters, and three resonances from the nucleoside triphosphates that are mainly composed of adenosine triphosphate (ATP) (Fig. 6). When the signal from phosphocreatine is seen, this is most likely due to the erroneous incorpora- tion of adjacent muscle in the voxel. Several studies have investigated the utility of phos- phorus (31P) MR spectroscopy in the diagno- sis of acute and chronic liver damage in animal models and humans [41–44]. Most studies show that phosphorus metabolite ratios (e.g., ATP/Pi) do not differ significantly between normal and damaged or diseased liver [42, 43]. Two problems are inherent to 31P MR spectroscopy. The spatial resolution, even with a 3-T magnet, is limited to 2–3 cm; thus, this technique is suitable only for assessing diffuse liver disease. In addition, several homeo static mechanisms maintain levels of phosphate metabolites and inorganic phos- phate at constant levels and limit the sensitiv- ity of 31P MR spectroscopy to pathophysio- logic changes. To overcome this problem, the
recovery of cellular energy mechanisms after the depletion of ATP by an IV infusion of fructose, a form of pharmacologic hepatic stress testing, has been investigated in rat models [41]. A study of nonalcoholic steato- hepatitis patients showed that, after fructose- induced depletion, recovery of hepatic ATP is severely impeded [42] (Fig. 6). This tech- nique may help differentiate patients with he- patic steatosis from those with nonalcoholic steatohepatitis. It also suggests that impair- ment of energy homeostasis may be an inte- gral part of the progression from fatty liver to nonalcoholic steatohepatitis.
Using energy-intensive processes, hepato- cytes maintain intracellular sodium concentra- tion (Nai+) of 20 mmol/L and an extracellular concentration (Nae+) of approximately 140 mmol/L. With hepatocellular damage, Nai+ concentration increases. Unfortunately, be- cause sodium is ionic, Nai+ and Nae+ have the same resonance and cannot be separated on routine MR spectroscopy. Several techniques have been proposed to overcome this, of which the most promising in humans is multiple- quantum transfer coherence filtering [45, 46]. Sodium (23Na) has three quantum energy lev- els and four possible spin orientations (Fig. 7). When the 23Na cation is transiently bound to macromolecules, as happens in the intracellu- lar compartment, double or triple quantum transitions can occur. In the normal extracellu- lar space, which is predominantly aqueous, most quantum transitions are single. In liver disease, the multiple-quantum transfer coher- ence filtering 23Na signal may increase step- wise from benign fatty liver to hepatitis and then to cirrhosis. Initially, the spectral signal increases due to elevated Nai
+ concentration from cellular damage in hepatitis (Fig. 7) lead
3/2
1/2
−1/2
SQ outer
A
B
Fig. 7—Sodium (23Na) MRI. A, Sodium (23Na) has spin quantum number (I) of 3/2 and four possible spin orientations (−3/2, −1/2, +1/2, +3/2). Three single-quantum (SQ) transitions are possible: an “inner” or −1/2 ↔ +1/2 transition, and two “outer” or −3/2 ↔ −1/2 and + 1/2 ↔ +3/2…
[7–9] (Fig. 1). The proportion of fat in the liver is not thought to alter the risk of develop- ing nonalcoholic steatohepatitis. Most cases of nonalcoholic steatohepatitis present de novo. The mortality rate from liver failure in nonalcoholic steatohepatitis is low, estimated to be 2–3% [8, 9]. Biopsy studies show that progression of nonalcoholic steatohepatitis (inflammation) to fibrosis occurs in approxi- mately 40% of patients but most of these pa- tients do not show clinical or biochemical deterioration [10]. The reported progression of nonalcoholic steatohepatitis to cirrhosis is 3–10% [8, 11, 12]. Thus, in comparison with alcoholic steatohepatitis, nonalcoholic ste- atohepatitis is a relatively benign disease. The 10-year survival rate of the former has been shown to be about 15%, compared with 60% for nonalcoholic steatohepatitis [10].
It is not clear why some patients with he- patic steatosis progress to nonalcoholic steato- hepatitis and most do not. Hepatic steatosis develops when the supply of fatty acids to the liver exceeds the requirements of triglyceride, phospholipid, and cholesterol synthesis, and mitochondrial oxidation. A “two hit” theory has been proposed [13] (Fig. 2).
Clinical Diagnosis and Management The original definition of nonalcoholic fat-
ty liver disease required liver biopsy findings as outlined and the absence of significant al- cohol intake or hepatic C seropositivity. Dis- agreement exists as to the threshold of alcohol intake that would be expected to cause alco- holic liver disease. Currently, nonalcoholic steatohepatitis is diagnosed if the alcohol in- take is less than 20g (≈ 2 drinks) per day.
Nonalcoholic Fatty Liver Disease
Chandana G. Lall1 Alex M. Aisen Navin Bansal Kumaresan Sandrasegaran
Lall CG, Aisen AM, Bansal N, Sandrasegaran K
1All authors: Department of Radiology, Indiana University School of Medicine, 550 N University Blvd., Ste. UH 0279, Indianapolis, IN 46202. Address correspondence to K. Sandrasegaran ([email protected]).
Hepatobi l iar y Imaging • Review
AJR 2008; 190:993–1002
© American Roentgen Ray Society
N onalcoholic fatty liver disease is a group of disease entities that are typified by macrovesicular fatty change in the liver, unrelated to
significant alcohol intake. The spectrum in- cludes hepatic steatosis, nonalcoholic steato- hepatitis, and chronic fibrosis and cirrhosis.
Epidemiology Autopsy studies suggest that approximately
20% and 3% of American adults have hepatic steatosis and nonalcoholic steatohepatitis, re- spectively [1]. In the United States, 7–10% of patients undergoing liver biopsy have nonalco- holic steatohepatitis [2], compared with only 1.5% of liver biopsy patients in Japan [3]. Obe- sity is the biggest risk factor for nonalcoholic fatty liver disease. It is estimated that 70–80% of obese individuals have hepatic steatosis and 15–30% have nonalcoholic steatohepatitis [1, 4]. In 1999, 35% of adults were overweight (as defined by a body mass index of 25–30 kg/m2) and 30% were obese (body mass index of > 30 kg/m2) [5]. Obesity shows an epidemic rise in prevalence, and it is estimated that 50% of adults will be obese by 2025 [6]. Given the prevalence of nonalcoholic steatohepatitis in obese individuals, it is probable that more than 25 million Americans will have nonalcoholic steatohepatitis in the next 20 years. As a result, nonalcoholic steatohepatitis is expected to be- come the most common cause of chronic liver disease, surpassing hepatitis C.
Natural History of Nonalcoholic Fatty Liver Disease
The progression of hepatic steatosis to non- alcoholic steatohepatitis is thought to be low
Keywords: CT, fatty liver, MRI, nonalcoholic fatty liver disease, nonalcoholic steatohepatitis, sonography
DOI:10.2214/AJR.07.2052
Received February 14, 2007; accepted after revision May 22, 2007.
CME This article is available for CME credit. See www.arrs.org for more information.
OBJECTIVE. The inflammatory subtype of nonalcoholic fatty liver disease, nonalco- holic steatohepatitis, is becoming one of the most important causes of chronic liver disease. In this article, we discuss the epidemiology, pathogenesis, and clinical and radiologic diagno- sis of the subtypes of nonalcoholic fatty liver disease.
CONCLUSION. We discuss the current and evolving imaging tests in the evaluation of hepatic fatty content, inflammation, and fibrosis.
Lall et al. Nonalcoholic Fatty Liver Disease
Hepatobiliary Imaging Review
Lall et al.
Liver biopsy finding of macrovesicular fat (single large vacuole) in the cytoplasm of hep- atocytes displacing the nucleus peripherally is the hallmark of hepatic steatosis caused by alcohol, diabetes, and obesity [14]. In contra- distinction, microvesicular steatosis is char- acterized by multiple small fatty inclusion bodies with a predominantly central nucleus and is associated with abnormalities of mito- chondrial fatty acid oxidation, such as in acute fatty liver of pregnancy and Reye’s syndrome [15]. In nonalcoholic steatohepatitis, addi- tional histologic features are seen, including Mallory’s bodies, cytoplasmic balloon degen- eration, perisinusoidal (zone III) fibrosis, and neutrophilic infiltrate [14] (Fig. 3). However, many atypical findings may be found, such as the presence of lymphocytic infiltration and periportal fibrosis. It is not clear if these are different clinical entities that have been
grouped as nonalcoholic steatohepatitis. Note that the histologic appearance of nonalcoholic steatohepatitis is identical to that of alcoholic liver disease, and the distinction between the two conditions has traditionally been made on the basis of the amount of alcohol intake.
Increasing evidence indicates that nonalco- holic fatty liver disease is the hepatic compo- nent of a systemic metabolic syndrome that includes obesity, insulin resistance, hyperlipi- demia, and hypertension. Thus, it is possible to have patients who have both nonalcoholic fatty liver disease (in view of their metabolic status) and alcoholic liver disease, or nonal- coholic fatty liver disease, and chronic liver disease due to hepatitis C [16].
The general goals of treating nonalcoholic steatohepatitis are to correct risk factors with exercise and appropriate diet and to avoid drugs such as alcohol, tamoxifen, and steroids that
exacerbate liver disease. Pharmacologic thera- pies that have been proposed include those that increase insulin sensitivity, such as glitazones and biguanides, and antihyperlipidemic drugs, such as gemfibrozil [17]. Bariatric surgery has also been shown to reduce the severity of he- patic steatosis and nonalcoholic steatohepatitis [18, 19]. Patients with nonalcoholic steatohepa- titis who progress to end-stage liver disease become candidates for liver transplantation.
Imaging Tests in Nonalcoholic Fatty Liver Disease Assessment of Hepatic Fat Content
Sonographic findings of fatty liver change include increased echogenicity of liver, blur- ring of vascular margins, and increased acous- tic attenuation. Hyperechogenicity is due to increased acoustic interfaces resulting from intracellular accumulation of lipid vesicles. Severe fatty liver, containing more than 30% fat by weight, is detected by sonography with a sensitivity and specificity of 67–84% and 77–100%, respectively [20, 21]. Hepatic ste- atosis may appear heterogeneous on sonog- raphy, with some areas of liver parenchyma spared by the steatosis. Sonography is poor in detecting smaller amounts of fat in liver. In addition, the degree of fatty change in the liver can only be subjectively classified as mild or severe on sonography. In general, severe ste- atosis is diagnosed when the hepatic paren- chymal echogenicity obscures visualization of the walls of the hepatic and portal veins.
Unenhanced CT can show fatty change by a reduced attenuation of liver density. Fatty change is suspected when the density of liver is more than 10 H below that of the spleen. When IV contrast material was used, a differ- ence in Hounsfield density of at least 20 units between liver and spleen (liver lower than spleen), at 80–100 seconds after the start of contrast injection, was found to have sensitiv- ity and specificity of 86–87% in diagnosing fatty liver [22]. That study used 150 mL of iothalamate meglumine (Conray 60, Mall- inckrodt) injected at a rate of 2 mL/s. How- ever, another study reported lower sensitivity and specificity values of 50–75% [23]. Dif- ferences between these studies may be due to confounding factors such as iron, copper, or fibrous tissue that alter the Hounsfield densi- ty of liver, and differences in the rate of con- trast injection and the timing of the scanning. In addition, CT is not sensitive in detecting mild or moderate elevations of hepatic lipid content (5–30%) [21]. Dual-energy CT has been reported to accurately determine the
Fig. 1—Natural history of nonalcoholic fatty liver disease. Small proportion of patients with fatty liver develop nonalcoholic steatohepatitis. Less than 10% of nonalcoholic steatohepatitis patients develop cirrhosis. Current research is aimed at detecting early stages of fibrosis that are potentially reversible. aProbable percentage of patients with hepatic steatosis progressing to nonalcoholic steatohepatitis. bDisease progression (%) of all nonalcoholic steatohepatitis patients.
Fig. 2—Pathogenesis of nonalcoholic steatohepatitis. Currently popular “two-hit” theory. First hit is insulin resistance, which leads to hepatic steatosis. Fatty liver is less able to cope with oxidative stress, which is second hit, leading to chronic liver inflammation. Purported factors causing liver damage include free radical formation, cytokine release, iron overload, and altered mitochondrial energy production [71].
Nonalcoholic steatohepatitis
Stable disease
Hepatic steatosis
Iron overload Impaired mitochondrial
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 70
.1 29
.2 0
on 0
3/ 23
/2 3
fr om
I P
ad dr
es s
27 .7
0. 12
9. 20
. C op
yr ig
ht A
R R
S. F
or p
er so
AJR:190, April 2008 995
Nonalcoholic Fatty Liver Disease
fatty content of liver in animals [24], although this finding has not been reproduced in hu- mans [25, 26].
Water and fat protons have slightly differ- ent precessional frequencies in a magnetic field. As a result, MRI has the potential for diagnosing fatty infiltration both qualitatively and quantitatively. The most commonly used quantitative method is the acquisition of the so-called in- and out-of-phase imaging, or chemical shift imaging, in which different TE parameters are used in gradient-echo images
so that the signal from fat protons is added or subtracted, respectively, from the signal from protons in water. With 1.5-T magnets, the fat and water protons are in phase or out of phase when the TE is an even or odd multiple, re- spectively, of 2.2 milliseconds. Reduction of signal on out-of phase T1-weighted images has been shown to be an accurate predictor of hepatic fat content (Fig. 4), with correlation (r) of 0.86–0.91 compared with the histologic assessment of liver fat [27–29]. MRI is supe- rior to sonography in assessing liver fat [30].
Chemical shift imaging has been shown to be useful in assessing reduction in fatty liver af- ter insulin sensitizing therapy [31].
With 3-T MRI, the order in which in-phase and out-of phase echoes are acquired varies depending on the scanner vendor. In our prac- tice, TEs of 2.5 and 6.2 milliseconds, respec- tively, are used for in-phase and out-of-phase imaging with the Magnetom Trio (Siemens Medical Solutions). With other 3-T MRI scan- ners, the out-of-phase echo may be acquired first. Susceptibility artifact from hepatic iron becomes more pronounced as the TE increas- es. Therefore, depending on the 3-T scanner, the presence of iron may variably affect the signal intensity of fatty liver. This confound- ing factor is less problematic when T2-weight- ed fast spin-echo sequences, which contain multiple refocusing pulses, are used. A study of cirrhotic and noncirrhotic patients per- formed at 1.5-T MRI found that the signal drop off on T2-weighted fast spin echo with- out and with fat saturation better correlated with fat content determined histologically than in- and out-of-phase T1-weighted gradi- ent-recalled images (r = 0.76 vs r = 0.25; p < 0.01 in cirrhotic patients; r = 0.92 vs r = 0.69; p < 0.01 in noncirrhotic patients) [32].
It is possible to use specialized MR pulse sequences to quantitatively assess the degree of fat in the liver with greater accuracy than is possible with CT. The original approach was based on the Dixon technique, a spin- echo method analogous to in-phase and out- of-phase imaging, in which a pair of spin- echo images is acquired. In one of these, the timing of the 180° focusing pulse is altered so that both in-phase and out-of-phase images
Fig. 3—Photomicrograph shows histology of nonalcoholic steatohepatitis in 62-year-old woman. Mallory’s hyaline bodies (pink filamentous structures, black arrowhead) are cytoplasmic inclusions in hepatocytes consisting of abnormal keratin, hyaline, and other proteins. They are usually found in hepatocytes that are ballooned (black arrow) and are morphologic hallmarks of alcoholic and nonalcoholic steatohepatitis. Mallory’s bodies are not cause but rather consequence of cellular injury. Usually hepatocytes with Mallory’s bodies do not contain large fat vacuoles, although microvesicular fat may be seen. In this frame, other hepatocytes are present, containing macrovesicular fat globules (white arrow), which occupy almost all cytoplasm, displacing nucleus (white arrowhead) to periphery. (H and E, × 400) (Courtesy of Romil Saxena, Department of Pathology, Indiana Universtiy School of Medicine)
A B Fig. 4—Chemical shift MRI detection of hepatic fat in 56-year-old man with nonalcoholic steatohepatitis. A and B, Liver appears diffusely hypointense on out-of-phase gradient-echo sequence (TR/TE, 130/2.2; flip angle, 70°) (B) compared with in-phase sequence (130/4.9; flip angle, 70°) (A), indicating presence of fat.
D ow
nl oa
de d
fr om
w w
w .a
jr on
lin e.
or g
by 2
7. 70
.1 29
.2 0
on 0
3/ 23
/2 3
fr om
I P
ad dr
es s
27 .7
0. 12
9. 20
. C op
yr ig
ht A
R R
S. F
or p
er so
Lall et al.
are obtained [33, 34]. These two images can then be mathematically added and subtracted to produce lipid-only and water-only images. With suitable calibration, quantification of fat is possible. Quantitative chemical shift imaging, based on an additional dimension of Fourier transformation to achieve separa- tion water and fat signals, can also be used to
measure lipid content [35]. Many additional approaches are described in the literature, including recent descriptions of quantitative gradient-echo methods that have been used to measure the degree of hepatic steatosis [36]. Many of these techniques are easy to perform and may be used as part of routine imaging assessment of the liver [32].
Proton (1H) MR spectroscopy has been found to have good correlation with hepatic lipid content (Fig. 5), as determined by liver biopsy, with correlation (r) values of 0.91–0.98 [37–39]. This technique is considered sensitive to small variation (as little as 0.5% change) in hepatic lipid content and has a potential use in assessing therapy of steatosis [39].
g
e d c a
A B Fig. 5—Proton (1H) hepatic MR spectroscopy. A and B, Single-voxel spectrum in 39-year-old-woman with hepatic steatosis (A) and 29-year-old asymptomatic male volunteer (B). In healthy volunteer, only resonances of water (f) and methylene (b), found in hepatic triglycerides and fatty acids, are discernible. Normal liver contains less than 5% fat by weight. In patient with hepatic steatosis, amplitude of methylene resonance (b) is much higher. Several other lipid resonances are now visible. Note chemical shifts of various resonances in 1H MR spectroscopy at 1.5 T: a, terminal methyl (CH3): 0.8 ppm; b, methylene (CH2)n: 1.2 ppm; c, CH2-C = C: 1.9 ppm; d, C = C-CH2-C = C: 2.6 ppm; e, CH2-O-COR: 4.15 ppm; f, water (H2O): 4.7 ppm; g, CH = CH and CH = O: 5.2 ppm.
b
f
α−ATP
β−ATP
A C Fig. 6—Phosphorus (31P) hepatic MR spectroscopy in healthy patient at 1.5 T. (Reprinted with permission from Cortez-Pinto H, Chatham J, Chacko VP, Arnold C, Rashid A, Diehl AM. Alterations in liver ATP homeostasis in human nonalcoholic steatohepatitis: a pilot study. JAMA 1999; 282:1659–1664 [42]. Copyright American Medical Association © 1999. All rights reserved.) A, There are six main resonances. PME = phosphomonoesters, Pi = inorganic phosphate, PDE = phosphodiesters; γ-ATP, α-ATP, and β-ATP = γ, α, and β phosphates of adenosine triphosphate (ATP). B, 15 minutes after fructose infusion, ATP resonances are reduced in amplitude but Pi is maintained. C, 60 minutes after infusion, ATP resonances recover. In patients with nonalcoholic steatohepatitis this recovery is impaired (not shown).
D20D15D10D5 ppm
Minimum
05
Pi
10
B
Recovery
0510
Detection of Inflammation in Nonalcoholic Steatohepatitis
Sonography, CT, and MRI are insensitive in differentiating hepatic steatosis from non- alcoholic steatohepatitis [21]. A small CT study of patients with nonalcoholic fatty liver disease found that those with nonalcoholic steatohepatitis had increased liver size and increased caudate lobe-to-right lobe size ra- tio, compared with those with steatosis only [40]. The caudate-to-right lobe size ratio was statistically higher in steatohepatitis (mean, 0.43; range, 0.31–0.55) compared with ste- atosis only (mean, 0.36; range, 0.22–0.47). However, measurements showed consider- able overlap in both categories, and it is un- likely that these measurements will be useful in individual patients.
In the future, MR spectroscopy may offer a role in the diagnosis of nonalcoholic steatohep- atitis. MR spectroscopic studies on hydrogen, phosphorus, and sodium have been reported. The latter two chemical species show promise.
The typical phosphorus (31P) hepatic MR spectroscopy at 1.5 and 3 T shows six dominant
resonances, denoting phosphomon oe sters, inorganic phosphate (Pi), phosphodiesters, and three resonances from the nucleoside triphosphates that are mainly composed of adenosine triphosphate (ATP) (Fig. 6). When the signal from phosphocreatine is seen, this is most likely due to the erroneous incorpora- tion of adjacent muscle in the voxel. Several studies have investigated the utility of phos- phorus (31P) MR spectroscopy in the diagno- sis of acute and chronic liver damage in animal models and humans [41–44]. Most studies show that phosphorus metabolite ratios (e.g., ATP/Pi) do not differ significantly between normal and damaged or diseased liver [42, 43]. Two problems are inherent to 31P MR spectroscopy. The spatial resolution, even with a 3-T magnet, is limited to 2–3 cm; thus, this technique is suitable only for assessing diffuse liver disease. In addition, several homeo static mechanisms maintain levels of phosphate metabolites and inorganic phos- phate at constant levels and limit the sensitiv- ity of 31P MR spectroscopy to pathophysio- logic changes. To overcome this problem, the
recovery of cellular energy mechanisms after the depletion of ATP by an IV infusion of fructose, a form of pharmacologic hepatic stress testing, has been investigated in rat models [41]. A study of nonalcoholic steato- hepatitis patients showed that, after fructose- induced depletion, recovery of hepatic ATP is severely impeded [42] (Fig. 6). This tech- nique may help differentiate patients with he- patic steatosis from those with nonalcoholic steatohepatitis. It also suggests that impair- ment of energy homeostasis may be an inte- gral part of the progression from fatty liver to nonalcoholic steatohepatitis.
Using energy-intensive processes, hepato- cytes maintain intracellular sodium concentra- tion (Nai+) of 20 mmol/L and an extracellular concentration (Nae+) of approximately 140 mmol/L. With hepatocellular damage, Nai+ concentration increases. Unfortunately, be- cause sodium is ionic, Nai+ and Nae+ have the same resonance and cannot be separated on routine MR spectroscopy. Several techniques have been proposed to overcome this, of which the most promising in humans is multiple- quantum transfer coherence filtering [45, 46]. Sodium (23Na) has three quantum energy lev- els and four possible spin orientations (Fig. 7). When the 23Na cation is transiently bound to macromolecules, as happens in the intracellu- lar compartment, double or triple quantum transitions can occur. In the normal extracellu- lar space, which is predominantly aqueous, most quantum transitions are single. In liver disease, the multiple-quantum transfer coher- ence filtering 23Na signal may increase step- wise from benign fatty liver to hepatitis and then to cirrhosis. Initially, the spectral signal increases due to elevated Nai
+ concentration from cellular damage in hepatitis (Fig. 7) lead
3/2
1/2
−1/2
SQ outer
A
B
Fig. 7—Sodium (23Na) MRI. A, Sodium (23Na) has spin quantum number (I) of 3/2 and four possible spin orientations (−3/2, −1/2, +1/2, +3/2). Three single-quantum (SQ) transitions are possible: an “inner” or −1/2 ↔ +1/2 transition, and two “outer” or −3/2 ↔ −1/2 and + 1/2 ↔ +3/2…
Related Documents
