Magnetic resonance (MR) imaging is a robust, noninvasive tool with the ability to perform a wide array of diagnostic functions within the abdomen. Over the last decade, there has been tremendous interest in the development of MR imaging–based biomarkers for the qualitative and quantitative assessment of liver fat and iron. The abnormal accumulation of intracellular fat and iron within the liver can lead to chronic liver disease and increase the risk of developing cancer. This article reviews MRI techniques used to assess liver fat and iron.
Key points
- •
Quantitative magnetic resonance (MR) imaging–based biomarkers for liver fat and iron have evolved rapidly over the last decade.
- •
Quantitative MR imaging–based biomarker techniques are accurate, reproducible, cost-effective, and noninvasive methods to both qualitatively and quantitatively assess liver disease and its related complications.
- •
The emerging pandemic of nonalcoholic fatty liver disease and the recognition of the role of the metabolic syndrome on the development of chronic liver disease remain driving forces for the ongoing technical development and validation of MR imaging–based biomarkers.
Introduction
Nonalcoholic fatty liver disease (NAFLD) is a form of chronic liver disease that encompasses a spectrum of hepatic pathologic abnormalities ranging from isolated steatosis, nonalcoholic steatohepatitis (NASH), fibrosis, and cirrhosis. Abnormal accumulation of fat in hepatocytes, largely in the form of triglycerides (unsaturated fatty acids), provides the foundation for the development of NAFLD. The pathogenesis of steatosis is likely multifactorial, involving genetic, environmental, and nutritional factors regulating lipid metabolism and the corresponding ebb and flow of free fatty acids (saturated fatty acids) from hepatocytes. Unlike isolated steatosis, NASH is characterized by steatosis with superimposed necroinflammation, ballooning degeneration, and fibrosis. Although isolated steatosis may be clinically stable, NASH can progress to cirrhosis in a significant percentage of patients. In fact, end-stage NASH accounts for a large percentage of patients with idiopathic or cryptogenic cirrhosis and can result in the development of hepatocellular carcinoma (HCC). It should be noted that there are several other causes for the excessive accumulation of triglycerides within hepatocytes, including alcohol, viral hepatitides including human immunodeficiency virus, genetic lipodystrophies, and treatment effect following chemotherapy ( Fig. 1 ).
Ferritin dysregulation, excess intestinal iron absorption, and repeated blood transfusions result in elevated hepatic iron content. Increased hepatic iron clearly contributes to cirrhosis and the development of HCC in patients with hereditary hemochromatosis (HH). Evidence also suggests that elevated hepatic iron results in progressive fibrosis in NAFLD and NASH and increases the risk for the development of HCC.
This article reviews magnetic resonance (MR) imaging techniques for the quantification of hepatic fat and iron. The content is divided into the following sections:
- •
Overview of the dynamic process of NAFLD and important related diseases resulting from elevated liver fat
- •
Overview of iron metabolism and deposition and important related diseases resulting from elevated liver iron
- •
Brief review of the qualitative and invasive quantitative assessment of liver fat and iron, with strengths and limitations
- •
Description of state-of-the-art noninvasive quantification of liver fat and iron with MR imaging
After reading this content, the reader should understand the scope of chronic liver disease related to elevated hepatic fat and iron and the limitations of both liver biopsy and qualitative assessment of liver fat and iron. The reader will become familiar with quantitative noninvasive methods for measuring liver fat and iron with MR imaging.
Introduction
Nonalcoholic fatty liver disease (NAFLD) is a form of chronic liver disease that encompasses a spectrum of hepatic pathologic abnormalities ranging from isolated steatosis, nonalcoholic steatohepatitis (NASH), fibrosis, and cirrhosis. Abnormal accumulation of fat in hepatocytes, largely in the form of triglycerides (unsaturated fatty acids), provides the foundation for the development of NAFLD. The pathogenesis of steatosis is likely multifactorial, involving genetic, environmental, and nutritional factors regulating lipid metabolism and the corresponding ebb and flow of free fatty acids (saturated fatty acids) from hepatocytes. Unlike isolated steatosis, NASH is characterized by steatosis with superimposed necroinflammation, ballooning degeneration, and fibrosis. Although isolated steatosis may be clinically stable, NASH can progress to cirrhosis in a significant percentage of patients. In fact, end-stage NASH accounts for a large percentage of patients with idiopathic or cryptogenic cirrhosis and can result in the development of hepatocellular carcinoma (HCC). It should be noted that there are several other causes for the excessive accumulation of triglycerides within hepatocytes, including alcohol, viral hepatitides including human immunodeficiency virus, genetic lipodystrophies, and treatment effect following chemotherapy ( Fig. 1 ).
Ferritin dysregulation, excess intestinal iron absorption, and repeated blood transfusions result in elevated hepatic iron content. Increased hepatic iron clearly contributes to cirrhosis and the development of HCC in patients with hereditary hemochromatosis (HH). Evidence also suggests that elevated hepatic iron results in progressive fibrosis in NAFLD and NASH and increases the risk for the development of HCC.
This article reviews magnetic resonance (MR) imaging techniques for the quantification of hepatic fat and iron. The content is divided into the following sections:
- •
Overview of the dynamic process of NAFLD and important related diseases resulting from elevated liver fat
- •
Overview of iron metabolism and deposition and important related diseases resulting from elevated liver iron
- •
Brief review of the qualitative and invasive quantitative assessment of liver fat and iron, with strengths and limitations
- •
Description of state-of-the-art noninvasive quantification of liver fat and iron with MR imaging
After reading this content, the reader should understand the scope of chronic liver disease related to elevated hepatic fat and iron and the limitations of both liver biopsy and qualitative assessment of liver fat and iron. The reader will become familiar with quantitative noninvasive methods for measuring liver fat and iron with MR imaging.
NAFLD
NAFLD is considered the hepatic manifestation of the metabolic syndrome, a product of insulin resistance. The metabolic syndrome comprises several clinical manifestations, including hypertension, hyperlipidemia, glucose intolerance and type 2 diabetes mellitus (T2DM), and obesity. NAFLD affects 30% of the adult population (more than 100 million Americans) and up to 80% of obese adults (body mass index >30), particularly the centrally obese (visceral adiposity) and diabetic population. In children, fatty liver is regarded as the most common cause of chronic liver disease with a prevalence of 9.6% of 2 to 19 year olds, representing greater than 6.5 million US children and 38% of obese children.
Although the prevalence of NAFLD is most strongly associated with visceral (central) obesity and features of the metabolic syndrome, age, gender, ethnicity, and race also play a role. In general, the rate of NAFLD increases with age with women, demonstrating a later peak, possibly related to hormonal changes of menopause. NAFLD more often affects men and in population studies is more common in Hispanic Americans compared with non-Hispanic white and black Americans. Although determining racial differences has been somewhat problematic, several studies suggest that the prevalence of NAFLD may be lower in black Americans. Browning and colleagues measured hepatic triglyceride content (with MR spectroscopy) in 2287 subjects and found that the incidence of steatosis in black American adults was significantly lower than in Hispanic and non-Hispanic white Americans, independent of obesity or diabetes.
In children, the effects of ethnicity and race may be even more strongly correlated than in adults. Even though non-Hispanic black American children have higher rates of risk factors for fatty liver, such as obesity and insulin resistance, the prevalence of fatty liver (1.5%) is significantly lower compared with Hispanic (11.8%) and non-Hispanic white (8.6%) American children.
Prior retrospective case series of patients with NASH and cryptogenic cirrhosis suggest heritability. Schwimmer and colleagues found that fatty liver is a highly heritable trait (h 2 = 1.0) and Abdelmalek and colleagues found that insulin resistance is more prevalent in first-degree relatives of patients with NAFLD than those without NAFLD.
NASH and cryptogenic cirrhosis
The presence and severity of NAFLD, NASH, and fibrosis increase with age, with most complications arising from cryptogenic cirrhosis occurring in the 6th to 8th decade. The incidence of NASH, the keystone element between hepatic steatosis and cryptogenic cirrhosis, may be as high as 30% to 40% in NAFLD. Visceral (central) obesity, as opposed to total adiposity, is most strongly linked to the conversion of NAFLD to NASH. Obesity, T2DM, and severity of baseline hepatic fibrosis are most closely associated with progressive fibrosis/cirrhosis. Similar to those with cirrhosis from untreated (or unresponsive to treatment) hepatitis C, those who develop NASH and/or advanced hepatic fibrosis are at risk to develop liver-related complications within 7 years ( Fig. 2 ). The age-adjusted mortality of cirrhosis is second only to cancer and exceeds that of cardiovascular disease according to the Verona diabetes study. Thus, it is becoming increasingly apparent that NAFLD complicated by fibrosis and NASH-related cirrhosis is correlated with increased mortality. In addition, NAFLD is associated with a several-fold increased risk of developing cardiovascular disease with a resulting increased risk of mortality.
HCC
Primary liver cancer is the fifth most common cancer and the third most common cause of cancer-related mortality, second to lung and gastric cancer. HCC accounts for most primary liver cancers and is typically associated with chronic hepatitis B and C infection. The incidence of HCC in developed countries has increased approximately 80% over the last 2 decades with hepatitis C accounting for approximately half of the cases. Cirrhosis is the single most important risk factor for the development of HCC and is present in approximately 80% of patients with HCC, regardless of underlying cause. This statement suggests that cryptogenic cirrhosis, most commonly the result of NAFLD, constitutes a large percentage of the underlying liver disease in patients who go on to develop HCC. Obesity, diabetes, and elevated hepatic iron content are independent risk factors associated with the development of HCC while also associated with the progression of NAFLD to NASH ( Fig. 3 ). In patients with chronic hepatitis C, hepatic steatosis, obesity, T2DM, and elevated hepatic iron content increase the risk for developing HCC.
Hepatic iron overload
Ferritin, the primary protein for intracellular storage of iron, acts as a buffer to maintain iron homeostasis within the body. In the liver, ferritin is primarily stored in Kupffer cells as part of the reticuloendothelial system (RES). However, in the setting of chronic liver disease, hepatic iron storage is more heterogeneous and can be found almost exclusively in hepatocytes, almost exclusively in Kupffer cells, or in a mixed pattern. Elevated hepatic iron increases cellular oxidative stress, resulting in the creation of reactive oxygen species that damages cell and organelle membranes, nucleic acids, and proteins, ultimately leading to cell death and progressive organ dysfunction.
Overexpression of ferritin is seen in both primary (HH) and secondary (hemosiderosis) iron overload disorders, resulting in elevated total body iron (circulating + stored). In addition, ferritin is an acute-phase protein that can be induced in the setting of systemic inflammation, resulting in hyperferritinemia. Elevated serum ferritin and hepatic iron have been reported in the setting of viral and alcohol-related cirrhosis as well as in obesity-related inflammatory conditions, such as diabetes, metabolic syndrome, and NAFLD. However, increased hepatic iron deposition may or may not be associated with hyperferritinemia, thus limiting the ability of simple blood tests to evaluate for elevated hepatic iron.
Elevated hepatic iron content clearly contributes to the progression and severity of liver disease, as seen in HH. Although the precise contribution remains unclear, there is evidence that supports the association between the presence of elevated hepatic iron content and the degree of hepatic fibrosis. More recent studies have shown that elevated hepatic iron content is associated with disease severity in NAFLD. In fact, Nelson and colleagues and Kowdley and colleagues showed that excessive accumulation of iron within the RES correlated with severity of fibrosis in NASH.
In addition to ferritin dysregulation, hepatic iron overload occurs in the setting of increased intestinal absorption and/or recurring blood transfusions. Iron homeostasis is regulated by intestinal absorption of iron; however, the body does not have the capacity to reduce excess total body iron effectively. With increased intestinal iron absorption (eg, HH and ineffective erythropoiesis such as untreated thalassemia), iron initially accumulates in periportal hepatocytes. Without treatment, excess iron continues to accumulate in hepatocytes throughout the liver, extending from the portal triad, into Kupffer cells, and into the biliary epithelium. Eventually the iron storage capacity of the liver is overwhelmed, resulting in spillage of iron into the systemic circulation, where iron is bound to transferrin. Transferrin then delivers excess iron to cells that contain high transferrin-receptor density, resulting in elevated tissue iron content within the pancreas, myocardium, thyroid, pituitary, and gonads ( Fig. 4 ). The RES does not sequester transferrin-bound iron; instead, senescent erythrocytes are phagocytized with the processed iron either stored as intracellular ferritin or released into circulation bound to transferrin. Thus, the spleen and marrow are spared in those diseases that result from increased intestinal absorption. In contrast, diseases that require repeated blood transfusions (eg, treated thalassemia, sickle cell disease [SCD], hemolytic anemia, and myelodysplastic syndromes) result in an accumulation of iron, in the form of ferritin, within the spleen, bone marrow, and hepatic Kupffer cells of the liver (organs/cells of the RES). Following 40 to 50 blood transfusions (approximately 10 g of iron), the ability of cellular ferritin to sequester iron in the RES is overcome. Excess iron then extends out into the liver hepatocytes and eventually spills over into the circulation, resulting in an accumulation of iron in the myocardium, pancreas, endocrine glands, and skin ( Fig. 5 ). Within the group of diseases requiring repeated blood transfusions, the degree of oxidative stress on the liver and resulting fibrosis is heterogeneous, thought to be the result of labile (or nontransferrin-bound) plasma iron. For example, patients with thalassemia major tend to have a greater degree of hepatic fibrosis compared with those with SCD at comparable levels of elevated liver iron.
Treatment of fatty liver disease and iron overload
The incidence of NAFLD is increasing worldwide, particularly in the developed world, paralleling the obesity epidemic. Although isolated fatty liver can be indolent, hepatic steatosis can progress to hepatitis, fibrosis, and cirrhosis with increased risk for malignancy. NAFLD is a herald for the metabolic syndrome and increases the risk for T2DM, obesity, and cardiovascular disease, resulting in increased morbidity and early mortality. Fortunately, with intervention, NAFLD and NASH are reversible, and lowering liver fat may reduce associated risks.
Lifestyle modifications, medications, and surgery have been used in the management of hepatic steatosis with variable degrees of success. Modest, sustained weight reduction (5%–10%) with or without exercise has been shown to reduce hepatic steatosis, suspend hepatic inflammation associated with NASH, and reverse fibrosis. In addition, the reduction of hepatic steatosis is correlated with a reduction in insulin resistance. A similar pattern of reduction in degree of hepatic steatosis, resolution of NASH, and reversal of hepatic fibrosis has been documented in patients who opt for surgery to assist in weight loss. It should be noted, however, that rapid weight reduction (>2 lbs/week) has been associated with worsening liver function, progressive hepatic fibrosis, and/or liver failure. Many pharmacologic agents have been used in an attempt to treat NAFLD and/or NASH, but only vitamin E and pioglitazone have demonstrated sustained efficacy. Metformin, which acts by suppressing hepatic glucose production and improving hepatic insulin resistance, is the first-line therapy for the treatment of T2DM and was once thought to improve insulin sensitivity and reduce hepatic steatosis; however, several recent randomized controlled trials have now proved otherwise. Vitamin E, which acts by reducing intracellular oxidative stress, has been shown to trend toward reducing hepatic steatosis, while significantly reducing ballooning degeneration at histology, resulting in regression of NASH in adults but not in children or adolescents. Thiazolidinediones, namely pioglitazone, which acts to reduce insulin resistance, have also been shown to decrease hepatic steatosis and reduce hepatic lobular inflammation, resulting in the regression of NASH in diabetics and nondiabetics with NASH. Ongoing studies are currently investigating the effect of combined vitamin E and pioglitazone and the role of diet plus pioglitazone. Glucagon-like peptide 1 analogues (GLP-1), newer pharmacologic agents used widely for the treatment of T2DM, augment glucose-dependent insulin secretion from the pancreas. Limited data from small or uncontrolled studies with GLP-1 suggest there are beneficial effects in patients with NAFLD.
The treatment of iron overload is based on the cause, the organs involved, and the severity of iron overload. With therapy, progression to liver fibrosis can be halted and liver fibrosis may regress. Life-long therapeutic phlebotomy is the standard of care for those with HH. The goal is to remove iron in the blood (stored within red blood cells, transferrin-bound iron, and labile plasma iron), allowing efflux of iron from tissues in an effort to achieve an acceptable liver iron concentration (LIC), the surrogate for total body iron content. Because phlebotomy is not an option in transfusion-dependent patients, iron chelation therapy is required. In iron chelation therapy, free serum iron is bound and excreted in either the feces (deferasirox) or the urine (deferoxamine and deferiprone). Each medication has a unique side-effect profile, so close monitoring of iron levels is required. Recently, Beaton and colleagues demonstrated that iron reduction by phlebotomy in patients with hepatic steatosis resulted in significantly improved NAFLD activity score. This finding suggests not only that iron overload contributes to hepatic inflammation in NAFLD, possibly leading to NASH, but also that the reduction of iron can reverse hepatic inflammation that leads to progressive fibrosis.
Methods of assessing of hepatic fat and iron
Ultrasound
Sonography is the most widely available and most commonly used modality to assess hepatic steatosis. Although ultrasound lacks the ability to quantify hepatic steatosis, fatty liver can be qualitatively assessed, based on parenchymal echogenicity and echotexture, conspicuity of portal triads and beam attenuation. Because ultrasound is operator-dependent and vendor-dependent, accuracy and reproducibility suffer. In addition, the patients at greatest risk, those with visceral obesity, are technically challenging to scan because body habitus leads to beam attenuation and suboptimal evaluation of the liver. Ultrasound carries a positive predictive value of 62% to 77% for the detection of hepatic steatosis.
Sonography has no role in the evaluation of hepatic iron content, although ultrasound can assess for the presence of complications related to elevated iron including the detection of fibrosis, cirrhosis, HCC, and portal hypertension. Thus, when hepatic steatosis and elevated hepatic iron coexist, sonography is only useful in the evaluation of the steatotic component and related complications.
Computed Tomography
Computed tomography (CT), like ultrasound, is widely available and carries with it the ability to assess hepatic fat content. CT provides a more objective assessment of hepatic fat content, the result of radiographic attenuation, which results in reduced attenuation of the hepatic parenchyma. Pickhardt and colleagues demonstrated that a liver attenuation of 48 HU at unenhanced CT was 100% specific for moderate to severe hepatic steatosis, although suffered poor sensitivity of only 54%. CT is relatively insensitive for the detection of mild to moderate hepatic steatosis and, unlike ultrasound, is confounded by the presence of elevated liver iron and other metals, such as copper, gold, and iodine (from amiodarone), which increase hepatic attenuation by way of radiation beam absorption. Last, ionizing radiation created by CT precludes widespread utilization for diagnosis and serial quantification of liver fat during and following medical and surgical management, particularly in children.
Nontargeted Liver Biopsy
Nontargeted percutaneous liver biopsy and direct histologic characterization are the currently accepted reference standards for the diagnosis of diffuse liver disease, including hepatic steatosis and iron overload. Comprehensive evaluation of tissue includes grading steatosis and iron deposition, hepatic fat and iron zonality, fat droplet size (macrovesicular or microvesicular), inflammation, cellular injury, and degree of fibrosis. A variety of scoring systems exist that use the histologic and/or clinical data to provide a semi-quantitative score of the fatty liver disease severity and response to therapy, including the Brunt scoring system, the NAFLD Activity Score, and the METAVIR scoring system (for hepatitis C). Although the semi-quantitative scale for assessing iron deposition proposed by Rowe and colleagues is the most accurate and reproducible histologic method for grading LIC, atomic absorption spectrophotometry is the most accurate and direct measure of LIC. Unfortunately, these systems are not widely available and preclude concurrent histologic evaluation of the liver specimen.
Liver biopsy is subject to substantial limitations. Although liver biopsy is not technically challenging, it does carry a low risk of significant bleeding (1%–3%), possibly requiring intervention, hospitalization, and very rarely, death (1:10,000). The right lobe of the liver (hepatic segments VIII, V, and/or VI) is more often targeted in random liver biopsies because of the assumption that the right liver is more frequently diseased. However, fat and iron deposition is often heterogeneous and can spare the right liver. Coupled with the fact that liver biopsy specimens represent 1/50,000th the overall liver volume, it is not difficult to understand the degree of sampling variability associated with liver biopsy, which has been borne out in the literature in a variety of ways, including low κ-reliability scores comparing 2 closely approximated biopsies, high sampling variability among samples, dependence on core sample length and number of specimens, and the effect of underlying liver disease on sampling error.
Phlebotomy
Phlebotomy can be a useful adjunct in the evaluation of total body iron content, an indirect measure of LIC. However, phlebotomy cannot be used in transfusion-dependent patients and has no role in the evaluation of hepatic steatosis, other than determining the presence or absence of serologic signs of metabolic syndrome. Serum markers used to follow trends in iron status include ferritin, serum iron, nontransferrin-bound iron (NTBI), total iron binding capacity, and transferrin saturation (TSAT). Elevated serum ferritin infers prognostic value for increased cardiac-related mortality among patients with thalassemia. However, the use of serum ferritin is limited because it is subject to fluctuations resulting from inflammation and vitamin C deficiency and has not been adequately documented to be predictive in disorders other than thalassemia. NTBI accumulates in the blood when transferrin is highly saturated and is responsible for liver, myocardium, and endocrine gland iron loading with resultant toxicity. Both NTBI (assayed as labile plasma iron) and TSAT can be quantified; however, both are subject to variation from inflammation and substantial inter-laboratory variability, among other limitations impeding their utility.
Qualitative MR Imaging Techniques
There are a variety of ways to volumetrically assess liver fat and parenchymal iron deposition at MR imaging. Conventional imaging (T1 and T2) without and with fat suppression is useful in the qualitative assessment of hepatic fat, whereas T2-weighted imaging including spin-echo/fast spin-echo and gradient-echo imaging can be used to indirectly evaluate for the presence of increased hepatic iron as well as increased parenchymal iron in the RES. Even though iron causes T1 shortening, T2* decay dominates signal augmentation at clinical field strengths, largely obscuring signal changes at T1. In-phase and opposed-phase (also referred to as “out-of-phase”) imaging (IOP), on the other hand, is useful for evaluation of both liver fat and parenchymal iron; however, when both increased fat and iron are present in the liver, the effect of T2* can lead to underdiagnosis of both liver fat and iron because of in-phase loss of hepatic signal, resulting in less or even no apparent liver signal change between IOP.
Conventional T1-weighted and T2-weighted fast spin-echo (T2FSE) MR imaging obtained without and with fat saturation is an approach that can be used to detect liver fat. In the presence of elevated liver fat, fat-suppression pulses suppress signal from fat, resulting in an overall decrease in hepatic signal. These methods, particularly with T2FSE, have been shown to correlate better than IOP imaging with steatosis grade at biopsy ( Fig. 6 ).
Conventional T2FSE MR imaging obtained with or without fat suppression can be useful in the evaluation of parenchymal iron. It should be noted that MR imaging does not and can not directly measure parenchymal iron; MR imaging detects the effects of parenchymal iron. Parenchymal iron accelerates T2 and T2* decay, resulting in signal loss on T2FSE and gradient echo MR imaging, respectively. The presence of elevated iron is inferred when signal from the parenchyma is lower than expected for normal. The echo time and degree of parenchymal iron loading affect MR imaging signal loss. In general, the longer the echo time and/or greater parenchymal iron overload, the greater the signal loss.
Patients with secondary hemosiderosis and HH can generally be distinguished at T2FSE based on the pattern of parenchymal involvement. Parenchymal iron loading in patients with secondary hemosiderosis (thalassemia, SCD, and other transfusion-dependent anemias) generally follows a predictable pattern involving the spleen, liver, bone marrow, and lymph nodes. The capacity of the RES has been exceeded when the pancreas, myocardium, and muscles demonstrate signal loss. In contrast, parenchymal iron loading with HH preferentially affects the liver, pancreas, and myocardium and spares the organs of the RES. The liver is generally spared of iron loading in patients with SCD who do not require blood transfusions; however, the renal cortices can demonstrate signal loss from iron loading resulting from intravascular hemolysis ( Fig. 7 ). The distribution of increased liver iron is usually homogeneous; however, it can be heterogeneous, focal, patchy, lobar, or segmental particularly in patients with cirrhosis who can frequently develop siderotic nodularity.