Nonalcoholic fatty liver disease (NAFLD) is a form of chronic liver disease with histopathologic features of typical alcohol-induced chronic liver disease that develops in individuals who do not consume alcohol excessively1). Since Ludwig et al.2) described nonalcoholic steatohepatitis (NASH) in 1980, there have been remarkable advances in diagnosing NAFLD. The disease spectrum of NAFLD ranges from simple steatosis to NASH, nonalcoholic hepatic fibrosis, and nonalcoholic cirrhosis, and NASH progresses towards cirrhosis, even in children3-5). Therefore, early detection of hepatic fibrosis and accurate diagnosis of NAFLD are essential, and NASH, fibrosis, and cirrhosis should be distinguished from benign simple steatosis in obese children suspected to have NAFLD. Typically, the diagnosis of NASH is based on clinical exclusion of alcohol or drug intake, followed by serologic exclusion of other underlying liver diseases, and, finally, percutaneous liver biopsy. Percutaneous liver biopsy is regarded as the gold standard for diagnosing NAFLD, even in children, because a definite NAFLD diagnosis requires histological data such as macrovesicular steatosis, necroinflammation, and hepatic fibrosis6).

As Roberts7) suggested in 2007, candidate criteria for liver biopsy in suspected pediatric NAFLD include factors such as an age of less than 10 years, hepatosplenomegaly, highly increased serum aminotransferase level, severe insulin resistance, comorbid chronic liver diseases, and family history. However, the use of liver biopsy is limited in the pediatric population because of its invasiveness and the fact that the histologic features of pediatric NASH are somewhat different from those of adult NASH8,9).

Recently, noninvasive diagnostic methods, including imaging studies for detecting steatosis or fibrosis, and biomarkers of oxidative stress, inflammation, hepatocyte apoptosis, and hepatic fibrosis have been introduced to diagnose the spectrum of NAFLD10,11). Additionally, noninvasive markers of hepatic fibrosis and noninvasive hepatic fibrosis scores have also been developed and applied for evaluating the severity of hepatic fibrosis, mainly in adults12-14).

To date, only a limited number of studies have examined noninvasive biochemical markers of hepatic fibrosis for pediatric NAFLD15-17) and only a few studies have been performed on previously suggested fibrosis scores or NAFLD-related hepatic fibrosis scores in children18-20). Although data and validation studies in the pediatric population are still lacking, these noninvasive modalities can also be applied for pediatric NAFLD.

In this review, noninvasive clinical and laboratory parameters as possible clues of NAFLD, noninvasive imaging studies for detecting hepatic steatosis (and possibly fibrosis), newly developed noninvasive biomarkers for diagnosing NAFLD, and noninvasive hepatic fibrosis scores indicating the severity of hepatic fibrosis will be discussed, with a primary focus on pediatric NAFLD.

Most cases of NAFLD in obese children and adolescents are incidentally detected by health checkups at schools or hospitals. Most patients with NAFLD are asymptomatic, except for nonspecific clinical symptoms or signs such as easy fatigability and right upper quadrant abdominal discomfort or hepatomegaly and acanthosis nigricans in some patients21,22). There are no clinical features definitely indicative of NAFLD in obese children. However, some clinical features such as obstructive sleep apnea or acanthosis nigricans may be significantly correlated with NAFLD23).

Increased levels of serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) in obese patients are well-known predictors of NAFLD, particularly NASH, although serum aminotransferase levels sometimes remain within a normal range despite the presence of NAFLD21). Mildly elevated levels of alkaline phosphatase or gamma glutamyl transpeptidase (GGT) have been associated with NAFLD21). However, the levels of these enzymes do not reflect the severity of steatosis or hepatic fibrosis in patients with NAFLD.

According to a recent study on pediatric NASH24), most NAFLD patients were obese children with a body mass index (BMI) higher than the 95th percentile, and other demographic features were not significantly different among the 3 groups examined (no NASH, adult-type NASH, and definite NASH), although age and ethnicity differed between pediatric-type NASH and definite NASH. Increased levels of serum AST and GGT were significant predictors of NAFLD among the 3 groups, as were age, prothrombin time, and insulin resistance between 2 groups24). Additionally, increased serum AST levels, white blood cell count, and decreased hematocrit levels were found to be significant predictors of moderate fibrosis in that study24). In contrast, a multicenter study in 2009 showed that none of the clinical features, including BMI or comorbid diseases, were significantly correlated with the stage of fibrosis; laboratory markers were also not correlated, except for serum AST levels25). The serum AST levels were correlated with the stage of fibrosis and could be used to distinguish significant fibrosis from no or mild fibrosis25). However, in this study, the authors concluded that AST is a good predictor of severe fibrosis in children with NAFLD, based on the low value (0.63) obtained for the area under the receiver operating curve (AUROC)25). In our recent studies, no single clinical or laboratory parameters reflected the presence or severity of hepatic fibrosis in children with NAFLD18), except for BMI, which was significantly different among children with simple steatosis, type 1 adult NASH, and type 2 pediatric NASH9).

Thus, noninvasive tools beyond clinical or laboratory markers are required to diagnose NAFLD in children, such as noninvasive imaging studies, biomarkers, or hepatic fibrosis scores.

Noninvasive imaging modalities for NAFLD include abdominal ultrasonography (USG), abdominal computed tomography (CT), magnetic resonance imaging (MRI), fibroscan (hepatic elastography), magnetic resonance (MR) elastography, and MR spectroscopy11).

Among noninvasive imaging techniques, abdominal USG has some advantages because it is a safe and economic modality showing a wide range of sensitivity and specificity in screening for the presence of hepatic steatosis in patients with NAFLD. Thus, abdominal USG is frequently used to detect the presence of fatty liver and to evaluate the degree of hepatic steatosis in the clinical field, although the diagnostic accuracy of liver USG in children has not yet been determined26). Ultrasonographic criteria for assessing hepatic steatosis include hepatorenal echo contrast, liver brightness, intrahepatic vessel blurring, and visualization of the liver parenchyma and diaphragm11). According to a recent study examining the clinical utility of ultrasonographic quantification of hepatic steatosis, liver USG is a useful tool for quantifying hepatic steatosis in children suspected to have NAFLD, because the degree of steatosis on USG was strongly correlated with the steatosis grade obtained on liver histopathological analysis26).

Unenhanced abdominal CT is useful for detecting the presence of fatty liver and visceral adiposity and sufficiently sensitive to quantify hepatic steatosis27). The diagnostic accuracy of abdominal CT was relatively high in cases where hepatic steatosis was found to be present in more than 30% of the liver when a liver-spleen house field unit ratio was applied28). However, CT is not recommended for the pediatric population due to the radiation hazard.

Compared with abdominal CT, MRI is a highly accurate imaging modality that is useful for identifying and quantifying hepatic steatosis safely, even in children, without radiation hazard29). The disadvantages of using MRI for diagnosing NAFLD include high costs and limitations in patients with claustrophobia, metallic devices, or iron overload30).

Clinically, abdominal USG, CT, and MRI have some limitations because these methods cannot accurately differentiate between simple steatosis and NASH and cannot confirm the presence or stages of hepatic fibrosis11). Recently introduced methods such as fibroscan, MR elastography, and MR spectroscopy are promising modalities expected to overcome the limitations of conventional imaging studies in detecting hepatic fibrosis or steatosis.

According to recent studies, fibroscan (transient elastography) using an ultrasonographic technique can be applied to noninvasively detect liver fibrosis by measuring liver stiffness in both adults and children, although this method has some disadvantages, including decreased performance with increase in BMI and also expensive equipment31,32). MR elastography, introduced for the assessment of hepatic fibrosis in adult patients with NAFLD, is a noninvasive imaging technique that can also be applied in children, despite some limitations33,34). Additionally, 1H-MR spectroscopy may be an alternative noninvasive method for identifying and measuring hepatic fat content in patients with NAFLD35).

Regarding the noninvasive diagnosis of pediatric NAFLD by using imaging techniques, in 2012, the Hepatology Committee of European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN) concurred that most studies examining various imaging methods had been performed with adult NAFLD patients and that only a few studies had been performed for a large group of pediatric NAFLD patients36). The committee reached a consensus in that abdominal USG is a safe method with a limited ability to quantify steatosis or fibrosis and that abdominal MRI is a very expensive, rapid, and reproducible method for measuring steatosis and fibrosis after some modifications36). Fibroscan is not yet universally recommended for pediatric NAFLD36).

The ideal biomarkers for NAFLD are those that would be simple and easy to measure, cost-effective, accurate in diagnosing NAFLD, capable of helping evaluate the stage of fibrosis and to monitor treatment response, and validated in a large or multicenter study10). There are 4 major categories of noninvasive biomarkers that can be used to indicate NAFLD not only in adults but also in the pediatric population: hepatic inflammation, oxidative stress, hepatocyte apoptosis, and hepatic fibrosis10,11,23,36).

To date, a variety of noninvasive hepatic fibrosis scores, involving of one or more clinical parameters and laboratory markers, have been developed and applied in adults with NAFLD to differentiate NASH from benign simple steatosis, particularly in terms of the presence of hepatic fibrosis14).

The AST/ALT ratio is a commonly used indicator of hepatic fibrosis because it is simple and easy to calculate; however, the diagnostic accuracy of an AST/ALT ratio>1 is very low in children with NAFLD, limiting its routine use as a hepatic fibrosis score18,24).

The AST-platelet ratio index (APRI), which includes AST and platelet count, has primarily been used to predict hepatic fibrosis in various chronic liver diseases in adults and children54,55). In our previous study on pediatric NAFLD, APRI was an excellent predictor of significant hepatic fibrosis (stage 2 or more) in children with biopsy-proven NASH18).

FibroTest, which involves total bilirubin, GGT, α2-macroglobulin, apolipoprotein A1, and haptoglobin levels, may be a good predictor of advanced fibrosis, but it has been evaluated only in adults with NAFLD and not in children56).

The FIB-4 score, which involves age, AST, ALT levels, and platelet count, has been applied in adults with NAFLD and has been found to be a better noninvasive method of scoring fibrosis than other scores57). In our recent study published in 2012, FIB-4 was first evaluated in children with biopsy-confirmed NAFLD and showed the highest diagnostic accuracy among various noninvasive hepatic fibrosis scores in differentiating significant fibrosis (stage 2 or more) from no or mild fibrosis in pediatric NAFLD18).

The enhanced liver fibrosis (ELF) test, involving hyaluronic acid, amino-terminal propeptide of type III collagen, and TIMP-1, was found to be a good predictor in adult patients with chronic liver disease58), and was then also applied to children with NAFLD in order to predict hepatic fibrosis19). The ELF test can differentiate significant fibrosis (stage 3 or more) from no or mild fibrosis, with an AUROC of 0.99 for a cutoff of 10.51 in children19).

The pediatric NAFLD fibrosis index (PNFI) is the only noninvasive hepatic fibrosis scoring system developed for children with NAFLD and involves age, waist circumference, and serum triglyceride levels20). According to previous studies, PNFI is useful for differentiating children with hepatic fibrosis from those without fibrosis with an AUROC of 0.85 and 0.761, respectively20,59). On combining the PNFI and ELF tests, the AUROC for predicting hepatic fibrosis increased up to 0.944 in children59).

Liver biopsy is regarded as the gold standard of diagnosis of NAFLD. However, the use of liver biopsy is clinically limited in children due to its invasiveness. Noninvasive diagnostic methods such as imaging studies, biomarkers of inflammation, oxidative stress, hepatic apoptosis, and hepatic fibrosis, and noninvasive hepatic fibrosis scores may be useful for overcoming these disadvantages.

Although these noninvasive methods have not been widely used or validated in pediatric NAFLD, some of them may be applicable for detecting hepatic steatosis and predicting hepatic fibrosis even in children with NAFLD.