AIM: To evaluate the possibility of treatment effect monitoring using hepatic fat quantification magnetic resonance (MR) in pediatric nonalcoholic steatohepatitis (NASH).

METHODS: We retrospectively reviewed the medical records of patients who received educational recommendations and vitamin E for NASH and underwent hepatic fat quantification MR from 2011 to 2013. Hepatic fat fraction (%) was measured using dual- and triple-echo gradient-recalled-echo sequences at 3T. The compliant and non-compliant groups were compared clinically, biochemically, and radiologically.

RESULTS: Twenty seven patients (M:F = 24:3; mean age: 12 ± 2.3 years) were included (compliant group = 22, non-compliant = 5). None of the baseline findings differed between the 2 groups, except for triglyceride level (compliant vs non-compliant, 167.7 mg/dL vs 74.2 mg/dL, P = 0.001). In the compliant group, high-density lipoprotein increased and all other parameters decreased after 1-year follow-up. However, there were various changes in the non-compliant group. Dual-echo fat fraction (-19.2% vs 4.6, P < 0.001), triple-echo fat fraction (-13.4% vs 3.5, P < 0.001), alanine aminotransferase (-110.7 IU/L vs -10.6 IU/L, P = 0.047), total cholesterol (-18.1 mg/dL vs 3.8 mg/dL, P = 0.016), and triglyceride levels (-61.3 mg/dL vs 11.2 mg/dL, P = 0.013) were significantly decreased only in the compliant group. The change in body mass index and dual-echo fat fraction showed a positive correlation (ρ = 0.418, P = 0.030).

CONCLUSION: Hepatic fat quantification MR can be a non-invasive, quantitative and useful tool for monitoring treatment effects in pediatric NASH.

Nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease in children and adolescents[1-3]. It varies in severity from simple fatty liver to nonalcoholic steatohepatitis (NASH), which may induce hepatic fibrosis, cirrhosis, and hepatocellular carcinoma[4,5]. Cirrhosis due to NAFLD has already been reported in 2 boys aged 10 and 14 years[6]. Therefore, prompt diagnosis and suitable treatment are mandatory. Proper diet, exercise, and vitamin E may help improve steatosis in liver histology[7-10].

Liver biopsy is the modality of choice to accurately diagnose NAFLD and to monitor disease progression during treatment, even in children and adolescents. However, this procedure is invasive and may result in life-threatening complications[11,12]. In addition, it has many drawbacks, including sampling error and inter- and intra-observer variability in interpretation. Liver biopsy can also be associated with complications such as abdominal pain, hypotension, hemobilia, and intraperitoneal hemorrhage, the last of which has an associated mortality rate of up to 0.5%. Furthermore, liver biopsy is not generally accepted by patients and especially not by children and adolescents[13-15]. Therefore, there is a need to develop noninvasive methods to accurately assess and monitor the progression of NAFLD in children.

Various imaging studies can be used for the noninvasive assessment of hepatic fat including ultrasonography, computed tomography (CT), and magnetic resonance (MR) imaging. Ultrasonography can detect fatty liver as increased liver parenchymal echogenicity. There is no risk of radiation exposure using this method. However, its ability to grade the fatty liver in children is limited[5], despite an attempt to quantitatively analyze hepatic fat with ultrasonography[16]. Moreover, positive ultrasonographic results in severely obese adolescents cannot be used to accurately predict the presence and severity of hepatic steatosis[17].

CT can also detect hepatic fat infiltration by analyzing hepatic parenchymal attenuation[18]. However, hepatic attenuation can also be affected by other conditions such as hemosiderin deposition[19]. Moreover, the risk of radiation exposure is a major disadvantage, especially in children. No studies have yet compared the CT assessment of hepatic steatosis with histologic grades of fatty infiltration in children[5].

Both MR imaging and MR spectroscopy can detect hepatic fat infiltration simply and accurately without radiation exposure[20], even though MR spectroscopy needs a long examination time for a small volume of the liver. When using MR imaging, quantitative measurement of the fat fraction using a chemical shift technique, which distinguishes resonant frequencies between fat and water, can easily evaluate fat in the liver. In addition, it can be obtained using the dual- or triple-echo gradient recalled-echo sequences, which can be used easily and quickly in children[21]. However, there is limited information regarding the utility of hepatic fat quantification MR in children with NASH, and no studies have used this technique to monitor treatment effects in these patients. Therefore, the purpose of this study was to evaluate the possibility of treatment monitoring with hepatic fat quantification MR as a non-invasive and quantitative method in pediatric NASH.

Dual- and triple-echo gradient recalled-echo sequences can easily perform noninvasive assessment and quantitative measurement of hepatic fat, even in children. Our study demonstrated that not only clinical and laboratory results but also fat fraction on MR improved in children with good treatment compliance. These results suggest the possible usefulness of fat quantification MR as a noninvasive and quantitative tool for monitoring treatment effects in children with NASH.

Ultrasonography is the favored method for evaluating hepatic disease in children because results are easily obtainable and it is noninvasive. Disturbed propagation of ultrasound waves can occur when there are lipid droplets within hepatocytes[5]. This causes the scattering of the waves, which creates more returning echoes to the transducer. Therefore, we can qualitatively determine fatty infiltration in the liver when the liver parenchymal echogenicity is increased when compared with the kidney. Shannon et al[16] performed a cohort study of pediatric patients with biopsy-proven NAFLD and demonstrated excellent correlation between the ultrasonographic steatosis score and the histologic grade of steatosis. However, Bohte et al[17] showed the limitations of ultrasonography in severely obese adolescents when evaluating the severity of hepatic steatosis. In addition, a recent systemic review for imaging liver fat in children concluded that ultrasonography had a low positive predictive value of 47%-62% for the diagnosis of fatty liver[5].

MR imaging is one of the most sensitive tools used to evaluate fat infiltration in the liver. Both MR imaging and MR spectroscopy can accurately and reproducibly measure hepatic fat[25]. However, MR spectroscopy evaluates only a small portion of the liver and requires a long examination time and expertise for data acquisition and analysis[23,25]. When using MR imaging, the degree of fat infiltration can be estimated by using chemical shift imaging[20]. The term chemical shift refers to the difference in precessional (or resonance) frequency between 2 proton MR signals, expressed in parts per million of the resonance frequency of the static magnetic field B0. This technique uses differences in the resonance frequencies of water and lipids to differentiate tissues containing only water from those containing both water and lipids[5]. The characteristic resonance frequencies of fat and water and the detection of a fat-specific frequency allows for quantitative measurement of hepatic steatosis. Therefore, while normal liver parenchyma exhibits similar signal intensity on IP and OP images, the fatty liver shows diminished signal intensity on OP images, as seen in our study.

The dual-echo technique is available on most clinical imagers with a field strength of 1.0T or higher. We can obtain both IP and OP images after a single radiofrequency excitation using this technique, allowing rapid acquisition of images during a single breath hold[23]. There are many reports about the utility of this technique in evaluating hepatic steatosis in adults[26]. However, there is only one report of the dual-echo technique being performed in children. In 2011, Pacifico et al[21] demonstrated a strong correlation between the dual-echo MR fat fraction and the histologic grade of hepatic steatosis among 25 children with NAFLD. No studies have yet evaluated the utility of this technique for monitoring treatment effects in children with NASH. This is the first study detecting longitudinal changes in hepatic fat in pediatric NASH, even though we did not perform histologic correlation as biopsy is not considered a proper method for treatment monitoring, especially in children. We demonstrated that not only clinical and laboratory results but also fat fraction on MR improved in children with good treatment compliance. We also showed that the BMI change and the dual-echo fat fraction change had a positive correlation. Therefore, the dual-echo technique can be a useful monitoring tool in children with fatty liver disease.

Triple-echo MR imaging is somewhat similar to dual-echo MR imaging. However, the triple-echo technique acquires a second IP image in addition to the first OP and IP images. The signal intensities of the first OP and IP images are corrected for the T2* effect using the T2* time estimated from the signal decay between the first and second IP images[25]. The dual-echo technique cannot correct for potentially confounding T2* relaxation effects, which should considered in cases with elevated liver iron content, such as cirrhosis and hemochromatosis[23,27]. Qayyum et al[28] demonstrated that fat quantification determined with chemical shift imaging correlated better with the histopathologically determined percentage of fat in non-cirrhotic patients compared with cirrhotic patients.

Several recent reports evaluated the triple-echo technique for quantification of hepatic steatosis in adults[29,30]. Hwang et al[29] demonstrated a correlation between both triple-echo MR imaging and MR spectroscopy with macrovesicular steatosis. They suggested cutoff values of 4.93% and 5.79% for the detection of substantial macrovesicular steatosis for both techniques, respectively, without a significant difference. Wu et al[30] compared dual-echo MR, triple-echo MR, and MR spectroscopy for the evaluation of hepatic steatosis and found that triple-echo MR imaging had the highest correlation. However, no studies evaluated triple-echo MR imaging in children with NASH. Our study demonstrated that both the dual-echo and triple-echo techniques showed a significant decrease in hepatic fat fraction after treatment. The mean fat fraction decreased from 34.9% to 15.8% with the dual-echo sequence and from 23.6% to 10.1% with the triple-echo sequence in the compliant group. However, we could not evaluate the effect of hepatic iron content or fibrosis between the 2 sequences in these patients. Further study is required to compare the accuracy of these sequences in children with NAFLD.

Our study has several limitations. First, diagnosis and improvement during NASH follow-up were not proven pathologically. We could not directly correlate or compare the MR fat fraction with histologic grades of fatty infiltration. We also could not evaluate or consider hepatic fibrosis in our patients. However, liver biopsy is not a routine examination in children with NASH and is not appropriate for treatment monitoring. The second limitation was the small number of patients with a short-term follow-up period. We included only children with NASH who underwent both a pre-treatment and a 1-year follow-up MR for hepatic fat quantification during the study period. Even though all children in the compliant group showed decreased hepatic fat fraction on both sequences in this short period of follow-up, long-term monitoring is mandatory in these patients. Additional prospective studies with a large number of children and long-term follow-up are needed to validate our results. The third limitation was that chemical shift MR imaging cannot differentiate the water dominant and fat dominant liver[23]. In fact, this technique cannot evaluate a fatty liver with more than a 50% fat component. However, a hepatic fat fraction of more than 50% is rare; in a study of over 2000 patients by Szczepaniak et al[31], no patients had a fat fraction exceeding 50%. Therefore, this may be only a theoretical concern.

In conclusion, hepatic fat quantification MR can be a useful tool to non-invasively and quantitatively monitor treatment effects in pediatric NASH. Furthermore, it can also help reduce the number of unnecessary biopsies in these patients.