We performed targeted metabolomics with machine learning (ML)-based interpretation to identify metabolites that distinguish the progression of nonalcoholic fatty liver disease (NAFLD) in a cohort. Plasma metabolomics analysis was conducted in healthy control subjects (n = 25) and patients with NAFL (n = 42) and nonalcoholic steatohepatitis (NASH, n = 19) by gas chromatography-tandem mass spectrometry (MS/MS) and liquid chromatography-MS/MS as well as RNA sequencing (RNA-seq) analyses on liver tissues from patients with varying stages of NAFLD (n = 12). The resulting metabolomic data were subjected to routine statistical and ML-based analyses and multi-omics interpretation with RNA-seq data. We found 6 metabolites that were significantly altered in NAFLD among 79 detected metabolites. Random-forest and multinomial logistic regression analyses showed that eight metabolites (glutamic acid, cis-aconitic acid, aspartic acid, isocitric acid, α-ketoglutaric acid, oxaloacetic acid, myristoleic acid, and tyrosine) could distinguish the three groups. Then, the recursive partitioning and regression tree algorithm selected three metabolites (glutamic acid, isocitric acid, and aspartic acid) from these eight metabolites. With these three metabolites, we formulated an equation, the MetaNASH score that distinguished NASH with excellent performance. In addition, metabolic map construction and correlation assays integrating metabolomics data into the transcriptome datasets of the liver showed correlations between the concentration of plasma metabolites and the expression of enzymes governing metabolism and specific alterations of these correlations in NASH. Therefore, these findings will be useful for evaluation of altered metabolism in NASH and understanding of pathophysiologic implications from metabolite profiles in relation to NAFLD progression.

Background: Several noninvasive tools are available for the assessment of nonalcoholic fatty liver disease (NAFLD) including clinical and blood biomarkers, transient elastography (TE), and magnetic resonance imaging (MRI) techniques, such as proton density fat fraction (MRI-PDFF) and magnetic resonance elastography (MRE). In the present study, we aimed to evaluate whether magnetic resonance (MR)-based examinations better discriminate the pathophysiologic features and fibrosis progression in NAFLD than other noninvasive methods.Methods: A total of 133 subjects (31 healthy volunteers and 102 patients with NAFLD) were subjected to clinical and noninvasive NAFLD evaluation, with additional liver biopsy in some patients (n=54).Results: MRI-PDFF correlated far better with hepatic fat measured by MR spectroscopy (r=0.978, P<0.001) than with the TE controlled attenuation parameter (CAP) (r=0.727, P<0.001). In addition, MRI-PDFF showed stronger correlations with various pathophysiologic parameters for cellular injury, glucose and lipid metabolism, and inflammation, than the TE-CAP. The MRI-PDFF and TE-CAP cutoff levels associated with abnormal elevation of serum alanine aminotransferase were 9.9% and 270 dB/m, respectively. The MRE liver stiffness measurement (LSM) showed stronger correlations with liver enzymes, platelets, complement component 3, several clinical fibrosis scores, and the enhanced liver fibrosis (ELF) score than the TE-LSM. In an analysis of only biopsied patients, MRE performed better in discriminating advanced fibrosis with a cutoff value of 3.9 kPa than the TE (cutoff 8.1 kPa) and ELF test (cutoff 9.2 kPa).Conclusion: Our results suggest that MRI-based assessment of NAFLD is the best non-invasive tool that captures the histologic, pathophysiologic and metabolic features of the disease.

Yoon CI, et al. J Korean Surg Soc. 2012 Jan;82(1):18-27. http://dx.doi.org/10.4174/jkss.2012.82.1.18

Park YH, et al. Yonsei Med J. 2017 Sep;58(5):1025-1030. https://doi.org/10.3349/ymj.2017.58.5.1025

Lee WK, et al. Yonsei Med J. 2014 Jan;55(1):149-156. http://dx.doi.org/10.3349/ymj.2014.55.1.149