Identification of a support vector machine-based biomarker panel with high sensitivity and specificity for nonalcoholic steatohepatitis

Artificial Intelligence Applications in Clinical Chemistry

Artificial intelligence (AI) applications are an area of active investigation in clinical chemistry. Numerous publications have demonstrated the promise of AI across all phases of testing including preanalytic, analytic, and postanalytic phases; this includes novel methods for detecting common specimen collection errors, predicting laboratory results and diagnoses, and enhancing autoverification workflows. Although AI applications pose several ethical and operational challenges, these technologies are expected to transform the practice of the clinical chemistry laboratory in the near future.

Applications of machine learning in the chemical pathology laboratory

Machine learning (ML) is an area of artificial intelligence that provides computer programmes with the capacity to autodidact and learn new skills from experience, without continued human programming. ML algorithms can analyse large data sets quickly and accurately, by supervised and unsupervised learning techniques, to provide classification and prediction value outputs. The application of ML to chemical pathology can potentially enhance efficiency at all phases of the laboratory's total testing process. Our review will broadly discuss the theoretical foundation of ML in laboratory medicine. Furthermore, we will explore the current applications of ML to diverse chemical pathology laboratory processes, for example, clinical decision support, error detection in the preanalytical phase, and ML applications in gel-based image analysis and biomarker discovery. ML currently demonstrates exploratory applications in chemical pathology with promising advancements, which have the potential to improve all phases of the chemical pathology total testing pathway.

The Nuanced Metabolic Functions of Endogenous FGF21 Depend on the Nature of the Stimulus, Tissue Source, and Experimental Model

Fibroblast growth factor 21 (FGF21) is a hormone that is involved in the regulation of lipid, glucose, and energy metabolism. Pharmacological FGF21 administration promotes weight loss and improves insulin sensitivity in rodents, non-human primates, and humans. However, pharmacologic effects of FGF21 likely differ from its physiological effects. Endogenous FGF21 is produced by many cell types, including hepatocytes, white and brown adipocytes, skeletal and cardiac myocytes, and pancreatic beta cells, and acts on a diverse array of effector tissues such as the brain, white and brown adipose tissue, heart, and skeletal muscle. Different receptor expression patterns dictate FGF21 function in these target tissues, with the primary effect to coordinate responses to nutritional stress. Moreover, different nutritional stimuli tend to promote FGF21 expression from different tissues; i.e., fasting induces hepatic-derived FGF21, while feeding promotes white adipocyte-derived FGF21. Target tissue effects of FGF21 also depend on its capacity to enter the systemic circulation, which varies widely from known FGF21 tissue sources in response to various stimuli. Due to its association with obesity and non-alcoholic fatty liver disease, the metabolic effects of endogenously produced FGF21 during the pathogenesis of these conditions are not well known. In this review, we will highlight what is known about endogenous tissue-specific FGF21 expression and organ cross-talk that dictate its diverse physiological functions, with particular attention given to FGF21 responses to nutritional stress. The importance of the particular experimental design, cellular and animal models, and nutritional status in deciphering the diverse metabolic functions of endogenous FGF21 cannot be overstated.

Fibroblast growth factor 21 in lipid metabolism and non-alcoholic fatty liver disease

Non-alcoholic fatty liver disease (NAFLD) is one of the most common liver diseases in several developed countries, ranging from simple non-alcoholic fatty liver (NAFL) to non-alcoholic steatohepatitis (NASH) and cirrhosis. Currently, NAFLD has been confirmed to be associated with dyslipidemia, insulin resistance, and pre-diabetes, which are always grouped together as metabolic syndrome. Fibroblast growth factor 21 (FGF21) plays an important role in liver pathophysiology with multiple metabolic functions. Accumulating evidence has shown that FGF21 could directly modulate lipid metabolism and reduce lipid accumulation in hepatocytes through an insulin-independent pathway, thus suppressing the pathogenesis of NAFLD. Furthermore, treatment with FGF21 could obviously reverse NAFLD and synergistically alleviate obesity and counteract insulin resistance. In this review, we summarize the current knowledge of FGF21 and the evidence of FGF21 as an important regulator in hepatic lipid metabolism. The mechanisms by which FGF21 affects the pathogenesis of NAFLD would also be proposed for the further understanding of FGF21.

Noninvasive Tests Do Not Accurately Differentiate Nonalcoholic Steatohepatitis From Simple Steatosis: A Systematic Review and Meta-analysis

BACKGROUND & AIMS

Nonalcoholic fatty liver disease is a rapidly increasing health problem. Liver biopsy analysis is the most sensitive test to differentiate between nonalcoholic steatohepatitis (NASH) and simple steatosis (SS), but noninvasive methods are needed. We performed a systematic review and meta-analysis of noninvasive tests for differentiating NASH from SS, focusing on blood markers.

METHODS

We performed a systematic search of the PubMed, Medline and Embase (1990-2016) databases using defined keywords, limited to full-text papers in English and human adults, and identified 2608 articles. Two independent reviewers screened the articles and identified 122 eligible articles that used liver biopsy as reference standard. If at least 2 studies were available, pooled sensitivity (sens_p) and specificity (spec_p) values were determined using the Meta-Analysis Package for R (metafor).

RESULTS

In the 122 studies analyzed, 219 different blood markers (107 single markers and 112 scoring systems) were identified to differentiate NASH from simple steatosis, and 22 other diagnostic tests were studied. Markers identified related to several pathophysiological mechanisms. The markers analyzed in the largest proportions of studies were alanine aminotransferase (sens_p, 63.5% and spec_p, 74.4%) within routine biochemical tests, adiponectin (sensp, 72.0% and spec_p, 75.7%) within inflammatory markers, CK18-M30 (sens_p, 68.4% and spec_p, 74.2%) within markers of cell death or proliferation and homeostatic model assessment of insulin resistance (sens_p, 69.0% and spec_p, 72.7%) within the metabolic markers. Two scoring systems could also be pooled: the NASH test (differentiated NASH from borderline NASH plus simple steatosis with 22.9% sens_p and 95.3% spec_p) and the GlycoNASH test (67.1% sens_p and 63.8% spec_p).

CONCLUSION

In the meta-analysis, we found no test to differentiate NASH from SS with a high level of pooled sensitivity and specificity (≥80%). However, some blood markers, when included in scoring systems in single studies, identified patients with NASH with ≥80% sensitivity and specificity. Replication studies and more standardized study designs are urgently needed. At present, no marker or scoring system can be recommended for use in clinical practice to differentiate NASH from simple steatosis.

Friends Turned Foes: Angiogenic Growth Factors beyond Angiogenesis

Angiogenesis, the formation of new blood vessels from pre-existing ones is a biological process that ensures an adequate blood flow is maintained to provide the cells with a sufficient supply of nutrients and oxygen within the body. Numerous soluble growth factors and inhibitors, cytokines, proteases as well as extracellular matrix proteins and adhesion molecules stringently regulate the multi-factorial process of angiogenesis. The properties and interactions of key angiogenic molecules such as vascular endothelial growth factors (VEGFs), fibroblast growth factors (FGFs) and angiopoietins have been investigated in great detail with respect to their molecular impact on angiogenesis. Since the discovery of angiogenic growth factors, much research has been focused on their biological actions and their potential use as therapeutic targets for angiogenic or anti-angiogenic strategies in a context-dependent manner depending on the pathologies. It is generally accepted that these factors play an indispensable role in angiogenesis. However, it is becoming increasingly evident that this is not their only role and it is likely that the angiogenic factors have important functions in a wider range of biological and pathological processes. The additional roles played by these molecules in numerous pathologies and biological processes beyond angiogenesis are discussed in this review.

Diagnostic Value of CK-18, FGF-21, and Related Biomarker Panel in Nonalcoholic Fatty Liver Disease: A Systematic Review and Meta-Analysis

Liver biopsy still remains the gold standard for diagnosing nonalcoholic steatohepatitis (NASH), but with limitations. There is an urgent need to develop noninvasive tests that accurately distinguish NASH from simple steatosis. The purpose of this meta-analysis was to evaluate the diagnostic value of serum biomarkers including cytokeratin 18 (CK-18), fibroblast growth factor 21 (FGF-21), and combined biomarker panel (CBP) in the diagnosis of NAFLD, especially NASH. A total of 25 studies met the inclusion criteria. Pooled sensitivity and specificity values for chosen serum markers for diagnosing NASH are as follows: CK-18 (M30), 0.75 and 0.77; CK-18 (M65), 0.71 and 0.77; FGF-21, 0.62 and 0.78; and CBP, 0.92 and 0.85. CBP demonstrated better accuracy with higher sensitivity and specificity than those tested individually. Furthermore, the AUROC of CBP was 0.94 (95% CI, 0.92-0.96), compared to CK-18 or FGF-21 assay, which showed the most significant ability to distinguish NASH from simple steatosis. The results suggest that increased circulating CK-18 and FGF-21 are associated with NASH and may be used for initial assessment, but not enough. Importantly, CBP is potentially used as accurate diagnostic tools for NASH. Further prospective designed studies are warranted to confirm our findings.

Roles of FGFs As Paracrine or Endocrine Signals in Liver Development, Health, and Disease

The liver plays important roles in multiple processes including metabolism, the immune system, and detoxification and also has a unique capacity for regeneration. FGFs are growth factors that have diverse functions in development, health, and disease. The FGF family now comprises 22 members. Several FGFs have been shown to play roles as paracrine signals in liver development, health, and disease. FGF8 and FGF10 are involved in embryonic liver development, FGF7 and FGF9 in repair in response to liver injury, and FGF5, FGF8, FGF9, FGF17, and FGF18 in the development and progression of hepatocellular carcinoma. In contrast, FGF15/19 and FGF21 are endocrine signals. FGF15/19, which is produced in the ileum, is a negative regulator of bile acid metabolism and a stimulator of gallbladder filling. FGF15/19 is a postprandial, insulin-independent activator of hepatic protein and glycogen synthesis. It is also required for hepatocellular carcinoma and liver regeneration. FGF21 is a hepatokine produced in the liver. FGF21 regulates glucose and lipid metabolism in white adipose tissue. Serum FGF21 levels are elevated in non-alcoholic fatty liver. FGF21 also protects against non-alcoholic fatty liver. These findings provide new insights into the roles of FGFs in the liver and potential therapeutic strategies for hepatic disorders.

Combined Serum Biomarkers in Non-Invasive Diagnosis of Non-Alcoholic Steatohepatitis

BACKGROUND

Non-alcoholic steatoheaptitis (NASH), the critical stage of non-alcoholic fatty liver disease (NAFLD), is of chronic progression and can develop cirrhosis even hepatocellular carcinoma (HCC). However, non-invasive biomarkers for NASH diagnosis remain poorly applied in clinical practice. Our study aims at testing the accuracy of the combination of cytokeratin-18 M30 fragment (CK-18-M30), fibroblast growth factor 21 (FGF-21), interleukin 1 receptor antagonist (IL-1Ra), pigment epithelium-derived factor (PEDF) and osteoprotegerin (OPG) in diagnosing NAFLD and NASH.

METHODS

179 patients with biopsy-proven NAFLD were enrolled as training group, 91 age- and gender-matched healthy subjects were recruited at the same time as controls. 63 other NAFLD patients were separately collected as validation group. 45 alcoholic fatty liver disease (AFLD) patients, 50 hepatitis B virus (HBV) patients, 52 hepatitis C virus (HCV) patients were also included. Serum biomarker levels were measured by enzyme-linked immunosorbent assay.

RESULTS

Serum levels of CK-18-M30, FGF-21, IL-1Ra and PEDF increased, while OPG decreased in a stepwise fashion in controls, non-NASH NAFLD patients and NASH patients (P < 0.01). The area under receiver-operating characteristics curve to diagnose NASH was 0.86 for CK-18-M30, 0.89 for FGF-21, 0.89 for IL-1Ra, 0.89 for PEDF and 0.89 for OPG. CK-18-M30 had 70% negative predictive value (NPV) and 79% positive predictive value (PPV) to diagnose NASH. A 5-step approach measuring CK-18-M30 followed by FGF21, IL-1Ra, PEDF and OPG gradually improved the NPV to 76% and PPV to 85%, which reached 80% and 76% respectively in the validation cohort.

CONCLUSION

Compared to single biomarker, stepwise combination of CK-18-M30, FGF-21, IL-1Ra, PEDF and OPG can further improve the accuracy in diagnosing NASH.

(Cleaved) CK18 serum and tissue expression levels differentiate acute HCV reinfection from acute rejection in liver allografts

BACKGROUND & AIMS

Orthotopic liver transplantation (OLT) is the sole therapeutic option to cure end-stage liver diseases including HCV-related cirrhosis. Timely and precise differentiation of relevant acute HCV reinfection from acute rejection after OLT is vital for appropriate therapy. Aim of this study was to evaluate the usefulness of (non-) invasive apoptosis (M30) and necrosis (M65) determination in the differential diagnosis of acute (and chronic) HCV reinfection vs. acute rejection in liver allografts.

METHODS

Serum samples and liver biopsy tissues were available from 76 patients including a control group (19× NAFL, 19× NASH, 16× acute rejection, 11× acute and 11× chronic HCV reinfection) and were analysed using M30- and M65 ELISAs (M30S, M65S) and M30-immunohistochemistry (M30H). Clinical and serological data were collected.

RESULTS

M30S, M65S and M30H were highly correlated with diagnostic groups in the total cohort (all P < 0.0001). M30S, M65S and M30H were independently able to differentiate acute HCV reinfection from acute rejection (P = 0.048, P = 0.001, P = 0.010) with moderate to excellent diagnostic accuracy (sensitivity, specificity, cut-off-value in M30S: 70%, 75%, 1025 U/L; M65S: 100%, 92%, 1308 U/L; M30H: 73%, 88%, 0.3%).

CONCLUSIONS

M30-, M65-ELISAs and M30-immunohistochemistry are potential useful tools in differentiating acute HCV reinfection from acute rejection facilitating both speed and accuracy of the diagnostic process for the clinician and hepatopathologist. In this context, M65S provided superior diagnostic characteristics compared to M30-based methods. However, being the first analysis of (cleaved) CK18 serum and tissue expression levels in this context, the results need to be verified in further studies.

The role of fibroblast growth factor 21 in the pathogenesis of non-alcoholic fatty liver disease and implications for therapy

Non-alcoholic fatty liver disease (NAFLD) includes a cluster of liver disorders ranging from simple fatty liver to non-alcoholic steatohepatitis (NASH) and cirrhosis. Due to its liver and vascular complications, NAFLD has become a public health problem with high morbidity and mortality. The pathogenesis of NAFLD is considered a "multi-hit hypothesis" that involves lipotoxicity, oxidative stress, endoplasmic reticulum stress, a chronic inflammatory state and mitochondrial dysfunction. Fibroblast growth factor 21 (FGF21) is a member of the fibroblast growth factor family with multiple metabolic functions. FGF21 directly regulates lipid metabolism and reduces hepatic lipid accumulation in an insulin-independent manner. Several studies have shown that FGF21 can ameliorate the "multi-hits" in the pathogenesis of NAFLD. The administration of FGF21 reverses hepatic steatosis, counteracts obesity and alleviates insulin resistance in rodents and nonhuman primates. Using several strategies, we show that the reversal of simple fatty liver and NASH is mediated by activation of the FGF21 signaling pathway. In this review, we describe the molecular mechanisms involved in the onset and/or progression of NAFLD, and review the current literature to highlight the therapeutic procedures associated with the FGF21 signaling pathway for simple fatty liver and NASH, which are the two most important types of NAFLD.

A Bayesian approach to an integrated multimodal noninvasive diagnosis of definitive nonalcoholic steatohepatitis in the spectrum of nonalcoholic fatty liver disease

BACKGROUND

The distinction between definitive nonalcoholic steatohepatitis (NASH) and other histological patterns of nonalcoholic fatty liver disease (NAFLD) continues to rely on liver biopsy, which has risks and limitations.

OBJECTIVES

We explored the usefulness of a Bayesian approach to establish the likelihood ratios (LRs) of different noninvasive diagnostic modalities (circulating cytokeratin-18 fragment levels, hydrogen breath test, transient elastography) for distinguishing definitive NASH from other forms of NAFLD.

PATIENTS AND METHODS

From a series of 235 consecutive patients with biopsy-proven NAFLD enrolled at a tertiary referral hepatology clinic, we identified 135 patients with definitive NASH and 100 with other forms of NAFLD. Bayesian probabilities in the form of LRs were estimated for predicting definitive NASH.

RESULTS

In conditional logistic regression models, the probability of having definitive NASH was associated with diabetes mellitus [odds ratio (OR) 2.89; 95% confidence interval (CI) 1.71-4.85], positive transient elastography results (OR 2.11; 95% CI 1.08-4.09), and apoptosis marker M30 (OR 2.57; 95% CI 1.78-3.93). We then estimated the LRs for variables independently related to definitive NASH, which were as follows: diabetes(+) LR=2.42; M30(+) LR=2.12; and transient elastography(+) LR=1.77. On the basis of the LR form of Bayes' theorem, the prediction model that took into account diabetes, M30 measurements, and the results of transient elastography at a tertiary hospital determined an 81% probability of distinguishing definitive NASH from other forms of NAFLD.

CONCLUSION

A Bayesian approach combining clinical, laboratory, and imaging data may be useful for identifying patients with the highest probability of having definitive NASH.

Limited utility of plasma M30 in discriminating non-alcoholic steatohepatitis from steatosis--a comparison with routine biochemical markers

Introduction

The utility of Cytokeratin-18 fragment, namely CK18Asp396 (M30), for the diagnosis of non-alcoholic steatohepatitis (NASH) is currently uncertain. We aimed to provide further data in this area among multi-ethnic Asian subjects with NAFLD.

Materials and Methods

The accuracy of M30 for detecting NASH was compared with serum alanine aminotransferase (ALT), aspartate aminotransferase (AST) and gamma glutamyl transpeptidase (GGT) levels in consecutive adult subjects with biopsy-proven non-alcoholic fatty liver disease (NAFLD).

Results

Data for 93 NAFLD subjects (mean age 51.0±11.1 years old and 51.6% males) and 20 healthy controls (mean age 50.2±16.4 years old and 33.3% males) were analyzed. There were 39 NASH subjects (41.9%) and 54 non-NASH subjects (58.1%) among the NAFLD subjects. Plasma M30 (349 U/L vs. 162 U/L), and serum ALT (70 IU/L vs. 26 IU/L), AST (41 IU/L vs. 20 IU/L) and GGT (75 IU/L vs. 33 IU/L) were significantly higher in NAFLD subjects than in healthy controls. Serum ALT (86 IU/L vs. 61 IU/L), AST (58 IU/L vs. 34 IU/L) and GGT (97 IU/L vs. 56 IU/L) were significantly higher in NASH subjects compared to non-NASH subjects, but no significant difference was observed with plasma M30 (435 U/L vs. 331 U/L). The accuracy of plasma M30, and serum ALT, AST and GGT was good for predicting NAFLD (AUROC 0.91, 0.95, 0.87 and 0.85, respectively) but less so for NASH (AUROC 0.59, 0.64, 0.75 and 0.68, respectively). Serum ALT and AST, but not plasma M30 showed a significant trend with increasing grades of ballooning and lobular inflammation.

Conclusion

The utility of M30 in the detection of NASH in clinical practice appears limited, in comparison to routine biochemical markers.