The diagnostic performance of a simplified blood test (SteatoTest-2) for the prediction of liver steatosis

Abstract

Supplemental Digital Content is available in the text.

Laboratory parameter‐based machine learning model for excluding non‐alcoholic fatty liver disease (NAFLD) in the general population

“…Third, the 70% PPV in NAFLD ridge score appeared to be modest when compared to the 96% NPV, yet already provides predictive power on the presence of NAFLD. The 70% PPV is also comparable to that of existing prediction models for NAFLD, in particular 63% PPV for SteatoTest . NAFLD ridge score is therefore a useful tool not inferior to existing prediction scores for the exclusion of NAFLD in epidemiological studies, especially for data without anthropometric parameters.…”

“…This limits the use of these models in large‐scaled epidemiological studies and health database research. In particular, NAFLD liver fat score includes fasting serum insulin; SteatoTest includes A2M, ApoA1 and haptoglobin, which are not routinely performed in many clinics and hospitals. In addition, existing NAFLD prediction models involved anthropometric parameters including waist circumference and BMI, which may not be easily captured in computerised data sources used for epidemiological studies.…”

“…Therefore, 1 H‐MRS can be treated as a better reference standard for NAFLD in model development. Also, some existing NAFLD prediction models used ultrasonography instead of histology to determine NAFLD . However, ultrasonography is often operator dependent, difficult in obese subjects and may not be sensitive enough to diagnose mild fatty liver with hepatic steatosis <30% .…”

“…NAFLD liver fat score estimates liver fat content in percentage with a formula comprising metabolic syndrome, type 2 diabetes, fasting serum insulin, aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels . SteatoTest is a logistic regression model of 12 predicting parameters: α2‐macroglobulin (A2M), apolipoprotein A1 (ApoA1), haptoglobin, total bilirubin, GGT levels, cholesterol, triglycerides, glucose, age, gender and BMI . An issue of SteatoTest is some of the predicting parameters (eg, A2M and ApoA1) may not be readily available in local laboratories.…”

Clinical risk scoring for predicting non‐alcoholic fatty liver disease in metabolic syndrome patients (NAFLD‐MS score)

“…NAFLD‐MS score developed in our study is different from previous tools in a number of aspect. Firstly, previous models were developed in general population, mixed population, or suspected liver disease patients and mostly with Caucasian individuals, while ours was developed in Asians people with MetS. Secondly, some of the previous models were complex, requiring a calculator to sum up the total score .…”