Etiology and prognosis are the foremost issues that we consider when treating liver disease, and I believe that hepatologists and patients are largely aligned on these important questions: What is the diagnosis, and what is the prognosis? Even today, we rely on liver biopsy in many patients, a procedure that is not liked by both patients and clinicians, given its invasiveness and the costs involved. There is a clear need for noninvasive tests (NITs) for diagnosis and, more importantly, for prognosis. 

While the burden from viral hepatitis C in the United States has been leveling off, the burden from nonalcoholic fatty liver disease (NAFLD)/nonalcoholic steatohepatitis (NASH) has been increasing, and the prevalence of NAFLD is now approximately 25% to 30%. It is not feasible to biopsy all of these patients, so there has been a focus on the use of NITs to identify individuals who are at high risk for advanced fibrosis or cirrhosis. 

Currently, the simple or nonpatented NITs (eg, the Fibrosis-4 [FIB-4] index or the NAFLD fibrosis Score) are the most widely used tests to assess the risk of fibrosis. These tests could easily be incorporated in electronic medical record systems, since the scores are derived using algorithms that incorporate data and laboratory values that are routinely performed for standard of care in patients with liver disease, such as age, liver function tests, and platelet counts. The goal of testing is to identify patients with NAFLD with fibrosis stage 2 or higher, as these individuals are more likely to develop progression with the development of liver-related events. Patients who are identified for high risk of fibrosis should be followed under hepatology specialist care, and they may also be considered for enrollment in appropriate clinical trials. 

We need accurate and simple NITs for use in routine clinical practice that are acceptable to patients and clinicians, and, at the same time, are cost effective. The validation and standardization of data on a large number of patients is required before an NIT is approved by the US Food and Drug Administration for use in clinical practice. An interesting abstract from the Non-Invasive Biomarkers of Metabolic Liver Disease (NIMBLE) consortium in collaboration with the NASH Clinical Research Network group was presented by Sanyal et al at The Liver Meeting 2021, and it is an excellent effort in the direction of standardization and validation (abstract LO1). The main goal of this study, which included 1073 patients with NAFLD (853 NASH with fibrosis stages 0-4), was to compare 5 blood-based biomarker panels, with a prespecified hypothesis of c-statistics of greater than 0.7 vs ALT for NASH diagnosis and vs FIB-4 for fibrosis assessment. The NIS4, the Enhanced Liver Fibrosis (ELF) test, and the FibroMeter vibration-controlled transient elastography (FM-VCTE) met prespecified criteria for NASH diagnosis and for fibrosis assessment. FM-VCTE is a test that combines a serum test with liver stiffness measurement. ELF is a blood-based test that is already authorized by the US Food and Drug Administration, and it generates a numeric score to assess the likelihood of progression to cirrhosis and liver-related clinical events in patients with advanced fibrosis due to NASH. 

Alcohol use disorder and metabolic syndrome are prevalent conditions in the general population. Hence, approximately 2% to 3% of the US population are potentially exposed to both risk factors. As the disease progression in alcohol-associated liver disease is faster compared with NAFLD, it becomes important to know the driver of liver disease in patients who consume alcohol and are obese. However, quite often, this can be challenging. In this regard, a poster abstract by Garrido et al from The Liver Meeting 2021 (abstract 361) is interesting and demonstrates the use of magnetic resonance imaging (MRI) scans of the liver to differentiate the 2 conditions. In this study, 83 patients with clinically and histologically confirmed steatohepatitis (57 NASH and 26 alcohol-associated steatohepatitis [ASH]) had an MRI of the abdomen within 13 months of the liver biopsy. The MR films were independently reviewed by 2 radiologists who were blinded to the clinical data, and the assessment focused on a specific pattern of heterogeneity that extended along the perivascular tissues. The investigators found that a diagnosis of ASH vs NASH as a cause for steatohepatitis was more likely to have this finding of heterogeneity, especially perivascular branching patterns. There was a moderate reliability between the 2 radiologists (interclass correlation coefficient of 0.68). Clearly, this study needs validation, but it opens avenues for machine learning approaches to bridge the knowledge gap in determining the risk factor driving the disease process in patients who consume alcohol and are obese. 

Other promising areas for the future using machine learning and an artificial intelligence approach include liquid biopsy, proteomic, lipidomic, transcriptomic, and gene expression analyses. There were several abstracts presented at this year’s The Liver Meeting using these approaches, and some of the examples are abstracts LO3, 74, 109, 294, 356, and 1591.