Several research groups showcased a mixture of new techniques and improvement in established methods to move beyond liver biopsy for identifying NAFLD.
Non-alcoholic fatty liver disease (NAFLD) has historically been diagnosed and prognosticated by performing a liver biopsy. However, this procedure is costly, invasive and carries a risk of morbidity, so developing effective non-invasive diagnostic methods is highly desirable.
During today’s ‘NAFLD – Non-invasive Assessment’ session, several groups presented advances in non-invasive assessment of liver health, including a mixture of new techniques and improvement in the use of established methods. For example, in a study of 40,729 individuals, results from repeated FIB-4 measurement within 5 years can better predict cirrhosis and complications in patients with NAFLD. Another team demonstrated how combining FIB-4 with magnetic resonance elastography can assist with identifying patients at risk of progressing to stage ≥2 fibrosis. A machine learning algorithm working with routine check-up data was also demonstrated as a feasible approach for pre-test risk stratification of liver fibrosis.
Additional reports of novel non-invasive NAFLD diagnostic methods were presented:
- Vibration-controlled transient elastography was used to investigate longitudinal changes in liver stiffness in a NAFLD cohort, informing the use of non-invasive endpoints in future clinical trials.
- Convolutional neural networks trained on liver biopsy images can create scores that correlate with key histological features of NASH: a novel strategy for quantitative histological analysis.
- Levels of the microRNA miR-34a in serum samples were successfully used to efficiently rule out NAFLD in healthy patients.
As the prevalence of NAFLD and NASH continues to increase worldwide, these developments in non-invasive testing will be a key tool in identifying people at risk and improving the standard of care in this important area, potentially leading to better outcomes for patients.