Abstract
Liver fibrosis is a key histologic marker of long-term outcome in chronic liver disease. Non-invasive tests (NITs) have been shown to have predictive value, but the superiority of "dynamic" versus "static" assessment remains controversial. This article systematically reviews the latest evidence to elucidate the association between longitudinal changes in NITs and hepatic adverse events and assess the incremental contribution of dynamic monitoring to the model. Additionally, it reveals that the dynamic monitoring of NITs is truly superior to single evaluation, but the evidence is limited and the heterogeneity is significant. Dynamic modeling approaches for NITs require a shift from traditional parameter estimation to time-series machine learning. Future studies should make breakthroughs in disease stratification, modeling method innovation, data quality improvement, and prediction ability assessment so as to promote the transition of NITs from "static risk label" to "dynamic individualized engine," which can truly serve clinical decision-making.