This study evaluated whether a joint latent class model (JLCM) using serial non-invasive fibrosis tests (LSM by VCTE, FIB-4, sex, age, platelets) could personalize prediction of liver-related events (LRE: cirrhosis decompensation or hepatocellular carcinoma) in adults with MASLD, across an international multicentre cohort with median follow-up of 4.0 years.
In 13,627 patients with 238 LREs, the JLCM achieved excellent discrimination with integrated AUROCs rising from 88.2% at baseline to 92.2% at 5-year follow-up — outperforming standalone LSM (39–56% sensitivity) and FIB-4 (32–60%), with the JLCM identifying 61–80% of patients who later experienced LRE as high risk.
- Outcome events were relatively rare (238 LREs in 13,627 patients), which may limit subgroup analyses. - Median follow-up of 4.0 years may be insufficient to capture long-term LRE risk trajectories in MASLD. - The JLCM requires serial measurements over time, limiting its applicability at a single baseline visit.
Tracking LSM and FIB-4 serially over time — rather than relying on a single snapshot — significantly improves identification of MASLD patients at high risk for cirrhosis decompensation or liver cancer. Clinicians should consider longitudinal monitoring with validated non-invasive tests as a routine strategy for personalized LRE risk stratification in MASLD.
Explore related topics
