Jeon MY, et al. - The clinicians undertook this study to investigate whether the use of risk prediction models could dynamically predict hepatocellular carcinoma (HCC) development at different time points in chronic hepatitis B (CHB) patients. At different time points, dynamic prediction of HCC development was achieved using 4 risk prediction models, but not using the changes in the absolute and percentage values between two-time points. Among 4 prediction models, the mREACH-B score was the most appropriately prediction model of HCC development.

• The clinicians recruited 1,397 CHB patients between 2006 and 2014.
• At intervals of >6 months, all patients underwent serial transient elastography.

• The median age of the study population (931 males and 466 females) was 49.0 years.
• At enrollment, the median CU-HCC, REACH-B, LSM-HCC, and mREACH-B score were 4.0, 9.0, 10.0, and 8.0, respectively.
• 87 (6.2%) patients developed HCC during the follow-up period (median, 68.0 months).
• In predicting HCC development, all risk prediction models were successful at both the first liver stiffness (LS) measurement [hazard ratio (HR)=1.067-1.467 in the subgroup without antiviral therapy (AVT) and 1.096-1.458 in the subgroup with AVT] and second LS measurement (HR=1.125-1.448 in the subgroup without AVT and 1.087-1.249 in the subgroup with AVT).
• On the other hand, neither the absolute nor percentage change in the scores from the risk prediction models predicted HCC development (all P>0.05).
• Compared to the other scores, the mREACH-B score performed similarly or significantly better (AUROCs at 5 years, 0.694-0.862 vs. 0.537-0.875).