Development of a Prediction Model and Risk Score for Self-Assessment and High-Risk Population Identification in Liver...

Area of Science:

• Hepatology
• Oncology
• Public Health

Background:

• Liver cancer presents a significant health challenge in China.
• Effective screening strategies require population stratification.
• Early detection is crucial for improving patient survival rates.

Purpose of the Study:

• To develop a simple prediction model for liver cancer screening.
• To create a risk score for stratifying the general population.
• To enhance early detection and survival rates for liver cancer.

Main Methods:

• A population-based cohort study of 153,082 residents aged 40-74.
• Prospective follow-up from 2014 to 2021 with data collected via interviews.
• Cox proportional hazards regression used to identify predictors and build a risk score system.

Main Results:

• Key predictors for liver cancer risk include age, male sex, cirrhosis, diabetes, and hepatitis B surface antigen (HBsAg) status.
• The prediction model demonstrated excellent discrimination (AUCs 0.707-0.831) and calibration.
• High-risk and moderate-risk groups showed significantly elevated liver cancer risks (11.88-fold and 3.51-fold, respectively).

Conclusions:

• A straightforward liver cancer prediction model and risk score were developed using accessible variables.
• The model effectively identifies asymptomatic individuals for prioritized liver cancer screening.
• The developed risk score system offers a valuable tool for population-based surveillance and early detection.