Development of a Prediction Model and Risk Score for Self-Assessment and High-Risk Population Identification in Liver Cancer Screening: Prospective Cohort Study
Bibliographic record
Abstract
Background: Liver cancer continues to pose a significant burden in China. To enhance the efficiency of screening, it is crucial to implement population stratification for liver cancer surveillance. Objective: This study aimed to develop a simple prediction model and risk score for liver cancer screening in the general population, with the goal of improving early detection and survival. Methods: This population-based cohort study focused on residents aged 40 to 74 years. Participants were enrolled between 2014 and 2019 and were prospectively followed until June 30, 2021. Data were collected through interviews at enrollment. A Cox proportional hazards regression was used to identify predictors and construct the prediction model. A risk score system was developed based on the weighted factors included in the prediction model. Results: A total of 153,082 study participants (67,586 males and 85,496 females) with a mean age of 55.86 years were included. During 781,125 person-years of follow-up (length of follow-up: median 6.07, IQR 3.07-7.09 years), 290 individuals were diagnosed with liver cancer. Key factors identified for the prediction model and risk score system included age (hazard ratio [HR] 1.06, 95% CI 1.04-1.08), sex (male: HR 3.41, 95% CI 2.44-4.78), education level (medium: HR 0.84, 95% CI 0.61-1.15; high: HR 0.37, 95% CI 0.17-0.78), cirrhosis (HR 11.93, 95% CI 7.46-19.09), diabetes (HR 1.59, 95% CI 1.08-2.34), and hepatitis B surface antigen (HBsAg) status (positive: HR 3.84, 95% CI 2.38-6.19; unknown: HR 1.04, 95% CI 0.73-1.49). The model exhibited excellent discrimination in both the development and validation sets, with areas under the curve (AUC) of 0.802, 0.812, and 0.791 for predicting liver cancer at the 1-, 3-, and 5-year periods in the development set and 0.751, 0.763, and 0.712 in the validation set, respectively. Sensitivity analyses applied to the subgroups of participants without cirrhosis and with a negative or unknown HBsAg status yielded similar performances, with AUCs ranging from 0.707 to 0.831. Calibration plots indicated an excellent agreement between the observed and predicted probabilities of developing liver cancer over the 1-, 3-, and 5-year periods. Compared to the low-risk group, participants in the high-risk and moderate-risk groups had 11.88-fold (95% CI 8.67-16.27) and 3.51-fold (95% CI 2.58-4.76) higher risks of liver cancer, respectively. Decision curve analysis demonstrated that the risk score provided a higher net benefit compared to the current strategy. To aid in risk stratification for individual participants, a user-friendly web-based scoring system was developed. Conclusions: A straightforward liver cancer prediction model was created by incorporating easily accessible variables. This model enables the identification of asymptomatic individuals who should be prioritized for liver cancer screening.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.020 | 0.025 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.003 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".