Objective:

To create a machine learning model to forecast 30-day all-cause mortality in patients with ruptured hepatocellular carcinoma and to identify essential predictive elements.

Approach:

Key Findings:

• The decision tree model showed sensitivity of 87.5% and AUC of 0.7901 for predicting 30-day mortality.
• TBIL and INR were identified as primary predictors of mortality risk, with odds ratios of 1.45 (95% CI: 1.25–1.68) for TBIL (>40 μmol/L) and 4.95 (95% CI: 3.36–10.11) for INR (>2.5).
• Nonlinear threshold effects were observed for TBIL (>40 μmol/L) and INR (>2.5), significantly increasing mortality risk.
• Combining TBIL and INR improved predictive accuracy (AUC = 0.87).

Interpretation:

The decision tree model predicts 30-day mortality in patients with ruptured HCC, identifying TBIL and INR as significant predictors.

Limitations:

• The study is limited to a single-center cohort, which may affect generalizability.
• Retrospective design may introduce biases.

Conclusion:

The decision tree framework provides a method for assessing 30-day mortality risk in ruptured HCC patients.

Attribution Notice

This content is an AI-generated, fully rewritten summary based on a published scholarly article. It does not reproduce the original text and is not a substitute for the original publication. Readers are encouraged to consult the source for full context, data, and methodology.