Objective:

To develop and validate a multimodal AI model for predicting sepsis and mortality in patients with enterocutaneous fistula (ECF) associated with complicated intraabdominal infection (CIAI), addressing a critical gap in early risk assessment.

Key Findings:

• The AI model demonstrated strong predictive accuracy for sepsis and mortality in ECF patients, achieving an AUC of X.
• Integration of multimodal data improved the identification of immune dysregulation and infection risk.
• The model's modular design allows for practical deployment even without transcriptomic data, maintaining predictive accuracy.

Interpretation:

The study highlights the potential of AI in improving early risk assessment and intervention strategies for high-risk ECF patients, which could significantly enhance clinical decision-making.

Limitations:

• The study is retrospective and relies on publicly available databases, limiting direct patient engagement and potentially introducing bias.
• The generalizability of the model may be affected by the specific patient population studied, necessitating further validation in diverse cohorts.

Conclusion:

This multimodal AI framework offers a promising tool for predicting sepsis and mortality in ECF patients, potentially guiding personalized interventions and improving clinical outcomes significantly.