Decoding fatal toxic effects in checkpoint inhibitor therapy using real-world pharmacovigilance data and machine...

Area of Science:

• Pharmacovigilance
• Oncology
• Machine Learning

Background:

• Immune checkpoint inhibitors (ICIs) improve cancer treatment outcomes.
• Immune-related adverse events (irAEs) are a significant clinical challenge associated with ICIs.

Purpose of the Study:

• To identify adverse event (AE) signals related to ICI therapies.
• To develop a machine learning model (SAFE-ICI) for predicting the risk and outcomes of fatal irAEs within the first 90 days of ICI therapy.

Main Methods:

• Observational pharmacovigilance analysis using the FDA Adverse Event Reporting System.
• Fatality kinetics simulation and multivariate logistic regression to analyze fatal AE patterns.
• Development and validation of the SAFE-ICI machine learning framework.

Main Results:

• Identified 358 significant AE signals across 18 organ systems linked to ICI therapies.
• PD-1/PD-L1 therapies were associated with 54 fatal irAEs; combination therapies with 20.
• The SAFE-ICI model accurately stratified patients into low- and high-risk groups for fatal irAEs, showing different survival benefits.

Conclusions:

• Machine learning holds potential for enhancing pharmacovigilance systems.
• The SAFE-ICI model can aid clinicians in improving patient outcomes during ICI therapy.