Related Experiment Video
Updated: Jun 14, 2025
07:21
Systems Analysis of the Neuroinflammatory and Hemodynamic Response to Traumatic Brain Injury
Published on: May 27, 2022
3.1K
Development of a biomarker prediction model for post-trauma multiple organ failure/dysfunction syndrome based on the
Ivan Duran1, Ankita Banerjee1, Patrick J Flaherty2
1Department of Surgery, Massachusetts General Hospital and Harvard Medical School, 50 Blossom St., Their 340, Boston, MA, 02114, USA.
Annals of Intensive Care
|August 28, 2024
Summary
Machine learning models using transcriptome data can predict multiple organ failure (MOF) in trauma patients within 24 hours. These genomic biomarkers significantly outperform traditional injury scores for early MOF outcome prediction.
Area of Science:
- Genomics
- Bioinformatics
- Trauma Research
Background:
- Multiple organ failure (MOF) is a significant cause of death and illness in severe trauma patients.
- Current diagnostic methods rely on physiological monitoring and clinical scores, often detecting MOF late.
- There is a critical need for early MOF prediction to improve patient outcomes.
Purpose of the Study:
- To develop an early prediction model for MOF outcome in trauma patients.
- To utilize machine learning analysis of genome-wide transcriptome data from blood samples.
- To compare the predictive performance of the developed model against traditional injury severity scores and infection detection.
Main Methods:
- Analysis of buffy coat transcriptome and clinical data from 141 adult blunt trauma patients.
- Application of Least Absolute Shrinkage and Selection Operator (LASSO) and eXtreme Gradient Boosting (XGBoost) algorithms.
- Development of predictive models using selected transcripts from blood samples collected within 24 hours of injury.
Main Results:
- LASSO model identified 18 transcripts with high predictive accuracy (AUROC 0.833 in test set).
- XGBoost model identified 41 transcripts with superior predictive accuracy (AUROC 0.907 in test set).
- Biomarker-based models significantly outperformed models based on injury severity scores (e.g., ISS, NISS) and sex.
Conclusions:
- Early MOF assessment using blood transcriptome data post-trauma can enhance clinical decision-making.
- Improved prediction may lead to reduced morbidity, mortality, and healthcare costs.
- Identifying key transcripts offers insights into MOF molecular mechanisms, potentially guiding novel therapeutic interventions.
