Development of a biomarker prediction model for post-trauma multiple organ failure/dysfunction syndrome based on the blood transcriptome
Study context
- Population
- Adult (age ≥ 16 years) blunt trauma patients with early buffy coat samples (≤ 24 h since injury), Glue Grant multi-center cohort58 MOF-cases, 83 non-cases
- Participants
- 141 participants
- Comparison
- Blood transcriptome biomarker models vs injury severity score + sex models for predicting MOF
- Outcome
- Early prediction of MOF/MODS outcome (AUROC)
Methodological detailsmachine learning (Least Absolute Shrinkage and Selection Operator (LASSO), XGBoost), multi-center cohort, training and test sets
- Machine learning (LASSO, XGBoost)
- Multi-center cohort
- Training and test sets
Performance
As reported in the abstract
AUROC [95% CI] for MOF prediction
| Test or model | AUROC (training) | AUROC (test) |
|---|---|---|
| LASSO | 0.938 [0.890-0.987] | 0.833 [0.699-0.967] |
| XGBoost | 0.999 [0.997-1.000] | 0.907 [0.816-0.998] |
| Overlapping transcripts | 0.935 [0.884-0.985] | 0.836 [0.703-0.968] |
| APACHEII + sex | 0.649 [0.537-0.762] | 0.493 [0.301-0.685] |
| ISS + sex | 0.630 [0.516-0.744] | 0.482 [0.293-0.670] |
| NISS + sex | 0.651 [0.540-0.763] | 0.525 [0.335-0.714] |
AUROC (training)
LASSO0.938 [0.890-0.987]
XGBoost0.999 [0.997-1.000]
Overlapping transcripts0.935 [0.884-0.985]
APACHEII + sex0.649 [0.537-0.762]
ISS + sex0.630 [0.516-0.744]
NISS + sex0.651 [0.540-0.763]
AUROC (test)
LASSO0.833 [0.699-0.967]
XGBoost0.907 [0.816-0.998]
Overlapping transcripts0.836 [0.703-0.968]
APACHEII + sex0.493 [0.301-0.685]
ISS + sex0.482 [0.293-0.670]
NISS + sex0.525 [0.335-0.714]
Further findings
Injury severity scores and sex
- Significantly associated with MOF
Original publication
Development of a biomarker prediction model for post-trauma multiple organ failure/dysfunction syndrome based on the blood transcriptome. Ann Intensive Care. Published 28 August 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →