Hemodynamics, fluids & transfusion
Machine learning-based prediction of significant intraoperative blood loss to guide risk-adapted blood management decisions in spinal surgery
Authors' conclusion
This 12-variable machine-learning model can accurately predict significant blood loss risk in spinal surgery. Risk-adapted Cell Saver use guided by predicted risk provides greater net clinical benefit than experienced-based real-world strategy.
Study context
- Population
- Spinal surgery patientsdevelopment cohort
- Participants
- 3,944 patients
- Comparison
- Test set vs External
- Externaln = 843
- Outcome
- Prediction of intraoperative significant blood loss (≥ 500 mL)
Methodological detailsmachine learning, decision-curve analysis, spline analysis
- Machine learning
- Decision-curve analysis
- Spline analysis
Performance
As reported in the abstract
Discrimination of the 12-variable ranger model
| Test or model | AUC |
|---|---|
| Test set | 0.814 (95% CI, 0.790-0.839) |
| External | 0.820 (0.785-0.854) |
AUC
Test set0.814 (95% CI, 0.790-0.839)
External0.820 (0.785-0.854)
Further findings
Net benefit (decision-curve analysis)
- Risk-adapted Cell Saver strategy gave greater net benefit than current practice
Cell Saver benefit across predicted risk (spline analysis)
- Cell Saver use associated with higher postoperative hemoglobin for risk above threshold
- >0.53
- Cell Saver use reduced allogeneic transfusion requirements for risk above threshold
- >0.58
- Among transfused patients, Cell Saver decreased red-cell unit volume at all risk levels, larger reductions in higher-risk patients
Original publication
Machine learning-based prediction of significant intraoperative blood loss to guide risk-adapted blood management decisions in spinal surgery. J Clin Anesth. Published 16 February 2026. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →