Key result
Machine learning models outperform ACS-NSQIP scoring for predicting postoperative risks, achieving an AUROC of ~0.90.
Why the study?
Because elderly patients often have multiple comorbidities and limited physiological reserve, individualized risk assessment using comprehensive geriatric assessment is needed to optimize surgical outcomes.
Do machine learning models based on comprehensive geriatric assessment improve the prediction of early postoperative complications and transfusion risk compared to the ACS-NSQIP scoring system in geriatric patients undergoing elective lumbar spinal stenosis surgery?
Population
261 patients 65 years or older undergoing elective lumbar spinal stenosis surgery
Comparison
Machine learning models vs ACS-NSQIP scoring system
Design
Prospective cohort study
Authors
Loading...
ML models may enhance risk prediction in geriatric spine surgery; hypothesis-generating, requires prospective validation before practice change.
Observational (n=261)
Do machine learning models based on comprehensive geriatric assessment improve the prediction of early postoperative complications and transfusion risk compared to the ACS-NSQIP scoring system in geriatric patients undergoing elective lumbar spinal stenosis surgery?
Absolute Event Rate: 0.9% vs 0.38%
Machine learning models incorporating comprehensive geriatric assessment features significantly outperformed the traditional ACS-NSQIP scoring system in predicting early postoperative complications and transfusion risk in elderly patients undergoing lumbar spinal stenosis surgery.
Rhee et al. (2025) conducted an observational in Lumbar spinal stenosis (n=261). Machine learning models (Compact model) vs. ACS-NSQIP scoring system was evaluated on AUROC and AUPRC for predicting complications and transfusion risk. Machine learning models achieved significantly greater AUROC values (nearing or surpassing 0.90) compared to the ACS-NSQIP scoring system (AUROC 0.38 and 0.22) for predicting postoperative risks.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: