Towards reliable prediction of intraoperative hypotension: a cross-center evaluation of deep learning-based and MAP-derived methods
Journal of Clinical Monitoring and Computing · 2025
DOI: 10.1007/s10877-025-01357-0 ↗Kurz gefasst
Multizentrische Studie in der Zeitschrift Journal of Clinical Monitoring and Computing, 2025. Verfasst von Nada Chaari und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Biomedical Engineering and Health Systems, KTH Royal Institute of Technology, Hälsovägen 11, Huddinge, 141 52, Sweden. nadcha@kth.se. Hauptthemen laut MEDLINE: Hypotension, Deep Learning, Intraoperative Complications, Monitoring, Intraoperative. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Royal Institute of Technology.
Abstract (Originalsprache)
Intraoperative hypotension (IOH) is associated with an increased risk of heart and kidney complications. Although AI tools aim to predict IOH, their real-world reliability is often overstated due to biased data selection. This study introduces a framework to enhance reliability by: (1) including borderline blood pressure cases (65–75 mmHg, the “Gray Zone”), (2) comparing AI model to simple blood pressure threshold, and (3) validating across diverse surgical cohorts, centers and demographics. Using datasets from Karolinska University Hospital (Sweden) and VitalDB (Korea), we found AI model performs better than MAP threshold method in more ambiguous cases. In contrast, when hypotensive and non-hypotensive cases had clearly separated MAP values, both methods performed similarly well. Cross-validation revealed asymmetric generalizability: models trained on datasets containing more borderline (Gray Zone) cases generalized better to datasets with clearer class separation, whereas the reverse struggled. To ensure fair model comparison and reduce dataset-specific bias, we standardized the MAP difference between positive (hypotension) and negative (non-hypotension) samples at the time of prediction. This virtually eliminated the class separation and demonstrated that inflated performance in some datasets can be attributed to selection bias rather than true model generalizability. Age also influenced generalization: Cross-age validation revealed models trained on older patients generalized better to younger cohorts, whereas differences in ASA classification had minimal effect. These findings highlight the need for realistic validation to bridge the gap between AI research and clinical practice.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Biomedical Engineering and Health Systems, KTH Royal Institute of Technology, Hälsovägen 11, Huddinge, 141 52, Sweden. nadcha@kth.se.
- Department of Perioperative Medicine and Intensive Care, Karolinska University Hospital, Huddinge, 141 86, Sweden.
- Department of Physiology and Pharmacology, Karolinska Institutet, Solnavägen 9, Solna, 171 77, Sweden.
- Maquet Critical Care AB, Röntgenvägen 2, Solna, 171 06, Sweden.
- Department of Biomedical Engineering and Health Systems, KTH Royal Institute of Technology, Hälsovägen 11, Huddinge, 141 52, Sweden.
Förderung
- Royal Institute of Technology
Zitieren
APA
Chaari, N., Winski, G., Hallbäck, M., Lundström, N., Björne, H., & Jacobsson, M. (2025). Towards reliable prediction of intraoperative hypotension: a cross-center evaluation of deep learning-based and MAP-derived methods. Journal of Clinical Monitoring and Computing. https://doi.org/10.1007/s10877-025-01357-0
Vancouver
Chaari N, Winski G, Hallbäck M, Lundström N, Björne H, Jacobsson M. Towards reliable prediction of intraoperative hypotension: a cross-center evaluation of deep learning-based and MAP-derived methods. Journal of Clinical Monitoring and Computing. 2025. doi:10.1007/s10877-025-01357-0
BibTeX
@article{Chaari2025,
author = {Chaari, Nada and Winski, Greg and Hallbäck, Magnus and Lundström, Niclas and Björne, Håkan and Jacobsson, Martin},
title = {Towards reliable prediction of intraoperative hypotension: a cross-center evaluation of deep learning-based and MAP-derived methods},
journal = {Journal of Clinical Monitoring and Computing},
year = {2025},
doi = {10.1007/s10877-025-01357-0}
}
EUVIMED