Classification of recurrence status after surgical treatment of chronic subdural hemorrhage – A machine learning approach
Hussam Hamou, Julius Kernbach, Hani Ridwan, Kimberley Fay-Rodrian, Hans Clusmann, Anke Hoellig, Michael Veldeman
Abstract
Chronic subdural hematoma (cSDH) recurrence requiring reoperation occurs in 5–33% of cases. Predicting recurrence could enable risk-stratified surveillance, reducing imaging in low-risk patients while maintaining monitoring for high-risk individuals.
Introduction
Chronic subdural hematoma (cSDH) has an estimated annual incidence ranging from 7 to 30 per 100,000 individuals in population-based surveys [1–3]. The incidence rises markedly with advancing age, reaching 58 per 100,000 in patients over 70 years old [4].
Materials and methods
All consecutive patients who underwent surgical treatment for chronic subdural hematoma at a single tertiary care university hospital (RWTH Aachen University Hospital, Aachen, Germany) between January 2015 and December 2023 were considered for inclusion.
Results
Between January 2015 and December 2023, a total of 630 consecutive patients underwent surgical treatment for chronic subdural hematoma at our institution. After excluding 66 patients (10.5%) with missing recurrence outcome data due to incomplete follow-up or unavailable imaging records, 564 patients (89.5%) constituted the final analytic cohort.
Discussion
This study systematically evaluated machine learning algorithms to predict postoperative recurrence in chronic subdural hematoma using readily available clinical, demographic, laboratory, and radiographic variables. We compared three fundamentally distinct modeling approaches, regularized logistic regression, Random Forest, and XGBoost, selecting the best-performing model for final evaluation on a held-out test set.
Conclusion
In this single-center study, using internal cross-validation and a held-out test set, machine learning algorithms did not achieve clinically actionable prediction of postoperative recurrence in chronic subdural hematoma using routinely available clinical, laboratory, and radiographic data. External validation in independent cohorts is required to determine whether this finding generalizes beyond our institution.
Citation: Hamou H, Kernbach J, Ridwan H, Fay-Rodrian K, Clusmann H, Hoellig A, et al. (2026) Classification of recurrence status after surgical treatment of chronic subdural hemorrhage – A machine learning approach. PLoS One 21(9): e0346756. https://doi.org/10.1371/journal.pone.0346756
Editor: Michael C. Burger, Goethe University Hospital Frankfurt, GERMANY
Received: March 27, 2026; Accepted: August 26, 2026; Published: September 18, 2026
Copyright: © 2026 Hamou et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data underlying the results presented in this study contain potentially identifiable patient information and cannot be shared publicly due to ethical and legal restrictions. Data are available from the RWTH Aachen University Hospital Institutional Data Access / Ethics Committee (contact via: ekaachen@ukaachen.de) for researchers who meet the criteria for access to confidential data.
Funding: This project was made possible by the generous funding of the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG), through which Michael Veldeman received a Walter Benjamin Scholarship (Grant Number: VE 1274/1-2). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
