Data-driven phenotyping of infective endocarditis using unsupervised machine learning: toward a personalized approach.
9 affiliations
- 1Department of Cardiology, Bitlis State Hospital, Beş Minare, Selahattin Eyyübi Caddesi, No:160, Merkez, Bitlis, Turkey. [email protected].
- 2Machine & Hybrid Intelligence Lab, Department of Radiology, Northwestern University, Chicago, IL, USA.
- 3Department of Cardiology, Tunceli State Hospital, Tunceli, Turkey.
- 4Department of Cardiology, Sultan II. Abdülhamid Han Training and Research Hospital, Health Sciences University, Istanbul, Turkey.
- 5Department of Cardiology, Kartal Kosuyolu Research and Education Hospital, Istanbul, Turkey.
- 6Department of Cardiology, İstanbul Education and Research Hospital, Istanbul, Turkey.
- 7Department of Cardiology, Sancaktepe Sehit Prof. Dr. İlhan Varank Training and Research Hospıtal, Istanbul, Turkey.
- 8Department of Cardiology, Van Training and Research Hospital, Van, Turkey.
- 9School of Medicine, University of Maryland, Baltimore, USA.
Abstract
Purpose
Infective endocarditis (IE) is a heterogeneous disease with diverse clinical presentations, microbiological profiles, and outcomes. Conventional classification systems may not fully capture this complexity. We aimed to identify data-driven clinical phenotypes in IE using an unsupervised machine learning (ML) approach.Methods
In this multicenter retrospective study, 335 patients with definite IE from three hospitals were included. A total of 55 clinical, microbiological, and echocardiographic variables were analyzed. Missing data were imputed using the MissForest algorithm. Clustering was performed using spectral clustering based on a Gower dissimilarity matrix and supported by a Bayesian non-parametric approach. Associations between cluster assignment and clinical outcomes were assessed using regression analyses.Results
Two distinct clusters were identified, reflecting different clinical profiles. Cluster 1 was characterized by native valve endocarditis, a higher burden of comorbidities, inflammatory activation, and larger vegetations. Cluster 2 was characterized by prosthetic valve and device-related IE, higher rates of atrial fibrillation, and larger cardiac dimensions. Cluster assignment was not significantly associated with in-hospital mortality in either unadjusted or adjusted analyses. Increasing age emerged as an independent predictor of in-hospital mortality. In the adjusted model, Cluster 2 was independently associated with longer hospital stay, whereas no significant difference was observed in ICU stay.Conclusion
Unsupervised ML identified two clinically distinct phenotypes in IE, reflecting disease heterogeneity. Although these phenotypes were not associated with major clinical outcomes, they provide a data-driven framework for understanding disease heterogeneity and may contribute to future personalized approaches in the management of IE. Larger prospective studies are needed to determine their clinical applicability.
Related
- Antibiotic Prophylaxis of Infective Endocarditis in Dentistry: An Update on the ESC, AHA, and NICE Guidelines
- Robotic tricuspid valve repair for infective endocarditis with leaflet perforation and severe tricuspid regurgitation
- Integrated Clinical, Biological, and Echocardiographic Factors Independently Associated With In-Hospital Mortality in Infective Endocarditis: A Single-Centre Cohort Study From Morocco
- Tricuspid Valve Infective Endocarditis Seen on Point-of-Care Ultrasound: A Case Report
- Incremental Predictive Value of Post-Baseline Surgery and the Impact of Temporal Coding on Mortality Associations in Infective Endocarditis: A Temporally Validated Cohort Study
BibTeX
@article{karacam2026datadriven,
title = {{Data-driven phenotyping of infective endocarditis using unsupervised machine learning: toward a personalized approach.}},
author = {Murat Karaçam and Vedat Cicek and Barkın Kültürsay and Almina Erdem and Seda Tanyeri Üzel and Şeref Berk Tuncer and Tezel Kovancı and Feryat İgit and Çağdaş Bulus and Mert Babaoğlu and Muhammet Bulut and Özgür Can Usta and Burak Tay and Muhammed Ensar Aslan and Mustafa Kamil Yemiş and Ahmet Öz and Tufan Çınar and Ulas Bagci},
journal = {BMC infectious diseases},
year = {2026},
month = jul,
doi = {10.1186/s12879-026-14119-3},
url = {https://europepmc.org/article/MED/42919191},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Data-driven phenotyping of infective endocarditis using unsupervised machine learning: toward a personalized approach.
AU - Murat Karaçam
AU - Vedat Cicek
AU - Barkın Kültürsay
AU - Almina Erdem
AU - Seda Tanyeri Üzel
AU - Şeref Berk Tuncer
AU - Tezel Kovancı
AU - Feryat İgit
AU - Çağdaş Bulus
AU - Mert Babaoğlu
AU - Muhammet Bulut
AU - Özgür Can Usta
AU - Burak Tay
AU - Muhammed Ensar Aslan
AU - Mustafa Kamil Yemiş
AU - Ahmet Öz
AU - Tufan Çınar
AU - Ulas Bagci
PY - 2026
DA - 2026/07/01
JO - BMC infectious diseases
DO - 10.1186/s12879-026-14119-3
UR - https://europepmc.org/article/MED/42919191
AB - Purpose Infective endocarditis (IE) is a heterogeneous disease with diverse clinical presentations, microbiological profiles, and outcomes. Conventional classification systems may not fully capture this complexity. We aimed to identify data-driven clinical phenotypes in IE using an unsupervised machine learning (ML) approach.Methods In this multicenter retrospective study, 335 patients with definite IE from three hospitals were included. A total of 55 clinical, microbiological, and echocardiographic variables were analyzed. Missing data were imputed using the MissForest algorithm. Clustering was performed using spectral clustering based on a Gower dissimilarity matrix and supported by a Bayesian non-parametric approach. Associations between cluster assignment and clinical outcomes were assessed using regression analyses.Results Two distinct clusters were identified, reflecting different clinical profiles. Cluster 1 was characterized by native valve endocarditis, a higher burden of comorbidities, inflammatory activation, and larger vegetations. Cluster 2 was characterized by prosthetic valve and device-related IE, higher rates of atrial fibrillation, and larger cardiac dimensions. Cluster assignment was not significantly associated with in-hospital mortality in either unadjusted or adjusted analyses. Increasing age emerged as an independent predictor of in-hospital mortality. In the adjusted model, Cluster 2 was independently associated with longer hospital stay, whereas no significant difference was observed in ICU stay.Conclusion Unsupervised ML identified two clinically distinct phenotypes in IE, reflecting disease heterogeneity. Although these phenotypes were not associated with major clinical outcomes, they provide a data-driven framework for understanding disease heterogeneity and may contribute to future personalized approaches in the management of IE. Larger prospective studies are needed to determine their clinical applicability.
SN - 1471-2334
ER -