EUVIMEDEUVIMED entdecken
Open Access

Predictive, preventive, and personalized medicine in statin-treated patients with NSTE-ACS, HFpEF, and type 2 diabetes: a data-driven model for long-term residual risk assessment

Zeyan Li, Hengyang Liu, Li Xu, Wenyuan Yin, Jiajun Zhu, Bowen Zhou, Bing Wu, Tongjian Zhu, Hui Li, Jinjun Liu, Yijun Wang, Jun Wang iD

EPMA Journal · 24.08.2026 · 17 (3) · 479-500

DOI: 10.1007/s13167-026-00467-2 ↗

Kurz gefasst

Fachartikel in der Zeitschrift EPMA Journal, online erschienen am 24.08.2026. Verfasst von Zeyan Li und 11 weiteren Autor:innen. Beteiligte Einrichtung: Department of Cardiology, The First Affilitaed Hospital of Bengbu Medical University, Bengbu, Anhui China. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 74 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: Clinical and Translational Research Project of Anhui Province, Research Funds of Joint Research Center for Regional Diseases of IHM, Anhui Provincial Health and Health Commission Scientific Research Project.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Despite significant advances in guideline-directed medical therapy (GDMT), statin-treated patients with non-ST-elevation acute coronary syndrome (NSTE-ACS), heart failure with preserved ejection fraction (HFpEF), and type 2 diabetes mellitus remain at substantial residual risk of major adverse cardiac and cerebrovascular events (MACCEs). Within the framework of predictive, preventive, and personalized medicine (3PM/PPPM), this study aimed to develop and externally validate a data-driven survival model for individualized long-term risk stratification to support the transition from reactive management to proactive and targeted prevention.

Method

In this multicenter retrospective cohort study, 1,206 NSTE-ACS patients with HFpEF and type 2 diabetes who underwent successful percutaneous coronary intervention (PCI) were enrolled from six Chinese tertiary hospitals and geographically divided into a training cohort (n = 763) and an independent validation cohort (n = 443). Sixty clinical, laboratory, and echocardiographic variables were screened using wrapper-based feature selection across five machine learning algorithms with stratified 5-fold cross-validation. Thirty-one survival models spanning four methodological categories were developed and compared. The primary endpoint was MACCEs, defined as a composite of recurrent acute coronary syndrome, stroke, and hospitalization for heart failure.

Results

Seven consensus biomarkers were identified: age, albumin, creatinine, monocyte-to-lymphocyte ratio, non-high-density lipoprotein cholesterol to high-density lipoprotein cholesterol ratio (NHHR), N-terminal pro-B-type natriuretic peptide (NT-proBNP), and non-high-density lipoprotein cholesterol (non-HDL-C). Among 31 survival models, Xgboost.Cox demonstrated the best overall performance, achieving a mean Uno's C-index of 0.879 across cohorts (training, 0.947; validation, 0.810). Time-dependent SurvSHAP(t) (survival Shapley additive explanations) was employed to provide instance-specific interpretability. An interactive web-based calculator implementing the Xgboost.Cox model was developed to facilitate clinical translation and individualized risk assessment. Conclusion and expert recommendations: This 3PM/PPPM-guided framework enables early identification of high-risk phenotypes among NSTE-ACS patients with HFpEF and type 2 diabetes before irreversible clinical deterioration occurs. We recommend: (1) implementing multidimensional baseline profiling including inflammatory, metabolic, and cardiac functional bio-marker panels for the statin-treated patients with NSTE-ACS to facilitate risk stratification at discharge; (2) integrating the web-based XGBoost-Cox calculator into clinical workflows to generate individualized risk estimates and support shared decision-making; and (3) tailoring follow-up intensity and preventive strategies to predicted risk, with closer monitoring and more intensive GDMT optimization considered for high-risk patients, while ensuring guideline-recommended therapy for all eligible individuals. Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1007/s13167-026-00467-2.

Schlagwörter

Stichwörter der Autor:innen

Heart failure with preserved ejection fractionLong-term risk stratificationMachine learningNon-ST-elevation acute coronary syndromePredictive preventive personalized medicine (PPPM / 3PM)Residual cardiovascular riskTherapy tailored to individualized patient profileType 2 diabetes mellitus

Autor:innen und Institutionen

  • Department of Cardiology, The First Affilitaed Hospital of Bengbu Medical University, Bengbu, Anhui China.
  • Joint Research Center for Regional Diseases of IHM, Bengbu Medical University, Bengbu, China.
  • Joint Research Center for Regional Diseases of IHM, The First Affiliated Hospital of Bengbu Medical University, Bengbu, China.
  • Department of Cardiovascular Medicine, Guiqian International General Hospital, Guiyang City, Guizhou Province 550024 China.
  • Department of Cardiology, Renmin Hospital of Wuhan University, Wuhan, Hubei China.
  • Electrocardiology Department, People's Hospital of Xinjiang Uygur Autonomous Region, Xinjiang Medical University, Urumqi, 830011 China.

ORCID-Profile: Jun Wang

Förderung

  • Clinical and Translational Research Project of Anhui Province
  • Research Funds of Joint Research Center for Regional Diseases of IHM
  • Anhui Provincial Health and Health Commission Scientific Research Project
  • The First Affiliated Hospital of Bengbu Medical University for Excellent Young Scholars
  • National Engineering Research Center of Science and Technology Information

Zitieren

APA

Li, Z., Liu, H., Xu, L., Yin, W., Zhu, J., Zhou, B., Wu, B., Zhu, T., Li, H., Liu, J., Wang, Y., & Wang, J. (2026). Predictive, preventive, and personalized medicine in statin-treated patients with NSTE-ACS, HFpEF, and type 2 diabetes: a data-driven model for long-term residual risk assessment. EPMA Journal, 17(3), 479-500. https://doi.org/10.1007/s13167-026-00467-2

Vancouver

Li Z, Liu H, Xu L, Yin W, Zhu J, Zhou B, et al. Predictive, preventive, and personalized medicine in statin-treated patients with NSTE-ACS, HFpEF, and type 2 diabetes: a data-driven model for long-term residual risk assessment. EPMA Journal. 2026;17(3):479-500. doi:10.1007/s13167-026-00467-2

BibTeX
@article{Li2026,
  author = {Li, Zeyan and Liu, Hengyang and Xu, Li and Yin, Wenyuan and Zhu, Jiajun and Zhou, Bowen and Wu, Bing and Zhu, Tongjian and Li, Hui and Liu, Jinjun and Wang, Yijun and Wang, Jun},
  title = {Predictive, preventive, and personalized medicine in statin-treated patients with NSTE-ACS, HFpEF, and type 2 diabetes: a data-driven model for long-term residual risk assessment},
  journal = {EPMA Journal},
  year = {2026},
  volume = {17},
  number = {3},
  pages = {479--500},
  doi = {10.1007/s13167-026-00467-2}
}
Kostenlos starten →