Machine learning-based prediction of one-year mortality after alloHCT identifies the impact of pre-transplant immunity and inflammation

Aug 4, 2026

Accurate prediction of mortality after allogeneic hematopoietic stem cell transplantation (alloHCT) is essential for individualized treatment decisions, yet existing clinical risk scores capture only a limited number of variables and show modest predictive performance. A single-center retrospective analysis included 909 adult patients with hematologic malignancies undergoing alloHCT. Using 31 features, machine-learning models predicted death within the first year. A random forest model achieved the best performance (AUC 0.773), with good generalizability (AUC 0.748). SHAP interpretation highlighted age plus pre-transplant lymphocyte subsets and inflammatory markers (CD4+, CD8+ and B-lymphocyte counts, albumin, CRP). The model outperformed established scores (HCTCI, EASIX, rDRI, mGPS), distinguishing low- and high-risk patients. The study supports pre-transplant inflammation and identifies lymphocyte subsets as previously underrecognized risk factors, with external multicenter validation needed.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages14

Reading Time35 min

AccessFree to download

Abstract

Study design and dataset

Machine-learning modeling and performance

Feature interpretation with SHAP

Comparison with established clinical risk scores

Key findings and validation needs

OPUS-HSO Repositorium der Hochschule Offenburg [https://opus.hs-offenburg.de](https://opus.hs-offenburg.de) Machine learning-based prediction of one-year mortality after alloHCT identifies the impact of pre-transplant immunity and inflammation Thomas Meyer, Robert Meyer, Maren Hackenberg, Daniela Oelke, Laura Gengenbach, Christoph Rummelt, Hauke Wilcken, Kristina Maas-Bauer, Ralph Wäsch, Justus Duyster, Hartmut Bertz, Jesús Duque-Afonso, Jürgen Finke, Robert Zeiser, Claudia Wehr Zitiervorschlag im APA Stil: Meyer, T., Meyer, R., Hackenberg, M., Oelke, D., Gengenbach, L., Rummelt, C., Wilcken, H., Maas-Bauer, K., Wäsch, R., Duyster, J., Bertz, H., Duque-Afonso, J., Finke, J., Zeiser, R., & Wehr, C. (2026) . Machine learning-based prediction of one-year mortality after alloHCT identifies the impact of pre-transplant immunity and inflammation. Frontiers in Immunology, 16, 01– 13. [https://doi.org/10.3389/fimmu.2025.1745873](https://doi.org/10.3389/fimmu.2025.1745873) Abstract Accurate prediction of mortality after allogeneic hematopoietic stem cell transplantation (alloHCT) is essential for individualized treatment decisions, yet existing clinical risk scores capture only a limited number of variablesand show modest predictive performance. In our single-center retrospective analysis, we included data from 909 adult patients with hematologic malignancies undergoing alloHCT. We used 31 features to build machine-learning models to predict death within the first year after alloHCT. These features included established clinical risk factors together with pre-transplant lymphocyte subsets and inflammatory markers. Among four models, a random forest algorithm showed the best performance (AUC = 0.773) and retained good generalizability in an independent test set (AUC = 0.748) . SHapley Additive exPlanations (SHAP)-based interpretation of the machine-learning models showed that age together with five easily measurable pre-transplant immunological and inflammatory parameters influenced the outcome: pre-transplant CD4 + , CD8 +, and B-lymphocyte counts, albumin, and C-reactive protein (CRP) levels. Based on these features, our random forest approach outperformed established clinical risk scores (HCTCI, EASIX, rDRI, mGPS) in predicting one-year mortality after alloHCT and more effectively distinguished patients at low and high risk of an adverse outcome. Our study shows that machine-learning-based models can not only predict patient outcomes after alloHCT but also serve as powerful tools for data exploration, confirming the prognostic relevance of pre-transplant inflammation while uncovering the critical role of lymphocyte subsets as previously unknown risk factors. External validation in independent multicenter cohorts will be required to confirm generalizability. Nutzungsbedingungen Dieses Dokument wird unter diesen Bedinungen zur Verfügung gestellt: Creative Commons-CC BY-Namensnennung 4.0 International Für weitere Informationen siehe: [https://creativecommons.org/licenses/by/4.0/deed.de](https://creativecommons.org/licenses/by/4.0/deed.de) Kontakt Hochschule Offenburg | Bibliothek Badstraße 24 77652 Offenburg Telefon: (0781) 205-240 [E-Mail: bibliothek@hs-offenburg.de](E-Mail: bibliothek@hs-offenburg.de)[ ](E-Mail: bibliothek@hs-offenburg.de)[www.hs-offenburg.de/bibliothek](www.hs-offenburg.de/bibliothek) TYPE Original Research PUBLISHED 19 January 2026 DOI 10.3389/fimmu.2025.1745873 OPEN ACCESS EDITED BY Kelley M. K. Hitchman, University of Texas Health Science Center San Antonio, United States REVIEWED BY Eman M. Elsabbagh, CITADEL Lab [Computational Immunology & Transplant AI Data Engineering Lab], United States Maximilian Alexander Röhnert, Technical University Dresden, Germany *CORRESPONDENCE Claudia Wehr [claudia.wehr@uniklinik-freiburg.de](claudia.wehr@uniklinik-freiburg.de) RECEIVED 13 November 2025 REVISED 18 December 2025 ACCEPTED 23 December 2025 PUBLISHED

FAQs

How does the proposed model compare with established risk scores?

The random forest approach outperformed established clinical risk scores (HCTCI, EASIX, rDRI, mGPS) in predicting one-year mortality and more effectively separated low- and high-risk patients. External multicenter validation is still required.

Aug 4, 2026

Upload by Valentina