Machine learning-based predictive model for the perioperative co-occurrence of T-cell-mediated rejection and pneumonia in liver transplantation - Research findings

Aug 5, 2026

A machine learning workflow predicts the perioperative co-occurrence of T-cell-mediated rejection (TCMR) and pneumonia after liver transplantation, addressing how both events can compromise graft function and patient survival. Recipient clinical data were retrospectively collected, key predictors were selected via LASSO regression, and five ML algorithms were trained and evaluated using ROC and calibration curves. The SVM model achieved strong discriminative and calibration performance, with SHAP used to interpret variable contributions and individual risk drivers.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages15

Reading Time38 min

AccessFree to download

Objective

Methods

Data and feature selection

Model development and evaluation

Interpretability with SHAP

Results

Cohort and incidence

Predictors identified by LASSO

Model performance

Conclusions

Keywords

TYPE Original Research PUBLISHED 17 September 2025 DOI 10.3389/fimmu.2025.1648993 OPEN ACCESS EDITED BY Antonio Sarasa-Cabezuelo, Complutense University of Madrid, Spain REVIEWED BY Xin Xue, Southeast University, China Jing Yu, Gansu Provincial Hospital, China *CORRESPONDENCE Xuyong Sun [sunxuyong@gxmu.edu.cn](sunxuyong@gxmu.edu.cn) †These authors have contributed equally to this work and share ﬁrst authorship RECEIVED 18 June 2025 ACCEPTED 03 September 2025 PUBLISHED 17 September 2025 CITATION Sun J, Zhu G, Liang Q, Wen N, Li H and Sun X (2025) Machine learning-based predictive model for the perioperative co-occurrence of T-cell-mediated rejection and pneumonia in liver transplantation. Front. Immunol. 16:1648993 . doi: 10.3389/fimmu.2025.1648993 COPYRIGHT © 2025 Sun, Zhu, Liang, Wen, Li and Sun. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Machine learning-based predictive model for the perioperative co-occurrence of T-cell-mediated rejection and pneumonia in liver transplantation Junjie Sun †, Guangyi Zhu †, Qingwen Liang, Ning Wen, Haibin Li and Xuyong Sun* Institute of Transplant Medicine, The Second Afﬁliated Hospital of Guangxi Medical University, Guangxi Clinical Research Center for Organ Transplantation, Guangxi Key Laboratory of Organ Donation and Transplantation, Nanning, China Objective: Perioperative T-cell-mediated rejection (TCMR) and pneumonia occurrence signiﬁcantly impair graft function and patient survival following liver transplantation (LT) . This article aims to develop a machine learning (ML) -based model to predict perioperative co-occurrence of TCMR and pneumonia. Methods: Recipient-related data were retrospectively collected. Predictive Variables were identiﬁed through LASSO regression analysis. Five machine learning algorithms, including support vector machine (SVM), were employed to develop predictive models. Model performance was appraised via the receiver operating characteristic (ROC) curve, and calibration curve. SHapley Additive exPlanations (SHAP) method was employed to visualize model characteristics and individual predictions. Results: This study enrolled 717 LT recipients, including 93 patients with perioperative co-occurrence of TCMR and pneumonia. LASSO regression identiﬁed postoperative direct bilirubin, postoperative international normalized ratio, high-density lipoprotein, postoperative alanine aminotransferase, natural killer cell, tacrolimus (FK506) concentration, Na+, operative time, anhepatic phase, induction regimen, and ICU stay as signiﬁcant predictors. The SVM model demonstrated superior predictive performance, with area under the curve values of 0.881 (95% CI: 0.83–0.93) and 0.786 (95% CI: 0.69–0.88) in the training and test sets, respectively. The calibration curve showed high agreement between the predicted and observed risks. The SVM model demonstrated superior speciﬁcity, sensitivity, F1 score, and recall compared to other models . SHAP analysis identiﬁed variables that contributed to the model predictions. Frontiers in Immunology 01 [frontiersin.org](frontiersin.org) Conclusions: This study constructed a robust predictive model for the perioperative co-occurrence of TCMR and pneumonia. The SVM model demonstrated superior predictive performance. KEYWORDS machine learning, liver transplantation, T-cell-mediated rejection, pneumonia, perioperative period, predictive model Introduction Liver transplantation (LT) has become the optimal treatment for end-stage liver disease. Although liver is considered an immunologi

FAQs

What does the proposed model predict in liver transplantation?

It predicts the perioperative co-occurrence of T-cell-mediated rejection and pneumonia after liver transplantation.

How were predictive variables selected for the model?

Predictive variables were identified using LASSO regression analysis from recipient-related retrospective data.

Which model performed best and how was it evaluated?

The SVM model showed superior predictive performance, assessed with ROC curves and calibration curves, and compared on metrics such as specificity, sensitivity, F1 score, and recall.

Aug 5, 2026

Upload by Clementine