Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 10, 2026Journal of Immunology ResearchOpen Access

Cytokine Signatures Outperform Immune Subsets in Machine Learning Models for Predicting Acute Graft‐Versus‐Host Disease at Neutrophil Engraftment

View Full Paper
Ask AI
Bookmark
Share

Authors

MMMohini MendirattaPPP K PandeySPS. K. Singh Pandey

Discussion

Loading...

Member takes

Overview

Research demonstrates cytokine profiles predict acute graft-versus-host disease better than immune subsets, suggesting improved early detection methods.

Key Points

  • The study aims to develop predictive models for acute graft-versus-host disease using cytokine and immune profiles.
  • Analyzed immune subsets and cytokines from peripheral blood samples of Allo-HSCT recipients.
  • Utilized machine learning algorithms, including support vector classifier, decision tree, and random forest.
  • Trained models on 48 features: 34 immune subsets and 14 cytokines.
  • Patients developing aGvHD had a reduced CD4+/CD8+ ratio and altered immune profiles.
  • Cytokine profiles achieved perfect predictive accuracy (1.00) compared to immune subsets.
  • Predictive performance for other models: T-cell (0.96), NK cell (0.93), DC (0.90), and B cell (0.86).

Cite This Study

Mendiratta et al. (2026) studied this question.

synapsesocial.com/papers/69af94fa70916d39fea4c133https://doi.org/10.1155/jimr/1066614
View Full Paper
Ask AI
Bookmark
Share