Dynamic Metabolic Profiling and Diagnostic Biomarkers of Acute Graft-Versus-Host Disease Based on Metabolomics

This study utilized LC-MS/MS-based metabolomics to characterize the dynamic metabolic changes associated with the onset and progression of acute graft-versus-host disease (aGVHD), aiming to identify potential noninvasive biomarkers for diagnosis and prognostic assessment. A total of 110 patients with AML who underwent allogeneic hematopoietic stem cell transplantation (allo-HSCT) were enrolled in the study. Plasma samples from these patients were analyzed using an untargeted metabolomics platform. Differential metabolites and their clinical relevance were investigated using multiple statistical approaches, including pathway enrichment, hierarchical clustering, and ROC curve analysis. The results showed that compared with patients without aGVHD, 36 plasma metabolites were significantly dysregulated in those with aGVHD (VIP > 1, p < 0.05). Among these, six glycerophospholipids-PC(P-16:0/18:1), PC(20:4/20:4), PC(16:0/16:0), PC(18:0/20:3), PE(P-18:0/20:5), and PC(o-16:1/18:0)-exhibited regular alterations across different stages of aGVHD. The combined AUC under the six metabolites was 0.960, highlighting their strong diagnostic potential. Multivariate analysis further indicated that this six-metabolite signature was predictive of 4-year overall survival, with several metabolites serving as independent prognostic factors. In conclusion, these six glycerophospholipid metabolites display dynamic changes during aGVHD progression and hold promise as noninvasive biomarkers for both diagnosis and prognostic stratification. Validation in larger patient cohorts is warranted to confirm these findings.