Related Experiment Video
Updated: May 8, 2026
Induction of Graft-versus-host Disease and In Vivo T Cell Monitoring Using an MHC-matched Murine Model
Published on: August 29, 2012
Diagnostic prediction of gastrointestinal graft-versus-host disease based on a clinical- CT- signs nomogram model
Qing Feng1,2, Fengming Xu1, Kaiming Guan1
1Department of Radiology, The First Affiliated Hospital of Guangxi Medical University, Shuangyong Road, Nanning, 530021, Guangxi Province, China.
Objective:
Gastrointestinal graft-versus-host disease (GI-GVHD) is one of the complications that can easily occur after hematopoietic stem cell transplantation (HSCT). Timely diagnosis and treatment are pivotal factors that greatly influence the prognosis of patients. However, the current diagnostic method lacks adequate non-invasive diagnostic tools.
Methods:
A total of 190 patients who suspected GI-GVHD were retrospectively included and divided into training set (n = 114) and testing set (n = 76) according to their discharge time. Least absolute shrinkage and selection operator (LASSO) regression was used to screen for clinically independent predictors. Based on the logistic regression results, both computed tomography (CT) signs and clinically independent predictors were integrated in order to build the nomogram, while the testing set was verified independently. The receiver operating characteristic (ROC), area under the curve (AUC), decision curve, and clinical impact curve were used to measure the accuracy of prediction, clinical net benefit, and consistency of diagnostic factors.
Results:
Four key factors, including II-IV acute graft-versus-host disease (aGVHD), the circular target sign, multifocal intestinal inflammation, and an increased in total bilirubin, were identified. The combined model, which was constructed from CT signs and clinical factors, showed higher predictive performances. The AUC, sensitivity, and specificity of the training set were 0.867, 0.787, and 0.811, respectively. Decision curve analysis (DCA), net reclassification improvement (NRI), and integrated discrimination improvement (IDI) showed that the developed model exhibited a better prediction accuracy than the others.
Conclusions:
This combined model facilitates timely diagnosis and treatment and subsequently improves survival and overall outcomes in patients with GI-GVHD.
Critical Relevance Statement:
GI-GVHD is one of the complications that can easily occur after HSCT. However, the current diagnostic approach lacks adequate non-invasive diagnostic methods. This non-invasive combined model facilitates timely treatment and subsequently improves patients with GI-GVHD survival and overall outcomes.
Key Points:
• There is currently lacking of non-invasive diagnostic methods for GI-GVHD. • Four clinical CT signs are the independent predictors for GI-GVHD. • Association between the CT signs with clinical factors may improve the diagnostic performance of GI-GVHD.
