Related Experiment Video
Updated: Jan 14, 2026
High Throughput Sequential ELISA for Validation of Biomarkers of Acute Graft-Versus-Host Disease
Published on: October 31, 2012
Assessing the Clinical Performance of a Presymptomatic Acute Graft-Versus-Host Disease Biomarker Test in
Lin Wang1, Rafael C Pulgrossi2, Jaclyn Anderson1
1Department of Pathology, Stanford University School of Medicine, Stanford, CA, United States.
Background:
Acute graft-versus-host disease (aGVHD) is a common complication of allogenic hematopoietic cell transplantation (HCT). The aGVHD presymptomatic algorithm is comprised of 2 serum biomarkers, suppression of tumorigenicity 2 and regenerating islet-derived 3-alpha. This test was developed to predict severe aGVHD and 6-month nonrelapse mortality (NRM) in allogeneic HCT recipients.
Methods:
The medical records of 223 patients tested with the aGVHD presymptomatic algorithm between January 2020 and June 2024 were retrospectively reviewed. Sensitivity and specificity for the prediction of severe aGVHD (grade III or IV) and 6-month NRM were determined. Logistic regression modeling was used to assess the odds ratios of risk factors for severe aGVHD and 6-month NRM. The likelihood ratio test was used to evaluate whether the addition of the aGVHD presymptomatic algorithm improved model fit.
Results:
Severe aGVHD occurred in 8.5% of patients, and 6-month NRM was 7.2%. The aGVHD presymptomatic algorithm demonstrated a sensitivity of 21.1% (95% CI, 8.0%-43.9%) and specificity of 83.3% (95% CI, 77.6%-87.9%) for severe aGVHD and a sensitivity of 43.8% (95% CI, 23.1%-66.8%) and specificity of 85.0% (95% CI, 79.5%-89.3%) for 6-month NRM. The addition of results from the aGVHD presymptomatic algorithm to conventional risk factors did not improve prediction of aGVHD but improved prediction of 6-month NRM.
Conclusions:
The aGVHD presymptomatic algorithm was suboptimal for routine clinical use. Further development of predictive aGVHD biomarkers may be required to aid in the management of allogeneic HCT recipients.
