Risk Evaluation Model For Kidney Transplant Rejection Using Peripheral Blood Gene Expression Profiling

Objective: This study is aimed to develop and validate a risk evaluation model for kidney transplant rejection for the purpose of working as a non-invasive auxiliary diagnosis method in clinical practice.Methods: This study conducted an integrated analysis of gene expression profiling in peripheral blood of kidney transplant recipients with rejection from different centers and regions in the GEO database.Gene selection was performed by LIMMA differential expression analysis and LASSO regression,followed by development of risk evaluation model for kidney transplant rejection through Logistic regression.The utility of the proposed model was evaluated using the receiver operating characteristic curve,calibration curve,and decision curve analysis.Results: The Risk of Rejection Score(RRS)generated from the Risk Evaluation Model for kidney transplant rejection using peripheral blood gene expression profiling of 15-gene panel from gene microarray data is a score ranging from 0 to 100 to evaluate the recipient’s risk of having rejection(including acute rejection,subclinical rejection and borderline changes).The greater the RRS,the higher risk of having rejection evaluated by this model,when the cut-off of RSS is set to 35.55,the AUC of this model reached0.836(0.812-0.859)and the overall accuracy was 76.3% with the sensitivity of 73.0% and specificity of 78.0%,while the positive and negative predictive value was 62.9% and 84.9%respectively.The decision curve analysis demonstrated that the application of this model will bring significant clinical benefits to kidney transplant recipients.Conclusion: This model is capable of evaluating the risk of rejection of kidney transplant recipients in the dataset of GEO by using 15-gene expression panel in peripheral blood,it also helps to avoid unnecessary invasive biopsy.However,this model requires additional clinical trial study for further validation.