Gene Signature-based Prediction of Infliximab Response in Patients with Crohn?s Disease

Objective: This study evaluated the predictive value of key genes in response to treatment with infliximab of patients with Crohn's disease (CD). Methods: Four datasets on CD treated with infliximab were downloaded from the GEO database. We integrated these datasets and analyzed the characteristics of differentially expressed genes, immune infiltration, and weighted gene co-expression network between the non-response and response groups of patients treated with infliximab. The key genes that could predict the infliximab response were determined using support vector machine (SVM) and least absolute shrinkage and selection operator (LASSO) regression. The receiver operating characteristic (ROC) curve, concordance index (C index), and GiViTi calibration band were used to evaluate the diagnostic performance of these genes. In addition, functional annotation and pathway enrichment analysis of key genes were performed. Finally, the correlation between key genes and immune cells was analyzed. Results: Seven genes with a high predictive value for infliximab response in patients with CD were selected from four integrated datasets. The area under the curve (AUC > 0.8), c-index (>0.8), and GiViTi calibration bands (p > 0.05) of the 7-gene signature showed its ability to predict the infliximab response in patients with CD. Differences existed in monocytes between non-response and response groups of infliximab in patients with CD. Three genes (SPP1, CHN1, and GAS1) were negatively proportional to the number of monocytes. Conclusion: The results indicated that the 7-gene signature could serve as a candidate biomarker for predicting the response to infliximab treatment in patients with CD.