Construction of three‐gene‐based prognostic signature and analysis of immune cells infiltration in children and young adults with B‐acute lymphoblastic leukemia

BACKGROUND: Although B‐acute lymphoblastic leukemia (B‐ALL) patients' survival has been improved dramatically, some cases still relapse. This study aimed to explore the prognosis‐related novel differentially expressed genes (DEGs) for predicting the overall survival (OS) of children and young adults (CAYAs) with B‐ALL and analyze the immune‐related factors contributing to poor prognosis. METHODS: GSE48558 and GSE79533 from Gene Expression Omnibus (GEO) and clinical sample information and mRNA‐seq from Therapeutically Applicable Research to Generate Effective Treatments (TARGET) database were retrieved. Prognosis‐related key genes were enrolled to build a Cox proportional model using multivariate Cox regression. Five‐year OS of patients, clinical characteristic relevance and clinical independence were assessed based on the model. The mRNA levels of prognosis‐related genes were validated in our samples and the difference of immune cells composition between high‐risk and low‐risk patients were compared. RESULTS: One hundred and twelve DEGs between normal B cells and B‐ALL cells were identified based on GSE datasets. They were mainly participated in protein binding and HIF‐1 signaling pathway. One hundred and eighty‐nine clinical samples were enrolled in the study, both Kaplan–Meier (KM) analysis and univariate Cox regression analysis showed that CYBB, BCL2A1, IFI30, and EFNB1 were associated with prognosis, CYBB, BCL2A1, and EFNB1 were used to construct prognostic risk model. Moreover, compared to clinical indicators, the three‐gene signature was an independent prognostic factor for CAYAs with B‐ALL. Finally, the mRNA levels of CYBB, BCL2A1, and EFNB1 were significantly lower in B‐ALL group as compared to controls. The high‐risk group had a significantly higher percentage of infiltrated immune cells. CONCLUSION: We constructed a novel three‐gene signature with independent prognostic factor for predicting 5‐year OS of CAYAs with B‐ALL. Additionally, we discovered the difference of immune cells composition between high‐risk and low‐risk groups. This study may help to customize individual treatment and improve prognosis of CAYAs with B‐ALL.