Genetic and transcriptomic analyses of diffuse large B‐cell lymphoma patients with poor outcomes within two years of diagnosis

Abstract

Background: R-CHOP immunochemotherapy is the first-line treatment for DLBCL, but approximately 30%–40% of patients still experience refractory or relapsed (R/R) disease. We aimed to characterize the molecular features of DLBCL derived from patients with early R/R disease and develop a prognosis model to identify those high-risk patients. Methods: We performed a comprehensive genomic and transcriptomic characterization on 161 R-CHOP-treated DLBCL samples. These patients were grouped by follow-up outcome status after R-CHOP treatment. The first group had R/R disease within two years of diagnosis (poor outcome; n = 50), and the second group remained in remission at two years following R-CHOP treatment (good outcome; n = 111). In an addition, external cohorts were included in the validation analysis (GSE117556, n = 374; GSE181063, n = 810). Results: Patients with poor outcomes more often had an advanced stage of the disease, a high international prognostic index (IPI) and a non-GCB subtype of the tumor. We identified a set of frequent somatically mutated targets (Figure 1A), including PIM1, TP53, MPEG1 and ROBO1, as well as a specific mutational signature (activation-induced cytidine deaminase related) in DLBCLs with poor outcomes. Transcriptomic analyses further showed a distinct gene expression pattern and a less inflamed tumor microenvironment in these patients (Figure 1B). Finally, we developed an 11-gene signature as an independent prognostic marker for DLBCL patients treated with R-CHOP (Figure 1C). The 11-gene signature could further stratify DLBCL patients based on their IPI scores (Figure 1D). Additionally, the use of 11-gene signature to identify high-risk patients was also validated in two independent cohorts, suggesting the robustness of the risk model. Furthermore, our model effectively identified high-risk patients, including those with double-hit and MCD genetic subtypes that are associated with poor survival outcomes. FIGURE 1 (A) The mutation pattern between poor outcome and good outcome DLBCLs. (B) Tumor-infiltrating immune cells in DLBCL patients with poor and good outcomes. (C-D) Kaplan–Meier survival analysis of PFS between high-risk and low-risk groups (C) and the association of 11-gene risk scores and IPI risk groups (D). Keywords: Aggressive B-cell non-Hodgkin lymphoma, Diagnostic and Prognostic Biomarkers, Genomics, Epigenomics, and Other -Omics No conflicts of interests pertinent to the abstract.