This study investigates the role of gene-expression signatures in predicting survival in patients with diffuse large B-cell lymphoma (DLBCL) treated with either CHOP or R-CHOP (rituximab plus CHOP). The researchers profiled gene expression in pretreatment biopsy specimens from 181 patients treated with CHOP and 233 patients treated with R-CHOP. They identified three gene-expression signatures: "germinal-center B-cell," "stromal-1," and "stromal-2." The "germinal-center B-cell" signature, reflecting the distinction between activated B-cell-like and germinal-center B-cell-like subtypes, was associated with favorable prognosis. The "stromal-1" signature, reflecting extracellular matrix deposition and histiocytic infiltration, was also prognostically favorable. In contrast, the "stromal-2" signature, reflecting tumor blood-vessel density, was prognostically unfavorable. These signatures were validated in a validation group, and a multivariate model incorporating these signatures predicted survival in both CHOP and R-CHOP cohorts. The study suggests that survival in DLBCL is influenced by differences in immune cells, fibrosis, and angiogenesis in the tumor microenvironment. The findings have implications for future clinical trials, highlighting the importance of considering biologic differences in patient cohorts to improve treatment outcomes.This study investigates the role of gene-expression signatures in predicting survival in patients with diffuse large B-cell lymphoma (DLBCL) treated with either CHOP or R-CHOP (rituximab plus CHOP). The researchers profiled gene expression in pretreatment biopsy specimens from 181 patients treated with CHOP and 233 patients treated with R-CHOP. They identified three gene-expression signatures: "germinal-center B-cell," "stromal-1," and "stromal-2." The "germinal-center B-cell" signature, reflecting the distinction between activated B-cell-like and germinal-center B-cell-like subtypes, was associated with favorable prognosis. The "stromal-1" signature, reflecting extracellular matrix deposition and histiocytic infiltration, was also prognostically favorable. In contrast, the "stromal-2" signature, reflecting tumor blood-vessel density, was prognostically unfavorable. These signatures were validated in a validation group, and a multivariate model incorporating these signatures predicted survival in both CHOP and R-CHOP cohorts. The study suggests that survival in DLBCL is influenced by differences in immune cells, fibrosis, and angiogenesis in the tumor microenvironment. The findings have implications for future clinical trials, highlighting the importance of considering biologic differences in patient cohorts to improve treatment outcomes.