High-grade B-cell lymphomas (HGBL) represent a heterogeneous group of aggressive mature B-cell malignancies characterized by clinical, morphologic, and cytogenetic features that bridge diffuse large B-cell lymphoma (DLBCL) and Burkitt lymphoma. Current classification frameworks define HGBL largely by morphology and MYC-based cytogenetics, yielding categories that are inconsistently applied and biologically incomplete. Recent gene expression profiling studies have identified a reproducible "dark-zone signature" (DZsig), present in most double-hit lymphomas and a subset of DLBCL, which portends poor outcomes and provides a more objective biological definition of HGBL. Genomic studies further demonstrate that HGBL can arise from diverse DLBCL subtypes, converging on MYC deregulation, genomic instability, and T-cell depleted immune microenvironment. Therapeutically, retrospective and phase 2 studies suggest improved outcomes with intensified regimens such as dose-adjusted EPOCH-R compared with R-CHOP, though randomized data remain limited.
This review focuses on the application of DZsig and molecular classifications, applicability of the POLARIX trial data to HGBL of germinal center B cell and activated B cell subtypes, and the use of novel T-cell engaging therapies in HGBL. CAR T-cell therapy has transformed the relapsed/refractory setting, with comparable efficacy in HGBL and DLBCL. In contrast, responses to CD20xCD3 bispecific antibodies appear attenuated in HGBL-DH, though combination regimens with antibody-drug conjugates or chemotherapy show promise.
Other options, including CD19-directed antibody-drug conjugates, offer some activity. Future progress will depend on replacing morphology-based criteria with biology-driven classification that incorporates DZsig and genomic subtyping, and on developing rational, mechanism-based therapies that address both tumor-intrinsic drivers and the hostile microenvironment.