Key Points

  • Sequencing-based approaches are readily used for molecular classifications of diffuse large B-cell lymphoma

  • Genetically distinct diffuse large B-cell lymphoma provides insights into unique lymphomagenesis, prognosis prediction, and combination of targeted treatments

  • Genetic classifiers identify novel vulnerabilities and inform clinical trial designs

Heterogeneity of Diffuse Large B-Cell Lymphoma. Diffuse large B-cell lymphoma (DLBCL) is the most common aggressive B-cell non-Hodgkin lymphoma and is thought to arise from antigen-exposed B cells. Although the majority of patients with DLBCL are curable with combination immunochemotherapy consisting of rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP), a substantial fraction of patients develop recurrent or progressive disease that is often fatal. The clinical heterogeneity of these tumors prompted the development of various classification schemes and included molecular classifiers that i.) improved the accuracy of diagnosis, ii.) identified relevant molecular subtypes with distinct biology, iii.) developed prognostic models for relevant clinical endpoints, and iv.) stratified patients for disease management.

The heterogeneity of DLBCL is partially captured in clinically and/or transcriptionally defined subtypes that provide insights into disease pathogenesis and candidate treatment targets. However, patient stratification for treatment-based transcriptional subtypes, including the activated B-cell type (ABC) and the germinal center B-cell type (GCB), has been largely unsuccessful thus far and suggested that additional heterogeneity had not been fully captured.

Technical advantages paved the way for next-generation sequencing (NGS)-based techniques to be included in clinical and/or low-throughput-based molecular classifiers. These technologies allowed detection and prioritization of all genetic alteration types—recurrent mutations, somatic copy number alterations (SCNAs), and structural variants (SVs) —followed by an integration and assessment of their temporal ordering and associated transcriptional profiles.

Genetic Heterogeneity of DLBCL. Recently, two independent comprehensive genomic efforts at base-pair resolution of hundreds of patients diagnosed with primary DLBCL have captured the complex intrinsic genetic heterogeneity of these lymphomas and have shed light on the genomic architecture of DLBCL. Integration of significant genetic alterations composed of recurrent somatic mutations, SCNAs and SVs, resulted in the discovery of at least 5 genetically distinct DLBCL subsets that predicted outcome to state-of-the-art frontline treatment, suggesting new insights into the lymphomagenesis of DLBCLs and rational combination treatments. These 5 subsets included: 1) a high-risk ABC DLBCLs with near-uniform BCL2 copy gain, frequent activating MYD88 and CD79B mutations, and extranodal tropism (C5 DLBCLs); 2) favorable-risk ABC-DLBCLs with genetic features of an extrafollicular, possibly marginal zone origin (C1 DLBCLs); 3) poor-risk GCB-DLBCLs with BCL2 SVs, inactivating mutations, and/or copy loss of PTEN and alterations of epigenetic enzymes (C3 DLBCLs); 4) a good-risk GCB-DLBCLs with distinct alterations in BCR/PI3K, JAK/STAT, and BRAF pathway components and multiple histones (C4 DLBCLs); and 5) an ABC/GCB-independent group of tumors with biallelic inactivation of TP53, 9p21.3/CDKN2A, and associated genomic instability (C2 DLBCLs).

The biology of these 5 DLBCL subsets was largely confirmed by an independent non-overlapping large-scale study that followed a completely different analytical approach and discovered similar groups with shared pathogenetic mechanisms. In particular, C1 DLBCLs were similar to BN2 (BCL6/NOTCH2), C3 DLBCLs were similar to EZB (EZH2/BCL2), and C5 DLBCLs were similar to MCD (MYD88/CD79B). In subsequent work, the biology of the remaining DLBCL subtypes (C2 and C4 DLBLCs) was recently independently validated by two other groups (C4 ~ ST2; C2 ~ A53 | C4 ~ SGK1; C2 ~ TP53).

These studies underscore that DLBCL is a genetically heterogenous disease with at least 5-7 different molecular subtypes. At the same time, they also highlight that despite having several hundreds of sequenced tumors, we still do not capture the full spectrum of DLBCL subsets, and for low-frequency subtypes in particular larger sample sizes are needed. More work is needed, to link genetically defined subtypes to actionable vulnerabilities, and eventually build molecular classifiers that link genetics to dependencies with the hope of using them for precision medical trials.

Citation Format: Bjoern Chapuy. Molecular classification of diffuse large B-cell lymphoma [abstract]. In: Proceedings of the AACR Virtual Meeting: Advances in Malignant Lymphoma; 2020 Aug 17-19. Philadelphia (PA): AACR; Blood Cancer Discov 2020;1(3_Suppl):Abstract nr IA06.