For patients diagnosed with diffuse large B cell lymphoma, standard molecular classification still leaves a significant proportion with unexplained treatment failure. A new multi-omic framework may finally explain why some tumors resist therapy regardless of how they look under conventional testing — and that has direct implications for how oncologists stratify risk and select treatment.
By layering proteomic, transcriptomic, and genomic data across 478 DLBCL tumor samples, investigators identified seven distinct proteogenotypes that cut across the established ABC/GCB cell-of-origin classification. One cluster, designated PG4, emerged as a high-risk group with significantly worse clinical outcomes independent of the international prognostic index, cell-of-origin subtype, and known genetic risk features. PG4 tumors — drawn from both activated B cell-like and germinal center B cell-like cases as well as genetically unclassified specimens — share a dark-zone B cell phenotype and are enriched for BTG1 mutations capable of activating MYC signaling. Single-cell and spatial transcriptomic analyses confirmed elevated MYC and TCF3/TCF4 transcriptional activity in PG4 even in the absence of canonical MYC chromosomal translocations. The PG4 tumor microenvironment is also marked by CD8+ T cell exhaustion, suggesting active immune suppression.
This work is analytically ambitious and scientifically meaningful. Multi-omic integration has been applied to other cancers — notably breast and lung — but DLBCL's extraordinary heterogeneity has made molecular stratification notoriously difficult. The discovery that MYC pathway activation can operate through BTG1 mutation rather than translocation fills a genuine explanatory gap: clinicians have long observed poor outcomes in patients without detectable MYC rearrangements, and this may partly explain that paradox. The T cell exhaustion signature in PG4 also raises the question of whether checkpoint immunotherapy might offer leverage in this subgroup. Key caveats: the cohort, while substantial at 478 cases, requires prospective validation, and proteomic profiling is not yet routine clinical infrastructure. This is incremental but potentially practice-shaping work if the PG4 classification proves reproducible in external cohorts.