Proteogenomic Analysis of 478 Tumors Identifies a High-Risk MYC-Driven Lymphoma Subtype

Article Impact Level: HIGH
Data Quality: STRONG
Summary of  Cancer Cell https://doi.org/10.1016/j.ccell.2026.05.008 
Dr. Julius C. Enssle  et al.

Points

Frankfurt medical investigators combined proteomic, transcriptomic, and genomic sequencing data from 478 tumor samples to resolve the clinical heterogeneity of diffuse large B-cell lymphoma.
Machine learning algorithms categorized the multi-omic data into seven distinct proteogenotypes that transcend standard genetic classification systems to predict patient survival outcomes.
Clinical tracking identified proteogenotype 4 as a highly aggressive tumor class associated with poor prognosis and standard therapy failure independent of traditional risk benchmarks.
Deep spatial transcriptomics proved that high-risk tumors display a dark-zone B-cell phenotype driven by hyperactivated MYC signaling pathways alongside a severe exhaustion of surrounding cytotoxic immune cells.
Experimental laboratory research successfully eliminated cultured high-risk lymphoma cells using targeted pharmacological inhibitors to establish a viable framework for future subspecialty precision medicine.

Summary

Initiated to decode the complex molecular heterogeneity of diffuse large B-cell lymphoma (DLBCL), this study integrated multi-layered omics data to identify novel biomarkers for therapy-resistant phenotypes. DLBCL is the most common aggressive lymphoma, with more than 150,000 new cases diagnosed globally each year. While standard chemoimmunotherapy regimens like R-CHOP or Pola-R-CHP achieve complete remission in nearly two-thirds of patients, more than one-third experience disease relapse or initial refractory failure. The research sought to overcome the limitations of purely genetic classification by combining multi-omic profiling with interpretable machine learning to map explicit oncogenic mechanisms governing high-risk cohorts.

Using an international patient cohort, investigators comprehensively analyzed biopsy tissue samples from 478 patients with DLBCL, simultaneously executing genomic, transcriptomic, and deep proteomic sequencing. Interpretable machine learning models classified these multi-omic data layers into seven distinct proteogenotypes (PGs) that cross cut traditional cell-of-origin boundaries. Within this framework, proteogenotype 4 (PG4) emerged as a critical high-risk cluster uniquely associated with poor clinical survival outcomes independent of established prognostic parameters, such as the International Prognostic Index, specific genetic mutations, or conventional molecular subtypes.

Mechanistic evaluations, validated via single-cell transcriptomics and spatial profiling, revealed that PG4 tumors exhibit a distinct dark-zone-related B-cell phenotype heavily enriched for BTG1 mutations and hyperactivated MYC and TCF3/4 signaling networks. Furthermore, spatial mapping demonstrated that the microenvironment of these high-risk lesions is immunologically “cold,” characterized by a profound depletion of active immune effectors and an enrichment of exhausted CD 8+ cells. In vitro experiments successfully eliminated PG4 lymphoma cells through targeted pharmacological inhibition of MYC-driven pathways, demonstrating that this proteogenomic framework represents a highly viable strategy to identify vulnerable patients upon diagnosis and guide precision therapeutic choices.

Link to the article: https://www.sciencedirect.com/science/article/pii/S1535610826002539?via%3Dihub

References

Enssle, J. C., Häupl, B., Qoku, A., Wang, B., Wright, G. W., Barrans, S., Zhou, Y., Care, M. A., Burton, C., Gribbin, C., Ziello, J., Weirather, J., Dai, Y., Kizhakeyil, A., Li, X., Phelan, J. D., Kanangat, S., Eckert, S., Scheich, S., … Oellerich, T. (2026). Pathogenesis of diffuse large B cell lymphoma proteogenotypes. Cancer Cell, S1535610826002539. https://doi.org/10.1016/j.ccell.2026.05.008