Uncovering the genomic landscape of DLBCL in Latin America: First report of two centers from Brazil and Chile

Abstract

Abstract Background: DLBCL is a biologically heterogeneous disease whose genomic landscape has been extensively mapped in high-income countries. However, these data are scarce from low- and middle-income countries (LMICs) like Brazil and Chile, where distinct demographics, healthcare settings and ancestry-related genetic backgrounds may influence disease biology. This knowledge gap limits the global applicability of existing risk models and targeted therapies. This study provides the first molecular characterization of a latin-american DLBCL cohort using NGS to compare its mutational landscape with established international profiles. Methods: We retrospectively analyzed Brazilian and Chilean patients from two centers (Einstein Hospital Israelita [EHA] and Clinica Alemana de Santiago [CAS]) diagnosed with aggressive B-cell lymphomas who underwent a targeted NGS panel. Molecular profiling was performed using either the TSO500 HT panel on NextSeq 500 or NovaSeq 6000 platforms (Illumina) at EHA or the Oncomine™ Comprehensive Assay Plus on the Ion GeneStudio™ S5 Plus platform (Thermo Fisher Scientific) at CAS. Patients were classified into molecular subtypes according to the LymphPlex classification system. Results: All 30 patients harbored at least one molecular alteration. The median tumor mutational burden (TMB) was 11.4 mut/Mb (range: 1.6 – 191), and microsatellite instability (MSI) was identified in three patients (10%). The most frequently mutated genes included TP53 (40%), PIM1 (20%), KMT2D (20%), CREBBP (17%), MYD88 (13%), CD79B (13%) and SOCS1 (13%). Pathway annotation revealed a predominance of mutations in apoptosis/DNA damage/cell cycle pathways (60%), followed by epigenetic/chromatin remodeling genes (57%), alterations in TCR/BCR/NF-κB signaling (43%) and other signaling (47%). Additional pathway involvement included transcription factor/differentiation (30%), immune response/cell trafficking (23%), and JAK/STAT signaling (13%). LymphPlex classification identified the following subtypes: TP53Mut (40%), Others (37%), MCD-like (10%), EZB-like without MYC rearrangement (10%), and EZB-like with MYC rearrangement (3%). Among GCB cases, the most frequent LymphPlex categories were TP53Mut (56%), EZB-like (19%) and Others (19%). In contrast, non-GCB cases were enriched for MCD-like (21%) and Others (57%), with TP53Mut observed in 21%. EZB-like subtypes were exclusive to GCB, and MCD-like was restricted to non-GCB. TP53Mut was enriched in GCB cases (75%), consistent with prior reports, while the Others category was predominantly non-GCB (72%). Subtype-specific profiles showed that EZB-like with MYC rearrangement (n=1) harbored co-occurring mutations in ID3, KMT2D, SOCS1, FUBP1 and CREBBP. EZB-like without MYC rearrangement was marked by TNFRSF14 (67%) and a broader spectrum including EZH2, PAX5, BCL10, ARID1A, SOCS1, and others (33%). MCD-like cases featured MYD88 and PIM1 (100%), with CD79B and BTG1 in 67% of patients. TP53Mut cases consistently had TP53 mutations (100%), often accompanied by TNFAIP3, KMT2D, RB1, PTEN, and others. Others (n=11) showed a heterogeneous profile, most commonly involving TET2, GNA13, NOTCH2, BTK, CCND3 and CDKN2A deletion. Conclusion: The molecular landscape of this latin-american DLBCL cohort, while sharing fundamental oncogenic drivers with global populations, is distinctly characterized by a significant enrichment of high-risk genetic features. The high prevalence of TP53 mutations and the aggressive MCD genetic subtype suggests this patient population may be skewed towards more aggressive disease biology. These findings highlight that risk-stratification models developed in other global populations may not fully capture the genetic risk profile of latin-american patients, underscoring the critical importance of generating robust local genomic data to advance precision oncology and improve the clinical management of DLBCL in LMICs.