GENOMIC PROFILING AND TARGETED THERAPY IN BRAZILIAN LARGE B-CELL LYMPHOMAS: A HIGH-RISK MOLECULAR LANDSCAPE

Diffuse large B-cell lymphoma (DLBCL) is biologically heterogeneous, with its genomic landscape mapped in high-income countries but scarcely characterized in low- and middle-income countries (LMICs) like Brazil. Distinct demographics, healthcare settings, and ancestry-related backgrounds may influence disease biology, limiting the applicability of existing risk models and targeted therapies. Incorporation of next-generation sequencing (NGS) into routine practice enables precise molecular classification, improved risk stratification, and identification of actionable alterations, directly informing clinical management. To report a molecular characterization of a Brazilian DLBCL cohort and illustrate how NGS-guided decisions can shape therapeutic strategies. We retrospectively analyzed Brazilian patients with DLBCL or high-grade B-cell lymphoma who underwent a targeted NGS for lymphoid malignancies. Mutational data were classified according to the LymphPlex algorithm. Twenty-two patients (median age 66) harbored ≥ 1 molecular alteration. Median tumor mutational burden was 11.4 mut/Mb (range 1.6–191); microsatellite instability was detected in one patient. Frequently mutated genes included TP53 (32%), PIM1 (23%), MYD88 (18%), CD79B (18%), SOCS1 (18%), KMT2D (18%), and CREBBP (18%). Mutations predominantly involved epigenetic/chromatin remodeling (64%), TCR/BCR/NF-κB signaling (59%), and apoptosis/DNA damage pathways (59%). LymphPlex classification identified Others (36%), TP53Mut (32%), MCD-like (14%), EZB-like without MYC rearrangement (14%), and EZB-like with MYC rearrangement (4.5%). High-risk categories (TP53Mut + MCD-like) accounted for 46% of patients. As expected, MCD-like was restricted to non-GCB tumors and EZB-like to GCB tumors, while the heterogeneous “Others” group included recurrent alterations such as TET2, GNA13, NOTCH2, BTK, and CDKN2A deletion. Two patients exemplify the clinical impact of these findings. One, a 60-year-old woman with relapsed/refractory non-GCB DLBCL and CNS involvement, harbored MYD88 (L265P) and CD79B mutations, defining the MCD subtype. Given its known sensitivity to BTK inhibition, acalabrutinib was initiated, inducing a rapid metabolic complete remission and enabling consolidation with anti-CD19 CAR-T therapy; she remains disease-free after two years. The other, a 50-year-old woman with CNS-involved DLBCL ineligible for ASCT, was also classified as MCD subtype and received maintenance ibrutinib, achieving and sustaining complete remission for 1.5 years. This Brazilian DLBCL cohort shares key oncogenic drivers with global series but shows a notable enrichment for high-risk subtypes, particularly TP53Mut and MCD-like, which are associated with poor prognosis. The case examples highlight how NGS not only refines molecular classification but also directly informs targeted therapy selection—enabling durable remissions in patients with advanced, refractory disease who otherwise have limited options. In conclusion, targeted NGS in DLBCL offers dual value: generating population-level insights that reveal a high prevalence of aggressive molecular subtypes in Brazil, and providing actionable data that can refine treatment selection and improve outcomes in individual patients. These findings underscore the need for robust local genomic datasets to guide precision oncology in LMIC settings.