2026 ASH®
Annual Meeting & Exposition December 12-15, 2026 ▪ New Orleans, Louisiana

Paper Details

Abstracts, authors, affiliations, and disclosures.

Saturday, December 12, 2026
09:30 a.m. - 11:00 a.m.
MCCNO - New Orleans Theatre A
Technology and Procedures, Assays, Molecular Testing, Omics Technologies, Pathology, Computational Biology, Profiling, Artificial Intelligence (AI), Machine Learning, Biological Processes, Genomics, Immune Mechanism, Diseases, Lymphoid Malignancies, Lymphomas, Measurable Residual Disease, Hodgkin Lymphoma, Non-Hodgkin Lymphoma, B Cell Lymphoma, Aggressive Lymphoma, Research, Fundamental Science, Translational Research, Clinical Research, Real-World Evidence, Study Population, Human, Adult, Pediatric, Young Adult
PAPER

70 Defining genetically distinct subtypes of central nervous system lymphoma and their association with pathogenic molecular programs

Paul Mueller, Fabian Hummel2, Jurik Andreas Mutter, PhD3, Roman Sankowski4, Lauritz Miarka5, Julia Allgaier6, Moritz Renner7, Lavanya Ranganathan8, Sabine Bleul9, Christian Klingler10, Lisa Toma11, Louisa von Baumgarten, MD12, Josefine Radke13, Stefan Alig, MD14, Bastian Sajonz15, Nicolas Neidert16, Gerald Illerhaus, MD17, Elisabeth Schorb, MD18, Juergen Beck19, Volker Coenen20, Ash A. Alizadeh, MD, PhD21, Justus Duyster22, Marco Prinz23, Peter Reinacher24, Julia Christina Kuehn, MD25, Florian Scherer, MD, PhD26
1 Medical Center – University of Freiburg, Freiburg, Germany
2 Medical Center – University of Freiburg, Freiburg, Germany
3 Medical Center – University of Freiburg, Freiburg, Germany, Stanford University, Stanford, California, United States
4 Medical Center – University of Freiburg, Faculty of Medicine, Freiburg, Germany
5 Medical Center – University of Freiburg, Faculty of Medicine, Freiburg, Germany
6 Medical Center – University of Freiburg, Freiburg, Germany
7 Medical Center – University of Freiburg, Freiburg, Germany
8 Medical Center – University of Freiburg, Freiburg, Germany
9 Medical Center – University of Freiburg, Freiburg, Germany
10 Medical Center – University of Freiburg, Freiburg, Germany
11 Medical Center – University of Freiburg, Freiburg, Germany
12 University Hospital, LMU Munich, Munich, Germany, University Hospital, LMU Munich, Munich, Germany
13 University Medicine Greifswald, Greifswald, Germany, University Medicine Greifswald, Greifswald, Germany
14 University Hospital Essen, University Duisburg-Essen, West German Cancer Center, Essen, Germany
15 Medical Center – University of Freiburg, Faculty of Medicine, Freiburg, Germany
16 Medical Center – University of Freiburg, Faculty of Medicine, Freiburg, Germany
17 Klinikum Stuttgart, Stuttgart, Germany
18 Medical Center – University of Freiburg, Freiburg, Germany
19 Medical Center – University of Freiburg, Faculty of Medicine, Freiburg, Germany
20 Medical Center – University of Freiburg, Faculty of Medicine, Freiburg, Germany
21 Stanford University, Stanford, California, United States
22 Medical Center – University of Freiburg, Freiburg, Germany
23 Medical Center – University of Freiburg, Faculty of Medicine, Freiburg, Germany
24 Medical Center – University of Freiburg, Faculty of Medicine, Freiburg, Germany
25 Medical Center – University of Freiburg, Freiburg, Germany, Stanford University, Stanford, California, United States
26 Medical Center – University of Freiburg, Freiburg, Germany

Introduction

Central nervous system lymphomas (CNSL) are aggressive extranodal non-Hodgkin B-cell lymphomas confined to the CNS compartment. Despite being perceived as genetically homogeneous with recurrent alterations affecting canonical B-cell receptor (BCR) and NF-kB signaling pathways, CNSL exhibit substantial clinical and biological heterogeneity, suggesting unrecognized underlying molecular diversity. Here, we define genetically distinct CNSL subtypes by unsupervised clustering and characterize their associated pathogenic mechanisms, clinicopathological features, and spatial microenvironmental architecture.

 

Methods

We performed comprehensive genetic profiling of 125 CNSL samples (n=107 PCNSL, n=18 isolated SCNSL) by targeted NGS (CAPP-Seq) and shallow whole-genome sequencing (sWGS). We then applied non-negative matrix factorization (NMF) consensus clustering to a subset of 68 cases (discovery cohort) and developed a novel molecular classifier for subtype assignment, leveraging 74 selected genetic features. This classifier was validated in two independent cohorts to assess its robustness and consistency: an internal validation cohort of the withheld 57 CNSL cases and an external WGS CNSL dataset (n=38; Radke, Nat Comm, 2022). Finally, the CNSL subtypes were further characterized by transcriptomic (RNA-Seq, n=89) and spatial proteomic profiling (Akoya PhenoCycler, n=10, 43-plex), and associated with biological, clinical, and pathological features.

 

Results

The NMF clustering framework identified two molecularly distinct subtypes within the discovery cohort, CNSL-1 (n=32, 47%) and CNSL-2 (n=36, 53%). The subtypes demonstrated divergent genetic profiles, with CNSL-1 being characterized by a significantly higher mutational burden (p=0.003), while CNSL-2 was enriched for copy number alterations (CNA, p=0.001). Furthermore, while both subtypes almost exclusively mapped to the C5 subtype by DLBclass (94%; Chapuy, Blood, 2025), CNSL-1 was significantly enriched for the MCD (78%) and CNSL-2 for the ‘Other’ subtype (67%) according to LymphGen (p=1.2e-5; Wright, Cancer Cell, 2020). Applying our classifier to the two independent validation cohorts assigned 33% and 45% of cases to the CNSL-1 subtype. Importantly, both validation datasets revealed genetic architectures highly consistent with those of the discovery cohort, with CNSL-1 again demonstrating significantly higher mutational burden (p=0.002 and p=2.2e-5) and higher rates of MCD (p=0.0003 and p=5.6e-7).

Next, we further deeply characterized the two subtypes and found fundamentally distinct molecular programs. While CNSL-1 exhibited enriched aberrant somatic hypermutation signatures together with an accumulation of variants in superenhancer regions (p=9e-5) and pathogenic mutations affecting proximal BCR signaling (CD79B) and anti-apoptotic regulators (BCL2, PIM1/2, MCL1), CNSL-2 showed substantial genomic instability with higher frequencies of CDKN2A deletions and BCL6 aberrations. Consistently, transcriptomic analyses revealed significant activation of BCR (p=0.047) and JAK-STAT signaling (p=0.009) in CNSL-1, and downregulation of p53 (p=4e-7) and DNA repair pathways (p=1e-10) in CNSL-2. Furthermore, CNSL-1 was enriched for pre-plasmablast-like B-cell states (60% vs. 32%), whereas CNSL-2 was dominated by pre-memory-lineage states (p=0.03, Steen, Cancer Cell, 2022), consistent with a mutational program favoring plasmablastic differentiation in CNSL-1. Finally, immune profiling by CNSL-focused spatial proteomics demonstrated substantial differences in brain lymphoma microenvironmental compositions.

Clinicopathological and radiological features as well as outcomes were not different between the two subtypes. However, cases displaying plasmablastic differentiation within CNSL-2 showed a particularly unfavorable PFS (p=0.003, HR=3.1) and OS (p=0.02, HR=2.9), while the plasmablast B-cell state did not influence prognosis in CNSL-1. Notably, immunodeficiency-associated (ID-CNSL) and EBV-positive CNSL were predominantly assigned to CNSL-2 (9/11, 82%), suggesting an association between CNA-driven genomic instability and pre-existing immune dysregulation.

 

Conclusion

Collectively, we identified two novel CNSL subtypes that are characterized by distinct genetic, transcriptomic, and spatial architectures. Our findings suggest divergent pathogenic mechanisms underlying CNS lymphomagenesis that may translate into subtype-specific therapeutic vulnerabilities.
Paul Mueller: No relevant conflicts of interest to declare.