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

Paper Details

Abstracts, authors, affiliations, and disclosures.

SESSION

622. Lymphomas: Translational – Non-Genetic: Poster I

Program: Poster Abstract Session
Saturday, December 12, 2026
05:30 p.m. - 07:30 p.m.
MCCNO - Hall F
Technology and Procedures, Assays, Imaging, Molecular Testing, Omics Technologies, Bioinformatics, Cell Expansion, Computational Biology, Emerging Technologies, Profiling, Treatment Considerations, Combination Therapy, Biological Therapies, Bispecific Antibody Therapy, (CAR)-T Cell Therapies, Immunotherapy, Monoclonal Antibody Therapy, Biological Processes, Pathogenesis, Genomics, Immune Mechanism, Microbiome, Molecular Biology, Diseases, Lymphoid Malignancies, Lymphomas, Hodgkin Lymphoma, Non-Hodgkin Lymphoma, B Cell Lymphoma, Indolent Lymphoma, Aggressive Lymphoma, Research, Fundamental Science, Translational Research, Clinical Research, Study Population, Human, Adult, Elderly, Animal Model
PAPER

1769 Cerebrospinal fluid proteomic scores for diagnosis, subtyping, and longitudinal monitoring of primary cns diffuse large B-cell lymphoma

Hannah Rohdjess, Benjamin-Leon Traub, MD2, Katharina Johanna Mueller, MD3, Uwe Warnken, PhD4, Sofia Doubrovinskaia, MD5, Michael Breckwoldt, MD, PhD6, Gina Marie Cebulla, MD7, Melissa Hahn8, Dirk Hoffmann, PhD9, Tim Windheim10, Georg Scheffler11, Brigitte Wildemann, MD12, Robin Wagener, MD, PhD13, Martin Bendszus, MD14, Jorg Dietrich, MD, PhD15, Tobias Kessler, MD16, Michael Platten, MD17, Felix Sahm, MD, PhD18, Louisa von Baumgarten, MD19, Wolfgang Wick, MD20, Leon Kaulen, MD, DSc21
1 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany, University of Heidelberg, Heidelberg, Germany
2 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany
3 Ludwig-Maximilians-University, Munich, Germany
4 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany
5 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany
6 University Hospital Heidelberg, Heidelberg, Germany
7 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany
8 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany, University of Heidelberg, Heidelberg, Germany
9 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany
10 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany
11 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, Ludwig-Maximilians-University, Munich, Germany
12 University Hospital Heidelberg, Heidelberg, Germany
13 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany
14 University Hospital Heidelberg, Heidelberg, Germany
15 Massachusetts General Hospital Cancer Center, Harvard Medical School, Boston, United States
16 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany
17 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, Medical Faculty Mannheim, Mannheim Center for Translation Neuroscience (MCTN), Heidelberg University, Mannheim, Germany
18 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany
19 Ludwig-Maximilians-University, Munich, Germany, Ludwig-Maximilians-University, Munich, Germany
20 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, University Hospital Heidelberg, Heidelberg, Germany
21 German Cancer Research Center (DKFZ) and German Consortium for Translational Cancer Research (DKTK), Heidelberg, Germany, Massachusetts General Hospital Cancer Center, Harvard Medical School, Boston, United States, University Hospital Heidelberg, Heidelberg, Germany

Background: Liquid biopsy using cerebrospinal fluid (CSF) has emerged as a powerful diagnostic tool for primary CNS diffuse large B-cell lymphoma (PCNS-DLBCL). However, current assays largely depend on recurrent genomic alterations present only in a subset of patients, require substantial CSF volumes, and do not capture biological subtypes. We therefore aimed to develop mutation-independent, diagnostic CSF proteomic signatures that reliably discriminate PCNS-DLBCL subtypes and distinguish them from other CNS tumors.

Methods: CSF from treatment-naïve patients with PCNS-DLBCL (Epstein-Barr virus (EBV)-negative, N=48; EBV-positive, N=12) was analyzed using mass-spectrometry-based proteomics from low input (median: 55 µL) samples. Data were integrated with an in-house reference CSF cohort from patients with glioblastoma (N=51), brain metastasis (N=81), and non-malignant controls (N=36). Differential protein expression, pathway enrichment, integrative transcriptomic analyses, and classifier development were performed. Diagnostic performance was assessed using bootstrap resampling, and independently collected prospective samples (N=13). Utility for treatment monitoring was assessed in 48 serial CSF samples from 21 patients.

Results: Dimensionality-reduction (t-SNE) and unsupervised hierarchical clustering of CSF proteomic profiles clearly separated EBV+ from EBV- PCNS-DLBCL (p?<?0.0001), irrespective of clinical, histopathological, and demographic characteristics. Differential abundance analysis identified 284 proteins (DEPs) that distinguished the two subtypes. In EBV+ PCNS-DLBCL patients, CSF was enriched for cytokine signaling pathways (e.g., IL6ST, AXL, CSF1R), whereas EBV- PCNS-DLBCL cases showed higher levels of IGHM and complement-associated proteins (e.g., C1QA, CFB). Pathway and kinase enrichment analyses confirmed cytokine and JAK-STAT signaling in EBV+, and enriched B-cell receptor signaling in EBV- lymphomas.

To determine whether EBV-associated CSF proteomic alterations reflected tumor transcriptional changes, CSF-derived DEPs were compared with tumor-tissue derived differentially expressed genes (DEGs). The tumor transcriptome recapitulated the principal signaling alterations identified in the CSF, including differential JAK-STAT and BCR signaling. Of 88 tumor-derived DEGs represented in the CSF proteomics dataset, 45 showed concordant transcript and protein changes (log2FC Pearson r=0.3, p < 0.05). Concordant DEG-DEP pairs were enriched for membrane-associated proteins, suggesting that shedding or extracellular release of tumor-associated proteins may contribute to the CSF proteomic signatures.

Comparison with brain tumor (N=153) and non-malignant (N=36) reference cohorts, followed by LASSO regression, identified 32 EBV+-associated and 41 EBV--associated proteins that were combined into subtype-specific proteomic scores. These scores robustly discriminated EBV+ (AUC=1.00) and EBV- (AUC=0.99) PCNS-DLBCL from all other brain-tumor and control CSF samples, with sensitivities of 100 % (EBV+) and 93.8?% (EBV-), and specificities?>?99?%. Importantly, scores were independent of clinical parameters, including total CSF protein, cell count, and tumor-to-CSF distance. Internal validation across 1,000 bootstrap resamples demonstrated stable diagnostic discrimination, with mean AUCs of 0.999 and 1.00 for the EBV- and EBV+ scores, respectively. Scores and classification thresholds were locked before evaluation in the independent prospective cohort, which supported the findings of the discovery cohort, with no malignant or non-malignant control sample classified as PCNS-DLBCL.

In serially sampled PCNS-DLBCL, CSF proteomic scores differed significantly across disease states (EBV+, p < 0.001, EBV- p < 0.0001), with highest levels at diagnosis and progressive disease and lower levels in partial and complete responders, thereby tracking radiographic treatment response and disease recurrence.

Conclusions: CSF proteomics from low input CSF samples captures disease-associated programs in PCNS-DLBCL observed in tumor tissue, recapitulating differences between EBV+ and EBV- PCNS-DLBCL. By linking these biological differences to quantitative proteomic scores, our findings establish a framework for minimally invasive PCNS-DLBCL diagnosis, subtyping, and longitudinal disease monitoring. The feasibility to derive CSF proteomic scores from low input samples and archived specimens further supports the translational potential of this approach.

Hannah Rohdjess: No relevant conflicts of interest to declare.