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

Paper Details

Abstracts, authors, affiliations, and disclosures.

SESSION

621. Lymphomas: Translational: Molecular and 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, Molecular Testing, Omics Technologies, Pathology, Bioinformatics, Emerging Technologies, Gene Editing, Profiling, Treatment Considerations, Combination Therapy, Drug Development, Non-Biological Therapies, Pharmacology, Biological Processes, Apoptosis, Pathogenesis, CHIP, Genomics, Hematopoiesis, Immune Mechanism, Molecular Biology, Diseases, Lymphoid Malignancies, Lymphomas, Plasma Cell Disorders, Hodgkin Lymphoma, Non-Hodgkin Lymphoma, B Cell Lymphoma, T Cell Lymphoma, Indolent Lymphoma, Aggressive Lymphoma, Research, Fundamental Science, Translational Research, Clinical Research, Clinical Trials, Real-World Evidence, Study Population, Human, Adult, Elderly, Animal Model
PAPER

1742 Prognostic association of cell-of-origin by digital gene expression profiling in relapsed/refractory large B-cell lymphoma: A biomarker-integrated cohort study

Inna Y Gong, MD, PhD, Michael Hong2, Ting Liu, PhD3, Peter Sabatini4, Tong Zhang5, David Scott, MD6, Tamar Uziel7, Wissam Assaily, PhD8, Tomohiro Aoki, MD, PhD9, Sita Bhella, MD10, Vishal Kukreti, MD11, Anca Prica12, Abi Vijenthira, MD, MS13, Chloe Yang, MD14, Michael Crump, MD15, John Kuruvilla, MD16, Robert Kridel, MD17
1 Princess Margaret Cancer Centre, Toronto, Ontario, Canada, University of Toronto, toronto, Ontario, Canada
2 Princess Margaret Cancer Centre, Toronto, Ontario, Canada
3 Princess Margaret Cancer Centre, Toronto, Ontario, Canada
4 Centre for Lymphoid Cancer, Vancouver, Ontario, Canada
5 Princess Margaret Cancer Centre, toronto, Ontario, Canada
6 BC Cancer and Division of Medical Oncology, Vancouver, British Columbia, Canada
7 Abbvie Inc., North Chicago, United States
8 Abbvie Inc., North Chicago, United States
9 Princess Margaret Cancer Centre, Toronto, Ontario, Canada, University of Toronto, toronto, Ontario, Canada
10 Princess Margaret Cancer Centre, Toronto, Ontario, Canada, University of Toronto, toronto, Ontario, Canada
11 Princess Margaret Cancer Centre, Toronto, Ontario, Canada, University of Toronto, toronto, Ontario, Canada
12 Princess Margaret Cancer Centre, Toronto, Ontario, Canada, University of Toronto, toronto, Ontario, Canada
13 Princess Margaret Cancer Centre, Toronto, Ontario, Canada, University of Toronto, toronto, Ontario, Canada
14 Princess Margaret Cancer Centre, Toronto, Ontario, Canada, University of Toronto, toronto, Ontario, Canada
15 Princess Margaret Cancer Centre, Toronto, Ontario, Canada, University of Toronto, toronto, Ontario, Canada
16 Princess Margaret Cancer Centre, Toronto, Ontario, Canada, University of Toronto, toronto, Ontario, Canada
17 Princess Margaret Cancer Centre, Toronto, Ontario, Canada, University of Toronto, toronto, Ontario, Canada

Background:

Cell-of-origin (COO) classification by gene expression profiling (GEP) is a robust prognostic marker in newly diagnosed large B-cell lymphoma (LBCL), but studies in the relapsed/refractory (R/R) setting have yielded conflicting results. Whether integrating molecular COO data with established clinical risk factors could improve outcome prediction in R/R LBCL treated in the modern era of cellular immunotherapy requires further study.

 

Methods:

Sequential patients with R/R LBCL prospectively treated at Princess Margaret Cancer Centre (2018-2025) with available FFPE relapse tissue were included. COO was classified by digital GEP using the NanoString LExA120 assay (n=116) and the Hans IHC algorithm (n=101); the dark zone signature (DZsig) was also derived from the LExA120 GEP data. Somatic mutation profiling was performed using LySeq v2, a sequencing panel covering 120 lymphoma-relevant genes (n=115). In patients with available bulk RNA sequencing data (n=85), exploratory tumor microenvironment (TME) classification was performed using three published transcriptomic classiers: EcoTyper, LymphoMAP, and the Lymphoma Microenvironment (LME). Univariable and multivariable Cox proportional hazards models evaluated associations between each biomarker platform and time to treatment failure (TTF, time from index date to stable or progression disease [SD/PD] or relapse, or death from any cause) and overall survival (OS), adjusting for IPI and primary refractory disease (PRD defined SD/PD at interim or end-of-treatment response). An integrated risk score was developed combining IPI 3-5, PRD, and ABC COO (one point each), stratifying patients into low (0 points), intermediate (1 point), and high risk (≥2 points). Among patients with a subsequent relapse, a further multivariable analysis evaluated the independent association of COO with OS, adjusting for IPI, PRD, and type of treatment received.

 

Results:

Of 116 patients, median age was 63 years (IQR 55-68); 67% male; 57% GCB, 27% ABC, 16% unclassified by digital GEP. Second-line treatment included ASCT-intent (n=85), CAR T (n=11), and other (n=20); 36 patients (31%) received CAR T at any time. In line with known biology, mutational profiles segregated by COO, with BCR/NF-κB pathway mutations (MYD88, ETV6, SETD1B, PIM1, OSBPL10, CD79B, IRF4, and NFKBIZ) enriched in ABC, and germinal centre mutations (CREBBP, BCL2, KMT2D, TP53, TNFRSF14, EZH2, and SGK1) enriched in GCB.

 

ABC COO was independently associated with inferior TTF in multivariable analysis (HR 2.98, 95% CI 1.38-6.44; p=0.006), and a directionally consistent trend for OS (HR 2.15, 95% CI 0.93-4.99; p=0.075). In contrast, Hans-classified COO, DZsig, and DH/TH status were not independently prognostic. The integrated risk score significantly improved TTF discrimination over clinical factors alone (C-statistic 0.69 vs 0.65; likelihood-ratio test p=0.009). Among digital GEP classified patients, 19 (24%) were low-risk, 36 (47%) intermediate-risk, and 22 (28%) high-risk. High-risk patients had markedly inferior TTF (HR 6.44, 95% CI 2.26-18.34; p<0.001) and OS (HR 10.60, 95% CI 3.15-35.71; p<0.001) compared to low-risk patients. In exploratory TME analyses, EcoTyper LE1-LE4 state was independently associated with inferior TTF and OS in multivariable analysis, while LymphoMAP and LME subtypes were not independently prognostic.

 

In patients with subsequent relapse, CAR T therapy was consistently associated with superior OS across all models, and ABC COO independently predicted inferior OS at time of subsequent progression (HR 4.25, 95% CI 1.20-15.03; p=0.025).

 

Conclusions: Digital GEP-based COO classification using relapse tissue independently refines risk stratification in R/R LBCL beyond IPI and PRD, retaining adverse prognostic significance across treatment lines. Integration with IPI and PRD into a simple three-factor score identifies a high-risk group comprising approximately one-quarter of patients with an approximately 11-fold higher hazard of death compared to low-risk patients. The lack of independent prognostic association with Hans IHC-based COO underscores the importance of GEP-based classification for accurate risk stratification in R/R LBCL.

Inna Y Gong, MD, PhD: Kite/Gilead - Honoraria