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

Paper Details

Abstracts, authors, affiliations, and disclosures.

SESSION

702. CAR-T Cell Therapies: Basic and Translational: Poster III

Program: Poster Abstract Session
Monday, December 14, 2026
06:00 p.m. - 08:00 p.m.
MCCNO - Hall F
Technology and Procedures, Assays, Imaging, Molecular Testing, Omics Technologies, Bioinformatics, Cell Expansion, Computational Biology, Gene Editing, Profiling, Treatment Considerations, Combination Therapy, Drug Development, Biological Therapies, Antibody Therapy, Bispecific Antibody Therapy, Chimeric Antigen Receptor, (CAR)-T Cell Therapies, Gene Therapy, Immunotherapy, Non-Biological Therapies, Radiation Therapy, Adverse Events, Artificial Intelligence (AI), Deep Learning, Biological Processes, Genomics, Hematopoiesis, Immune Mechanism, Microbiome, Molecular Biology, Diseases, Immunology, Autoimmune Disorders, Immunodeficiency, Infectious Diseases, Viral, Lymphoid Malignancies, Lymphoid Leukemias, Lymphomas, Plasma Cell Disorders, Myeloid Malignancies, Acute Myeloid Malignancies, Chronic Myeloid Malignancies, ALL, Non-Hodgkin Lymphoma, B Cell Lymphoma, Aggressive Lymphoma, Research, Fundamental Science, AML, Translational Research, CML, Clinical Research, Clinical Trials, Epidemiology, Health Outcomes Research, Health Disparities Research, Real-World Evidence, Study Population, Human, Adult, Elderly, Young Adult, Animal Model
PAPER

5865 Gut microbiome dynamics shape CAR T-cell outcomes in relapsed/refractory large B-cell lymphoma

Beatrice Casadei, MD, PhD, Alice Bernardi2, Gabriele Conti3, Cinzia Pellegrini, MD, PhD4, Francesca Bonifazi5, Silvia Turroni6, Claudia Romano7, Stella Blandino8, Marianna Gentilini, PhD9, Valentina Idda10, Giulia Gabrielli, MD11, Andrea Farolfi12, Marcello Roberto13, Francesco De Felice14, Margherita Ursi15, Alessandro Desiderio16, Giorgia Rossetti17, Enrico Maffini18, Patrizia Brigidi19, Pier Luigi Zinzani, MD, PhD20
1 IRCCS Azienda Ospedaliero-Universitaria di Bologna, Istituto di Ematologia “Seràgnoli', Bologna, Italy
2 University of Bologna, Bologna, Italy
3 University of Bologna, Bologna, Italy
4 IRCCS Azienda Ospedaliero-Universitaria di Bologna, Istituto di Ematologia “Seràgnoli', Bologna, Italy
5 IRCCS Azienda Ospedaliero-Universitaria di Bologna, Bologna Italy Università di Bologna, Bologna, Italy
6 University of Bologna, Bologna, Italy
7 University of Bologna, Bologna, Italy
8 IRCCS Azienda Ospedaliero-Universitaria di Bologna, Istituto di Ematologia “Seràgnoli', Bologna, Italy
9 University of Bologna, Bologna, Italy
10 University of Bologna, Bologna, Italy
11 University of Bologna, Bologna, Italy
12 IRCCS Azienda Ospedaliero-Universitaria di Bologna, Bologna, Italy
13 University of Bologna, Bologna, Italy
14 IRCCS Azienda Ospedaliero-Universitaria di Bologna, Bologna, Italy
15 IRCCS Azienda Ospedaliero-Universitaria di Bologna, Bologna, Italy
16 University of Bologna, Bologna, Italy
17 University of Bologna, Bologna, Italy
18 IRCCS Azienda Ospedaliero-Universitaria di Bologna, Bologna, Italy
19 University of Bologna, Bologna, Italy
20 University of Bologna, Bologna, Italy

CD19-directed CAR T-cell (CART) therapy has substantially improved the prognosis of relapsed/refractory large B-cell lymphoma (rrLBCL), although durable responses remain limited to a subset of patients. Host-related determinants of CART efficacy are poorly understood. Given the pivotal role of the gut microbiome (GM) in regulating systemic immunity, we prospectively investigated the longitudinal relationship between GM composition and functionality, metabolic tumor burden, and clinical outcomes during CART journey.

In this real-life, single-center study (AIRC IG2022-27350; NCT05725720), 59 consecutive rrLBCL patients treated with CART (55 axicabtagene ciloleucel, 4 lisocabtagene maraleucel) underwent fecal sampling from leukapheresis to post-treatment follow-up. GM was characterized by 16S rRNA amplicon sequencing and shotgun metagenomics. Response was assessed by FDG-PET/CT according to the Lugano classification (Deauville 5-point score). FDG-PET/CT was also analyzed using semiquantitative parameters as maximum standardized uptake value (SUVmax), metabolic tumor volume (MTV) and total lesion glycolysis (TLG). Feature selection was performed using the Boruta random-forest algorithm, followed by Firth logistic regression. Univariate Cox model associations with progression-free survival (PFS), overall survival (OS), disease-free survival (DFS), relapse (RD) and time to next treatment (TTNT) were investigated.

Median age was 58 years (range, 29–74); 76% of patients had stage III-IV disease, 93% were refractory to their last therapy, and 95% received bridging treatment. Best overall response rate was 76%, median PFS was 6.2 months, while median OS was not reached after a median follow-up of 18.2 months.

Boruta feature selection consistently identified microbiome taxa together with PET-derived parameters among the variables associated with response and survival endpoints. SUVmax, MTV, TLG and Eggerthella relative abundance were significantly associated with inferior PFS, OS, RD and TTNT. Among these parameters, MTV emerged as the most consistent prognostic marker across clinical endpoints.

GM longitudinal characterization showed that  alpha diversity remained stable throughout treatment in responders, whereas it progressively decreased in non-responders. Beta diversity analysis further demonstrated significant differences between the two groups over time. The GM compositional and functional structure was more stable in responders, while it fluctuated considerably in non-responders. Taxonomic profiling revealed a consistent presence of commensal taxa in responders, including short-chain fatty acid (SCFA) producers genera Ruminococcus, Monoglobus and Christensenellaceae R-7 group, and species Faecalibacterium prausnitzii, Anaerostipes hadrus and Blautia wexlerae. These taxa progressively decreased in non-responders, who instead showed an increase in species associated with dysbiosis and intestinal inflammation: Enterocloster clostridioformis and Enterocloster bolteae. Functional metagenomics revealed an enrichment of pathways involved in respiratory metabolism, ubiquinone biosynthesis, nitrate reduction and lipopolysaccharide biosynthesis, along with a decrease in purine/pyrimidine biosynthetis and reductive acetyl-CoA pathways in non-responders. This suggests a metabolically impaired, pro-inflammatory GM ecosystem that is less capable of providing essential services, such as anaerobic fermentation and SCFA production. Notably, resistome profiling revealed a trend toward broader antibiotic resistance potential in non-responders before chemotherapy, further supporting the presence of a less favorable GM configuration.

This study demonstrates that CART response in rrLBCL is associated with distinct GM signatures. Responders exhibited a greater ecological stability, preservation of commensal taxa and functions involved in GM-host interactions, and a lower burden of antibiotic resistance genes before treatment. GM profiling may improve prognostic discrimination by providing complementary biological information beyond conventional imaging. To our knowledge, this is the first prospective longitudinal study integrating GM taxonomic and functional profiling, FDG-PET biomarkers and machine-learning approaches to characterize the biological determinants of CART efficacy in rrLBCL. These findings support GM-informed risk stratification and future GM-modulating interventions to optimize CART therapy.

Beatrice Casadei, MD, PhD: Roche - Consultancy (Includes expert testimony); Takeda - Consultancy (Includes expert testimony); Abbvie - Consultancy (Includes expert testimony); Novartis - Consultancy (Includes expert testimony); Janssen - Consultancy (Includes expert testimony); Kite-Gilead - Consultancy (Includes expert testimony); Celgene-BMS - Consultancy (Includes expert testimony)