Citations

Abstract Introduction: CAR-T therapy has transformed cancer therapy. with its first Food and Drug Administration (FDA) approved product in August 2017. Given the expanding indications and growing experience, particularly with management of immediate toxicities requiring specialized expertise such as Cytokine Release Syndrome (CRS) and Immune Effector Cell-Associated Neurotoxicity syndrome (ICANS), FDA recently removed the Risk Evaluation and Mitigation Strategies (REMS) for CD19 and BCMA targeting CAR-T on June 26th2025. In particular, given that CRS and ICANS typically occur in the first 2 weeks post CAR-T infusion, the minimum required stay near CAR-T treatment center is reduced from one month to 2 weeks. While this could reduce treatment burden for patients (pts) who need to travel away from home to receive CAR-T, there is a need to update the transition of care from CAR-T center to pts' local oncology practice. Pts carry in their rate of recovery and have varying supportive care needs in the second half of the month post CAR-T infusion. To assist with care transition and planning, we examine the use of AI multimodal large language model (LLM) to predict patients' supportive care needs 2 weeks after CAR-T infusion based on the clinical data at the time of CAR-T evaluation, prior to leukapheresis. This timepoint was selected to also maximize travel, lodging and caregiver planning for pts and family as well. Methodology: Pts who were treated with FDA-approved CAR-T at Mayo Clinic Rochester between December 2022 and June 2025 were examined. Based on experiences with referral practices in the Midwest region, we defined patients with high monitoring needs as those who between day 14-30 post CAR-T had either hospital admission due to recurrent CRS or ICANS or required transfusion of more than 1 blood product per week (red blood cell or platelets). Others were classified as low monitoring needs. Clinical notes, labs, vital signs and electrocardiogram at the time of CAR-T evaluation were abstracted and summarized by Gemini LLM for association with patients with high versus low monitoring needs post CAR-T. Model was developed and tested for predicted needs and concordance with actual outcome of these pts. Results: Among the 362 pts who received FDA approved CAR-T in the timeframe, 84 pts had high monitoring needs. An additional 84 pts with low monitoring needs were randomly selected from 278 pts. These were divided into a training cohort and test cohort of 84 pts each, with 42 pts with high and 42 pts with low monitoring needs. Demographics of the two cohorts displayed no statistical difference. LLM analysis of the clinical data at the time of CAR-T eval identified 5 categories associated with risk for high monitoring needs: 1) Disease status (clinical notes about bulky, progressive disease or CNS involvement); 2) inflammatory and tumor burden markers (CRP, LDH); 3) hematologic status (platelet, plt; neutrophil count ANC); 4) renal function; 5) performance status. Using common clinical lab parameter convention (platelets <100, ANC< 1, CrCl< 30ml/min), and prioritizing for not missing pts with high monitoring needs, LLM model was defined to predict patients for high monitoring needs if high risk feature was identified in category 1 or any category combinations; whereas patients with no high risk feature or only 1 category that is not disease status was predicted as low monitoring needs. In this model, the sensitivity was 85.7%, specificity 23.8%, F1 score 65.5%. Applying this model to test cohort 2, the sensitivity was 83.3%, specificity 28.6%, and F1 score 65.4%. Discussion: We demonstrate proof-of-concept that LLM can be used to analyze clinical data at the time of evaluation for CAR-T eligibility to predict for supportive care needs post treatment. Not surprisingly, features identified by LLM to be associated with high monitoring needs are those previously reported by us and others at the time of CAR-T evaluation and or pre-lymphodepletion to be associated with risk for cytopenia, infection, hospitalization and non-relapse mortality. Based on the summary output generated by LLM, additional opportunities for classification methods were identified to improve specificity and balanced accuracy of the LLM model. This is a promising tool for further development to assist clinicians and patients to plan their treatment journey.

Abstract Chimeric antigen receptor T (CART) cell therapy has shown unprecedented success in the treatment of hematological malignancies. However, the immunosuppressive tumor microenvironment of both solid and liquid tumors is a major barrier for CART efficacy. Immunosuppressive myeloid cells inhibit CART efficacy. Colony-stimulating factor 1 receptor (CSF1R) promotes macrophage proliferation, survival, and anti-inflammatory response and is upregulated on myeloid cells in many cancers, making CSF1R an appealing therapeutic target. We aimed to test the interactions between CSF1R, myeloid cells, and CART response, and to study CSF1R targeting in the context of CART cell therapy. We utilized CD19-targeted CART cell therapy (CART19) in lymphoma preclinical models and thyroid-stimulating hormone receptor (TSHR)-targeting CART cell therapy (TSHR-CART) in thyroid cancer models. We first measured CSF1R expression on myeloid cells in vitro. Suppressive myeloid cells, polarized from healthy donor monocytes using an anaplastic thyroid cancer (ATC) cell line, 8505c, or mantle cell lymphoma cell line, JeKo-1, significantly upregulate CSF1R (p=0.0002) compared to media control. After polarization, myeloid cells significantly inhibit TSHR-CART and CART19 antigen-specific proliferation compared to freshly isolated monocytes (p=0.012, p<0.01, respectively) indicating that suppressive myeloid cells contribute to CART failure in multiple tumor types. We then performed single cell RNA sequencing on peripheral blood taken at baseline from patients with mantle cell lymphoma treated with CART19. We found a significant increase of CSF1R on monocyte subsets in CART19 non-responders vs responders (padj=0.0037). Next, we determined the impact of CSF1R blockade on CART cell therapy in vivo. Using lymphoma models, B6 mice were engrafted with the luciferase+ CD19+ murine lymphoma cell line TBL-12 (1x106 cells i.v. 1 day after 250 mg/kg cyclophosphamide). Tumor burden was significantly reduced after 7 days of treatment with low dose CART19 (0.5x106 i.v.) + anti-mouse CSF1R antibody (16mg/kg i.p. twice weekly) vs CART19 + IgG control (p=0.0146). We then tested the impact of CSF1R blockade in mice experiencing tumor relapse. After TBL-12 engraftment, mice were treated with anti-mouse CSF1R or IgG control antibody followed two days later with high dose CART19 (1x106 i.v.) to induce remission. Mice were re-challenged with TBL-12 14 days later. Mice treated with anti-mouse CSF1R antibody showed increased tumor control and significantly higher survival (p=0.0416). Using thyroid cancer models, first we identified high macrophage infiltration in biopsy specimens of patients with ATC compared to normal thyroid tissue and benign thyroid tumors through immunohistochemistry (IHC) of patient samples (p<0.01, p<0.0001, respectively). To test if macrophages are associated with CART failure in ATC, NSG mice were subcutaneously engrafted with 1x106 TSHR+ THJ529 tumors, an ATC cell line derived from patient biopsy samples from the Mayo Clinic thyroid cancer biobank, followed by i.v. treatment with 1x107 TSHR-CART or untransduced T cell control (UTD). IHC staining of tumor tissue collected at experimental endpoints revealed a dose-dependent, significant infiltration of human CD3+ T cells into the tumors which inversely correlated with murine F4/80+ myeloid cell and murine YM-1+ M2 macrophage infiltration. This led us to hypothesize that myeloid cell targeting with CSF1R blockade can improve TSHR-CART efficacy. We utilized the same ATC model described above to test CSF1R targeting. When TSHR+ THJ529 tumors reached approximately 100 mm3, mice were randomized by tumor volume and treated with CSF1R inhibitor PLX5622 (600mg/kg) or vehicle control, followed by either UTD or TSHR-CART cells 5x106 i.v. Mice treated with TSHR-CART + PLX5622 showed significantly improved tumor control vs TSHR-CART + vehicle control (p<0.0001) or UTD + PLX5622 (p<0.0001). Together, our results highlight immunosuppressive myeloid cells which express CSF1R as a major cell type limiting CART efficacy in both solid and liquid tumors. We also identified a strategy to overcome myeloid-mediated suppression of CART cells by targeting CSF1R with antibody or small molecule inhibitors.

Abstract Background Chimeric antigen receptor T cell (CART) therapy has revolutionized treatment for hematologic malignancies. However, durable remission following anti-CD19 CART therapy (CART19) remains limited to 40–50% of patients, often due to tumor relapse and CART dysfunction (PMID:31501612). We previously demonstrated that tumor-derived extracellular vesicles (EVs) from patients with chronic lymphocytic leukemia (CLL) can directly impair CART function (PMID:33388419). In this study, we aimed to identify tumor-derived EV microRNA (miRNA) cargo associated with clinical CART failure in B cell malignancies and define how they mechanistically drive CART cell dysfunction. Methods To achieve these aims, we profiled tumor-derived EV miRNA cargo using small RNA sequencing (RNAseq) in two independent patient cohorts. Cohort 1 compared plasma-derived EVs from untreated CLL patients to those from healthy donor EVs to identify miRNAs enriched in CLL (n=23). Cohort 2 included patients with diffuse large B cell lymphoma (DLBCL) treated with axicabtagene ciloleucel (CART19, n=12). Here, plasma EV miRNAs were analyzed at baseline (pre-CART treatment) and 1-month post-CART treatment and compared between CART responders (complete remission at 6 months) and nonresponders (no response at 1 month). Enriched miRNAs from both cohorts were identified following miRNA count normalization and differential expression analysis using adjusted p values of ≤0.05. Functional studies were conducted by transfecting CART19-28ζ cells with synthetic miRNA mimics or scrambled controls. CART cells were also generated with stable overexpression of miRNA via lentiviral shRNA delivery. CART function was assessed via in vitro killing, proliferation, apoptosis assays, RNAseq, and in vivo tumor xenograft models. Results In Cohort 1, we identified several miRNAs including miR-125a to be significantly enriched in CLL-derived EVs relative to healthy controls. Similarly, in Cohort 2, miRNA-125a was also identified among the miRNAs enriched in nonresponders vs. responders to CART treatment; Notably, miR-125a was significantly enriched in nonresponders at baseline (pre-CART treatment) and 1-month post-CART treatment (p=0.0177, p=0.0196). To define the functional effects of miR-125a, we delivered synthetic mimics or stable shRNA constructs encoding miR-125a into CART cells. Delivery of miR-125a was confirmed by RT-qPCR (p<0.0001 for mimic vs. scramble, p=0.0069 for shRNA vs. scramble). Here, miR-125a delivery and overexpression impaired CART cytotoxicity/killing of target tumor cells (~25% reduction, p= 0.0039), decreased antigen-specific proliferation (~47% reduction, p=0.0002), and increased apoptosis following antigen stimulation (1.8-fold increase, p=0.0002) as compared to scramble controls in vitro. We next studied how miR-125a impacts CART functions in vivo. Here, tumor-bearing mice treated with miR-125a–overexpressing CART cells (CART sh125a) exhibited accelerated tumor progression (4.1-fold increase, p=0.0019) and significantly reduced survival as compared to mice treated with scramble control CART treatment (CART shScr) (20% vs 80% survival at day 35, p=0.0064), underscoring the dramatic detrimental impact of miR-125a on therapeutic efficacy. To uncover changes in CART gene expression induced by miR-125a delivery, RNAseq was also performed across stimulated CART sh125a vs. CART shScr cells. CART sh125a cells demonstrated broad transcriptional changes, with gene set enrichment of E2F targets, G2/M checkpoint, and apoptosis pathways (FDR p<0.0001, p<0.0001, and p=0.006) further validating the role of miR-125a in disrupting CART cell cycle control, proliferation, and survival. Conclusion In summary, our study identifies miR-125a as a tumor-derived EV cargo that drives CART dysfunction in two separate B cell malignancies (CLL and DLBCL). miR-125a was enriched in both CLL EVs and nonresponders to CART19 therapy at baseline prior and post CART treatment. Functional studies demonstrate that miR-125a delivery and overexpression directly impair CART cytotoxicity and proliferation, increase CART apoptosis, reduce therapeutic efficacy in tumor xenograft models, and disrupt key pathways related to CART cell cycle regulation and survival. These findings position miR-125a as both a key mechanistic driver of CART failure and a potential predictive biomarker for treatment response, while highlighting it as a possible therapeutic target to improve CART outcomes in the clinic.

Abstract Background Nodal peripheral T-cell lymphomas (nPTCLs) are a challenging group of non-Hodgkin lymphomas that frequently recur after or are refractory to first line (1L) treatment. Chemotherapy options in the relapsed or refractory (R/R) setting are limited. The incorporation of non-cytotoxic agents including antibody drug conjugates, immune checkpoint inhibitors, epigenetic modifiers, and phosphatidylinositol 3-kinase inhibitors offers new therapeutic potential and possibly better outcomes. We report on outcomes for patients (pts) with R/R nPTCL enrolled in the multi-center LEO-MER prospective cohort study (NCT02736357) who received second line (2L) therapy. Methods Six hundred and three adult pts with newly diagnosed nPTCL were prospectively enrolled in the University of Iowa/Mayo Clinic MER observational cohort (2002-2015) or the expanded LEO cohort (2015-2020). Secondary overall survival (sOS) was calculated from date of 2L treatment initiation to date of death or last follow-up. Secondary event-free survival (sEFS) was calculated from date of 2Ltreatment initiation to disease progression, initiation of 3rd line therapy, or death from any cause. sEFS and sOS were evaluated using Cox model and Kaplan-Meier estimator. Log-rank test was used to test the significance of difference between groups. Pts with concomitant B-cell lymphomas were excluded from this analysis. Results Thus far, 169 pts with nPTCL initiating 2L therapy were evaluated. Of these, 75 (44.4%) had PTCL-NOS, 63 (37.3%) had angioimmunoblastic T-cell lymphoma (AITL, inclusive of nodal T-follicular helper cell lymphoma), 20 (11.8%) had ALK-negative anaplastic large cell lymphoma (ALCL), and 11 (6.5%) had ALK-positive ALCL. At 2L treatment initiation, the median age was 64 with 86.1% of pts having advanced stage disease. Of the 108 pts with available international prognostic index (IPI) scoring, 46.3% had an IPI of ≥3. Thirty-four (20.1%) pts had received a consolidative autologous stem cell transplant (ASCT) in the 1L. A total of 86 pts (50.1%) had primary refractory disease (partial response, stable or progressive disease as best response to 1L therapy), 45 (26.6%) experienced early relapsing disease within 6 months of 1L therapy, and 38 (22.5%) had relapsed disease >6 months from 1L therapy. After a median follow-up of 7.1 years (yrs), median sOS was 1.52 yrs, and sEFS was 3.2 months (mos). There was no difference in sOS or sEFS regardless of timing of relapse. In keeping with other studies, ALK-positive ALCL pts had the best 5-year OS (70.1%) compared to other subgroups (ALK-negative ALCL 45.0%, AITL 30.4%; PTCL-NOS 21.3%). Lymphoma was the leading cause of death following 2L therapy, with a 2-year cumulative incidence of 49.2%. Eighty pts (47.3%) received only cytotoxic chemotherapy as second line (2L) therapy; ICE (41 (51.3%)) and gemcitabine-containing regimens were most common (14 (17.5%)). Seventy-five pts (44.4%) were treated with non-cytotoxic agents given as monotherapy or in combination with chemotherapy. The most common non-cytotoxic therapies were brentuximab vedotin (33 (44.0%)), romidepsin (17 (22.7%)) and immune checkpoint inhibitors (9 (12.0%)). Fourteen pts (8.3%) underwent palliative interventions involving radiation, steroids, or cyclosporine alone. Intent to proceed to transplant in the 2L was known in 158 pts (101 (63.9%) not intended, 57 (36.1% intended)). Following 2L treatment, 18 (10.7%) pts proceeded to consolidative ASCT and 10 (5.9%) to allogeneic transplant. Only 21 (12.9%) were treated on a clinical trial. Pts who received a non-cytotoxic agent had superior outcomes compared to pts receiving only chemotherapy; the median sOS was 30.4 mos (95% confidence interval [CI] 15.3, 53.52 mos) vs 11.7 mos ((95% CI 7.95,19.58); p=0.04; HR 0.67) and the median sEFS was 4.83 mos (95% CI 3.38, 10.09) vs 2.33 mos ((95% CI 1.71,3.02); p=0.02; HR=0.65). Conclusions We report here one of the largest prospective cohorts in R/R nPTCL. The LEO-MER cohort, despite its association with academic centers, shows poor outcomes regardless of refractory or relapsed disease, frequent inability to proceed to intended consolidative transplant, and low clinical trial participation in 2L, which corroborates findings from other registries (Jain 2025, Stuver 2019). Outcomes improved for pts treated with non-cytotoxic agents. Updated data including additional pts and subgroup analyses will be reported at the meeting.

Abstract Introduction: Elranatamab (Elra) is a bispecific antibody (BsAb) targeting BCMA that is approved for patients with triple class exposed relapsed refractory myeloma (RRMM) on the basis of the MagnetisMM-3 phase 2 trial. For BCMA naïve patients, an overall response rate (ORR) of 61.0% and median progression free survival (PFS) of 17.2m was observed with a median overall survival (OS) of 24.6m. However, efficacy and safety information in the real world is lacking. Methods: This was an international retrospective study of patients with RRMM treated with Elra monotherapy outside of clinical trials. Data was collected through the International Myeloma Foundation Immunotherapy Database from 7 countries (UK, USA, Canada, Greece, Spain, Czech Republic and Singapore). High-risk cytogenetic abnormalities (HRCA) were defined as 1q+, t(4;14), t(14;16), t(14;20), and/or del(17p). Response was assessed using IMWG consensus criteria. Patient characteristics were summarized by frequency (percentage) or median (range). PFS, OS were evaluated using the Kaplan-Meier method and reported with 95% confidence intervals. Results: 79 patients were included. Median age was 67 years (range: 41-95y) with 20% ≥75 years. Racial distribution was: 75% White, 16% Black/African American, 8% Asian/Pacific. 55% were ECOG performance status (PS) 1-2. Patients received a median of 5 prior lines (range 1-11), and 35% were BCMA exposed (9% CAR-T, 5% ADC, 7% BsAb, 8% BsAb plus CAR-T, 4% ADC plus BsAb). Of those with available data, 45% had HRCA,13% had extramedullary disease (EMD) and 32% had penta-drug refractory RRMM. Overall, 48% would not have been eligible for the MagnetisMM-3 trial. Median follow-up was 6.3m (range 0.5-35.0). ORR was 53% (PR 11%, VGPR 33%, CR 8%, sCR 1%) and 67% for BCMA naïve patients. Responses were significantly lower for BCMA exposed patients (ORR 31%) but did not differ by age (51% <75yrs vs 60% ≥75yrs). On univariate analysis, prior BCMA exposure, platelets <50, presence of HRCA and penta-refractoriness were significantly associated with a lower response rate. PFS for the whole cohort at 6 and 12m was 64% (95%CI, 54-77%) and 52% (40-67%); for BCMA naïve patients, it was 77% (65-90%) and 67% (54-84%), respectively. Within the whole cohort, PFS was shorter for: BCMA exposed (median 4.0m vs not reached (NR)), platelets <50 (2.0 vs 21.7m), HRCA (7.2m vs NR), penta-refractory (4.3m vs NR). PFS did not differ significantly by EMD, PS, age, or creatinine clearance. OS for the whole cohort at 6 and12m was 74% (64-85%) and 58% (45-74%); for BCMA-naïve patients, it was 80% (43-85%) and 70% (57-87%), respectively. Within the whole cohort, OS was shorter for: BCMA exposed (median 10.3m vs NR), platelets <50 (4.3m vs NR), HRCA (10.7m vs NR) and penta-refractory (10.3m vs NR). Univariate regression analysis demonstrated that prior BCMA exposure, platelets <50 and penta-refractoriness were significantly associated with a worse PFS and OS, and HRCA predicted poorer survival CRS was observed in 41% of patients (grade(G)1: 35%, G2: 5%, ≥G3: 0%) with most occurring with the first step up dose (SUD): SUD1: 29%, SUD2: 15%, full dose: 4%. Tocilizumab was used in 20% and steroids in 3%. ICANS was observed in 4 cases (5%): G1: 2, G2: 1, G4: 1. Treatment was with steroids (n=3) and tocilizumab (n=2). G3-4 thrombocytopenia was observed in 19% and 14% required platelet transfusions. G3-4 neutropenia occurred in 25% with overall 31% requiring GCSF. Infections were reported in 33 (42%) patients with median time to infection of 41 days. A total of 94 infection episodes were observed over 581 personal months. There were 2 infection-related deaths reported (1 had IVIG, 1 did not). 64% had hypogammaglobulinemia (IgG < 400mg/dl) and IVIG was given in 65% of patients. Conclusions: In this first international real world analysis, Elranatamab achieved a response rate of 67% and a 12m PFS of 67% in BCMA naïve patients, similar to the MagnetisMM-3 trial despite almost half not meeting trial eligibility. However, prior BCMA exposure, platelets <50, high risk cytogenetics and penta refractory status were predictors of poorer outcomes. Acknowledgements: The International Myeloma Foundation, International Myeloma Working Group (IMWG)

Abstract Chimeric antigen receptor T (CART) cell therapy is potentially curative in certain hematological malignancies, resulting in several FDA approvals since 2017. However, widespread adoption is limited by 1) cost and complexity of an autologous platform, 2) severe toxicities, and 3) limited durable remissions. To better predict CART efficacy and toxicity, live CART imaging would allow the realtime assessment of in vivo CART expansion and trafficking. Incorporation of the sodium iodide symporter (NIS)in CART cells (NIS+ CART cells) allows the CART cells to be tracked via PET imaging at high sensitivity. PET signal correlates with CART expansion and toxicity in mouse models (PMID: 34244299). Building on this platform, we aimed to track xenogeneic CART cells to monitor for their rejection, utilizing dogs with spontaneous lymphoma. Canine diffuse large B-cell lymphoma (DLBCL) affects over 250,000 dogs a year and has shared pathologic, biologic, phenotypic, genetic, immunotherapeutic targets, and treatment responses with human DLBCL, providing strong rationale for studying canine lymphoma as a pre-clinical model. We therefore initiated a USDA-CVB regulated pilot study and treated companion dogs with naïve canine CD20 (cCD20)+ DLBCL with human-derived TRAC- and B2M-KO NIS+ CART cells which targeted cCD20. First, we tested and optimized a chemotherapy regimen to achieve optimal lymphodepletion in dogs. Healthy beagles were treated with pentostatin (pen) on days 2-4 (0.125, 0.25, or 0.5 mg/kg i.v.), cyclophosphamide (cy) on day 4 (400 mg/m2 i.v.), or combination cy/pen. The most robust decrease in lymphocytes was seen with combination cy/pen (all cy/pen doses vs all other treatments, p≤0.0001) and in neutrophils with cy or cy/pen (all cy/pen doses vs all pen doses, p≤0.0001; cy vs 0.25 mg/kg pen, p≤0.001; cy vs 0.5 mg/kg pen, p≤0.0001). All lymphodepletion was transient and recovered within two weeks. All regimens were well-tolerated, and we therefore used 0.5 mg/kg pen plus 300 mg/m2 cy for the companion dog pilot study. Next, we assessed the feasibility of generating cCD20-targeted NIS+ human CART cells (CART20-NIS) with double KO (DKO) of TRAC and B2M. Primary T cells from healthy human donors were transduced with bicistronic lentivirus encoding 41BB-costimulated cCD20-targeted CAR and NIS. Two days after transduction, cells were transfected with CRISPR Cas9-gRNA targeted to TRAC and B2M. The resulting CART20-NIS DKO cells were expanded until day 11 and then were further depleted for CD3 and cryopreserved. CART20-NIS DKO cells demonstrated high CAR expression (untransduced T cells (UTD) vs CART20-NIS DKO, p<0.0001) and KO efficiency (TCRαβ and HLA-ABC expression of UTD vs CART20-NIS DKO, p<0.0001). NIS functionality was confirmed via a qualitative in vitro I-125radioactive uptake assay. We then assessed antigen specificity and activity in vitro. CART20-NIS DKO or UTD were co-cultured with either wildtype or cCD20-overexpressing K562 cells for three days. Target cell killing was then quantified via flow cytometry. CART20-NIS DKO cells demonstrated significant cytotoxicity against cCD20+ K562 cells (UTD vs CART20-NIS DKO, p<0.0001). Lastly, we treated two companion dogs in a USDA-CVB regulated pilot study of CART20-NIS DKO cells in naïve cCD20+ DLBCL. Following lymphodepletion on days -5 to -3, on day 0, dogs were i.v. administered 1 x 108 CART20-NIS DKO cells and intratumorally administered 2.5 x 107 CART20-NIS DKO cells in two lymph nodes, with contralateral saline injection as controls. Dogs underwent 18F-TFB-PET imaging on days 1, 3, and 7, which showed evidence of CART20-NIS DKO cells both in the CART-injected lymph nodes and distal lymph nodes, as determined by SUVmax. Both dogs sustained clinical remission for the duration of the 21-day study, as determined by sum of longest diameters and accompanying x-ray radiology and abdominal ultrasound. Histology showed marked reduction of CD20 cells in the lymph nodes. Organ and hematological toxicities were minimal, and dogs remained in remission for 95 and 101 days post-CART. In summary, this study demonstrated that xenogeneic CART cells were well-tolerated, induced remission, and were visualized via PET imaging in companion dogs with lymphoma. These results are promising both for the treatment of companion dogs with novel cell therapies and for the translation of novel CART cell therapies into human clinical trials.

Abstract Introduction: Reporting patient outcomes by CAR-T treatment centers is part of the accreditation requirement under the FACT (Foundation for Accreditation of Cell Therapy) Immune Effector Cell (IEC) Program standard. Currently, this is a time-consuming and manual process, making it increasingly difficult to sustain with the growing CAR-T applications and patient volume. The objective of this study is to evaluate how a Large Language Model (LLM) can be used to automate the abstraction and summarization of medical record for CAR-T toxicities. Methods: Clinical data from patients treated with CAR-T between January 2018 and June 2025, including hematology clinical notes, lab results, flowsheets, vital signs, and medication administration record, were used as the source of patient information. The American Society of Transplant and Cell Therapy (ASTCT) CAR-T toxicity grading guideline for cytokine release syndrome (CRS) and immune effector cell associated neurologic syndrome (ICANS) were used to define these toxicities. With the large context window of the LLM (Gemini 2.5 pro), we created a prompt by embedding both the patient data and the ASTCT guidelines for CRS and ICANS, along with an outline of specific information we aimed to abstract from the data. Additionally, the LLM was instructed to provide contextual justification for the data it extracted from patient information. The LLM-generated output included event onset date, resolution date, maximum grade, date of maximum grade, and medications given for management. The CRS and ICANS outcome reports generated by the LLM were then compared to the IEC compliance program data. Any discrepancies identified were reviewed manually. Results: Among patients (pt) treated with CAR-T in the specified timeframe, 47 were selected as training cohort. This cohort included pt with toxicities events, no events, and a mix of events of varying grades. For the test cohort, 66 pts who received CAR-T from March to June 2025 were included. Demographics for the training and test cohorts are reflective of evolving clinical practice (training, test cohorts: median age 69, 65; male 55%, 74%; lymphoma/leukemia 66%, 48%; multiple myeloma 34%, 52%; CD19 CAR-T 46%, 50%). Compared to data from the IEC compliance program, the LLM-generated summaries for CRS events demonstrated an accuracy of 90%, precision 95%, sensitivity 100%, specificity 80%, F1 score 97%, and Matthew's correlation coefficient 87%. For ICANS event, the LLM achieved an accuracy of 91%, precision 92%, sensitivity 92%, specificity 90%, F1 score 92%, Matthew's correlation coefficient 83%. Maximum grade, for both CRS and ICANS, were identified with >90% accuracy. However, dates for CRS and ICANS were less accurate, with the lowest accuracy noted in identifying the event resolution date, 45% for CRS and 67% for ICANS. Medical records were reviewed for all discrepancies between the LLM and the IEC Compliance Program data. The most common discrepancy was a mismatch in event resolution date, determined to be due to differing criteria used by the compliance team and the LLM prompt. Of note, a few discrepancies were attributed to missed information by the IEC Compliance Team, which was correctly identified and categorised by the LLM. An updated LLM prompt (LLMv2) was developed and applied to the test cohort. In version 2, the prompt was revised to include clearer instructions on how to summarize CRS and ICANS events, as well as new instructions on additional data elements to be summarized. The performance of LLMv2 on this cohort for CRS event identification was 100% across all categories (accuracy, precision, sensitivity, specificity, F1 score, and Matthew's correlation coefficient). For ICANS the LLMv2 achieved accuracy of 96%, precision 69%, sensitivity 100%, specificity 93%, F1 score 82%, Matthew's correlation coefficient 80%. Accuracy in identifying event resolution date improved significantly from 64% to 99%. Conclusions: Our study demonstrates that LLMs can significantly reduce the time required to abstract and summarize CAR-T toxicities, while maintaining high accuracy. One limitation of our study is that it is based on data from a single institution. However, plans are already underway to evaluate the use of LLMs in a multi-institutional setting.

Abstract Introduction: Patients (pts) with soft tissue plasmacytomas noncontiguous with bone (true extramedullary disease [EMD]) have poor outcomes with standard therapies shown by low overall response rates (ORRs), which are not durable. Talquetamab (Tal; anti-GPRC5D×CD3) and teclistamab (Tec; anti-BCMA×CD3) are first-in-class bispecific antibodies (BsAbs) approved as monotherapies for triple-class exposed (TCE) relapsed/refractory multiple myeloma (RRMM). Primary analysis from the RedirecTT-1 (NCT04586426) dedicated phase 2 EMD cohort (March 2025 data cut; median follow-up [mFU] 12.6 mo) showed Tal + Tec elicited an ORR of 78.9% (assessed by independent review committee) and a 12-mo progression-free survival (PFS) of 61.0%. Here, we report updated efficacy, using investigator assessment, and safety results from the phase 2 RedirecTT-1 EMD cohort with Tal + Tec. Importantly, we report EMD location and a novel analysis of tumor burden as a prognostic indicator of ORR. Methods: Pts had TCE RRMM and true EMD defined as ≥1 nonradiated soft tissue plasmacytoma noncontiguous with bone ≥2 cm in 1 dimension (with or without paramedullary plasmacytomas). Nonsecretory/oligosecretory disease was permitted. Prior CAR-T and non-BCMA/-GPRC5D BsAb therapies were permitted. Pts received Tal 0.8 mg/kg + Tec 3.0 mg/kg Q2W, preceded by step-up doses, with a permitted switch to monthly dosing at investigator's discretion after cycle 6 or after cycle 4 with confirmed ≥VGPR. Response was assessed per IMWG; EMD response was assessed by PET-CT or MRI whole-body scans. Tumor burden was assessed by total EMD tumor volume. Results: As of July 2025, 90 pts received Tal + Tec, with a median follow-up of 16.3 mo (range 0.5–23.7). Baseline characteristics were as previously reported; 39% had nonsecretory/oligosecretory disease, 20% had prior anti-BCMA CAR-T therapy, and 9% had prior BsAb therapy (all anti-FcRH5). The median number of plasmacytomas was 2 (range 1–7); 17 (18.9%), 29 (32.2%), and 74 (82.2%) pts, respectively, had ≥1 nodal, organ, or soft tissue EMD, while 11 (12.2%) also had ≥1 additional paramedullary plasmacytoma. Of 268 EMDs across all pts, 95 (35.4%) were nodal EMD, 77 (28.7%) were organ EMD (most commonly liver), and 64 (23.9%) were soft tissue EMD; 32 (11.9%) were paramedullary. Baseline EMD tumor volume was <25 cm2in 43 (47.8%) pts, 25–50 cm2 in 21 (23.3%) pts, and >50 cm2 in 26 (28.9%) pts. Per investigator assessment, ORR (95% CI) was 77.8% (67.8–85.9); 50.0% had a ≥CR. Median duration of response (DOR) was not estimable (NE; 12-mo DOR rate, 60.1%), 12-mo PFS was 55.6%, and median overall survival (OS) was NE (12-mo OS rate 73.8%). ORR was 90.7% (77.9–97.4; ≥CR 60.5%) in pts with EMD tumor volume <25 cm2, 66.7% (43.0–85.4; ≥CR 52.4%) for 25–50 cm2, and 65.4% (44.3–82.8; ≥CR 30.8%) for >50 cm2. Common adverse events (AEs) included CRS (77.8%; no grade [gr] 3/4) and neutropenia (72.2%; gr 3/4 62.2%). Taste changes (78.9%), non-rash skin AEs (68.9%), and nail AEs (55.6%) were all gr 1/2, and rash AEs (30.0%) were mostly gr 1/2. ICANS occurred in 11 (12.2%) pts (gr 3, 1.1%; gr 4, 1.1%). Infections occurred in 72 (80.0%) pts (gr 3/4 40.0%), most commonly upper respiratory tract infection (26.7%; gr 3/4 5.6%); IVIG was highly recommended to prevent and manage hypogammaglobulinemia and infection. Ten (11.1%) pts discontinued Tal + Tec due to treatment-emergent AEs (4 pts due to infections and 6 due to gr 5 AEs). Two pts discontinued Tal only due to AEs. In total, 11 (12.2%) pts had gr 5 AEs (6 due to infections), 6 of which were deemed to be drug related by investigators. Data will be updated with an additional 4 mo of follow-up for the presentation. Conclusions: With longer follow-up in pts with TCE RRMM with true EMD regardless of baseline tumor characteristics, Tal + Tec efficacy exceeded all approved therapies, including T-cell redirecting and cellular therapies, noting limitations of cross-study comparisons. Lower total EMD tumor volume was associated with a higher ORR but low pt numbers in each group and lack of statistical testing limits robust interpretation. The safety profile of Tal + Tec was generally consistent with each monotherapy; AEs were not exacerbated with the combination. These data continue to highlight the clinical benefit of the novel combination of Tal + Tec in pts with true EMD, a population with high disease burden and significant unmet need.

Abstract Background: CD19-directed CART cell (CART19) therapy has transformed the treatment of B-cell malignancies but is limited by life-threatening toxicities, including cytokine release syndrome (CRS) and immune effector cell–associated neurotoxicity syndrome (ICANS), as well as low durable remission rates. Our understanding of the pathogenesis of CART-associated toxicities is limited, and although current treatments are effective, refractory cases remain challenging. Methods and Results: To investigate CRS/ICANS pathogenesis and identify new therapeutic targets, we analyzed plasma cytokine kinetics within the first month of CART19 cell therapy in 25 relapsed/refractory (r/r) large B-cell lymphoma (LBCL) and 14 r/r follicular lymphoma (FL) patients (NCT02030834). In this cohort, 22 (88%) patients developed any-grade CRS (73% LBCL, 27% FL) and 4 (16%) developed any-grade ICANS (50% LBCL, 50% FL). Across 30 analyzed cytokines, IL-5 emerged as the second most enriched cytokine based on the median peak-to-baseline (Day 0) ratio in patients who experienced CRS vs. those who did not. The IL-5 ratio was 6.6-fold higher (p<0.05) and 11.9-fold higher (p=0.093) in patients who experienced CRS or ICANS, respectively. Moreover, IL-5 was elevated in patients experiencing CRS on Days 1 and 2 (p<0.05). Prompted by our results, we evaluated the effect of IL-5 neutralization with a monoclonal antibody (mAb) on CART-associated toxicity in a patient-derived B-acute lymphoblastic leukemia (B-ALL) xenograft model. After engraftment (average of 110 human CD19+ cells/µL peripheral blood), NSG mice were randomized based on tumor burden and treated with 3.5×10⁶ CART19 plus either IL-5 mAb or IgG control (10 mg/kg, weekly), or no treatment. Mice receiving IL-5 mAb showed reduced weight loss (p<0.005, Days 13–16) and neuroscore (p<0.05, Day 15), as assessed by a daily, blinded 18-point assessment, vs. the IgG group. Additionally, while CRS-associated cytokines (GM-CSF, IFN-γ, MIP-1β) were significantly elevated in the IgG group (p<0.05) vs. the untreated group, they were not in the IL-5 mAb group. In this model, all mice treated with CART19 cells were able to clear the tumor burden. However, in a follow-up challenge model generated by engrafting NSG mice with 1x106 luciferase+ JeKo-1 cells, mice treated with a combination of 1x106CART19 cells + IL-5 mAb showed reduced weight loss (p<0.0001) and enhanced overall survival (p<0.01) vs. mice treated with CART19 + IgG. Next, we evaluated IL-5 neutralization in an immunocompetent toxicity model. BALB/c mice were treated with 10×10⁶ murine untransduced T cells, CART19 + IL-5 mAb, or CART19 + IgG (10 mg/kg, weekly). Mice receiving IL-5 mAb exhibited significantly reduced weight loss (p<0.05, Day 1) and lower neuroscore (p<0.05, Days 1, 6, and 8) vs. those receiving IgG control, with a consistent trend observed on other days, supporting a protective role for IL-5 blockade in mitigating CART-associated toxicities. To explore the direct effects of IL-5 on CART19 cells, we disrupted either IL-5 (IL-5KO-CART19) or its receptor (IL-5RαKO-CART19) using CRISPR-Cas9. In NSG mice engrafted with 1×10⁶ luciferase+NALM6 cells, mice receiving 1×10⁶ IL-5KO-CART19 vs. IL-5WT-CART19 showed a trend toward reduced weight loss (p=0.08), significantly increased antitumor activity by Day 20 (p<0.05), and improved overall survival (p<0.05), indicating that IL-5 production can significantly impair CART19 cell efficacy. We next evaluated changes in IL-5Rα gRNA representation in a published genome-wide CRISPR screen (PMID: 39266501). CART19 cells with IL-5Rα-targeting gRNAs were enriched following chronic stimulation, indicating a protective effect for IL-5 pathway interruption. In subsequent studies, IL-5RαKO-CART19 cells showed enhanced killing of JeKo-1 cells following 48 hours of co-culture (p<0.05) and reduced co-expression of multiple inhibitory markers (p<0.01) following chronic stimulation with JeKo-1 cells for one week vs. IL-5RαWT-CART19 cells, suggesting an improvement in CART19 cell activity and phenotype following disruption of IL-5Rα. Conclusion: Our data identify IL-5 as a previously unrecognized mediator of CART–associated toxicities and show significant improvement of CART19 activity with disruption of IL-5 signaling either indirectly with an IL-5 neutralizing antibody or directly with CRISPR gene editing. Thus, IL-5 could serve as a dual-purpose target to improve both safety and efficacy of CART cell therapy.

Abstract Introduction: A subset of patients with chronic lymphocytic leukemia (CLL) experiences disease progression from CLL to Richter transformation (RT), characterized by the occurrence of a high-grade lymphoma. The outcomes of patients with RT are typically dismal and identifying better therapeutic strategies remains critically important for these patients. To discern the biologic underpinnings driving this transformation, we studied immune changes and genetic evolution in samples from CLL patients who developed RT. Methods: We identified 18 patients with CLL who developed RT; and performed single cell (sc) RNA-sequencing(seq) on sorted B and T cells from peripheral blood mononuclear cells (PBMCs) at the time of RT (n=18) and at the earlier time of CLL diagnosis (n=11). The median time between the CLL sample and the one closest to RT was 4.3 years (range, 0.2-7.3); and patients received a median of 1 CLL directed treatment (range, 0 – 7). A total of 749,906 cells and 36,601 genes were captured during scRNAseq and 720,580 cells and 21,892 genes were retained for downstream analysis, respectively. Analysis was performed using the immunopipe (https://github.com/pwwang/immunopipe) that wraps around Seurat (v5.1.0) and identified enriched pathways using Enrichr service. The sc-T cell receptor (TCR)-repertoire(R)-seq data was analyzed and visualized using scplotter (). The whole exome sequencing (WES) analysis with tumor B-cells from same patients of CLL & RT stages were done using the Agilent sureselect Human all exon v5 + UTR capture kit. Reads from paired end sequencing were processed with the sarek pipeline. Results: Analysis of 720,580 cells from 11 CLL and 18 RT samples showed 16 distinct cell types- 3 B and 13 T cell subtypes. We observed an increasing trend of CD8 TEM [mean 0.548 vs. 0.384, p = 0.12 (Wilcoxon test)] but a decreasing trend of CD4 TCM (mean 0.236 vs. 0.337, p = 0.08) and B-memory (mean 0.0085 vs. 0.0195, p = 0.51) subsets in RT compared to their CLL stage. When comparing transcriptional changes, we found that among the top 18 highly variable genes, the expression of FP236383.3 (lncRNA) and TCL1A was significantly higher (p<0.0001) in the B cell subsets of patients with RT vs. CLL stage. Additionally, in T cell subsets—primarily dnT, γδT, and CD8/CD4 naïve cells—the expression of FP236383.3, NKG7, GNLY, and CCL5 was significantly higher (p<0.0001), while the expression of LDHB, GZMK, and IL32 was significantly lower (p<0.0001) in RT vs. CLL stage samples. Comparison of specific gene set enrichment analysis showed significantly increased expression of GZMB in gdT (p<0.001) and decreased expression of GZMA, GZMK, TIGIT and TCF7 in dnT and CD8 Naïve cells (p<0.001) for RT vs. CLL stage. The top 10 differentially expressed pathways in RT included oxidative phosphorylation, allograft rejection, IFN-α/-γ responses, TNF-α signaling via NF-κβ, complement, Myc targets V1, mTORC1, hypoxia, and p53 signaling across various B and T cell subsets compared to the CLL stage. Initial analysis of TCR-R data showed that patients with RT stage have lower number of unique clones (p<0.5) and reduced clonal diversity (p<0.5), and a higher number of large and hyperexpanded clones compared to their CLL stage. Additionally, WES data indicated that the most frequently mutated gene is TP53 in tumor B cells of both CLL and RT stages. However, a higher percentage of RT-stage samples carry mutations in MGA, ATM, and IRF2BP2 genes compared to their CLL-stage. WES data also showed a trend toward differing frequencies of copy number amplification at 19q12 (p=0.10) and 4q13.1 (p=0.52) between RT and CLL stages. Conclusions: The main findings of our study are as follows:The transcriptional landscape showed significant changes in multiple genes and pathway enrichment related to cell growth, differentiation and immune response in RT vs. CLL stage.A decreased fraction of CD4 TCM cells and reduced TCR clonal diversity were observed in samples at the RT stage and may indicate a poor prognosis for the patients at this stage. WES data of tumor B cells also indicated possible genetic alterations in the RT vs. CLL stage. Data from in-depth analyses on understanding of the dynamics of T cell response and additional experiments on these patients are expected to provide valuable insights into tumor genetic heterogeneity, transcriptional and immune regulation in blood-based immune subsets.