Corraes, Andre De Menezes Silva

Abstract Introduction: Liver cancer remains a leading cause of cancer mortality globally with limited treatment options for patients (pts) with unresectable hepatocellular carcinoma [HCC] and intrahepatic cholangiocarcinoma [iCCA][1,2,3,4]. We hypothesize that combining the intratumoral dendritic cell (DC) vaccination after external beam radiotherapy (EBRT) without (phase I) and with (phase II) systemic atezolizumab and bevacizumab (atezo/bev for HCC) would enhance tumor-specific immunity and improve clinical outcomes, and present preliminary results from our phase I/II trial (NCT03942328). Methods: Peripheral blood mononuclear cells (PBMCs) were collected at baseline (PRE), after radiation (EBRT), and after completion of DC without or with atezo/bev (POST). Single-cell RNA sequencing with TCR sequencing were analyzed using CellRanger v7.0.1, Immunopipe, Seurat v5.3.0, & Gene-set enrichment analysis (GSEA). Long EFS (EFS-L) is defined as >=12 months (mo) for phase I and >=18mo for phase II, and short EFS (EFS-S) below that. Results: Seventeen patients have been analyzed to date (Phase I: n=8 [4 HCC, 4 iCCA]; Phase II: n=9 HCC). 674,106 cells were sequenced (146,843 PRE; 168,838 EBRT; 114,437 POST). GSEA for EFS-L vs EFS-S were comparable at PRE, EBRT and POST between phase I and phase II. EFS-S was associated with enrichment of interferon alpha (IFNα; B naïve, monocytes, pDC, cDC2, NK, Treg, CD8 Tcm and Tem at PRE and POST; B intermediate, CD4 Tem and NK at EBRT) and reactive oxygen species (ROS; PRE: pDC, CD8 Tcm, Tem. EBRT: CD4 Tem. POST: CD8 T naïve Tem, CD4 Tcm & Tem, cDC2) pathways. In contrast, EFS-L was associated with enrichment of WNT-β-catenin (WNT; CD14 mono at EBRT & POST, B intermediate at EBRT), TGFb (PRE: NK; EBRT: B intermed, CD4 Tem; POST: monocytes, CD4 naïve & Tem, pDC, cDC2), and angiogenesis (EBRT: NK; POST: monocytes, cDC2) pathways. Interestingly, enrichment of protein secretion pathway in NK cells was associated with EFS-L, but among other cells were associated with EFS-S (PRE: CD16 mono, Treg; POST: CD16 mono, CD4 & D8 Tem). Compared to phase I, addition of atezo/bev in phase II pts was associated with a significant increase in Gini coefficient (p<0.0001), suggesting expansion of certain TCR clones, along with a trend for increased Gini coefficient from Pre to Post in both phase I and II. Conclusions: Preliminary analysis of this study combining EBRT with intratumoral DC vaccination +/- PD-L1/VEGF blockade identified distinct treatment-related changes in systemic cellular immune profiles. Enhanced TCR clonal diversity in phase II suggests added immunologic benefit from PD-L1/VEGF blockade. As accrual continues, integrated analyses will identify systemic immune correlates of clinical benefit and resistance to this multimodal immunotherapy approach. Citation Format: Melody Wu, Panwen Wang, Chen Wu, Ying Li, Christopher L. Hallemeier, Thomas D. Atwell, Nguyen H. Tran, Andre de Menezes Silva Corraes, Kevin Regan, Zuoyi Shao, Rayaan Kamal, Haidong Dong, Lewis R. Roberts, Sean Park, Yi Lin, Lionel Aurelien Kankeu Fonkoua. Radiation-dendritic cell combination therapy drives systemic cellular immune shifts in liver cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 3741.

Abstract Large Language Models (LLM) are being widely adopted into the medical field for their impressive ability to analyze and summarize large amounts of text data. These models enable clinicians and researchers to extract meaningful insights from complex datasets and may assist with decision making. Here we present our workflows and application of LLM for the interpretation and summarization of clinical data related to CAR-T cell therapy. Using an LLM (Gemini 2.5 pro) in the Google Cloud Computing (GCP) environment, two applications were developed to analyze CAR-T cell therapy clinical data: 1) extracting and summarizing CRS and ICANS event-related data to streamline the compliance team workflow, and 2) identifying features available at time of CAR-T infusion able to classify patients into high- or low-monitoring needs 14 days post CAR-T. Patient data (vitals, labs, hematology notes, and EKGs) were extracted from the electronic medical record (EMR) using Google BigQuery into SQL tables in GCP. For each application, relevant data fields were retrieved, formatted into JSON objects, and embedded in the LLM prompt for context-aware processing. Both applications have undergone iterative prompt engineering after analyzing the LLM output against ground truth data in the EMR. For the application extracting and summarizing CRS and ICANS events, the LLM was optimized on Mayo Clinic Rochester data. When compared to the IEC compliance database, the LLM achieved 100% accuracy and F1 score for CRS events, and 96% accuracy and 82% F1 score for ICANS events. The match rate for CRS and ICANS grades were 83% and 89% respectively. We applied the same LLM to Mayo Clinic Arizona (MCA) and Mayo Clinic Florida (MCF) cases, who document clinical notes and toxicity flowsheet differently, and achieved an accuracy (MCA: 93%, MCF:93%) and F1 score (MCA: 96%, MCF: 96%) for CRS and accuracy (MCA: 81%, MCF:82%) and F1 score (MCA: 76%, MCF: 75%) for ICANS. LLM was able to capture events missed by manual reviews. For most of the discrepancies where compliance team final adjudication is needed, LLM will be updated to flag discrepancies for review by the compliance team. For the application related to the monitoring needs 14 days post CAR-T, the LLM identified 5 categories predictive of high or low monitoring needs post CAR-T infusion (disease status, inflammatory and tumor burden markers, hematologic status, renal function and performance status). Our model achieved a sensitivity of 85.7%, specificity of 23.8%, and F1 score of 65.5% for our first cohort, compared with data extracted by the LLM from the EMR. For our second cohort, with demographics statistically similar to cohort 1 and using the same 5 categories, the LLM achieved a sensitivity of 83.3%, a specificity of 28.6%, and F1 score of 65.4%. Our workflow and applications of LLM’s provide examples and guidance to others interested in applying LLM’s for clinical and research applications. Citation Format: Emmanuel Contreras Guzman, Matthew Jankowski, Andre De Menezes Silva Corraes, Malvika Gupta, Monica L. Shaw, Madiha Iqbal, Talal Hilal, Saurabh Chhabra, Ricardo Daniel Parrondo, Jody K. Mclean, Kim R. Riester, Kayla Joseph, Melinda Tan, Holly Ross, Cleyonia Barnett, Sylvia Carter, Semy Girmay, Rachel Wolan, Milana Ramsey, Christian Downhour, Kristy Morgan, Shae Sibley, Erica Rushing, Lucy Holmes, Allison Burgstahler, Stephen M. Ansell, Hassan Alkhateeb, Matthew Hathcock, Ramona Bruno, Allison C. Rosenthal, Hemant Murthy, Patrick B. Johnston, Jonas Paludo, Yi Lin. Applications of large language models to CAR-T cell therapy clinical data using Google cloud computing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2740.

Abstract Chimeric antigen receptor (CAR) T-cell therapy achieves high response rates in relapsed/refractory aggressive B-cell non-Hodgkin lymphoma, yet only 40% of patients achieve durable remission. Identifying biological programs linked to long-term response remains critical. As metabolic fitness underlies T-cell persistence and effector function, we profiled metabolic pathways associated with clinical outcomes following commercial CD19 CAR T-cell therapy. Single-cell RNA sequencing of peripheral blood mononuclear cells collected at baseline (BL), peak CAR-T expansion (PK), and one-month post-infusion (M1) was analyzed using CellRanger v7.0.1, immunopipe. Seurat v4.3.0 was applied for unsupervised clustering to delineate cell subsets based on top differentially expressed genes. Patients were categorized as durable complete remission (CR ≥6 months; n = 16), primary refractory (PD1; n = 4), or relapse after initial response (PD2; n = 12). Gene-set enrichment analysis defined metabolic pathway activity across T-cell, monocyte, dendritic cell (DC), and natural killer (NK) subsets. Oxidative phosphorylation (OXPHOS) emerged as the dominant metabolic program distinguishing clinical outcomes. At BL and PK, OXPHOS was consistently enriched in PD1 relative to CR across T-cell, monocyte, and DC subsets, suggesting early oxidative activation in non-responders. By M1, this pattern inverted, with higher OXPHOS activity in CR. PD2 largely paralleled PD1, but several subsets including CD8 T central memory, classical monocytes TGFβ, intermediate monocytes CD38, monocytic myeloid-derived suppressor cell (mMDSC) HIF1A, and mMDSC SIRPA, showed OXPHOS enrichment in CR at BL and/or PK that persisted through M1. In contrast, DC and NK subsets exhibited the opposite pattern: OXPHOS was enriched in CR at PK (conventional DC 2 (cDC2), plasmacytoid DC (pDC), NK, proliferating NK), but shifted toward enrichment in PD2 at M1 (cDC2, NK, NK CD56bright). Glycolysis (GLY) followed a similar trajectory in monocytes and NK cells, with enrichment in PD1/PD2 at BL and PK, followed by enrichment in CR at M1. In T-cells, PD1 maintained GLY enrichment from BL through PK, with no significant differences observed at M1. The inositol-phosphate metabolism pathway showed a more static pattern and was consistently enriched in multiple PD2 effector and memory T-cell subsets without reversal at M1. Taken together, these data reveal distinct, lineage-specific metabolic signatures that differentiate durable remission from early and late progression. Dynamic OXPHOS and GLY programming, characterized by lower activity early and enhanced activity at M1 in CR, may reflect adaptive metabolic programming that supports sustained antitumor immunity. Immune-metabolic profiling may therefore serve as a biomarker of response and highlight actionable metabolic pathways to enhance CAR T-cell durability. Citation Format: Melinda S.Y. Tan, Panwen Wang, Patrizia Mondello, Jacqueline Turner, Andre de Menezes Silva Corraes, Chen Wu, Zuoyi Shao, Kevin Regan, Ma Audrey, Arushi Khurana, Nora N. Benanni, Yucai Wang, Paul Hampel, Jonas Paludo, Saad J. Kenderian, Urshila Durani, Patrick B. Johnston, Jose Caetano Villasboas, Stephen M. Ansell, Ying Li, Haidong Dong, Hu Zeng, Yi Lin. Metabolic pathway signatures defining response to CD19 CAR T-Cell therapy in aggressive B-cell non-hodgkin lymphoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 3278.

Abstract Introduction: Therapeutic outcomes in B-cell non-Hodgkin’s lymphoma (B-NHL) have substantially improved with the introduction of chimeric antigen receptor T-cell (CAR-T) therapy. Yet, persistent immune dysfunction and underlying metabolic dysregulation remain major obstacles to achieving durable clinical responses in B-NHL. We hypothesize that mitochondrial pathways play a critical role in mediating anti-tumor immunity during CAR-T therapy. Methods: Peripheral blood mononuclear cells were collected from healthy controls (CNTRL, n=5) and patients with advanced stage lymphoma (LYM) who received FDA approved CAR-T (n=32). Samples were collected before lymphodepletion (BL), at peak CAR-T expansion (PK) and one-month post-infusion (M1). Single-cell RNA sequencing was used to interrogate differential mitochondrial gene expression. Samples were obtained from LYM patients with complete remission for >6 months (CR), primary refractory (PD1), or relapsed (PD2) disease. Results: Compared to CNTRLs, LYM patients exhibited reduced expression of cytochrome c oxidase. At BL, CR patients demonstrated broader mitochondrial differential gene expression across all T-cell subsets relative to PD1 and PD2 and showed higher expression of MT-ATP-6, MT-ATP-8, and MT-CO3. Following CAR-T infusion, global mitochondrial gene expression increased across patients corresponding with expansion of CD8 effector T cells (Tem tumor circulating) and activated peripheral memory T cells (Tpm). Certain mitochondrial genes, such as MT-CO1, were durably expressed across LYM patients at BL, PK, and M1 in peripheral and central memory T cells (Tpm, Tcm). At PK, cytochrome c oxidase and ATP synthase genes were upregulated in CR Tpm and Tcm. At M1, MT-ATP-6, MT-ATP-8, MT-CO3 were significantly upregulated in CD8 memory precursor effector T-cells (MPECs) of CR patients compared to both PD1 and PD2. While CR and PD2 patients appeared clinically responsive at M1, CR samples maintained significantly higher MT-ATP-8 expression across T cell subsets (CD8 MPECS, Tpm, Tcm) compared to PD2 (p values 0.008; <0.0001; <0.0001). Conclusions: We report a temporal shift in immunometabolism with CAR-T therapy in LYM. Clinical responders exhibited sustained mitochondrial activity characterized by persistent expression of ATP synthase in CD8 MPECs and cytochrome c oxidase in Tpm and Tcm. In sum, mitochondrial programming may distinguish effective versus dysfunctional T cells and identify targets for therapeutic development. Citation Format: Jacqueline Turner, Panwen Wang, Patrizia Mondello, Melinda Tan, Christoph Schaefers, Chen Wu, Andre de Menezes Silva Corraes, Kevin Regan, Zuoyi Shao, Ma Audrey, Arushi Khurana, Nora N. Benanni, Yucai Wang, Paul Hampel, Saad J. Kenderian, Jonas Paludo, Urshila Durani, Patrick B. Johnston, Jose Caetano Villasboas, Stephen M. Ansell, Haidong Dong, Ying Li, Zeng Hu, Yi Lin. Mitochondrial transcriptome profiles associated with clinical responses to CAR-T therapy in aggressive lymphoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 1213.

Abstract Background: Tarlatamab, a Delta-like ligand (DLL3)/CD3-targeted bispecific T-cell engager (TCE) is FDA approved in patients (pts) with extensive stage small cell lung cancer (ES-SCLC), after progression on frontline chemoimmunotherapy. We aimed to evaluate the immune cell profile in pts who received this therapy in standard-of-care (SOC) practice with progression free survival (PFS) less than and greater than two months (mo). Methods: Patients who received tarlatamab at Mayo Clinic Rochester and consented to immuno phenotyping of blood are included in the present study. Immune phenotyping was performed on whole blood by flow cytometry and analyzed by Kaluza. Data analysis was performed with Microsoft Excel and PRISM. Results: Eighteen patients with median age 64 (range 37-79) were included in the study. 66% of our cohort were women and 83% had present or past history of smoking, with an average of 40 pack years (range 26-60). With a median follow-up of 2 months, the median PFS for the cohort was 1.5 mo (range 0.33-2.76 months). 61% (13/18) of patients had PFS <2mo. 39% (5/18) had PFS>2mo and of which, three patients had partial response with one maintaining stable disease and one patient with mixed response. At baseline (BL), there were no difference in T cell, CD4 or CD8 cell count between patients in the PFS<2mo and PFS>2mo groups. Patients with PFS<2mo had higher exhausted CD8 T cells compared to those in the PFS>2mo group (CD8+PD1+TIGIT+CD57+, PFS<3mo vs >3mo, cells/mL: 7.14±4.28, 1.58±1.37, p=0.009). Additionally, when analyzing the B-cell population, the PFS<2mo group had lower percentage of B cells compared to the group with PFS>2mo at baseline (6.0±7.5, 10.4±14, p=0.02). At day 7, the group with PFS<2mo had lower Treg compared to the group with PFS>2mo (2.88±1.38, 4.60±1.47, p=0.02). Also, the PFS<2mo group had an increase of B cells compared to the group with PFS>2mo at day 7(<2mo vs >2mo: 10.0±14, 4.37±3.07, p=0.04). Finally, the group with PFS>2mo had decrease in total monocytes (mono), classical mono, and immunosuppressive cells (PFS<2mo vs >2mo, Mono: 340±347, 253±106, p=0.04. Classical mono: 305±283, 155±109, p=0.02. CD14+HLA-DRneg: 105.8±96, 12±39, p=0.03) compared to the group with PFS<2mo by day 7. While no changes were seen between BL and day 7 for intermediate mono in PFS<2mo group, PFS>2mo group had decreased post treatment intermediate monocytes (BL vs day 7, cells/mL, intermed mono: 32.5±15,17.7±6.2, p=0.03). Conclusion: In this study investigating the SOC outcomes of tarlatamab, early progression was associated with higher presence of exhausted CD8 T cells, B cells, and immunosuppressive monocytes. Analysis of additional patients will be shared at AACR meeting. Citation Format: Dhauna Karam Prasad, Andre De Menezes Silva Corraes, Malvika Gupta, Audrey Ma, Chen Wu, Zuoyi Shao, Kevin Reagan, Rayaan Kamal, Ashley Potter, Abdullah Al-Ajmi, Syeda Mina, Sykler J. Taylor, Antonious Hazim, Anastasios Dimou, Kaushal Parikh, Mohammed Shanshal, Ailsa Luce, Anna Schwecke, Julian Molina, Aaron Mansfield, Katherine Smith, Lucy Holmes, Haidong Dong, Yi Lin, Konstantinos Leventakos. Immune cell profile changes in patients treated with tarlatamab for extensive stage small cell lung cancer in real world practice [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 3931.

Abstract Recent therapeutic advances, including bispecific antibodies (bsAbs), have improved outcomes for patients with relapsed/refractory multiple myeloma (RRMM). However, early mortality (EM) remains a significant challenge, even in the era of novel immunotherapies. We conducted a multicenter, retrospective analysis of real-world patients with RRMM treated with commercial bsAbs to evaluate the incidence and causes of EM. Between May 2022 and June 2025, a total of 441 patients were treated across 7 countries. EM, defined as death within 12 months of therapy initiation, occurred in 148 patients (34%), with a median time from bsAb initiation to death of 2.87 months (interquartile range, 1.38-6.85). The leading cause of EM was progressive disease (72%), followed by infection (13%); regardless of the cause of death, 82% had active disease at the time of death. Our findings highlight that despite encouraging efficacy, EM remains high among real-world bsAb recipients, emphasizing the need for improved disease control strategies and supportive care interventions.

Abstract Background Chimeric antigen receptor T-cell therapy (CAR-T) has transformed care for relapsed or refractory B-cell malignancies. Cytokine-release syndrome (CRS) complicates most infusions, yet the ensuing infection risk is ill-defined.Table 1Profile of 230 patients undergoing with CRS following CAR-T therapyTable 2Cytokine Release Syndrome Severity, Management, and Early Outcomes by 90-Day Infection Status Methods CAR-T recipients who developed CRS at our institution from 1 Jan 2018 to 3 June 2024 were retrospectively reviewed. Variables were abstracted into REDCap. CRS was graded with American Society for Transplantation and Cellular Therapy (ASTCT) criteria. All infections occuring after CRS onset to day 90 were adjudicated. Continuous data are reported as medians with interquartile ranges (IQR); categorical data as counts and percentages. Table 2 comparisons used Mann-Whitney U for continuous and χ² or Fisher exact tests for categorical variables (two sided, P < 0.05). Multivariable logistic regression was used to identify independent predictors of infection.Table 3Infectious Syndromes, Pathogens, and Timing of First Infection Following CRS Onset Results Among 230 CRS cases, median age was 64 yr (IQR 57–70) and 59% were male. Non-Hodgkin lymphoma accounted for 62% and multiple myeloma for 38%. Median prior therapy lines were 3. CRS was grade 1–2 in 97%. CAR-HEMATOTOX was low in 156 (68%) (Table 1). Within 90 days, 51 patients (22%) developed infection; median onset was 15 days (IQR 5–26). Infected patients more often received anakinra (22% vs 8%, p = 0.02) and required ICU care (18% vs 7%, p = 0.05); grade ≥3 CRS trended higher (10% vs 3%, p = 0.06). In multivariable analysis adjusting for age, CAR-HEMATOTOX, ICU admission, and anakinra, both ICU admission (aOR 2.06; 95% CI 0.70–5.81; p = 0.18) and anakinra use (aOR 2.18; 95% CI 0.83–5.53; p = 0.10) were associated with higher infection odds, though not statistically significant. 90-day survival was 88% with infection versus 96% without (p = 0.09) (Table 2). 51 infectious episodes were documented: bacteremia 22; gastrointestinal 9; central line–associated bloodstream infection 7; other sites 22. Bacteria caused 59%, viruses 37% (15 CMV, 4 SARS-CoV-2), and fungi 4%. No infection-related deaths occurred (Table 3). Conclusion Infection was common after CRS, but did not alter early survival. Anakinra exposure and ICU admission were associated with a higher likelihood of infection, though not statistically significant. Targeted surveillance should focus on patients requiring intensified CRS management. Disclosures Yi Lin, M.D., BMS: Board Member|Caribou: Advisor/Consultant|Genentech: Board Member|Jassen: Board Member|Jassen: Steering Committee|Kite/Gilead: Steering Committee|Legend: Board Member|NexImmune: Advisor/Consultant|NexT Therapeutics: Board Member|Pfizer: Board Member|Regeneron: Board Member|Sanofi: Board Member|Tessera: Board Member Paschalis Vergidis, MD, MSc, Current Fungal Infection Reports: Honoraria|F2G: Grant/Research Support|Merck Manuals: Honoraria|Mundipharma: Grant/Research Support|Scynexis: Advisor/Consultant|Scynexis: Grant/Research Support

Abstract Chimeric antigen receptor (CAR) T-cell therapy has demonstrated remarkable response rates in relapsed/refractory aggressive B-cell non-Hodgkin's lymphoma. However, durable remissions occur in only 40% of treated patients, highlighting mechanisms of therapeutic resistance and immune escape. While cytotoxicity is principally executed by CAR T-cells, monocytes shape CAR-T cell efficacy and orchestrate cytokine-driven immunity. In this study, we defined distinct metabolic profiles of T-cell and monocyte subsets based on clinical response. Single-cell RNA sequencing (scRNA-seq) and T-cell receptor sequencing (scTCR-seq) were performed on peripheral blood mononuclear cells at three timepoints: before lymphodepletion (BL), at peak CAR-T expansion (PK), and one-month post-infusion (M1). Data were processed with CellRanger v7.0.1 and analyzed with the immunopipe pipeline (Wang et al., 2025). Seurat v4.3.0 was utilized for unsupervised hierarchical clustering analysis; subsets identified included classical monocytes (CM), intermediate monocytes (IM), non-classical monocytes (NM), and monocytic myeloid derived suppressor cells (mMDSC). Metabolic profiles were compared across patients achieving ≥6-month complete remission (CR, n=16), primary refractory disease (PD1, n=4), or relapsed after a prior response (PD2, n=12). A total of 107035 T cells and 129981 monocytes from 77 serial samples in 32 patients were analyzed. CR patients had a higher T-cell:monocyte ratio than PD1/PD2 at all time points, especially at BL and PK (BL 1.2 vs 0.4, p=0.011; PK 1.3 vs 0.3, p=0.027). Clinical responses were compared with metabolic pathway analysis and identified oxidative phosphorylation (OXPHOS) as the dominant pathway. OXPHOS was consistently enriched in PD cohorts across T-cell subsets, particularly in PD1 vs CR at BL (mucosal-associated invariant T-cell (MAIT) and T peripheral memory (Tpm)), and at PK (CD4/CD8 memory precursor effector cells (MPEC) and activated MPECs (MPECa), CD4 T-helper 1 proliferating (Th1p), CD4 Th2 activated (Th2a), regulatory T-cell (Treg), CD8 naive senescent (CD8NS), CD8 T effector memory (Tem) tumor recirculating (TR), CD8 Tpm, CD8 T resident memory (Trm) TR and MAIT Tpm). By M1, OXPHOS enrichment in CD4/CD8MPECa, CD4Th2a, CD8NS, CD8TemTR reversed, now predominant in CR although CD8Tpm and CD8TrmTR remained enriched in PD1. In the monocyte subsets at BL, PD2 showed OXPHOS enrichment in CD163+ monocytes (mono), IM IL1, mMDSC SIRPA, mMDSC TGFβ, and CM IL6 subsets compared to PD1 and CR, while PD1 showed OXPHOS enrichment in IM CD38, CM TGFβ, and mMDSC HIF1A vs CR and PD2. This persisted at PK. By M1, the pattern reversed and mirrored that of T-cells, with CR patients showing OXPHOS enrichment in CD163+mono, CM IL6, CM TGFβ, IM CD38, mMDSC HIF1A, mMDSC SIRPA and mMDSC TGFβ, and additionally in CM EMT, CM p53, NM, NM CD36 vs PD1/PD2. Glycolysis/gluconeogenesis (GLY) was also enriched across subsets. In T-cells, PD1 vs CR showed GLY enrichment in CD4Th2a, CD8Tpm, CD8Tpm proliferating and MAIT Tpm at PK. In monocytes, GLY was enriched in PD1/2 vs CR at BL (CD163+mono, CM IL6) and PK (CD163+mono, IM CD18, mMDSC SIRPA). By M1, GLY metabolism was enriched in CR vs PD1/PD2 in CD163+mono, CM IL6, mMDSC SIRPA. Additional GLY enriched subsets at CR vs PD1 included CM, CM p53, EMT, IM CD38, IM IL1, mMDSC HIF1A, NM, NM CD36. Inositol phosphate metabolism (IP) was predominantly enriched in PD2 vs CR at BL (CD8NS, CD8Tcm, CD8Tpm activated), PK (CD4Th2, CD8NS, CD8Tem TR, CD8Tpm proliferating, γδT Tem) and M1 (CD4/CD8 MPECa, CD4 Th1/Th2, CD8NS, CD8Tpm, CD8 Tpm proliferating, Treg and MAIT Tpm). Notably, Treg was the only T-cell subset enriched for IP at PK in CR vs PD1/PD2. Additionally, the TCA cycle was enriched in PD1 Tregs and Th2 cells at BL and PK vs CR/PD2. A higher T-cell:monocyte ratio in CR vs PD1/PD2 supports prior evidence that elevated pre-lymphodepletion lymphocyte:monocyte ratios predict better outcomes. We identified distinct metabolic signatures in T-cells and monocytes between CR and PD1/PD2. OXPHOS enrichment in T-cells and monocytes at M1 in CR patients may reflect oxidative programming as a hallmark of sustained anti-tumor immunity. Conversely, GLY enrichment in PD1 T-cells at PK may reflect short-lived effector states lacking persistence. Overall, these findings highlight metabolic profiling as a promising predictive biomarker and therapeutic programming target to enhance CAR T-cell outcomes.

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 Introduction: Bispecific T-cell engagers (TCE) targeting BCMA (teclistamab, elranatamab) and GPRC5D (talquetamab) are important therapeutic options in relapsed/refractory MM (RRMM). However, data on the independent prognostic impact of individual cytogenetic abnormalities and functional high-risk MM on outcomes with TCEs remains limited. Methods We included RRMM patients treated with BCMA TCEs (teclistamab, elranatamab) and talquetamab across 15 centers in the US. Patients where TCEs were used only as bridging therapy to CAR-T were excluded. High-risk cytogenetic abnormalities (HRCA) included del(17p), t(4;14), t(14;16), 1q gain/amplification (1q21+), del(1p) before initiation of TCE. Functional high-risk disease (FHRMM) was defined as PFS <18 months with frontline therapy. Extramedullary disease (EMD) included visceral or soft tissue disease and excluded paraskeletal disease. Progression-free survival (PFS) was estimated from TCE initiation using the Kaplan–Meier method and compared using the log-rank test. Univariate and multivariable Cox proportional hazards models were used to identify predictors of PFS in the entire cohort and the subcohorts of BCMA TCEs and talquetamab (Talq). Results We analyzed 943 treatment initiations with TCEs in RRMM, comprising 577 BCMA-directed TCEs [teclistamab (n=436) and elranatamab (n=141)] and 366 talquetamab initiations. The median follow-up from TCE initiation was 13.4 months (95% CI: 12.5–14.2) and median age was 68 years (interquartile range 61-75 years). Two ore more HRCAs were noted in 30% of the cohort (34% in Talq vs. 28% in BCMA TCEs, p=0.04), EMD was present in 21% of patients (26% in Talq vs. 18% in BCMA TCEs, p=0.002), and 45.5% of patients had FHRMM (47% in Talq vs. 45% in BCMA TCE, p=0.54). The median prior lines of therapy was 6 (interquartile range: 5-8), with 92% being triple-class refractory (TCR); 60% received prior BCMA-directed therapy (85% in Talq vs. 45% in BCMA TCE, p<0.0001) and 45% had prior CAR-T in 45% (57% in Talq vs. 37% in BCMA TCE, p<0.001). The best overall response rate (ORR) was 67.1 % for the entire cohort (66.4% for BCMA cohort and 68.2% for Talq), with 32.5% patients achieving a ≥ complete response. The median PFS was 7.7 months (95% CI: 6.7–8.6) for the entire cohort, with a median PFS of 10.1 months (95% CI: 7.5-13.4) for teclistamab and 10.1 months (95% CI: 3.9-NR) for elranatamab. In the more heavily pretreated talq cohort, the median PFS was 6.4 months (95% CI: 5.1-7.5). In a univariate analysis (UVA), several HRCAs were associated with inferior PFS−del(17p) (HR 1.59), t(4;14) (HR 1.43), del(1p) (HR 1.54), and ≥2 HRCAs (HR 1.64), all p<0.01. Apart from HRCAs, adverse disease characteristics included FHRMM (HR 1.20) and EMD (HR 1.46); all p<0.05. Among laboratory markers, hemoglobin <9 g/dL (HR 2.10), platelet count <75 ×10⁹/L (HR 1.96), and elevated ferritin (≥600 ng/mL; HR 2.17) (all p<0.01) were prognostic for PFS. ECOG ≥2 (HR 1.62), prior BCMA-directed therapy (HR 1.56), and prior CAR-T (HR 1.44) (all p<0.001) were also associated with inferior PFS. In a multivariable analysis (MVA), prior BCMA therapy (HR 1.72, p<0.0001), ferritin >600 ng/mL (HR 1.68, p<0.0001), hemoglobin <9 g/dL (HR 1.53, p=0.001), ECOG PS ≥2 (HR 1.52, p=0.0009), del(17p) (HR 1.44, p=0.0034) and FHRMM (HR 1.26, p=0.047), were associated with inferior PFS. Within the talq cohort (n=366), EMD (HR 1.64), del(17p) (HR 1.42), del(1p) (HR 1.54), FHRMM (HR 1.34), ECOG ≥2 (HR 1.56), ferritin >600 (HR 1.88), hemoglobin <9 g/dL (HR 1.6), platelet count <75 ×10⁹/L (HR 1.36) and leukopenia <0.25 x 10⁹/L (HR 1.52) were adverse prognostic markers on UVA. A MVA confirmed the independent impact of ferritin >600 ng/mL (HR: 1.77), ECOG ≥2 (HR 1.61) and functional high-risk status (HR 1.52) on PFS, whereas del(17p) and EMD were trending toward significance. Within the BCMA TCE cohort (n=577), prior BCMA therapy (HR 1.9), ferritin >600 (HR 1.79), ECOG ≥2 (HR 1.55), Hb <9 g/dL (HR 1.54), and del(17p) (HR 1.52) were independently associated with inferior PFS on a MVA.Conclusions: In this large real-world cohort of TCE, outcomes varied by fitness, disease biology, and treatment history. The ECOG PS, elevated ferritin, functional high-risk status and del(17p) were consistently associated with inferior PFS. Prior BCMA exposure was an independent risk factor for inferior PFS in the BCMA TCE cohort, but not with talquetamab.