Abstract Recurrent glioblastoma (rGBM) is an aggressive brain tumor with median survival under one year after standard chemoradiation. Antigen heterogeneity, immune exclusion, and a suppressive tumor microenvironment (TME) limit responses to immunotherapy. A first-in-human phase 1 trial of intracerebroventricular EGFR/IL13Rα2 CAR T cells (CART-EGFR-IL13Rα2) in EGFR-amplified rGBM was feasible, produced manageable neurotoxicity, and induced radiographic tumor regressions in a subset of patients (NCT05168423).To understand how this therapy reshapes the local TME, we analyzed paired tumor resections from 6 patients enrolled in the phase 1 trial, with specimens obtained from the primary intracranial disease site at trial enrollment (pre-treatment) and at radiographic progression after CART-EGFR-IL13Rα2 infusion. Multimodal spatial profiling included regional transcriptomic and protein mapping (GeoMx), single-cell whole-transcriptome imaging (CosMx), and high-resolution spatial transcriptomics (Visium HD). We annotated tumor, myeloid, lymphoid, and stromal compartments and derived composite scores for stemness, invasion, cell death, and immune regulation. Neighborhood- and interaction-based analyses were used to compare cellular states and cell-cell communication.Across patients, post-treatment samples showed reduced expression of CAR target antigen and a shift in tumor-intrinsic programs toward less stem-like, less migratory, and more apoptotic states, despite radiographic progression. The post-treatment TME was remodeled, with fewer suppressive myeloid- and B-cell-rich niches and increases in interferon-responsive and T cell-associated activation programs. Spatial interaction analyses indicated that pre-treatment rGBM contained dense networks of myeloid-tumor and myeloid-T-cell contacts consistent with impaired antigen presentation and effector function. Post-treatment specimens, in contrast, showed partial disruption of these suppressive circuits and the emergence of microenvironments more permissive to T-cell infiltration and activity.In the parent phase 1 trial, CART-EGFR-IL13Rα2 was feasible & induced radiographic tumor regressions in a subset of patients. This correlative spatial analysis suggests that prior EGFR/IL13Rα2 CAR T exposure can leave a less suppressive, more immunologically engaged TME at the primary site, even in resections obtained at radiographic progression. Together, these data support the idea that intracerebroventricular CAR T therapy may condition rGBM for subsequent immunotherapy. Myeloid and B-cell interactions are highlighted as candidate targets for armoring next-generation CAR T cells and for designing rational combination and sequencing strategies. Citation Format: Wesley V. Wilson, MacLean P. Nasrallah, Nakial Cross, Yael A. Day, Vanessa Gonzalez, Rachel M. Leskowitz, Amy Marshall, Julie K. Jadlowsky, Gabriela Plesa, Donald L. Siegel, Elizabeth O. Hexner, Joseph A. Fraietta, Carl H. June, Stephen J. Bagley, Donald O’Rourke, Zev Binder, Andrew J. Rech. Spatial profiling of recurrent glioblastoma in a Phase I clinical trial reveals favorable immune remodeling induced by intracerebroventricular CAR T therapy [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 3444.

Abstract Patients with relapsed or refractory B-cell lymphoma who experience disease progression after CD19-directed CAR T-cell therapy have poor outcomes and few effective treatment options. Subsequent therapies provide only modest benefit, with complete remission (CR) rates near 20% and limited durability. huCART19-IL18, an IL-18-secreting CD19 CAR T-cell product, produced an overall response rate of 81% and a CR rate of 52% after prior anti-CD19 CAR T failure in a phase I trial, but the spatial determinants of durable response within the tumor microenvironment (TME) remain incompletely defined. We performed multimodal spatial profiling of paired pre- and post-huCART19-IL18 lymph node biopsies from 11 patients (6 with CR, 5 without CR) enrolled in the phase I study. GeoMx whole-transcriptome atlas and protein profiling were integrated with CosMx 6000-plex single-cell spatial imaging to map transcriptional and cellular remodeling across tumor, myeloid, and lymphoid compartments. CoPro, a computational framework for detecting coordinated progression of cell states in space, was applied to identify spatially coordinated gene expression programs within and between myeloid and T-cell compartments. Post-infusion samples from responders showed increased T-cell and NK-cell infiltration, frequent tertiary lymphoid structures, and induction of interferon (IFN)- and tumor necrosis factor-responsive chemokine and immune effector programs, including enhanced antigen presentation, together with coordinated loss of B-cell identity and signaling. These changes localized to regions enriched for CAR T cells, effector memory CD4+ and CD8+ T cells, and IFN-polarized macrophages, consistent with IL-18-driven recruitment and reprogramming of myeloid and T-cell compartments. Pre-infusion TMEs in responders showed baseline myeloid chemokine signatures associated with clinical response. CoPro revealed that myeloid functional heterogeneity is organized along spatial gradients that are independent of lineage identity and coordinated with T-cell effector programs in CAR T-infiltrated regions. These findings support a model in which IL-18 armoring promotes spatially organized remodeling of the TME through coordinated myeloid and T-cell activation associated with durable remission after CD19 CAR T failure, nominating myeloid recruitment, antigen presentation, and T-cell exhaustion programs as critical biomarkers and rational engineering targets for next-generation armored CAR T strategies in lymphoma. Citation Format: Nakial C. Cross, Yael A. Day, Sonia Ndeupen, Vanessa E. Gonzalez, Zhen Miao, Sam I. Kim, Rachel M. Leskowitz, Amy Marshall, Julie K. Jadlowsky, Gabriela Plesa, Donald L. Siegel, Elizabeth O. Hexner, Jakub Svoboda, Stephen J. Schuster, Nancy R. Zhang, Joseph A. Fraietta, Andrew J. Rech, Carl H. June. Spatial remodeling of the tumor microenvironment by IL-18-armored CD19 CAR T cells is associated with durable remission in relapsed or refractory 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 6470.

CD19-directed CAR T-cells (CTL019) can produce durable remissions in chronic lymphocytic leukemia (CLL), but therapeutic success depends on whether autologous T-cells expand, persist, and retain cytotoxic function after manufacturing. Failure of CLL T-cells is often attributed to exhaustion, although many dysfunctional CLL T-cells retain inflammatory cytokine production. We tested whether this paradox reflects immunosenescence, an aging-like program defined by costimulatory loss, DNA damage, inflammatory secretion, repertoire restriction, and proliferative arrest. In response-linked preinfusion CTL019 products, nonresponders and short partial responders had higher senescence and senescence-associated secretory phenotype (SASP) programs than the functional-responder group, comprising complete responders and partial responders with transformed disease. These programs were detectable before infusion and became more prominent during manufacture in susceptible products. They tracked with weak in vivo CAR T-cell expansion and were associated with inferior survival. Serum proteomics revealed a circulating, CLL-associated SASP-like inflammatory milieu in nonresponders. Flow cytometry and T-cell receptor (TCR) profiling showed that poor responders carried CD27⁻CD28⁻ and KLRG1⁺ CD8 T-cells at apheresis, lower product CD27, and reduced product TCR diversity. Research-manufactured CAR T-cells from treatment-naïve CLL samples displayed the same state, most prominently in CD4 CAR T-cells, with SA-β-gal, p16, p53, DNA damage, depletion of less-differentiated states, and senescence-enriched transcriptomes. Under repeated CD19 stimulation, CLL-derived products reached a proliferative ceiling, and p53-high products lost cytotoxic reserve. Ibrutinib improved proliferative fitness, attenuated senescence-associated features and SASP output in paired patient and direct-exposure assays, and enhanced CAR T-cell expansion in an ibrutinib-resistant CLL model. Together, these data identify immunosenescence as a measurable and functionally consequential barrier to CAR T-cell efficacy in CLL and a candidate for therapeutic modulation.

Solid tumors often evade TCR-engineered αβ T cells when antigen expression varies or when the restricting Human Leukocyte Antigen (HLA) allele is lost. γδ T cells, in contrast, detect cellular dysregulation through non-peptide/Major Histocompatibility Complex (MHC) cues, including phosphoantigens and stress ligands, and can be developed as allogeneic therapies. Although intratumoral γδ T cell signatures are associated with improved outcome across cancers, γδ recognition itself is broad and still selected within the thymus just as αβ T cell receptors (TCRs) are. It does not, however, anchor specificity to a defined driver-mutation pMHC epitope. We therefore asked whether a high-affinity, co-receptor–independent αβ TCR could graft oncogenic-driver specificity onto γδ T cells while leaving the endogenous γδ TCR intact. We knocked the KRASG12V/HLA-A*11:01 TCR A11v into primary human γδ T cells. Engineered cells co-expressed the transgenic αβ TCR and the endogenous γδ TCR and lysed KRASG12V/HLA-A*11:01+ tumor cells in vitro and in vivo. To cover potential resistance through loss of HLA-A*11:01, we delivered an mRNA lipid nanoparticle (LNP) encoding a secreted mesothelin×CD3 (M5) bispecific T cell engager (TCE). LNP-M5 produced circulating TCE that redirected γδ A11v T cells and polyclonal bystander T cells to kill mesothelin+ targets, accompanied by development of higher γδ A11v T cell counts in vivo. In humanized mice bearing mixed HLA-A*11:01+ and HLA-A*11:01 − KRASG12V tumors, γδ A11v T cells produced transient control, whereas adding LNP-M5 yielded complete responses and prolonged survival. Thus, this two-part therapy couples invariant driver targeting to tunable redirection and addresses loss of the restricting HLA allele, a central escape route for TCR-based therapy. It provides an off-the-shelf reagent to enable KRAS-anchored treatment with the ability to redeliver the reagent.

We show continuous tumor exposure results in a loss of chimeric antigen receptor (CAR) T cell (CART) endocytic activity due to downregulation of Rab5. Loss of endocytic activity exacerbates the effects of trogocytosis, the bidirectional transfer of tumor target antigens and CARs between malignant cells and CARTs, resulting in CART dysfunction and fratricide. Constitutive expression of Rab5 within the CARTs reduced fratricide by reducing the amount of trogocytosed antigens on the cell surface, while simultaneously enhancing CAR availability through dissociation of CAR from target, recycling unbound CAR back to the plasma membrane, and limiting CAR capture by tumor cells. Rab5-expressing CARTs exhibited superior antitumor activity in both BCMA-CARTs isolated from the bone marrow of treated patients and mesothelin-specific CARTs in a solid tumor model. These studies uncover an unexpected relationship between endocytosis and CART function and suggest that pairing Rab5 with CAR expression could improve the clinical efficacy of CART therapy.

2013 Background: We previously reported the safety, bioactivity, and preliminary efficacy of bivalent CAR T cells targeting EGFR and IL13Rα2 in 18 patients with recurrent GBM (Bagley, et. al, Nat Med 2025). Given that median follow-up time was only 8.1 months at the time of the data cut-off for that publication, we now report updated overall survival (OS) and safety data with longer follow-up time, as well as neurologic function outcomes. Methods: Eighteen patients with recurrent EGFR-amplified GBM were enrolled using a 3+3 design (dose levels: 5.0 x 10 6 , 1.0 x 10 7 , and 2.5 x 10 7 cells). Patients received a single intracerebroventricular (ICV) dose of CART-EGFR-IL13Rα2 cells. OS was defined as the time from the date of initial study treatment to the date of death from any cause, or censored at the last date of contact if the patient was not known to have died at the date of analysis cut-off (January 16, 2026). To capture changes in neurologic function over the course of the trial, patients were assessed prospectively using the Neurologic Assessment in Neuro-Oncology (NANO) scale. NANO is an objective clinician-reported outcome of neurologic function scored on nine domains, with a higher score indicating worse neurologic function. NANO scores were assessed at baseline prior to CART (day 0), on days 1, 4, 7, 10, 14, 21, and at the 1-month and 2-month visits post-CART. Results: With median follow-up time of 18.5 months, median OS was 12.0 months (95% confidence interval, 7.4 – 23.2 months) with 3 patients surviving greater than 18 months. Other than one patient with prolonged grade 1 neurotoxicity as previously described, no prolonged, late-onset, or unexpected toxicities were observed within this longer follow-up time, including no evidence of on-target off-tumor toxicity or secondary cancers. On day 1 post infusion, an increase in NANO scores was observed (mean increase in NANO score from baseline of 1.9; p=0.09), which resolved to baseline by day 4. Patient NANO scores were unchanged from baseline at one month (mean change 0.41; p=0.998) and two months (mean change 0.54; p=0.992) post infusion. Those who underwent retreatment with a second dose of CAR-T cells at the time of tumor progression (n=7) did not experience an increase in NANO score on day 1 (mean change 0.33, p=0.99), despite having a significant increase on day 1 following the first infusion (mean change 2.9, p=0.008). Conclusions: With over 18 months of median follow-up time, ICV delivery of CART-EGFR- IL13Ra2 cells in patients with recurrent GBM has not demonstrated long-term or delayed toxicities and is associated with promising OS outcomes that warrant continued clinical development. Neurologic function worsens immediately following infusion but recovers to pre-treatment baseline by one month following CAR T cell infusion. Clinical trial information: NCT05168423 .

Analysis of EGR2 and type I IFN pathway regulation in CAR T-cells. (A-C) ATAC-seq tracks of genes associated with dysfunction, type I IFN signaling, and memory differentiation are shown, with differentially accessible regions indicated. D, Heatmap showing expression of exhaustion genes in tonically signaling GD2 and functional CD19 naïve CD8+ CAR-T cells (GSE136891). E, Comparison of EGR2 and type I IFN gene signature scores in central memory CD8+ T-cells expressing a tonically signaling GD2 CAR and control CD19 CAR. CAR T-cells were generated using T-cells from a healthy donor (GSE136891). F, Top transcription factor co-expression signatures overexpressed in CD19 CAR T-cell products of CLL complete responders. G, Comparison of EGR2 and type I IFN signature scores in unstimulated CD19 CAR T-cell products from CLL patients (CR: complete response, PRTD: very good partial response, PR: conventional partial response, NR: no response). H, Comparison of type I IFN and EGR2 module scores in PSMA CAR T-cells from lymphodepleted prostate cancer patients, and their association with in vivo CAR T-cell proliferation and PSA decline.

BACKGROUND:Relapse of B-cell acute lymphoblastic leukemia (B-ALL) with CD19-antigen loss after CD19-targeted chimeric antigen receptor (CAR) T-cell therapy has a dismal prognosis. Novel immunotherapeutic strategies for this patient population are urgently needed. METHODS:We tested a novel, fully human anti-CD22/4-1BB CAR T-cell construct, CART22-65s, in parallel phase I studies for pediatric and adult B-ALL. After lymphodepletion, CART22-65s was infused using a 3-day fractionated dosing scheme, allowing for omission of the second and third doses in cases of early cytokine release syndrome (CRS). RESULTS:Twenty-two patients, all with relapse after prior CD19-directed immunotherapy, were enrolled. Of 19 infused patients (pediatric, n=17; adult, n=2), 14 (74%) achieved a complete remission (CR), including 4 of 6 (67%) patients refractory to prior inotuzumab. Five of 14 patients in a CR proceeded to consolidative hematopoietic cell transplantation (HCT). With a median follow-up of 38 months, the 12-month relapse-free survival rate was 38.4% (95% CI 19.3% to 76.5%) and overall survival rate was 52.6% (95% CI 34.3% to 80.6%). Two patients received additional CART22-65s treatments for subsequent CD22-positive relapses; one achieved another CR. All CRS (n=17, 89%) and neurotoxicity (n=4, 21%) events after initial infusion were grades 1-2. The only grade 3 CRS/neurotoxicity and the only high-grade immune effector cell-associated hemophagocytic lymphohistocytosis-like syndrome occurred in the retreatment setting. In vivo cellular kinetic data revealed robust CART22-65s proliferation by quantitative PCR peaking at a median of 20 days postinfusion, with the cells persisting out to month 42 in one patient who achieved a long-term remission with CART22-65s alone. CONCLUSIONS:The favorable safety profile and high remission rates in exceedingly refractory B-ALL support the continued development of CART22-65s but also highlight the need to use the product in combination with HCT or other novel strategies. TRIAL REGISTRATION NUMBERS:NCT02650414 and NCT03620058.

Patients with advanced pancreatic ductal adenocarcinoma (PDAC) have a median survival of less than a year, highlighting the urgent need for treatment advancements. We report on a phase 1 clinical trial assessing the safety and feasibility of intravenous and local administration of anti-mesothelin CAR T cells in patients with advanced PDAC. While therapy is well tolerated, it demonstrates limited clinical efficacy. Analyses of patient samples provide insights into mechanisms of treatment resistance. Single-cell genomic approaches reveal that post-infusion CAR T cells express exhaustion signatures, including previously identified transcription factors ID3 and SOX4, and display enrichment for a GZMK+ phenotype. Single knockout of ID3 or SOX4 enhances efficacy in xenograft models, though with donor-dependent variability. However, single-knockout cells eventually fail. Conversely, ID3 and SOX4 double-knockout CAR T cells exhibit prolonged relapse-free survival, demonstrating a sustained therapeutic effect and a potential avenue for engineering more potent CAR T cells in PDAC. This study was registered at ClinicalTrials.gov (NCT03323944).

Analysis of survival outcomes and EGR2 gene expression in CD19 CAR T-cell products. The figure presents the P values and hazard ratio of different EGR2 molecular marker stratification points in relation to A, overall survival and B, event-free survival The black arrows indicate the stratification points used in the study. C, EGR2-targeted gene expression scores in CD19 CAR T-cell products from responders and non-responders in pediatric ALL. D, Summary of how EGR2 regulates resistance to CAR T-cell therapy through the type I IFN pathway.

Marker gene expression in CAR T-cell clusters. A, Uniform manifold approximation and projection (UMAP) plot of AAVS1 and EGR2 knockout (KO) CAR T-cell samples is shown. B, UMAP plots showing expression levels of CD4 and CD8. C, Violin plot depicting expression of cluster-defining markers in CD4+ T-cells. D, Differentially expressed genes in IL7R+ versus CTLA4+ CD4+ T-cells. E, Violin plots showing expression levels of cluster-defining markers in CD8+ T-cells. F, Heatmap displaying differentially expressed genes between CD8+ cell clusters. G, Cell cycle scores mapped on UMAP plots.

Epigenetic remodeling of CAR T-cells by EGR2 knockout and effect of type I IFN signaling on the development of memory and exhaustion. A, Volcano plots showing differentially accessible chromatin regions within genes between KLF2+ and MKI67+ CD8+ T-cells. B, Volcano plots depicting differentially accessible chromatin regions within genes between EGR2 and AAVS1 knockout (KO) CD8+ CAR T-cells. C, Representative contour plots showing frequencies of TIM3- and LAG3-expressing CD8+ CAR-T cells after exposure to IFNβ (1ng/mL) following chronic CAR stimulation. D, Proportions of CD27+ (left) or CD62L+ (right) CD8+ CAR-T cells after exposure to IFNβ. E, Representative contour plots showing frequencies of CD45RO+CD27+ CD8+ CAR-T cells after IFNAR blockade (Anifrolumab, 1µg/mL) during chronic antigen stimulation. F, Frequencies of TIM3+LAG3+ CD8+ CAR-T cells after IFNAR blockade. G, Cytolytic capacity of CAR T-cells as measured by normalized cell index kinetics using the xCELLigence real-time cytotoxicity assay following chronic stimulation with target cancer cells in the setting of either IFNβ or IFNAR blockade. H, Normalized cell index at 75 hours after challenge with target cancer cells. All experiments were conducted using healthy donor T-cells from independent donors (Mann-Whitney test, n = 4). *P < 0.05, *P < 0.01, ***P < 0.001, ns.: not significant.

EGR2 knockout enhances the long-term potency and memory function of CAR T-cells in vivo. A, Schematic of the tumor rechallenge experiments. B, Tumor growth after the first (left) and second (right) NALM6 challenges (Mann-Whitney test, n = 10). C, Mouse survival after high-dose second tumor challenge (Log-rank test). D, Tumor growth inhibition in a prostate cancer model using PSMA-targeting CAR T-cells with EGR2 or AAVS1 knockout (KO) compared to control CD19 CAR T-cells (19BBz). The tumor growth was monitored over time using caliper-based measurements in NSG mice (n = 8).

Genes comprising the TCF7 regulon, the type I IFN signature, and the EGR2 Co-expression signatures. Gene names and symbols are listed.

Pathways regulated by EGR2 in CD8+ CAR T-cells. (A-C) Top pathways differentially expressed in EGR2 knockout CD8+ CAR-T cells compared to AAVS1 knockout CAR T-cells. Libraries used in this enrichment analysis: A, Reactome 2016. B, NCI-Nature Pathway Interaction Database 2015. C, ARCHS4 transcription factor (TF) co-expression.

Long-term risks of gene therapy are not fully understood. In this study, we evaluated safety outcomes in 783 patients over more than 2,200 total patient-years of observation from 38 T cell therapy trials. The trials employed integrating gammaretroviral or lentiviral vectors to deliver engineered receptors to target HIV-1 infection or cancer. Eighteen patients (2.3

BACKGROUND:Chimeric antigen receptor (CAR) T cells targeting CD19 have transformed the treatment of B-cell cancers, but many patients do not have long-term remission. We designed an anti-CD19 enhanced (armored) CAR T-cell product (huCART19-IL18) that secretes interleukin-18 to enhance antitumor activity. METHODS:In this study, we assessed the safety, feasibility, and preliminary efficacy of huCART19-IL18 in patients with relapsed or refractory lymphoma after previous anti-CD19 CAR T-cell therapy. Using a 3-day manufacturing process, we administered huCART19-IL18-positive cells in doses ranging from 3×106 to 3×108. RESULTS:A total of 21 patients received huCART19-IL18. Cytokine release syndrome occurred in 62% of the patients (47% with grade 1 or 2), and immune effector-cell-associated neurotoxicity syndrome occurred in 14% (all grade 1 or 2). No unexpected adverse events were observed. Robust CAR T-cell expansion was detected across all dose levels. At 3 months after infusion, a complete or partial response was seen in 81% of the patients (90% confidence interval [CI], 62 to 93) and a complete response in 52% (90% CI, 33 to 71). With a median follow-up of 17.5 months (range, 3 to 34), the median duration of response was 9.6 months (90% CI, 5.5 to not reached). CONCLUSIONS:In this small study, huCART19-IL18 had a safety profile consistent with other CAR T-cell treatments and showed promising efficacy at low cell doses in patients with lymphoma after the failure of previous anti-CD19 CAR T-cell therapy. (ClinicalTrials.gov number, NCT04684563.).

Chromatin accessibility profiles for EGR2 knockout (KO) and AAVS1 knockout (AAVS1) CD8+ CAR T-cells. The table provides information on the locus mapping, p-value, average log2 fold change, percentage of cells, adjusted p-value, transcript ID, gene name, gene ID, gene biotype, region type, closest region, query region, and distance.

Gene expression and pathway enrichment analysis of CD8+ T cell clusters. A, Heatmap showing differentially expressed genes between memory-like KLF2+ and exhausted-like MKI67+ CD8+ T-cells. Gene signature scores related to cell cycle and clinical response are indicated on the top bars. B, Top downregulated GO biological processes in EGR2 compared to AAVS1 knockout CAR T-cells.

102 Background: Outcomes in patients with rGBM are poor, with historical median overall survival (OS) of 6-9 months. Here we report the results from the dose exploration phase of a phase 1 trial investigating ICV-delivered, bivalent CAR T-cells targeting EGFR epitope 806 and IL13Rα2 in rGBM. Methods: Patients with EGFR-amplified GBM that was recurrent/progressive following front-line radiotherapy were enrolled using a 3+3 design (dose levels: 5.0 x 10 6 , 1.0 x 10 7 , and 2.5 x 10 7 cells). Patients underwent surgery for (1) maximal safe resection and confirmation of viable tumor and (2) Ommaya reservoir placement. Post-operatively, patients received a single ICV dose of CART-EGFR-IL13Rα2 cells without lymphodepleting chemotherapy. Primary endpoints included dose-limiting toxicity (DLT) and determination of the maximum tolerated dose (MTD). Secondary endpoints included objective radiographic response, progression-free survival (PFS), and OS. Serial CSF samples were analyzed for CAR T-cell pharmacokinetics and single-cell RNA sequencing (scRNAseq). Results: Eighteen patientsreceived CART-EGFR-IL13Rα2 cells (n=6 per dose level). Median age was 57, 15 (83%) were male, 13 (72%) had MGMT unmethylated tumors, and 7 (39%) had >1 prior relapse. One DLT (grade 3 lethargy/fatigue) was observed at the MTD (2.5 x 10 7 cells). Acute neurotoxicity related to CAR T-cells occurred in all patients. Using immune effector cell-associated neurotoxicity syndrome (ICANS) grading, 10 of 18 patients (56%) experienced grade 3 neurotoxicity; none had grade 4-5 neurotoxicity. Using tumor-inflammation associated neurotoxicity (TIAN) grading, 2 of 18 patients (11%) had grade 3 and 1 patient (6%) had grade 4 neurotoxicity. Grade 1-2 fever occurred in all patients. Eleven of 13 patients (85%) with measurable disease at time of CAR T-cell infusion experienced tumor shrinkage, ranging from 1-62% reductions (median 35%, IQR 12 – 39%) in target lesions and with one confirmed partial response by modified RANO criteria. PFS and OS continue to mature and will be presented. CAR T-cell expansion in CSF was robust with a dose-response relationship observed. The CAR transgene remained detectable in CSF and blood 12 months post-CART infusion in a patient experiencing durable stable disease lasting for 17 months (ongoing at data cut-off). In patients undergoing repeat resection following treatment, CART-EGFR-IL13Rα2 cell infusion markedly increased the number of tumor-infiltrating lymphocytes. scRNAseq in post-treatment CSF revealed higher cytotoxicity and exhaustion scores in CD8+ CAR T-cells as compared to the infusion product, indicative of target cell engagement. Conclusions: ICV delivery of CART-EGFR-IL13Rα2 is feasible and appears safe. CART-EGFR-IL13Rα2 cells are bioactive and exhibit an encouraging early efficacy signal in rGBM. Clinical trial information: NCT05168423 .