A phase 1 study of intracerebroventricular (ICV) delivery of bivalent chimeric antigen receptor (CAR) T-cells targeting EGFR and IL13Ra2 in patients with recurrent glioblastoma (rGBM).

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 .