Longer follow-up from a phase I trial of intracerebroventricular bivalent CAR T-cell therapy in recurrent glioblastoma showed no long-term or delayed-onset toxicities, with a median overall survival of 12.0 months, according to updated data presented at the 2026 ASCO Annual Meeting.

The investigational CAR T-cell therapy, CART-EGFR-IL13Rα2, targets epidermal growth factor receptor (EGFR) epitope 806 and IL13Rα2. At a median follow-up of 19.1 months, 5 of 18 patients (28%) were alive beyond 18 months. One patient had an ongoing partial response at 33 months.

“We observed cytokine-release syndrome and acute neurotoxicity in 100% of patients,” said Stephen J. Bagley, MD, MSCE, section chief of Neuro-Oncology and associate professor of Hematology-Oncology and Neurosurgery at the Perelman School of Medicine at the University of Pennsylvania. However, he reported that neurologic function returned to pretreatment baseline between days 4 and 7 after CAR T-cell infusion. With longer follow-up, there was no evidence of late-onset or unexpected toxicities, including on-target/off-tumor toxicity or secondary cancers.

Phase I Trial Design

The updated analysis included 18 patients with recurrent EGFR-amplified glioblastoma treated in a 3+3 phase I trial. Patients received a single intracerebroventricular dose of autologous bivalent CAR T cells delivered through an Ommaya reservoir. Lymphodepleting chemotherapy was not used.

The study evaluated dose levels of 5.0 × 10⁶, 1.0 × 10⁷, and 2.5 × 10⁷ cells. The maximum tolerated dose was established at 2.5 × 10⁷ cells.

The trial was previously reported after a median follow-up of 8.1 months. In that initial report, cytokine-release syndrome and acute neurotoxicity occurred in all patients. Among 13 patients with measurable disease, 8 (62%) had some reduction in tumor size, and 2 had partial responses by Response Assessment in Neuro-Oncology criteria, for an objective response rate of 15%.

The current update, with a data cutoff of January 16, 2026, focused on longer-term safety, updated overall survival, and neurologic function.

No Delayed Toxicity Observed

With longer follow-up, one patient (5%) had prolonged grade 1 neurotoxicity lasting 10 months, as previously reported. No other cases of prolonged neurotoxicity were observed.

Neurologic function worsened transiently after infusion, as measured by the Neurologic Assessment in Neuro-Oncology (NANO) scale, a clinician-reported assessment of nine domains in which higher scores indicate worse function. In this study, NANO scores were assessed before CAR T-cell therapy and repeatedly after infusion, including on days 1, 4, 7, 10, 14, and 21 and at 1 and 2 months.

NANO scores increased during the acute neurotoxicity period after infusion but returned to baseline between days 4 and 7. Scores remained close to baseline at 1 and 2 months.

Updated Survival and Long-Term Responder

At a median follow-up of 19.1 months, median overall survival was 12.0 months (95% confidence interval = 7.4–23.2 months). Five of 18 patients were alive beyond 18 months, corresponding to an 18-month overall survival rate of 28%.

Dr. Bagley also noted that two additional patients were alive at 16 and 17 months, respectively, at the time of the data cutoff.

One patient had an ongoing partial response at 33 months. The patient, a 60-year-old man with MGMT-methylated glioblastoma, had received standard Stupp protocol therapy in 2021, followed by lomustine and bevacizumab at first recurrence. At second recurrence, he enrolled in the bivalent CAR T-cell trial but experienced rapid disease progression with leptomeningeal involvement and abrupt clinical decline between pre–CAR T-cell surgery and CAR T-cell infusion.

He received a single intracerebroventricular dose of 10 million CAR T cells on August 16, 2023, and achieved a partial response as his best response, which remained ongoing at 33 months. CAR T cells remained detectable in both cerebrospinal fluid and peripheral blood at 24 months.

During the discussion, Dr. Bagley was asked whether leptomeningeal disease may have contributed to the durable response. He said he suspected it may have played a role, possibly reflecting a different immune environment. Still, he cautioned that the patient entered the trial in very poor clinical condition, underscoring the difficulty of treating patients with advanced disease.

Retreatment and Next Steps

In response to a question about antigen loss at progression, Dr. Bagley said antigen loss did not appear to be the primary mechanism of resistance after initial CAR T-cell therapy. Most patients who relapsed maintained EGFR amplification and target alterations when present, he said.

Seven of the 18 patients were retreated with CAR T cells at progression. According to Dr. Bagley, immune-mediated rejection of the cells appeared to be a key limiting factor, providing a rationale for adding lymphodepleting chemotherapy in an ongoing trial in recurrent glioblastoma.

Several related studies are underway or planned. Accrual was recently completed in a study of CART-EGFR-IL13Rα2 cells in newly diagnosed glioblastoma (ClinicalTrials.gov ID: NCT06973096). A trial in recurrent glioblastoma evaluating the addition of lymphodepleting chemotherapy is currently enrolling (ClinicalTrials.gov ID: NCT07209241), and a second-generation version of the product, armored with a dominant-negative TGF-β receptor, is also in development for a phase I trial.

DISCLOSURE: Dr. Bagley reported honoraria from Jazz Pharmaceuticals and Physicians’ Education Resource; consulting or advisory roles with Modifi Bio, Nexus NeuroTech, Servier, Telix Pharmaceuticals, and Third Rock Ventures; institutional research funding from GlaxoSmithKline, Incyte, Lilly, and Novocure; research funding from Kite, a Gilead company; and intellectual property as co-inventor of University of Pennsylvania intellectual property licensed to Novartis: Combination Therapies of EGFRvIII Chimeric Antigen Receptors and PD-1 Inhibitors, U.S. Patent No. 62/809,245.

REFERENCE

1. Hollander EE, Desai AS, Marshall A, et al: Updated overall survival, safety, and neurologic function outcomes from a phase 1 trial of bivalent chimeric antigen receptor (CAR) T-cell therapy in recurrent glioblastoma (GBM). 2026 ASCO Annual Meeting. Abstract 2013. Presented May 31, 2026.