LAG-3 Plus PD-1 Blockade Shows Early Immune Signal in Recurrent Glioblastoma

Once glioblastoma recurs, treatment options are limited; immune checkpoint inhibitors, which have reshaped the treatment landscape in other cancers, have yet to deliver a clear survival improvement in this brain tumor. Now, a multicenter phase 1 trial shows that simultaneously releasing the LAG-3 and PD-1 immune brakes may alter the immune state within the tumor and has also revealed a survival signal worthy of further validation.

This open-label study enrolled 46 patients with a first recurrence who had previously received radiotherapy and temozolomide. Using sequential cohorts, 23 received the anti-LAG-3 antibody relatlimab, while another 23 received relatlimab plus the anti-PD-1 antibody nivolumab. The trial’s primary objective was not to compare efficacy, but to identify tolerable doses; the doses established by the study were relatlimab 800 mg as monotherapy and relatlimab 160 mg plus nivolumab 240 mg.

Safety results differed between the two groups. No grade 3–4 treatment-related adverse events occurred in the monotherapy group; in the combination group, six patients experienced events of this severity, including cerebral edema, decreased muscle strength, hypertension, syncope, and thyroiditis. The research team assessed the overall safety profile as acceptable, but serious events occurred in about one-quarter of those receiving combination therapy, indicating that future studies will still need to carefully weigh immune activation against neurological risks.

The trial also included a “window-of-opportunity” surgical component: 13 patients received one dose of treatment approximately 10 days before tumor resection, allowing researchers to directly examine tissue after drug exposure. Compared with existing tumor samples, tumors treated with either monotherapy or combination therapy contained more CD8-positive T cells; combination therapy was also accompanied by interferon signaling and activation of genes related to antigen presentation, showing that the drugs could indeed affect the highly immunosuppressive tumor microenvironment.

However, the increase in T cells itself did not directly correspond to longer survival. An exploratory analysis found that the small number of patients who survived longer already had stronger interferon signaling and higher T-cell clonality before treatment, suggesting that whether a tumor already possesses an immune response that can be reawakened may be more important than simply increasing the number of immune cells. Because the relevant tissue analyses covered only a very small number of patients, these features cannot yet serve as biomarkers for treatment selection.

In terms of clinical outcomes, the one-year overall survival rate was 34.8% in the relatlimab monotherapy group and 52.2% in the combination group; median overall survival was 16.7 months in the preoperative-treatment subgroup receiving combination therapy. However, the study was nonrandomized, used sequential cohorts, and its sample size was not designed for an efficacy comparison; subsequent radiotherapy, chemotherapy, or other immunotherapy received by patients may also have affected survival. Central imaging review confirmed no complete or partial tumor responses, and delayed imaging changes may also have included pseudoprogression caused by immunotherapy.

Therefore, the study’s most robust conclusion is that LAG-3 blockade can be administered in patients with recurrent glioblastoma and that preoperative dosing provided biological evidence that the drug entered the tumor immune environment. Whether dual blockade can truly prolong life, and which patients are most likely to benefit, must still be answered by larger follow-up trials with control groups.