ID #698 Therapeutic DC vaccines combined with double checkpoint blockade in pediatric HGG: Results of the HIT-HGG Rez Immunovac trial.

Abstract

Children, adolescent and young adults (CAYA) with relapsed high-grade gliomas (rHGG) share a dismal prognosis. Singular immunotherapeutic interventions like therapeutic vaccines have shown safety and immunogenicity, yet limited clinical efficacy. In the HIT-HGG Rez Immunovac phase I/II study (Eudra-CT 2013-000419-26) we optimzed induction and persistance of immune responses by combining upfront Treg-depletion using metronomic cyclophosphamide followed by therapeutic dendritic cell-based vaccines and subsequent checkpoint inibitor blockade (4x Nivolumab 3 mg/kg and Ipilimumab 1 mg/kg q3wk followed by Nivo mono 6 mg/kg q4wk for up to one year). Primary efficacy endpoint was to reach a 6-month post-relapse overall survival (prOS) of at least 82%. Twenty-five CAYAs with a mean age of 13.8 years [4.3-19.5] were enrolled between 2018-2024. Five patients were excluded due to progression before vaccine start or insufficient tumor material, 20 patients received therapy as per protocol. Fifteen SAEs were noted with no unexpected toxicities. The primary end point was confirmed with a 6-month prOS of 84% and one patient being alive after 31.2 months follow-up. PrOS was significantly improved over a historical control group (HCG, n = 87) from the HIT-HGG database (median OS 9.0 months, range 3.3-31.2 vs. 4.98 months, range 0.26-47.77 in Immunovac vs. HCG patients respectively, p=0.0003). Patients with stable disease at the end of vaccination had a better prOS than those relapsing during the vaccine schedule. Immunomonitoring demonstrated a decline in Tregs and a shift from naïve to memory T-cell subsets under treatment. Vaccine-specific CD4+ and CD8+ T-cell responses were observed in almost all patients, including responses against tumor-specific HLA-class I/II epitopes from the respective immunopeptidomes. Furthermore, spatial omics using cyclic immunofluorescence imaging technology at baseline revealed an immune-cold tumor microenvironment (TME) dominated by myeloid cells (∼50%), predominantly M2-polarized macrophages. T- and NK-cells contributed ∼10% and 15%, respectively. Immune effector cells were spatially dispersed, displayed an exhausted phenotype and were typically embedded within immunosuppressive glial niches or stromal compartments. The tumor compartment exhibited aggressive features, including high Ki67, strong GFAP expression, and frequent p53 alterations, but showed robust HLA-class II expression, revealing potential vulnerability to CD4+ T-cell-mediated responses. In conclusion, our data demonstrate safety and clinical efficacy of an optimized immunotherpeutic regime in CAYA with rHGG. Our immunomonitoring data show substantial T-cell responses and changes in the peripheral immune compartments. Together with insights from the TME these results will shape the design of future immunotherapy trials in HGG.