A phase I clinical trial on intracranial administration of autologous myeloid dendritic cells (myDC) in combination with ipilimumab and nivolumab in patients with recurrent glioblastoma (rGB).

2033 Background: Intracerebral administration of ipilimumab (IPI) and nivolumab (NIVO) following resection of rGB was demonstrated to be safe and resulted in encouraging survival (Duerinck, Schwarze et al. JITC 2021; Neyns et al. ESMO 2021). CD1c(BDCA-1) + and CD141(BDCA-3) + myDC play a pivotal role in initiating an adaptive anti-tumor immune response by re-licensing cytotoxic T lymphocytes within the tumor microenvironment. Methods: Eligible patients (pts)(diagnosed with rGB following radiation and temozolomide treatment; not in need of steroids) underwent a leukapheresis followed by immunomagnetic bead isolation and cryopreservation of CD1c (BDCA-1) + / CD141(BDCA-3) + myDC. At the time of surgery, an escalating number of myDC (1, 10, and 20x10 6 myDC) were injected into the brain tissue lining the resection cavity following maximal safe resection of the rGB (ICer) or injected intratumorally (ITum) following stereotactic biopsy (STx). IPI (5 mg) plus NIVO (10 mg) were co-injected with myDC. NIVO was administered intracavitary (ICav, 10mg) using an Ommaya port and intravenously (IV, 10mg) Q2w (max 12x). Results: Fourteen pts (9 male; median 48y [range 20-78]) were recruited (resection n = 11; STx n = 2) and underwent a successful leukapheresis and isolation of myDC; peroperative administration of myDC was preceded by resection in 10 pts (1 pt did not undergo surgery due to clinical deterioration/cerebral edema), and by STx in 2 pts. Respectively 6 (incl both pts who underwent a STx), 3, and 4 pts were treated at the 3 dose levels. All pts received ITum/ICer/IV-administrations of IPI and NIVO as planned. Median number of postoperative ICav/IV NIVO-administrations was 7 (range 2-11). Most frequent adverse events (AE) were headache (n = 11), fatigue (n = 9), transient dysphasia (n = 6), and nausea (n = 5). Bacterial colonization of the Ommaya occurred in 3 pts necessitating removal. Immune-related AE were infrequent and mild. No G5 AE occurred. No dose-limiting toxicities were seen with increasing numbers of myDC. After a median follow-up of 54w, 3 pts remain progression-free (after 42+, 51+, 54+ weeks of FU), 6 (46%) pts have died; median PFS is 13w (95% CI 0-26), median OS has not been reached; 6-months PFS- and OS-rate are respectively 30% and 84%, 12-months PFS- and OS-rate are respectively 23% and 51%. OS compares favorably to an historical cohort of Belgian rGB patients (n = 469; Log-Rank p = 0.018). Gene expression profiling of resected tissue, analysis of cellular counts, cytokines, NIVO/IPI-concentrations in on-treatment cerebrospinal fluid samples is ongoing. Conclusions: Intracranial administration of autologous myDC plus IPI and NIVO in combination with IV NIVO was found to be feasible, sufficiently safe and associated with encouraging survival justifying further investigation in pts with resectable rGB. Clinical trial information: NCT03233152.