Circulating PD-L1 levels change during bevacizumab-based treatment in recurrent glioma

PURPOSE: In primary brain tumors, the efficacy of immune-modulating therapies is still under investigation as inflammatory responses are restricted by tight immunoregulatory mechanisms in the central nervous system. Here, we measured soluble PD-L1 (sPD-L1) in the plasma of patients with recurrent glioblastoma (GBM) and recurrent WHO grade II–III glioma treated with bevacizumab-based salvage therapy. METHODS: Thirty patients with recurrent GBM and 10 patients with recurrent WHO grade II–III glioma were treated with bevacizumab-based salvage therapy at the Medical University of Vienna. Prior to each treatment cycle, EDTA plasma was drawn and sPD-L1 was measured applying a sandwich ELISA with a lower detection limit of 0.050 ng/ml. Leukocyte counts and C-reactive protein (CRP) levels were measured according to institutional practice. RESULTS: Median number of sPD-L1 measurements was 6 per patient (range: 2–24). At baseline, no significant difference in sPD-L1 concentrations was observed between WHO grade II–III glioma and GBM. Intra-patient variability of sPD-L1 concentrations was significantly higher in WHO grade II–III glioma than in GBM (p = 0.014) and tendentially higher in IDH-mutant than in IDH-wildtype glioma (p = 0.149) In WHO grade II–III glioma, sPD-L1 levels were significantly lower after one administration of bevacizumab than at baseline (median: 0.039 ng/ml vs. 0.4855 ng/ml, p = 0.036). In contrast, no significant change could be observed in patients with GBM. CONCLUSIONS: Changes in systemic inflammation markers including sPD-L1 are observable in patients with recurrent glioma under bevacizumab-based treatment and differ between WHO grade II–III glioma and GBM. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s00262-021-02951-2.