428PPembrolizumab (Pem) in recurrent high-grade glioma (HGG) patients with mismatch repair deficiency (MMRd): An observational study

Abstract Background Pem, an immune checkpoint inhibitor, has been demonstrated to be active in various neoplasms with MMRd. No data exist about its efficacy in MMRd glioma patients. Methods MMRd HGG patients relapsed after receiving RT and CT were treated with Pem. MMR status was analyzed by immunohistochemistry, including the MLH1, MSH2, MSH6, and PMS2 markers. MMR deficiency was defined as presence of a weak (wMMRd) or absent (aMMRd) signal on immunohistochemistry for at least one MMR protein. Other inclusion criteria were: ECOG PS 0-2, histologically confirmed gliomas, dexamethasone ≤4mg. Pem was administered at 200mg every 3 weeks until disease progression or unacceptable toxicity. Tumor response was evaluated by brain MRI every 10 weeks according to the RANO criteria. OS and PFS were evaluated by Kaplan-Meier curves. CTCAE v4.0 was used for toxicity assessment. Results Among 167 glioma patients, we found 22 MMRd gliomas. 12 PTS were treated with Pem: 8 wMMRd and 4 aMMRd. According to Bethesda criteria, all PTS had microsatellite stability. Tumor histologies included 5 anaplastic astrocytoma, 1 anaplastic oligodendroglioma, 6 glioblastoma (GBM). MSH2 deficiency was found in 6 cases, MSH6 deficiency in 9 cases, PMS2 and MLH1 deficiency in 2 cases. Median number of prior lines of chemotherapy was 1 (range 1-5). Stable disease (SD) was reported in 4 PTS (33%); 8 PTS showed progressive disease (PD). PTS with anaplastic gliomas showed a statistically significant association with SD (p=0.03, OR=3); all GBM PTS reported PD; status of MMRd (weak/absent), IDH (mutated/wild-type), MSH2 and MLH6 (deficient/proficient) were not associated with SD. Median follow up was 14.7ms. OS was 5.6ms (95% CI 0.1-13.8), PFS 2.4ms (95% CI 1.8-2.9). OS was 2.8ms and 5.6ms (p=0.9), PFS was 1.8ms and 3.1ms (p=0.5) in PTS with wMMRd and aMMRd. PTS reporting SD and PD had PFS of 7.4ms (95% CI 4.6-10.2) and 1.8ms (95% CI 0.2-3.4), p=0.002; OS was "not reached" and 2.8ms in PTS having SD vs PD (p=0.04). Grade ≥3 adverse events were reported in 8% of PTS. Conclusions A subgroup of recurrent MMRd HGG patients might benefit from Pem, especially those with anaplastic gliomas. There was a trend for a longer PFS and OS in PTS with aMMRd. Analyses for identifying additional molecular predictive factors is ongoing. Legal entity responsible for the study The authors. Funding Has not received any funding. Disclosure All authors have declared no conflicts of interest.