Combined Immunotherapy Improves Outcome for Replication-Repair-Deficient (RRD) High-Grade Glioma Failing Anti-PD-1 Monotherapy: A Report from the International RRD Consortium

Abstract

Immune checkpoint inhibition (ICI) is effective for replication-repair-deficient, high-grade gliomas (RRD-HGG). The clinical/biological impact of immune-directed approaches after failing ICI monotherapy is unknown. We performed an international study on 75 patients treated with anti-PD-1; 20 are progression free (median follow-up, 3.7 years). After second progression/recurrence (n = 55), continuing ICI-based salvage prolonged survival to 11.6 months (n = 38; P < 0.001), particularly for those with extreme mutation burden (P = 0.03). Delayed, sustained responses were observed, associated with changes in mutational spectra and the immune microenvironment. Response to reirradiation was explained by an absence of deleterious postradiation indel signatures (ID8). CTLA4 expression increased over time, and subsequent CTLA4 inhibition resulted in response/stable disease in 75%. RAS-MAPK-pathway inhibition led to the reinvigoration of peripheral immune and radiologic responses. Local (flare) and systemic immune adverse events were frequent (biallelic mismatch-repair deficiency > Lynch syndrome). We provide a mechanistic rationale for the sustained benefit in RRD-HGG from immune-directed/synergistic salvage therapies. Future approaches need to be tailored to patient and tumor biology.

Significance: Hypermutant RRD-HGG are susceptible to checkpoint inhibitors beyond initial progression, leading to improved survival when reirradiation and synergistic immune/targeted agents are added. This is driven by their unique biological and immune properties, which evolve over time. Future research should focus on combinatorial regimens that increase patient survival while limiting immune toxicity. This article is featured in Selected Articles from This Issue, p. 201.

Figure 1.

Patients with RRD-HGG treated with ICI (n = 75). A, Cohort characteristics. B, Flow of patients with RRD-HGGs treated with ICIs and salvage regimens. MEK-i, MEK inhibitor. C, Progression-free (PFS1) and overall survival (OS1) on anti–PD-1/PDL1 monotherapy. D, Overall survival (OS2) for 55 patients progressing on monotherapy stratified by continued (n = 38) or no ICI (n = 17). PPAP, polymerase proofreading associated polyposis syndrome; CMMRD, constitutional mismatch repair deficiency syndrome.

Figure 2.

Genomic features of patients with RRD-HGG treated with salvage therapies (n = 38). A, Onco-plot summarizing clinical and genomic features of patients who progressed on anti–PD-1 monotherapy and continued ICI treatment. B, Impact of TMB on survival. C, Impact of genomic MSI as measured by the tumor MMRDness score on survival (Methods).

Figure 3.

Patients receiving dual-checkpoint inhibition with ipilimumab and anti–PD-1 after failing ICI monotherapy (n = 24). A, Swimmer's plot for each patient, showing the best documented radiologic response at any time during treatment. Inset, representative radiologic image showing response to dual-checkpoint inhibition. B, Progression-free (PFS2) and overall survival (OS2). C, Normalized CTLA4 expression counts generated using NanoString platform (Methods) for in-house non-RRD, nonmalignant brain controls, and at different time points for RRD-HGG. D, Paired analysis of normalized CTLA4 expression counts before and after anti–PD-1 therapy for the same patient. E, Toxicities (≥CTCAE grade 3) observed in CMMRD and Lynch syndrome patients on dual ICI.

Figure 4.

Patients treated with nivolumab and trametinib (n = 5). A, Objective responses in bifocal glioblastoma progressing on nivolumab. B, Swimmer's plot summarizing events in each patient. C, The patient progressing at primary and distant sites on nivolumab had complete response on being simultaneously treated with radiation and addition of trametinib to nivolumab. T-cell receptor (TCR) clonotype analysis shows an initial increase in TCRB after starting on nivolumab, and invigoration of response at salvage treatment postprogression. This was observed both in terms of increased absolute clonal counts after adding trametinib, as well as clonal selection, plausibly for more tumor-specific TCR clones (as shown by the colored bars), as this correlated with the response seen in radiology. MEK-i, MEK inhibitor.

Figure 5.

Impact of radiotherapy. A, Patient progressing on nivolumab received reirradiation, following which further progression prompted the addition of ipilimumab. This was followed by radiologic flare, and continued treatment with supportive care led to delayed response. B, Impact of reirradiation on survival. C, Impact of additional reirradiation (RT) in patients who received nivolumab and ipilimumab. D, Relative contribution of the radiation-induced indel signature (ID8) in RRD-HGG and controls from the GLASS cohort (Methods). E, Paired analysis of the relative contribution of ID8 before (at diagnosis) and after the second progression (post primary radiation, and then immunotherapy at first progression) in five patients.

Figure 6.

Delayed response after initial progression on continued anti–PD-1 monotherapy. A, Delayed response after initial progression on nivolumab monotherapy. B, Cyst decompression and biopsy at progression showed a more mitotically active and higher grade tumor with higher TMB and increased CD8 T-cell infiltration. C, Mutation overlap showed minimal overlap of the somatic mutational spectrum between the two time points (200X magnification; scale bars represent 50 µm). D, Normalized expression counts of immune markers before initiation and after progression on anti–PD-1 monotherapy using the NanoString platform (100X magnification, scale bars represent 100 µm; Methods).