Meeting update—Society for Neuro-Oncology 2016 annual meeting Free

2016 SNO Annual Meeting Summary

The 21st annual scientific meeting of the Society for Neuro-Oncology (SNO) occurred November 17–20, 2016 in Scottsdale, Arizona. This year’s meeting was one of the most heavily attended annual meetings, with 2166 attendees at the main meeting representing 42 countries. Over 970 attendees participated in the SNO Education Day, which focused on precision medicine—the current state and future directions for precision approaches to brain tumor clinical trials and care. Abstract submissions for this year’s meeting reached a new record, with 1024 total submissions, up from 914 in 2015. Webcast, e-Posters, and i-Talks offered new interactive ways to experience meeting content. In addition to the 640 traditional posters, 219 oral presentations and 54 e-Talks were presented. “Best of the Day” video highlights, the SNO Twitter feed, and the SNO annual meeting app helped connect attendees and link to important updates. Pulitzer Prize‒winning author and hematologist/oncologist Dr Siddhartha Mukherjee delivered a keynote address highlighting successes and challenges in the field. The following summarizes just a few of the many important updates and advances presented at this year’s meeting.

2016 WHO Classification of Brain Tumors

For the first time, the 2016 World Health Organization (WHO) Classification of Tumors of the Central Nervous System combines molecular and histologic parameters to define brain tumors. In his keynote presentation, Dr David Louis (Boston, MA) reviewed the restructuring in the 2016 WHO Classification, including major changes to diffuse gliomas, medulloblastomas, and embryonal tumors as well as several other additions and removals. In this new classification, infiltrating astrocytomas and oligodendrogliomas are grouped into the family of “diffuse gliomas,” which are defined by molecular alterations in the isocitrate dehydrogenase 1 (IDH1) gene and chromosome 1p/19q codeletion. Newly recognized entities include diffuse midline glioma H3K27M (WHO grade IV pediatric glioma), v-rel avian reticuloendotheliosis viral oncogene homolog A (RELA) fusion-positive ependymoma, and epithelioid glioblastoma multiforme (GBM). Gliomatosis cerebri, among other entities, was removed.

Epigenetic Mechanisms of Glioma

Mutations in the IDH1 gene are common in low-grade gliomas (LGGs). Keynote presentations from Drs Joseph Costello (San Francisco, CA) and Bradley Bernstein (Boston, MA) highlighted genetic and epigenetic mechanisms underlying IDH1 mutant glioma progression. Dr Costello demonstrated striking parallels between the branching tumor evolution observed using both genetic (DNA sequencing) and epigenetic (DNA methylation) data in paired primary and recurrent LGG specimens. Although the great majority of LGGs retain IDH1 mutation upon progression/recurrence, both deletion and amplification of the IDH mutant allele coincided with malignant progression in a subset of progressive LGG patients. Data from Dr Bernstein showed that IDH1 mutant-associated DNA hypermethylation disrupted insulator protein (CTCF) binding at specific sites. Compromised insulator function affects chromosome topology, leading to increased expression of platelet derived growth factor receptor alpha (PDGFRA), a receptor tyrosine kinase and glioma oncogene, elucidating the oncogenic role of IDH mutation.

Update on Outcomes from EF-14

Final progression-free survival (PFS) and overall survival (OS) data from the phase III EF-14 trial showed significant survival benefit for patients receiving tumor treatment field (TTF) therapy. This prospective, multicenter study was initiated in 2009 to test the efficacy and safety of combining TTF with temozolomide (TMZ) compared with TMZ alone following radiation therapy in newly diagnosed GBM. Results of the intent-to-treat analysis of all 695 enrolled patients with a mature minimum follow-up of 18 months (median 36) show a median PFS of 6.7 months (95% CI: 6.1–8.1) in the TTF/TMZ experimental group compared with 4.0 months (95% CI: 3.8–4.3) for the TMZ alone control group (hazard ratio [HR] 0.63, P < .00005). Median OS from randomization was 20.8 months versus 16 months, respectively (HR 0.65, 95% CI: 0.54–0.79, P < .00062), and 2-year survival was 42.5% (95% CI: 38–47) versus 30.0% (95% CI: 24–37, P < .001). Quality of life analyses to determine perceived patient burden are forthcoming.

The ACT IV Study: EGFR VIII–Targeted Vaccine in Newly Diagnosed GBM

Despite favorable results in single-arm phase II trials, the epidermal growth factor receptor variant III (EGFR VIII)–targeting vaccine rindopepimut failed to show a survival advantage in patients with newly diagnosed GBM. This international, double-blind, phase III trial enrolled 745 patients at 165 centers. The trial was terminated for futility in early 2016 after a second planned interim analysis. Median OS was 20.1 months (95% CI: 18.5–22.1) in the rindopepimut group compared with 20.0 months (95% CI: 18.1–21.9) in control patients with minimal residual disease (HR 1.01, P = .93). Anti–EGFR VIII humoral response was similar to that observed in prior phase II studies. Surprisingly, in patients with bulky disease, a trend toward an OS advantage was observed with rindopepimut (HR 0.79, 95% CI: 0.61–1.02, P = .066); however, this was not supported by PFS analysis (PFS HR 0.86, 0.66–1.12).

KEYNOTE-028 Glioblastoma Cohort

The first efficacy results of immune checkpoint inhibitor therapy in patients with recurrent GBM were presented. The KEYNOTE-028 study evaluated the safety and efficacy of the anti‒programmed death 1 (PD-1) monoclonal antibody pembrolizumab as monotherapy in an expansion cohort of 26 recurrent glioblastoma patients with positive PD-1 ligand (PD-L1) expression. Pembrolizumab had a manageable safety profile with grade 3–4 treatment-related adverse events (TRAEs) observed in 15.4% of patients (lymphopenia, type 2 diabetes mellitus, arthritis, and syncope). While only 1 partial response (PR) was observed, 12 patients (46%) experienced stable disease (SD) at a median duration of 39.4 weeks (95% CI: 7.1–85.9), median PFS 2.8 months (95% CI: 1.9–9.1), and median OS 14.4 months (95% CI: 10.3–not reached). Durable response was suggested in 4 patients who continued therapy >54 weeks following enrollment.

MEDI4736 in Bevacizumab-Naïve Recurrent GBM

Preliminary safety and efficacy data from the ongoing phase II multicenter, open-label study of the anti–PD-L1 antibody MEDI4736 (durvalumab) were presented for cohort B (bevacizumab-naïve first recurrent GBM). In these 31 patients treated with durvalumab monotherapy, no grade 4/5 serious TRAEs were observed; grade 3 TRAEs were reported in 9.7%. Response rate was 13%, median PFS was 13.9 weeks (95% CI: 8.1–24.0), and 6-month PFS was 20% (90% CI: 9.7–33.0) with 5 of these 6 patients remaining progression free at 1 year.

Interim Data of First-in-Class IDH Inhibitor

Interim data from the phase I study of AG120, a first-in-class IDH1 inhibitor, were reported. The agent was very well tolerated with no dose-limiting toxicities, and the maximum tolerated dose was not reached. An overall radiographic response rate of 3% (95% CI: 0.4–10.7) was reported. Of the 65 patients with recurrence of progressive IDH1 mutant glioma, no objective radiographic response was observed in the enhancing disease cohort (n = 31), and a 9% PR and 83% SD was observed in the non-enhancing disease expansion cohort. Median treatment duration was 8.1 months (95% CI: 1.4–17.8) in patients with non-enhancing disease, with 42% of patients still on treatment, compared with 1.9 months (95% CI: 0.4–10.4) in enhancing disease. Using volumetric measurement of fluid attenuated inversion recovery tumor volume, 58% of patients with non-enhancing tumor had SD or decreasing slope on drug.

Top-Scoring Abstracts

Soeren Mueller (Diaz Laboratory, San Francisco, CA) presented results of single-cell profiling of glioblastoma biopsies using RNA sequencing data, which suggest that mutations affecting PDGFRA enhance proliferation and are found in 16% of GBM.

Yi Fan (Philadelphia, PA) presented data on the contribution of cellular plasticity to chemotherapy response and showed that tumor endothelial cells undergo c-Met‒mediated transformation into mesenchymal stemlike cells, leading to TMZ resistance.

Mariella Filbin (Suva Laboratory, Boston, MA) profiled adult and pediatric glioblastomas using single-cell RNA sequencing and showed that most cancer cells differentiate along specialized glial programs with a subpopulation of undifferentiated cells expressing neural stem cell‒like programs.

Elias Sayour (Mitchell Laboratory, Gainesville, FL) used a preclinical cellular immunotherapy model to show that RNA nanoparticles can be combined with tumor RNA and supersede dendritic cells in mediating antitumor activity.

Cameron Herting (Hambardzumyan Laboratory, Atlanta, GA) used RCAS/t-va technology to generate adult proneural and mesenchymal GBM mouse models that more closely resemble human tissue samples. These models recapitulated the microenvironmental and molecular features of human tumors and may allow for accurate preclinical testing of subtype-specific targeted therapies.

Fumiharu Ohka (Natsume Laboratory, Nagoya, Japan) reported on the in vivo mechanisms of IDH wild-type grade III glioma formation. Using an IDH wild-type grade III glioma mouse model, the researchers showed that upregulation of enhancer of zeste homolog 2 (EZH2), a methyltransferase of histone H3K27, is a major alteration. These data suggest that IDH wild-type grade III gliomas may depend on EZH2 dysregulation, a potential novel therapeutic target.

Daniel Orringer (Ann Arbor, MI) presented data on the intraoperative use of stimulated Raman scattering microscopy, a simple, automated method for rapid intraoperative histopathological imaging of fresh, unprocessed human surgical specimens. This method may ultimately contribute to more thorough histological sampling, particularly at the brain‒tumor interface.

Christine Hoeman (Becher Laboratory, Durham, NC) developed the first genetically engineered mouse model of diffuse intrinsic pontine glioma harboring both H3.1 K27M and activin A receptor 1 (ACVR1) mutations. Using the RCAS/t-va system, the researchers showed that activating bone morphogenic protein pathway mutations in ACVR1, which encodes for the activin A receptor, contributes to the tumor growth in this model. Treatment with an ACVR1 inhibitor decreased tumor neurosphere viability.