PL02.3.A Survival and T-cell tumor reactivity in patients treated with nivolumab and bevacizumab for recurrent glioblastoma in the clinical trial CA209-9UP

Background

Glioblastoma (GBM) is an aggressive brain tumor with a median survival of 14.6 months. We have no standard treatment for relapse and current treatment options have limited effect. Novel treatments are necessary to improve survival and quality of life.

Material and Methods

We present data from; a phase II open label, two-armed clinical trial studying nivolumab and bevacizumab in treatment of recurrent GBM, with progression after Stupp’s regime. Patients were included in two arms depending on the possibility of salvage neurosurgical resection. All patients had biopsies for genome sequencing at primary tumor and recurrence. Both arms received nivolumab and bevacizumab administrated every second week and the surgical arm also received neoadjuvant nivolumab 7 days prior to surgery. Fresh tissue samples were collected for tumor digest, TILs (tumor infiltrating lymphocytes) for phenotype exploration and intracellular staining to test reactivity. Patients were treated until progression, death, or intolerable side effects. Toxicity screens were reported, and follow-up ended in Marts 2022.

Results

Forty-four patients were included from November 2018 to January 2022; 20 in each arm (four screen-failures). Treatment was overall well tolerated. Median (m) age at inclusion was 57,5 years (arm A) and 50,5 years (arm B), and the groups had an even distribution. The surgical and non-surgical arm had an mPFS of 5.95 and 3.83 months respectively, while the mOS was 13.96 months and 6.77 months, respectively. Multivariate analysis was performed by variables such us steroid, MGMT, gender, age at diagnosis, resection extent and arm. Steroid at inclusion was a significant negative predictor of outcome (p = 0.0378). Controls from our GBM registry (N=140), which were treated with neurosurgical resection and then bevazicumab and irinotecan in recurrent setting had an mOS of 8.64 months (log-rank p=0.0181).Furthermore, reactive tumor infiltrating lymphocytes (TIL) were detected in four of the patients who presented with a longer mOS and mPFS of 16.75 months and 9.18 months, while the 16 patients without TIL reactivity had mOS and mPFS of 12.63 months and 5.13 months, respectively (not significant).

Conclusion

We found an increased mOS in patients treated with nivolumab and bevacizumab at recurrence, compared to our controls: 13.96 months and 8.64 months, respectively. Four patients with T-cell reactivity towards tumor cells showed an even longer mPFS and mOS. Though not significant, these results warrant further research evaluation in larger patient cohorts. We are currently investigating proteomics and sequencing data to identify predictive biomarkers.

CTIM-22. NIVOLUMAB AND BEVACIZUMAB FOR RECURRENT GLIOBLASTOMA; T-CELL REACTIVITY AGAINST AUTOLOGOUS TUMOR CELLS

INTRODUCTION

Glioblastoma is an aggressive brain tumor with a median survival of 14.6 months. We have no standard treatment for relapse and known options have limited effect. Novel treatments are necessary to improve survival and quality of life.

METHODS

We present our trial; phase II open label, two-armed translational study of Nivolumab and Bevacizumab for recurrent GBM, who have failed Stupp’s regimen. Patients are included in two arms depending on the possibility of salvage neurosurgical resection. Both arms receive Nivolumab and Bevacizumab administrated every second weekend, and the surgical arm also receive Nivolumab 7 days prior surgery. Forty-four patients were included by January 2021; 20 in each arm (four screen-failures). In the surgical arm, 20 fresh tumor samples as well as paired tissue from primary tumor were available. Tumor infiltrating lymphocytes (TILs) and tumor digest were produced in vitro from recurrent settings. Young TILs were expanded from fresh tumor fragments after minimal-culture, whereas rapidly expanded TILs (REP TILs) were obtained after massive expansion. By intracellular cytokine staining, we investigated the TIL reactivity after exposure to autologous tumor digest in order to evaluate whether the TILs were tumor-reactive, non-reactive or bystanders. RNA and whole exome sequencing were available before and after treatment.

RESULTS

Material from 19 patients was analyzed (one out of the 20 collected biopsies was limited in size, therefore no tumor digest could be produced). Four out of 19 TIL samples showed tumor reactivity after exposure to the autologous tumor digest. Tumor reactivity was ranged between 1,2 to 13,6 tox% in CD8+ TILs and between 2,8 to 10,9 tox% in CD4+ TILs. By flowcytometry we found, IgG4+ CD3+ TILS from tumor biopsies, meaning that Nivolumab were found in the brain. Currently controls are included to evaluate these results.

CONCLUSIONS

Updated results will be presented at SNO.

ATIM-01. NIVOLUMAB AND BEVACIZUMAB FOR RECURRENT GLIOBLASTOMA; A TRANSLATIONAL TRIAL IN PROGRESS

Glioblastoma multiforme (GBM) is an aggressive brain tumor with a poor prognosis. Standard of care at diagnosis is surgical resection, followed by radiation and temozolomide. Receiving this therapy, the median survival is 14.6 months [1]. We have no standard treatment for relapse and known options have limited effect. There is an urgent need for novel treatment interventions to improve clinical outcomes and quality of life. Recently, improved overall survival has been achieved with immune therapeutics in melanoma and renal cell carcinoma. Accordingly, it has been posited that immunotherapy may offer promise in other difficult cancers such as GBM [2]. We present our translational study; a phase II open label, two-armed translational study of Nivolumab and Bevacizumab for recurrent GBM, who have failed Stupp’s regime [1]. Patients are included in two arms depending on possibly salvage neurosurgical resection. Both arms receive Nivolumab and Bevacizumab administrated every second weekend, but the surgical arm also receive Nivolumab 7 days prior surgery. We expect 40 patients; 20 in each arm. Enrollment period is expected to 20 months, started October 2018. Our primary objective is to make preliminary assessment of immune related biomarkers, including PD-L1; therefore, we perform full genome sequencing on tumor biopsies from the surgical arm and on blood samples from both arms. We evaluate changes in the transcriptomic landscape caused by the check-point inhibition and relation to response as compared with baseline sequencing data, as well as the impact of tumor mutation burden and neoepitope load. We investigate the tumor microenvironment by harvesting tumor infiltrating lymphocytes and study the composition by flow-cytometry. The patients are evaluated by blood samples, FET-PET as wells as clinical examinations to evaluate PFS and OS. Overall the study will provide us with a unique possibility to investigate and thereof predict which patients will profit from the treatment.

EPID-06. IMMUNE-RELATED PLASMA BIOMARKERS IN GLIOBLASTOMA

BACKGROUND

Glioblastoma (GBM) is an aggressive and malignant primary brain tumor with limited treatment options and a dismal prognosis. Immunotherapy is advancing in other types of cancer but has not yet proven effective in glioblastoma. Prognostic and predictive biomarkers and new treatment targets are urgently needed.

MATERIAL AND METHODS

158 patients with glioma WHO grade II-IV were included in the study. Plasma was analyzed for interleukin (IL)-6, YKL-40 (Gene: CHI3L1) and 91 other immune-related protein biomarkers by ELISA or/and Olink technology. CHI3L1 rs4950928 genotyping was analyzed on DNA extracted from whole-blood.

RESULTS

Fourteen of 92 immune-related biomarkers in plasma detected with the Olink immuno-oncology array were related with tumor grade or/and type (p< 0.05) (corrected for age), whereas plasma IL-6 and YKL-40 did not change with tumor grade. In univariate analysis of plasma from 94 patients with newly diagnosed GBM higher baseline IL-6 was associated with short OS (HR=1.19 (per 2-fold change in IL-6), p=0.04), YKL-40 showed a similar trend (HR=1.20 (per 2-fold change in YKL-40), p=0.056) and high/low baseline levels of additional 15 immune-related biomarkers (e.g. CD244, Granzyme B, ICOS ligand, IL-8 and Pleiotrophin) were associated with short OS (p< 0.05). Genetic variation in YKL-40 was associated with plasma YKL-40 levels (+/+ vs. -/+ vs. -/-, p=0.0004) but not with OS in patients with GBM (HR= 0.78 (+/+ vs. -/+), p=0.32).

CONCLUSION

High levels of several immune-related biomarkers in plasma including IL-6, IL-8, CD244, Granzyme B, ICOS ligand, and Pleiotrophin from patients with glioblastoma were related or inversely related to OS.