HGG-11. Clinical characteristics and clinical evolution of a large cohort of pediatric patients with primary central nervous system (CNS) tumors and tropomyosin receptor kinase (TRK) fusion

BACKGROUND: TRK fusions are detected in less than 3% of CNS tumors. Given their rarity, there are limited data on the clinical course of these patients. METHODS: We contacted 166 oncology centers worldwide to retrieve data on patients with TRK fusion-driven CNS tumors. Data extracted included demographics, histopathology, NTRK gene fusion, treatment modalities and outcomes. Patients less than 18 years of age at diagnosis were included in this analysis. RESULTS: Seventy-three pediatric patients with TRK fusion-driven primary CNS tumors were identified. Median age at diagnosis was 2.4 years (range 0.0–17.8) and 60.2 % were male. NTRK2 gene fusions were found in 37 patients (50.7%), NTRK1 and NTRK3 aberrations were detected in 19 (26.0%) and 17 (23.3%), respectively. Tumor types included 38 high-grade gliomas (HGG; 52.1%), 20 low-grade gliomas (LGG; 27.4%), 4 embryonal tumors (5.5%) and 11 others (15.1%). Median follow-up was 46.5 months (range 3-226). During the course of their disease, a total of 62 (84.9%) patients underwent surgery with a treatment intent, 50 (68.5%) patients received chemotherapy, 35 (47.9%) patients received radiation therapy, while 34 (46.6%) patients received NTRK inhibitors (3 as first line treatment). Twenty-four (32.9%) had no progression including 9 LGG (45%) and 9 HGG (23.6%). At last follow-up, only one (5.6%-18 evaluable) patient with LGG died compared to 11 with HGG (35.5%-31 evaluable). For LGG the median progression-free survival (PFS) after the first line of treatment was 17 months (95% CI: 0.0-35.5) and median overall survival (OS) was not reached. For patients with HGG the median PFS was 30 months (95% CI: 11.9-48.1) and median OS was 182 months (95% CI 20.2-343.8). CONCLUSIONS: We report the largest cohort of pediatric patients with TRK fusion-driven primary CNS tumors. These results will help us to better understand clinical evolution and compare outcomes with ongoing clinical trials.

Clinical characteristics and outcome of a large cohort of patients with primary central nervous system (CNS) tumors and tropomyosin receptor kinase (TRK) fusion

2052

Background: TRK fusions are detected in less than 3% of central nervous system (CNS) tumors. Given their rarity, there are limited data on the clinical course of affected patients. Methods: We contacted 166 oncology centers worldwide to retrieve data on patients with TRK fusion-driven CNS tumors. Data extracted included demographics, histopathology, TRK gene fusion, treatment modalities and outcomes. Results: Ninety-two patients with TRK fusion-driven primary CNS tumors were identified including 76 pediatric patients (82.6%), 15 adults (16.3%) and 1 not specified (1.1%). Median age at diagnosis was 4.4 years (range 0.0–78.3) and 58.7 % were male. NTRK2 gene fusions were found in 45 patients (48.9%), NTRK1 and NTRK3 aberrations were detected in 27 (29.3%) and 20 (21.7%), respectively. Tumor types included 56 high-grade gliomas (HGG; 60.9%), 20 low-grade gliomas (LGG; 21.7%), 4 embryonal tumors (4.3%) and 12 others (13.0%). Median follow-up was 40.5 months (range 3–226). During the course of their disease, 75 (81.5%) patients underwent surgery with a treatment intent, 67 (72.8%) patients received chemotherapy, 50 (54.3%) patients received radiation therapy, while 47 (51.1%) patients received NTRK inhibitors (6 as first line treatment). There were significant differences in the median progression-free (PFS) and overall survival (OS) between pediatric patients compared to adults. The pediatric median PFS was 32 months (95% CI: 15.5–48.5) compared to 8 months for the adult (95% CI: 4.5–11.5, p = 0.015). The pediatric median OS was 182 months (95% CI: 25.1–338.9) compared to 24 months (95% CI: 18.3–29.7 p < 0.001) for adult patients. There was no difference in the PFS of LGG compared to HGG. However, the OS was significantly worse for the HGG when compared to LGG (p = 0.039). The median OS for LGG was not reached and the median OS for HGG was 70 months (95% CI 7.5–132.5). Nineteen patients with HGG (38.0 % 19/50 evaluable patients) died compared to only one patient with LGG (5.6% 1/18 evaluable patients, p = 0.014). Conclusions: We report the largest cohort of patients with TRK fusion-driven primary CNS tumors. These results will help us to better understand clinical evolution and compare outcomes with ongoing clinical trials.

INNV-04. CONGENITAL SPINAL GLIOBLASTOMA WITH SUSTAINED RESPONSE TO TROPOMYOSIN RECEPTOR KINASE (TRK) INHIBITION: A CASE REPORT

Large scale genomic efforts have shown that a significant percentage of infant high-grade gliomas (HGG) carry targetable gene fusions. Given the rarity of congenital spinal HGGs, little is known about the molecular features of these tumors and their response to treatment. We describe a case of a patient with congenital spinal glioblastoma. A full-term female infant with symmetric IUGR was noted to have weakness with little movement of her upper extremities. MRI of the spine showed an infiltrative and hemorrhagic intraspinal cervicothoracic neoplasm with a large hemorrhagic intramedullary component. A biopsy was done and revealed a high-grade glial neoplasm with pan-TRK expression. Next generation sequencing (NGS) confirmed a MEF2D-NTRK1 gene fusion. Prior to the result of her molecular testing, the patient clinically deteriorated with decreased movement and increased apneic episodes despite steroid treatment. She was then started on treatment with carboplatin and etoposide per Macy et al 1. After one cycle of chemotherapy and given the results of NGS, targeted therapy with larotrectinib was added to her treatment. The patient showed a response (clinical and imaging) following cycle 1 with chemotherapy, and a marked response following cycle 2 with both chemotherapy and targeted TRK inhibition. The patient completed six cycles of chemotherapy and has continued on daily oral larotrectinib. She continues to tolerate the targeted therapy well and currently is 8 months old, growing and developing well. She has regained much of her movement and is eating and growing with apparent normal development. This case provides an example of a patient with congenital glioblastoma with TRK gene fusion that showed an impressive response to targeted therapy combined with more conventional chemotherapy. 1Macy M, et al. Clinical and molecular characteristics of congenital glioblastoma. Neuro Oncol. Jul;14 (7) 931–941, 2012).

DDRE-03. INTERNATIONAL PRECLINICAL DRUG DISCOVERY AND BIOMARKER PROGRAM INFORMING AN ADOPTIVE COMBINATORIAL TRIAL FOR DIFFUSE MIDLINE GLIOMAS

INTRODUCTION

DMG-ACT (DMG- multi-arm Adaptive and Combinatorial Trial) aims to implement a highly innovative clinical trial design of combinatorial arms for patients with diffuse midline gliomas (DMGs) at all disease stages that is adaptive to pre-clinical data generated in ten collaborating institutions. Novel drug and drug combination were tested, predictive biomarkers were identified and incorporated in clinical trial design.

METHODS

In vitro (n=15) and in vivo (n=8) models of DMGs across ten institutions were used to assess single and combination treatments with ONC201, ONC206, marizomib, panobinostat, 5-Azacytidine, Val-083, GDC0084 and TAK228. In vivo drug toxicity screenings were conducted using larval zebrafish model and murine PDX models. Predictive biomarkers for ONC201 and ONC206 were identified using meta-analysis, and extensive molecular assays including CRISPR, RNAseq, FACS, and IHC.

RESULTS

Inhibitory concentrations (IC50) were established and validated multiple preclinical models. ONC201 and ONC206, ONC201 and TAK228, ONC201 and GDC0084 showed synergism. In vivo survival assays showed increased survival for: ONC201 (p=0.01), ONC206 (p=0.01), ONC201+ONC206 (p=0.02), and ONC201+panobinostat (p=0.01). Marizomib showed toxicity in murine/zebrafish PDXs models. Murine pharmacokinetic analysis showed peak brain levels of ONC201 and ONC206 above pre-clinical IC50. Molecular testing and analyses of existing drug screen across 537 cancer cell lines validated mitochondrial protease ClpP and ATF4 as ONC201/6 targets. Predictive biomarkers of response to drug were identified.

CONCLUSION

Thorough preclinical testing in a multi-site laboratory setting is feasible and identified ONC201 in combination with ONC206, TAK228 and GDC0084 as promising therapeutics for DMGs.