CTNI-59. RARE AND NOVEL KIAA1549-BRAF FUSION BREAKPOINTS PREDICT POOR CLINICAL OUTCOME FOR PATIENTS WITH RECURRENT OR PROGRESSIVE PEDIATRIC LOW-GRADE GLIOMA: PHASE II PNOC001 TRIAL RESULTS

Abstract

Abstract OBJECTIVES PNOC001 sought to estimate progression-free survival (PFS) associated with everolimus for progressive/recurrent pediatric low-grade glioma (pLGG) and to determine if activated PI3K/AKT/mTOR pathway as measured by positive phosphorylated-ribosomal protein S6 (p-RPS6) was associated with response. A secondary objective was to assess whether somatic drivers influence median PFS. METHODS Patients, 3-21 years of age, with recurrent/progressive pLGG received everolimus orally at 5 mg/m2 daily. Somatic drivers, when available, were associated with PFS and compared to four independent molecular pLGG cohorts with newly-diagnosed and progressive/recurrent patients. RESULTS From 2012 to 2019, 65 subjects enrolled [median age 9.6 years, 3.0-19.9; 46% female]; Median PFS for all-comers was 11.1 months (95%-CI 7.6-19.8, n = 64). PFS was not modulated by PI3K/AKT/mTOR pathway activation (median PFS 11.2 vs 9.4 months, P = 0.83, n = 62). BRAF alterations were identified in 67.8% (40/59) of patients. Rare/novel KIAA1549-BRAF fusion breakpoints were enriched in PNOC001 compared to newly-diagnosed pLGGs in independent cohorts (28.6% vs 2.6%, P = 1.4e-4, n = 214). In PNOC001, patients with rare/novel KIAA1549-BRAF fusion breakpoints had worse clinical outcomes than patients with common fusion breakpoints (median PFS 6.1 vs 16.7 months, P < 0.05, n = 20). Rare/novel KIAA1549-BRAF fusion breakpoints were also enriched in patients with progressive/recurrent disease in independent pLGG cohorts (17.6% vs 2.6%, P = 2e-3, n = 227). Integration of BRAF alterations and DNA methylation-based tumor subtypes further revealed that rare/novel KIAA1549-BRAF fusion breakpoints are more frequent in supratentorial midline (14.1%) compared to infratentorial (2.0%) pilocytic astrocytomas (P = 1e-4). KIAA1549-BRAF fusion breakpoints remained risk factors for progressive/recurrent disease in PNOC001 (P = 0.05) and independent pLGG cohorts (P < 0.05) after accounting for tumor location and DNA methylation subtypes. CONCLUSIONS Rare/novel KIAA1549-BRAF fusion breakpoints are more frequent in supratentorial midline pLGGs and define independent genomic risk factors for disease progression/recurrence. Our results indicate that future clinical trials must interrogate and define KIAA1549-BRAF fusion breakpoints for informative results and conclusions.