EPEN-19. Impact of molecular classification on prognosis in children and adolescents with spinal ependymoma: Results from the HIT-MED database

Abstract

Abstract PURPOSE: Ependymomas of the spinal cord are rare among children, and individual risks of disease progression are difficult to predict. This study aims at evaluating the prognostic impact of DNA methylation-based classification in children with spinal ependymoma. METHODS: Eighty-two patients with spinal ependymoma <22 years registered in the HIT-MED database between 1992 and 2021 were included. Clinical, radiological, and histopathological data were collected retrospectively. DNA methylation profiles of 46 tumors were classified according to the Heidelberg Brain Tumor Classifier. RESULTS: Spinal myxopapillary ependymoma (SP-MPE, n=27) was the most common methylation group followed by spinal ependymoma (SP-EPN, n=15). Two cases belonged to MYCN-amplified subgroup, one had no match, and one was re-classified as anaplastic pilocytic astrocytoma (the latter excluded from final analysis). WHO grade I and III ependymomas (according to the WHO 2016 classification) classified predominantly as SP-MPE, whereas grade II ependymomas clustered into SP-MPE and SP-EPN. 6/15 patients with SP-EPN (40%) suffered from Neurofibromatosis type 2. Among patients with SP-MPE, 23 underwent gross-total and four a subtotal resection (GTR/STR). Relapses of SP-MPE were more common following STR (5-year progression-free survival (5y-PFS) [STR] 25.0% [95% confidence interval: 0.0-68.4], [GTR] 75.0% [53.4-96.6], p=0.003). In the SP-EPN group, 2/8 patients relapsed after STR (5y-PFS 64.3% [22,3-100]) and 0/7 after GTR (n.s.). WHO I° ependymoma had significantly inferior PFS than II° and III° ependymoma (5y-PFS [I°] 39.0% [5.8-62.2], [II°] 82.4% [67.8-97.0], [III°] 50.5% [18.9-82.1], p=0.009). However, PFS did not significantly differ between SP-MPE and SP-EPN (5y-PFS 65.9% [44.9-86.9], 76.9% [46.3-100], respectively). CONCLUSION: Spinal ependymomas of WHO grade I go along with relatively poor PFS in our cohort, while DNA methylation profiling does not segregate patients into distinct risk groups. Still, larger cohorts and further investigations of methylation class heterogeneity in pediatric spinal ependymomas are needed to complete the basis for future clinical decision-making.