Abstract 31: Subgroup-specific molecular alterations may explain the tremendous clinical heterogeneity of intracranial ependymoma

Abstract

Abstract Intracranial ependymoma comprises the second most common malignant brain tumor in childhood. The prognosis of these tumors remains generally poor and its biological behavior is unpredictable based on current stratification approaches. Neither clinical variables nor histopathological grading or molecular markers have so far been successful in defining a well circumscribed group of high-risk patients. Thus, an innovative staging model for ependymoma is desperately needed. We studied 122 samples from patients with intracranial ependymoma with a median follow-up of circa 8 years by genome-wide assessment of DNA copy-number aberrations using array-CGH (10K BAC array). Aberrations with a potential prognostic value were validated in an independent cohort of 170 patients by FISH analysis. Consecutively, we investigated genome-wide mRNA expression profiling (Agilent 44k) in 65 primary ependymomas and performed unsupervised clustering to identify potential transcriptome-based subgroups. We compared these findings with the previously identified DNA copy-number profiles. For validation of single molecular markers, selected candidate genes were investigated by QRT-PCR on transcriptional level, and protein expression was measured by immunohistochemistry on tissue microarrays (n=170). We were able to define a novel molecular staging system comprised of three genetically distinct subgroups of ependymoma based on DNA copy-number aberrations: i) a low risk group (34% of patients) including tumors with gain of chromosomes 9, 15q, 18, or loss of chromosome 6, or a combination thereof with patients showing a 5-year OS of 100%; ii) an intermediate risk group (41% of patients) characterized by a balanced cytogenetic profile especially for aberrations of chromosomes 1q, 9, 15q, 18, 6 and without a homozygous deletion of CDKN2A which was associated with a 5-year OS of 77%; iii) a high risk group (25% of patients) defined by tumors harbouring a gain of 1q and/or a homozygous deletion of CDKN2A, which was concurrent with a 5-year OS of only 33%. Interestingly, these cytogenetic risk-groups showed a significant overlap with transcriptome-based subgroups identified by unsupervised clustering. Thus, we aimed at the identification of interesting candidate genes which show subgroup-specific expression and have the potential to be used as surrogate marker for certain biological subgroups. The most robust subgroup-specific molecular markers for poor and good outcome were SHC1 and WDR16, respectively. In summary, we could decipher a novel stratification model for intracranial ependymoma consisting of three subgroups based on cytogenetic aberrations. By integrative genomics looking at DNA aberrations and mRNA levels in a large subset of samples, we were able to identify novel biomarkers in ependymoma, which have high potential to be useful for stratifying patients in future clinical trials. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 31.