EPEN-28. HETEROGENEITY WITHIN THE PFB EPENDYMOMA SUBGROUP. Issue 2 (22nd June 2018)

• Record Type: Journal Article
• Authors: Cavalli, Florence; Hübner, Jens-Martin; Sharma, Tanvi; Sill, Martin; Luu, Betty; Zapotocky, Michal; Korshunov, Andrey; Pfister, Stefan; Pajtler, Kristian; Taylor, Michael; Aldape, Kenneth; Kool, Marcel; Ramaswamy, Vijay
• Is Part Of: Neuro-oncology. Volume 20:Issue 2(2018)supplement 2
• Page Start: i79
• Page End: i79
• Publication Date: 2018-06-22
• Subjects: Brain Neoplasms -- Periodicals; Brain -- Tumors -- Periodicals; Brain -- Cancer -- Periodicals; Nervous system -- Cancer -- Periodicals; 616.99481
• DOI: 10.1093/neuonc/noy059.228
• Languages: English
• ISSNs: 1522-8517

Abstract

Abstract: BACKGROUND: Posterior fossa ependymoma comprise two distinct molecular groups, termed EPN_PFA and EPN_PFB. Clinically they are very disparate and EPN_PFB are currently being explored for de-escalation of therapy. However, to move forward, a risk stratification within EPN_PFB would be highly desirable. METHODS: To discern the molecular heterogeneity within EPN_PFB, we performed an integrated analysis consisting of DNA methylation profiling, copy number profiling and clinical correlation across a cohort of 217 primary EPN_PFB. RESULTS: DNA methylation data were analyzed using various methods including spectral clustering, unsupervised consensus clustering and t-distributed stochastic neighbor embedding analysis. The integrated analyses revealed four distinct subgroups with distinct age distributions, copy number alterations, and survival rates. 1q gain was strongly enriched for one of the subgroups and is associated with a worse progression free survival. A univariable analysis revealed that 1q gain, incomplete resection and no upfront radiation were significant predictors of poor progression free survival, and in a multivariable cox regression model, 1q gain was a highly predictive marker of 5 and 10 year progression free survival (HR 3.534 95% CI 1.59–7.87). CONCLUSION: There is significant intertumoral heterogeneity within EPN_PFB, revealing at least four distinct molecular subgroups. Identification of these subgroups may lead to a better understanding what is driving these tumors. The biological heterogeneity must be accounted for in future pre-clinical modeling and personalized therapies. 1q gain is a significant risk factor of poor progression free survival in EPN_PFB and may represent a prognostic marker in future trials of de-escalation of therapy for EPN_PFB.