-
PDF
- Split View
-
Views
-
Cite
Cite
Kaspar Draaisma, Aikaterini Chatzipli, Martin Taphoorn, Melissa Kerkhof, Astrid Weyerbrock, Marc Sanson, Ann Hoeben, Lukacova Slavka, Giuseppe Lombardi, Monique Hanse, Ruth Fleischeuer, Sieger Leenstra, Colin Watts, Thierry Gorlia, Vassilis Golfinopoulos, Johan Kros, Martin van den Bent, Ultan McDermott, Pierre Robe, Pim French, GENE-01. STABILITY OF ACTIONABLE MUTATIONS IN PRIMARY AND RECURRENT GLIOBLASTOMAS, Neuro-Oncology, Volume 19, Issue suppl_6, November 2017, Page vi92, https://doi.org/10.1093/neuonc/nox168.376
- Share Icon Share
Abstract
Efforts to improve patient survival in recurrent glioblastomas (GBMs) are often based on targeting the tumors acquired genetic changes. However, resections on recurrent GBMs are infrequently performed and molecular data is extrapolated from the initial tumor. We therefore initiated a large study to assess whether genetic changes of initial GBMs are still present at recurrence.
DNA was isolated from pairs of initial and recurrent GBM (FFPE) tumor samples and sequenced on a panel of 419 cancer associated genes. We also determined EGFRvIII expression status and MGMT promoter methylation status. All patients were treated according to the Stupp protocol.
Tumor pairs from 226 patients from 10 medical centers in six countries were included in the analysis. Median survival was 21.1 months and median time to second surgery 11.1 months. Higher age, greater lesion size at second surgery and steroid treatment were associated with poorer survival in the recurrent tumor. ~4% of tumors were IDH1 mutated (median OS 33.7 months). Of the 114 tumor pairs sequenced to date, 73 patients (64%) gained or lost a variant in a recurrent driver oncogene while the median retention rate of driver variants was 80%. Common genes with gained mutations include TP53 (13 patients) and NF1 (8 patients), common losses of mutations were found in EGFR (12 out of 23). EGFRvIII was expressed in 25.5% of initial tumors and was lost at recurrence in 50% of cases. Two patients developed a hypermutated tumor at recurrence (276 and 431 coding variants after 12 and 17 cycles of temozolomide respectively).
Preliminary analysis indicates that there are distinct gains and losses in each tumor, including in some driver mutations at tumor recurrence. This has important consequences on the need for re-biopsies in targeted treatment studies in recurrent glioblastoma.