PDTM-36. WHOLE EXOME SEQUENCING (WES) OF DIPG PATIENTS FROM THE BIOMEDE TRIAL REVEALS NEW PROGNOSTIC SUBGROUPS WITH SPECIFIC ONCOGENIC PROGRAMMES

Abstract

Abstract BACKGROUND & METHODS. The BIOlogical MEdicines for DIPG Eradication (BIOMEDE) trial is a randomized multi-arm multi-stage program (drop-the-loser adaptive design). The first trial was an open-label phase-II trial comparing three drugs (everolimus, dasatinib, erlotinib) combined with irradiation, allocated according to the presence of their specific targets. All patients underwent a biopsy at diagnosis to confirm the DIPG molecularly and establish the expression of pre-specified biomarkers. Patient samples were explored by WES with a sequencing depth average >100X for the tumor and >60X for the blood. RESULTS. Unsupervised clustering of copy-number-variations (CNV) identified 4 groups corresponding to a new stratification of DIPG: cluster 1 (n=42) with high CNV, cluster 2 (n=23) with chromosome 1q + chromosome 2 gains, cluster 3 (n=9) with low CNV and cluster 4 (n=21) with isolated 1q gain. Clusters 1 had higher hazard-ratio for death than the others (mean HR 1.959, range 1.054–3.643, p< 0.0001, adjusted Cox model). Extensive structural rearrangements/chromotrypsis were significantly more frequent in TP53-mutated samples. Indeed, when stratiftying patients in 4 groups based on type of histone H3 mutated at K27 and the presence of a TP53 pathway alteration the subgroup with TP53 altered pathway had a significantly higher hazard-ratio for death than the others (mean HR 3.450, p=0.0017, adjusted Cox model). Additional drug targets were identified, especially in the DNA repair machinery that could be exploited for new targeted therapies. CONCLUSION. WES at diagnosis was feasible in most patients and brings new prognostic and theranostic informations. This allows a better patients stratification and the development of personalized medicine in DIPG.