10P - Comparison of the mutational landscape of breast cancer during pregnancy and non-pregnant controls

Date 04 May 2017
Event IMPAKT 2017
Session Welcome reception and Poster Walk
Topics Anticancer Agents
Breast Cancer
Biological Therapy
Presenter Sibylle Loibl
Authors S. Loibl1, N. Pfarr2, K. Weber1, S. Villegas-Angel3, A. Stenzinger4, J. Furlanetto1, J. Budczies3, F. Marmé5, C. Denkert3, W. Weichert2
  • 1Medicine And Research, German Breast Group, 63263 - Neu-Isenburg/DE
  • 2Institute Of Pathology, Technical University Munich, 81675 - Munic/DE
  • 3Pathology, Charité Hospital, 10437 - Berlin/DE
  • 4Institute Of Pathology, University Hospital Heidelberg, 69120 - Heidelberg/DE
  • 5National Center For Tumor Diseases, University Hospital Heidelberg, 69120 - Heidelberg/DE



Background: Currently, breast cancer during pregnancy (BCP) is not believed to be biologically different from breast cancer (BC) unrelated to pregnancy based on limited datasets mainly obtained by immunohistochemistry. Therefore, the aim of the study is to compare the pattern of somatic mutations between pregnant and non-pregnant patients (pts) with BC using a dataset of pregnant pts enrolled in the BCP study and non-pregnant controls obtained from the TCGA database.

Methods: The BCP study (GBG 29; BIG 03-02) is a multicentre observational study for BC during pregnancy. FFPE core biopsies taken before therapy were retrospectively analysed for somatic mutations by an Ion Torrent Proton/PGM sequencing platform. The samples were assayed on a custom designed BC panel (BCPv2) including 236 amplicons split into two primer pools and covers hotspot regions of 138 exons of 25 genes. Only non-synonymous mutations without germline origin were processed.

Results: Comparison of the mutational patterns between BCP and TCGA cohorts showed 1.03 mutations per pt in average in BCP vs. 1.27 in TCGA. The most frequent somatic mutations for both cohorts were TP53 (65% vs. 37%), PIK3CA (11% vs. 29%) and GATA3 (6% vs. 18%). Exact matching of BCP and TCGA cohorts was performed based on age, HR, HER2 and grading and yielded 41 pts from both datasets. After matching BCP pts had significantly less frequently N+ tumors compared to TCGA pts (p=0.046) with no significant difference for TP53 (p=0.502) and GATA3 (p=1.000) mutational status; PIK3CA mutations were detected in only 2.4% of the pregnant vs. 22.0% of the non-pregnant pts (p=0.015). Within HR subgroups, overall TP53 was the most frequently mutated gene with higher mutational rate in HR-negative pts (52.4% vs. 75.0% for BCP; 23.8% vs. 85.0% for TCGA).


Overall the mutational landscape does not seem to differ between pregnant and non-pregnant pts. Imbalances in the PIK3CA mutational rate after matching might be explained by a remaining bias caused by differences in sensitivity or specificity of methods used to detect mutations or differences in variables not used for matching. Further comparisons using other datasets and looking into gene expression patterns are currently conducted.

Clinical trial identification