258PD - Investigating the clinical relevance of genomic characteristics in luminal A and B breast cancer (BC)

Date 27 September 2014
Event ESMO 2014
Session Breast cancer, early stage
Topics Pathology/Molecular Biology
Translational Research
Breast Cancer
Basic Scientific Principles
Basic Principles in the Management and Treatment (of cancer)
Presenter Vassiliki Kotoula
Citation Annals of Oncology (2014) 25 (suppl_4): iv85-iv109. 10.1093/annonc/mdu327
Authors V. Kotoula1, F. Zagouri1, E. Timotheadou1, Z. Alexopoulou1, R. Wirtz2, A. Lyberopoulou1, S. Lakis1, H. Gogas1, E. Charalambous1, G. Pentheroudakis1, D.G. Pectasides1, A. Koutras1, P. Papakostas1, C. Christodoulou1, P. Kosmidis1, K.T. Kalogeras1, G. Fountzilas1
  • 1Data Office, Hellenic Cooperative Oncology Group (HeCOG), 11524 - Athens/GR
  • 2Stratifyer Molecular Pathology Gmbh (stratifyer, STRATIFYER Molecular Pathology GmbH (STRATIFYER, 50935 - Cologne/DE




To address the major challenge of understanding the clinical relevance of the increasingly available large amount of tumour genomic data in BC.


Histologically reviewed, paraffin tumour DNA samples (N = 1092) from patients who had received anthracycline-based adjuvant chemotherapy in the frame of two randomized trials by HeCOG (HE10/00, pre-trastuzumab; HE10/05, post-trastuzumab era), were investigated with targeted massively parallel sequencing (Ion Torrent systems) for variants in 58 genes implicated in BC. Upon multiple stringent quality filters, pathogenic mutations (mut) and allelic imbalance (AI) were evaluable in 844 cases (77.3%). IHC4 was used for BC subtyping.


Mut were observed in 499 and AI in 497 tumors (59%), reaching up to 55 and 20 affected genes in single tumours, respectively. Mut and AI often coexisted (p < 0.0001), while AI was positively related to nodal status (p = 0.0241). Mut were more frequent in TP53 (25%, excluding non-pathogenic p.P72R), PIK3CA (24.6%), GATA3 (8.8%), CDH1 (6.3%), MLL3 (5.5%), ARID1B (5.1%), TBX3 (5.1%) and PTEN (4.7%), while AI in EGFR (51.2%), TERT (41.4%), TP53 (40.8%), CASP8 (33.5), PARP2 (32.5%), GATA3 (28.9%), and BRCA1 (22.2%). TP53mut were more frequent in triple-negative and HER2-enriched BC (45% each), while PIK3CAmut (32%) in luminal A tumours (p < 0.0001). In luminal HER2-negative tumours, TP53mut were significantly associated with shorter disease-free survival (DFS) (log-rank, p = 0.0032), aggravating the prognosis of patients with any nodal status (p < 0.0001), irrespectively of Ki67 status that was used for distinguishing between luminal A and B subtypes. In these luminal HER2-negative tumours, AI in less than 6 genes in the presence of PIK3CAmut conferred the longest, while AI in more than 6 genes in the absence of PIK3CAmut the shortest DFS (p = 0.0156); this AI/PIK3CAmut combination was associated with worse prognosis in patients with 0-3 positive nodes, which fared similarly to those with higher nodal involvement (p < 0.0001).


Combined assessment of AI and mutation status may provide useful prognostic information in patients with luminal A and B tumours treated with adjuvant anthracycline-based chemotherapy.


R. Wirtz: Dr R.W. has pending patent applications; G. Fountzilas: On behalf of the Hellenic Foundation for Cancer Research, Prof. G.F. has pending patent applications. All other authors have declared no conflicts of interest.