172O - ER+ /HER2+ and ER-/HER2+ breast cancers are molecularly distinct but immune gene signatures are prognostic and predictive in both groups

Date 29 September 2012
Event ESMO Congress 2012
Session Biomarkers in breast and colorectal cancer
Topics Biomarkers
Breast Cancer
Presenter Takayuki Iwamoto
Authors T. Iwamoto1, L. Pusztai2, J. Matsuoka3, M. Callari4, C.M. Kelly5, Y. Qi6, T. Motoki1, N. Taira1, L. Santarpia7, H. Doihara1, L. Gianni8, G. Bianchini8
  • 1Breast And Endocrine Surgery, Okayama University Hospital, 700-8558 - Okayama/JP
  • 2Dept. Of Breast Medical Oncology, MD Anderson Cancer Center, TX 77030-1439 - Houston/US
  • 3Palliative Medicine, Okayama University, Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, 700-8558 - Okayama/JP
  • 4Experimental Oncology And Molecular Medicine, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan/IT
  • 5Department Of Medical Oncology, Mater Misericordiae University Hospital, Dublin/IE
  • 6Department Of Bioinformatics And Computational Biology, MD Anderson Cancer Center, Houston/US
  • 7Department Of Oncology, Translational Research Unit and Istituto Toscano Tumori, IT-59100 - Prato/IT
  • 8Medical Oncology, Ospedale San Raffaele, Milan/IT

Abstract

Objectives

We examined if gene expression patterns differ among HER2 positive breast cancers by estrogen receptor (ER) status. We also assessed the prognostic and chemotherapy response predictive values of over 3000 gene sets separately in the ER subsets.

Methods

Publically available, clinically annotated Affymetrix gene expression data of 537 HER+ (ER-/HER2+ n = 278, ER + /HER2+ n = 259) and 2985 HER2- (ER thinsp;+ /HER2- n = 1923, ER-/HER2- n = 1062) patients were studied including 121 node-negative, HER2+ cases with no systemic adjuvant therapy (ER+ n = 61, ER- n = 60) and 86 HER2+ patients treated with neoadjuvant taxane or anthracycicline chemotherapy (ER+ n = 27 and ER- n = 59).

Results

Genes that distinguished ER+ from ER- cases differed for HER2+ and HER2- cancers. HER2+ cancers showed less difference by ER status compared to HER2- cancers (differentially expressed genes by ER status: n = 194 and n = 6750, p < 1.0E10-6, shared genes 84%). Significant differences were also observed when HER2+ and HER2- cancers were compared stratified by ER status (HER2-related differentially expressed genes in ER+ n = 242, in ER- n = 1200, p < 1.0E10-6, shared genes 20%). In ER + /HER2+ tumors, 67 gene sets (GSs) were associated with good prognosis (e.g. B, T and NK cells and interferon-gamma production) and 2 GS with poor prognosis (PTEN dependent cell cycle arrest and apoptosis) (p ≤ .001). The same GSs were also prognostic in ER-/HER2+ cancers. In the neoadjuvant series, 11 and 2 GSs were associated in both ER+ and ER-/HER2+ groups with pathologic complete response (pCR) (e.g. T-cell activation and differentiation) and residual disease (RD) (e.g. Cell-cell junction) respectively (combined p ≤ .001). We also noted ER specific associations with pCR or RD. For instance, 18 GSs were associated with pCR in ER-/HER2+ (p ≤ .001) but not in ER + /HER2+ tumors (e.g. chemochine C-C binding and activity, phospholipid catabolic process).

Conclusions

Among HER2+ tumors, ER- and ER+ cancers represent distinct molecular subtypes. Immune signatures predict for good prognosis and higher chemotherapy sensitivity in HER2+ cancers regardless of ER status.

Disclosure

All authors have declared no conflicts of interest.