189P - Breast cancer prognosis after neoadjuvant chemotherapy for breast cancers: molecular downstaging, proliferation, and endocrine sensitivity importance

Date 10 October 2016
Event ESMO 2016 Congress
Session Poster display
Topics Breast Cancer
Presenter Marc-Antoine Benderra
Citation Annals of Oncology (2016) 27 (6): 43-67. 10.1093/annonc/mdw364
Authors M. Benderra1, S. Richard2, M. Antoine3, D. Buob3, S. Zilberman4, A. Esteso2, J.P. Lotz2, K. Kerrou5, J. Gligorov2
  • 1Medical Oncology, APHP, CancerEst, Tenon University Hospital, 75020 - Paris/FR
  • 2Medical Oncology, APHP, CancerEst, Tenon University Hospital, Paris/FR
  • 3Pathology, APHP, CancerEst, Tenon University Hospital, 75020 - Paris/FR
  • 4Gynaecology, APHP, CancerEst, Tenon University Hospital, Paris/FR
  • 5Nuclear Medicine, APHP, CancerEst, Tenon University Hospital, Paris/FR



Pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) is a questionable prognostic factor for all breast cancer (BC) subtypes. We assessed the importance of other clinical and pathological parameters as prognostic factors after NAC.


From 2005 to 2014, 236 non metastatic BC patients were consecutively treated in our institution with the same NAC combining sequentially anthracyclines-cyclophosphamide followed by taxanes (trastuzumab was added in HER2 positive population). All the data concerning patient population, initial tumor stage, pathological characteristics on biopsy, as also after surgery were collected. Pathological analysis was centralized. pCR was defined according to UICC criteria, disease free survival (DFS) was calculated since the day of first surgery. Analyses employed logistic regression and Cox proportional hazard models.


Of 236 patients, 140 (59%) were ER positive, 44 (19%) were triple negative and 52 (22%) were Her2 positive. The overall pCR rate was 26.3%. We found that High Ki67, low PgR and low ER are associated with a pCR. In the group of ER positive population, a cut-off of Ki67 at 30% was associated in a multivariate analysis with the probability of pCR (> 30% versus 30% - odds ratio= 3.09, IC 95% = 1.13-8.91, p= 0.028) as well as ER 90% (>90% vs 90% - OR = 0.24, IC 95% = 0.07-0.68, p = 0.007). The association of these two factors gave a pCR with an odds ratio of 7.8 (p = 0.0015). Even if pCR was not achieved, we found a molecular down-staging particularly among the luminal B population (defined by IHC criteria) (58.1% change into luminal A). Among patients that did not achieve pCR, several parameters are associated with better survival after surgery. In multivariate analysis, the clinical stage and the decrease of the Ki67 (defined by (final Ki67-initial Ki67)/initial Ki67) are associated with a reduced risk of disease relapse (hazard ratio = 2.93, IC 95% = 1.19-7.24, p = 0.02).


Our data suggest that particularly in HR positive population, pathological characteristics of residual tumor (ER, PgR, Ki67) might be very informative for clinical outcome of patient that did not achieve pCR after NAC.

Clinical trial identification

Legal entity responsible for the study





J. Gligorov: Consulting or advisory role: Eisai, Roche/Genentech, Novartis, Teva, Prizer Honoraria: Eisai, Roche/Genentech, Novartis, Teva, Genomic Health, Prizer Speaker's bureau: Eisai, Roche/Genentech

All other authors have declared no conflicts of interest.