2015 San Antonio Breast Cancer Symposium

1488 Pages·2015·3.56 MB·English

2015 San Antonio Breast Cancer Symposium
Publication Number: S1-01
Title: Analysis of molecular prognostic factors associated with tumor immune and stromal microenvironment in BEATRICE, an
open-label phase 3 trial in early triple-negative breast cancer (eTNBC)
Molinero L, Yu J, Li C, Deurloo R, Dent RA A, Bell R, Brown J, Parmar M, Toi M, Suter T, Steger G, Pivot X, Mackey J, Jackisch
C, Hall P, Hegde P, Bais C and Cameron D. Genentech Inc., South San Francisco, CA; F Hoffmann-La Roche Ltd, Basel,
Switzerland; National Cancer Center, Singapore, Singapore; Sunnybrook Health Sciences Center, University of Toronto,
Toronto, Canada; Deakin University, Geelong, Australia; Clinical Trials Research Unit, University of Leeds, Leeds, United
Kingdom; Medical Research Council Clinical Trials Unit, London, United Kingdom; Kyoto University, Kyoto, Japan; Swiss
Cardiovascular Center, Bern University Hospital, Bern, Switzerland; Medical University of Vienna, Vienna, Austria; University
Hospital Jean Minjoz, Besançon, France; Cross Center Institute, Edmonton, Canada; Klinikum Offenbach, Offenbach, Germany;
Leeds Institute of Health Sciences, University of Leeds, Leeds, United Kingdom and University of Edinburgh and Cancer
Services, NHS Lothian, Edinburgh, United Kingdom.
Body: Background: TNBC is a mutationally complex heterogeneous breast cancer subtype. In BEATRICE, adding bevacizumab
to standard adjuvant chemotherapy for eTNBC improved neither invasive disease-free survival (IDFS; primary endpoint) nor
overall survival (OS) [Cameron 2013; Bell SABCS 2014]. We explored prognostic effects of tumor-associated immune and
stromal gene signatures.
Methods: Gene expression (RNA) was assessed in pretreatment archival tumor tissue using an 800-gene nanostring platform.
Given the low event rates and lack of bevacizumab effect in BEATRICE, treatment arms were pooled. The biomarker-evaluable
population (BEP; all patients with an evaluable biomarker sample and ‡1 postbaseline efficacy assessment) was dichotomized
using median gene expression level as the cutoff. Prognostic associations between IDFS/OS and prespecified candidate gene
sets/de novo identified clusters were assessed using univariate Cox proportional hazards models.
Results: Baseline characteristics and efficacy were similar in the BEP (988/2591 randomized pts; 38%) and the overall study
population. In hierarchical cluster analysis based exclusively on immune gene expression, immune genes were enriched in 33%
of samples, intermediate in 38%, and weak in 28%. Further characterization suggested differential prognostic value of distinct
immune and stromal cell gene sets (Table). A significant prognostic effect for IDFS and OS was seen for CD8 effector T cell (T )
eff
and regulatory T cell (T ) gene signatures, but not for the T :T ratio. A less pronounced positive prognostic effect was seen
reg eff reg
for other gene sets representing immune cells, including macrophages, CD4 T cells, and B cells (data not shown). Activated T
helper (Th)-1 cell-derived chemokines and negative immune modulators of T cell activity (eg PD-L1) were highly prognostic for
IDFS and OS. Both the cytokine IL-8 and ESM1 (target of VEGF-A pathway activation) were associated with worse IDFS and OS.
No association was seen between outcome and markers for classic microvasculature (CD31, CD34), cancer-associated
fibroblasts (FAP, BGN, DCN), VEGF-A, or VEGF-C.
IDFS OS
Gene signature HR (95% CI) Interaction p-value HR (95% CI) Interaction p-value
T 0.40 (0.28-0.57) 7.2x10-7 0.29 (0.17-0.49) 4.2x10-6
eff
T 0.38 (0.26-0.54) 1.6x10-7 0.23 (0.13-0.40) 2.9x10-7
reg
T :T ratio 0.80 (0.58-1.12) 0.2 0.89 (0.57-1.39) 0.6
eff reg
Th1 0.45 (0.31-0.64) 8.1x10-6 0.43 (0.27-0.70) 5.8x10-4
PD-L1 0.42 (0.29-0.60) 1.8x10-6 0.24 (0.14-0.41) 3.4x10-7
IL-8 1.48 (1.06-2.08) 0.022 1.89 (1.18-3.01) 0.0076
ESM1 1.73 (1.23-2.43) 0.0017 2.22 (1.38-3.58) 0.001
Conclusions: These molecular gene signature analyses in eTNBC confirm that markers of cytotoxic CD8 T cells are associated
with good prognosis. This is the first report of a positive prognostic effect of regulatory T cell markers, immune checkpoint
modulators, and macrophage-associated markers in the adjuvant TNBC setting. High VEGF-A activity, but not its expression, was
associated with worse prognosis. The strong prognostic effect of immune checkpoint modulators suggests equilibrium between
cytotoxic T cells and their inhibitors in eTNBC, supporting further exploration of immune checkpoint inhibitors in this therapeutic
context.
2015 San Antonio Breast Cancer Symposium
Publication Number: S1-02
Title: Lymphocytic infiltration in invasive lobular breast cancer
Desmedt C, Salgado R, Buisseret L, Zoppoli G, Fornili M, Van den Eynden G, Garaud S, Gundem G, Rothé F, Brown D,
Kheddoumi N, Rouas G, Galant C, Bertucci F, Piccart M, Campbell P, Viale G, Larsimont D, Willard-Gallo K, Biganzoli E, Pruneri
G and Sotiriou C. Institut Jules Bordet, Université Libre de Bruxelles, Brussels, Belgium; University of Genoa, Genoa, Italy;
University of Milan and Istituto Nazionale Tumori, Milan, Italy; Universiteit Antwerpen, Antwerp, Belgium; Cancer Genome
Project, Wellcome Trust Sanger Institute, Hinxton, United Kingdom; Université Catholique de Louvain, Brussels, Belgium; Institut
Paoli-Calmettes, Marseille, France and European Institute of Oncology, Milan, Italy.
Body: Background: The presence and prognostic value of tumor infiltrating lymphocytes (TILs) in invasive breast carcinoma has
been demonstrated in several studies, especially in the triple-negative and HER2-positive subtypes. So far, TILs have not been
investigated with sufficient detail in invasive lobular breast cancer (ILBC). Here we therefore aimed at: first, assessing the
distribution of stromal TILs in ILBC; second, correlating the presence of TILs with standard clinical and pathological markers;
third, exploring associations of TILs with recurrent genomic alterations; and, fourth, comparing the lymphocytic composition of
ER-positive/HER2-negative lobular to ER-positive/HER2-negative ductal tumors.
Material and methods: The percentage of stromal TILs was independently assessed according to Salgado et al. (Ann Oncol 2015)
by three pathologists on full-face hematoxylin and eosin slides in a well-annotated retrospective series of 614 primary ILBCs
previously characterized at the genomic level. The median value of TILs was used for the analyses. For the association analyses,
we focused on the more homogeneous group of ER-positive/HER2-negative ILBC (555/614). Breast cancer-free interval was
used as survival endpoint and the analyses were censored at 12 years of follow-up. The comparison of the lymphocytic
composition (relative percentage of CD45+ TILs which are CD4+, CD8+ or CD19+) was assessed by FACS in a separate
prospective cohort of 51 ER-positive/HER2-negative lobular and 112 ER-positive/HER2-negative ductal tumors.
Results: The intraclass correlation coefficient between the three pathologists was 0.71 (95%CI:0.65-0.76). The median
percentage of stromal TILs was 5% and the interquartile range 5-10%, with only 9% of the samples having ‡ 20%. Greater
numbers of TILs were significantly associated with younger age at diagnosis, axillary lymph node involvement, high proliferative
tumors as assessed by Ki67, and with the mixed non-classic ILBC subtypes. Greater numbers of TILs were associated with
worse prognosis (HR=1.22; 95%CI:1.07-1.38, p=0.003) only in the unadjusted analysis, as it lost significance after adjustment for
standard clinical and pathological variables. Greater numbers of TILs were observed in tumors harboring ARID1A, BRCA2,
KMT2C and TP53 mutations, as well as chr3p21.31 and chr8q24.23 (PTK2) loss; whereas lower numbers were observed in
tumors with ERBB3 mutations as well as chr7p and chr11q14.1 (PAK1) gains. There were no significant differences in the relative
proportion of CD4+, CD8+ or CD19+ lymphocytes between ER-positive/HER2-negative lobular and ductal tumors.
Conclusion: In this work, which reports to our knowledge on the largest series of ILBC ever assessed for TILs, we showed that
most ILBCs were characterized by low lymphocytic infiltration. Besides the association of TILs with clinical and pathological
features of ILBC patients, we found that higher TIL levels were observed in the presence of specific mutations and copy number
alterations. Higher numbers of TILs were associated with worse prognosis at the univariate analysis. Finally, based on the
assessed markers, we have no evidence of differential lymphocytic composition between ER-positive/HER2-negative lobular and
ductal tumors.
2015 San Antonio Breast Cancer Symposium
Publication Number: S1-03
Title: Pooled individual patient data analysis of stromal tumor infiltrating lymphocytes in primary triple negative breast cancer
treated with anthracycline-based chemotherapy
Loi S, Drubay D, Adams S, Francis PA A, Joensuu H, Dieci MV Vittoria, Badve S, Demaria S, Gray R, Piccart MJ J,
Kellokumpa-Lehtinen P-L, Andre F, Dufaure-Gare I, Denkert C, Salgado R and Michiels S. Peter MacCallum Cancer Centre,
East Melbourne, VIC, Australia; Gustave Roussy, Villejuif, France; New York University School of Medicine, NY; Helsinki
University Central Hospital, Helsinki, Finland; University of Padova, Padova, Italy; Indiana University, Indianapolis, IN;
Dana-Farber Cancer Institute, Boston, MA; Institut Jules Bordet, Brussels, Belgium; Tampere University Hospital, Tampere,
Finland and Charite Universite Hospital, Berlin, Germany.
Body: Background:
Retrospective analyses from individual clinical trials have suggested that host anti-tumor immunity as measured by stromal tumor
infiltrating lymphocytes (TILs) is important for the outcomes of the primary triple negative breast cancer (TNBC) subgroup, but the
clinical utility of TILs in day-to-day management of primary TNBC is still limited. Our objective was to conduct a pooled analysis of
the clinical trials that have investigated TILs in TNBC patients treated by anthracyclines-based (A) chemotherapy regimens in
order to gain a robust understanding of the prognostic value of TILs in this setting.
Material and methods:
Methods were predefined in a protocol. Eligible studies were randomized clinical trials that have evaluated the prognostic
associations of TILs (evaluated in the same manner) in patients diagnosed with early stage TNBC treated with A or A plus
taxanes (A+T). Cox regression models stratified by trial for invasive disease-free survival (IDFS, primary endpoint) and overall
survival (OS), fitting stromal TILs as a continuous variable.
Results:
We collected individual data from 991 TNBC patients included in 6 randomized clinical trials (ECOG2197, ECOG1199, BIG2-98,
FinHER, 2 from Gustave Roussy): 62% of patients were treated by A+T and 38% by A alone; 32% of patients had no nodal
involvement, 43% of patients had 1-3 nodes and 25% patients more than 3 nodes involved. The average age was 49 years
(range 22.6-85 yrs) and the average tumor size 3.0 cm (sd 1.7).
Across the entire data set, the average value of stromal TILs was 20% (sd 17%); 90% of patients had at least 1% stromal TILs.
After adjusting for trial, stromal TILs were significantly lower with increasing tumor size (linear model, p<0.0001) but not
significantly associated with nodal status categories (p=0.52 and p=0.37) nor age (p=0.25). With a median follow-up of 6.6 years
for IDFS and 7.3 years for OS, a total of 363 IDFS events and 273 deaths were observed. Each 10% increase in stromal TILs
was associated with a 14% relative reduction in IDFS events (HR=0.86, 95% 0.80 to 0.93, p<0.0001) and a 17% relative
reduction in deaths (HR=0.83, 95% CI 0.76 to 0.91, p=0.0001). There was no significant evidence for heterogeneity between trials
for IDFS (chi2=4.55, p=0.34) nor for OS (chi2=4.45, p=0.34).
In a multivariable analysis adjusted for age, nodal status, tumor size and chemotherapy regimen, stromal TILs added significant
independent prognostic information for both IDFS and OS (likelihood chi2=17.9 for IDFS, p<0.0001 and chi2=16.7 for OS,
p<0.0001). The adjusted hazard ratio for each 10% increase in stromal TILs was HR=0.86 (0.76-0.92) for IDFS events and
HR=0.84 (0.76-0.92) for death.
Conclusion:
This large pooled individual patient data analysis confirms the strong prognostic role of stromal TILS in primary TNBC treated with
A or A+T. TILs should now be strongly considered for incorporation as a stratification factor in future clinical trials enrolling TNBC
patients. Given the important prognostic role of pre-existing immunity, patients with TNBC are rational candidates for
immunotherapy clinical trials.
Funding:Ligue Nationale Contre le Cancer.
2015 San Antonio Breast Cancer Symposium
Publication Number: S1-04
Title: Avelumab (MSB0010718C), an anti-PD-L1 antibody, in patients with locally advanced or metastatic breast cancer: A phase
Ib JAVELIN solid tumor trial
Dirix LY Y, Takacs I, Nikolinakos P, Jerusalem G, Arkenau H-T, Hamilton EP P, von Heydebreck A, Grote H-J, Chin K and
Lippman ME E. Sint Augustinus - University of Antwerp, Antwerp, Belgium; Semmelweis University, Budapest, Hungary;
University Cancer & Blood Center, LLC, Athens, GA; CHU Sart Tilman Liege and Liege University, Liege, Belgium; Sarah
Cannon Research Institute, London, United Kingdom; Sarah Cannon Research Institute, North Nashville, TN; Merck KGaA,
Darmstadt, Germany; EMD Serono, Billerica, MA and University of Miami Miller School of Medicine, Miami, FL.
Body: Background: The programmed death-1 receptor (PD-1) and its ligand (PD-L1) are key therapeutic targets in the
reactivation of the immune response against cancer. Avelumab* (MSB0010718C) is a fully human anti-PD-L1 IgG1 antibody
being investigated in clinical trials. We report clinical activity of avelumab in a cohort of patients (pts) with locally advanced (LA) or
metastatic breast cancer (MBC) refractory to or progressing after standard-of-care therapy (NCT01772004).
Methods: Pts received avelumab at 10 mg/kg Q2W until confirmed progression, unacceptable toxicity, or any criterion for
withdrawal occurred. Tumors were assessed every 6 wks (RECIST 1.1). Unconfirmed best overall response was evaluated.
Adverse events (AEs) were graded by NCI-CTCAE v4.0. Biopsy or surgical specimens were collected within 90 days prior to 1st
dose of avelumab for biomarker analyses. Tumor PD-L1 expression was assessed by immunohistochemistry using various cutoff
criteria.
Results: As of 27 Feb 2015, 168 pts (167 female, 1 male) with MBC, including ductal (56.5%), carcinoma NOS (9.5%), lobular
(3.6%), or other (30.4%), were treated with avelumab and followed for a median of 10 mo (range 6-15). Median age was 55y
(range 31-81), ECOG performance status was 0 (49.4%) or 1 (50.6%), and pts had received a median of 3 prior therapies for
LA/M disease (range 0-10; pts must have received prior treatment with taxane and anthracycline, unless contraindicated). Pts
were HER2–/ER+ or PR+ (69 [41.1%]), triple negative (TNBC = HER2–/ER–/PR–; 57 [33.9%]), HER2+ (26 [15.5%]), or had
unknown biomarker status (16 [9.5%]). Median duration of treatment was 8 wks (range 2-50), and 9 pts (5.4%) remained on
avelumab. Any grade treatment-related treatment-emergent AEs (TEAEs) occurred in 120 pts (71.4%); the most common (>10%)
were fatigue (33 [19.6%]), nausea (24 [14.3%]), and infusion-related reactions (20 [11.9%]). Treatment-related grade ‡3 TEAEs
occurred in 24 pts (14.3%) and included (‡1%) fatigue, anemia, increased GGT, and autoimmune hepatitis (each 3 [1.8%]), and
arthralgia (2 [1.2%]). There were 2 treatment-related deaths (acute liver failure, respiratory distress). Unconfirmed objective
response rate (ORR) in the entire cohort was 5.4% (9 pts; 95% CI: 2.5, 9.9), with 1 CR and 8 PRs. Five of 9 responses were
ongoing at time of cutoff. Stable disease was observed in additional 40 pts (23.8%), for an overall disease control rate of 29.2%.
Evidence of tumor reduction by ‡30% was seen in 15 pts (8.9%). There were responders in all biomarker subgroups, including 5
PRs in TNBC (n=57 [8.8%; 95% CI: 2.9, 19.3]). PD-L1 expression was evaluable in 136 pts. Among all pts with PD-L1 expressing
immune cells within the tumor, 33.3% (4 of 12) had PRs. In pts with TNBC who had PD-L1+ immune cells within the tumor, 44.4%
(4 of 9) had PRs, compared with 2.6% (1 of 39) for TNBC and PD-L1– immune cells.
Conclusions: Avelumab showed an acceptable safety profile and had clinical activity in a subset of pts with MBC. In pts with
TNBC, presence of PD-L1 expressing immune cells within the tumor may be associated with clinical responses to avelumab.
Further analyses of PD-L1 expression and clinical activity of avelumab in MBC are ongoing. *Proposed INN.
2015 San Antonio Breast Cancer Symposium
Publication Number: S1-05
Title: Moved to S1-05
Pituskin E, Mackey JR R, Koshman S, Jassal D, Pitz M, Haykowsky MJ J, Thompson R, Oudit G, Ezekowitz J and Paterson I.
University of Alberta, Edmonton, AB, Canada; University of Manitoba, Winnipeg, MB, Canada; Mazankowski Alberta Heart
Institute, Edmonton, AB, Canada; Cross Cancer Institute, Edmonton, AB, Canada; Bergen Cardiac Care Centre, Winnipeg, MB,
Canada and Cancer Care Manitoba, Winnipeg, MB, Canada.
Body:
2015 San Antonio Breast Cancer Symposium
Publication Number: S1-05
Title: Abstract Withdrawn
Body:
2015 San Antonio Breast Cancer Symposium
Publication Number: S1-06
Title: The miR-424/503 cluster is a breast cancer tumor suppressor with a role in chemoresistance
Silva J, Llobet-Navas D, Rodriguez-Barrueco R, Sanchez-Garcia F and Pe'er D. Mount Sinai School of Medicine, NYC, NY and
Columbia University, NYC, NY.
Body: Recently, we have identified the miR-424(322)/503 cluster as an important regulator of mammary epithelial homeostasis.
The miR-424(322)/503 cluster was identified as one of the few miRNAs significantly upregulated during involution after
pregnancy. By generating a knock-out mouse model, we found that regression of the mammary epithelium after pregnancy was
compromised in the absence of miR-424(322)/503. Mechanistically, our studies unveiled that miR-424(322)/503 is induced by the
canonical TGF-b-SMAD pathway, and that it orchestrates changes in the mammary epithelium by downregulating the expression
of key components of signal transduction (IGF1R) and apoptosis (BCL2) Llobet et al. Genes&Development 2014).
Remarkably, our new studies have revealed that miR-424(322)/503-/- female mice develop hyperplasia and mammary tumors
that are promoted by pregnancy. Thus, we investigated the status of this cluster in human breast cancers. For this we analyzed
the METABRIC dataset. These studies revealed that the miR-424(322)503 cluster was heterozygously deleted in ~16% of breast
cancers and that its deletion correlates with lower expression levels of the mature miRNA forms. Importantly, miR-424(322)/503 is
located on the X-chromosome and we have confirmed that is monoallelically expressed due to X-chromosome inactivation. Thus,
the mutation of the active allele strongly impacts the expression of the miR-cluster. Deletions of the miR-424(322)/503 locus were
more frequent in molecular subtypes with aggressive behavior (Luminal B, HER2+ and Basal) and both deletions and low
expression of the cluster were associated with poor prognosis and reduced survival.
Some of the miR-424(322)/503 targets that we have previously validated (BCL2 and IGF1R) are involved in resistance to
chemotherapy. Thus, we investigated in vivo, utilizing our knock-out mouse model, if loss of miR-424(322)/503 induces resistance
to chemotherapeutic drugs. For this we crossed our miR-424(322)/503-/- animals with the HER2+ model
FVB/N-Tg(MMTVneu)202Mul/J. Tumors emerging in HER2+/miR-424(322)/503-/- animals presented higher levels of BCL2 and
hyperactivation of the IGF1R-AKT signaling compared to HER2+/miR-424(322)/503+/+ counterparts. Furthermore, these tumors
were resistant to standard chemotherapy. Importantly, inhibition of BCL2 with ABT-199 and IGF1R with BMS-754807, two
compounds currently in clinical trial, completely reverted chemotherapy resistance.
Overall, our data present evidence supporting a tumor suppressor role for miR-424(322)/503 cluster and its implication in
resistance to chemotherapy.
2015 San Antonio Breast Cancer Symposium
Publication Number: S1-07
Title: A phase III trial of adjuvant capecitabine in breast cancer patients with HER2-negative pathologic residual invasive disease
after neoadjuvant chemotherapy (CREATE-X, JBCRG-04)
Toi M, Lee S-J, Lee ES, Ohtani S, Im Y-H, Im S-A, Park B-W, Kim S-B, Yanagita Y, Takao S, Ohno S, Aogi K, Iwata H, Kim A,
Sasano H, Yokota I, Ohashi Y and Masuda N. Yeungnam University Hospital, Daegu, Korea; Kyoto University Hospital, Kyoto,
Japan; National Cancer Center, Seoul, Korea; Hiroshima City Hospital, Hiroshima City, Hiroshima, Japan; Samsung Medical
Center, Seoul, Korea; Seoul National University Hospital, Seoul, Korea; Severance Hospital, Yonsei University College of
Medicine, Seoul, Korea; Asan Medical Center, Seoul, Korea; Gunma Prefectural Cancer Center, Ota, Gunma, Japan; Hyogo
Cancer Center, Akashi, Hyogo, Japan; National Kyusyu Cancer Center, Fukuoka, Japan; NHO Shikoku Cancer Center,
Matsuyama, Ehime, Japan; Aichi Cancer Center, Nagoya, Aichi, Japan; Korea University Guro Hospital, Seoul, Korea; Tohoku
University, Sendai, Miyagi, Japan; Kyoto Prefectural University of Medicine, Kyoto, Japan; Chuo University, Tokyo, Japan and
NHO Osaka National Hospital, Osaka, Japan.
WITHELD PENDING PRESS CONFERENCE
2015 San Antonio Breast Cancer Symposium
Publication Number: S1-08
Title: High risk premenopausal luminal A breast cancer patients derive no benefit from adjuvant chemotherapy: Results from
DBCG77B randomized trial
Nielsen TO O, Jensen ‎ M-B, Gao D, Leung S, Burugu S, Liu S, Tykjær Jørgensen CL L, Balslev E and Ejlertsen B. Pathology
and Laboratory Medicine, University of British Columbia, Vancouver, BC, Canada; Danish Breast Cancer Cooperative Group,
Copenhagen, Denmark; Genetic Pathology Evaluation Centre, Vancouver, BC, Canada; Canadian Immunohistochemistry
Quality Control, Vancouver, BC, Canada and University of Ottawa, Ottawa, ON, Canada.
WITHELD PENDING PRESS CONFERENCE