Annual Meeting Proceedings Part 1 - American Society of Clinical ...

614 General Poster Session (Board #11A), Sat, 8:00 AM-12:00 PM
Quantitative HER2 measurement and PI3K mutation profile in matched
primary and metastatic breast cancer tissues. Presenting Author: Weidong
Huang, Monogram Biosciences, South San Francisco, CA
Background: HER2 status of primary breast cancer (PBC) is routinely used
to determine systemic treatment for metastatic breast cancer (MBC)
patients. Discordance rates of HER2 status between PBC and MBC range
from 5.5% to 29% based on published meta-analyses. The clinical benefit
of re-assessing HER2 in MBC tissues remains controversial. In this study,
we measured quantitative HER2 expression in matched PBC and MBC
tissues and correlated changes of HER2 with mutations in the catalytic
domain of PI3 kinase(PIK3CA). Methods: Total HER2 protein expression
(H2T) was quantified by the HERmark assay in 41 matched PBC and MBC
formalin-fixed, paraffin-embedded specimens. PIK3CA mutation status in
exons 9 (E545K and E542K) and 20 (H1047R) was determined using a
validated pyrosequencing assay. Results: MBC samples included 5 lymph
node, 13 viscera, 6 brain, and 17 soft tissue lesions (N�41). 27 (66%)
cases showed higher H2T in MBC than in matched PBC; and 14 (34%)
cases had higher H2T in PBC than in matched MBC, indicating an overall
increase of H2T in matched MBC lesions (fold change 0.25-17.57;
p�0.005, paired Wilcoxon rank sum test). HER2 positive conversion
(HERmark negative/equivocal in PBC, but positive in matched MBC) was
found in 6 (15%) cases, while HER2 negative conversion (HERmark
positive in PBC, but negative/equivocal in matched MBC) was seen in 2
(5%) cases. HER2 status was unchanged in 33 (80%) cases. PIK3CA
mutations were detected in 13 (32%) of PBC and 19 (46%) of MBC
samples. Among the HER2 positive conversion cases, PIK3CA mutation
was identified in 50% (3/6) PBC and 67% (4/6) MBC, compared to 0%
(0/2, PBC or MBC) in the HER2 negative conversion cases. Among cases
with unchanged HER2 status, PIK3CA mutation was observed in 30%
(10/33) PBC and 42% (14/33) MBC. Conclusions: Quantitative HER2
assessment revealed a 20% discordance in HER2 status between matched
PBC and MBC tissues, with more frequent conversion from low HER2 in
PBC to high HER2 in MBC. PIK3CA mutation was observed more frequently
in patients who converted from HER2 negative PBC to HER2 positive MBC.
These results suggest that re-assessment of biomarkers in MBC tissues may
better inform the selection of therapeutic options for patients with MBC.
616 General Poster Session (Board #11C), Sat, 8:00 AM-12:00 PM
Can MRI accurately identify which patients with operable breast cancer will
have a pathologic complete response after neoadjuvant therapy? Presenting
Author: Carolyn Nessim, Sunnybrook Odette Cancer Centre, University of
Toronto, Toronto, ON, Canada
Background: With the introduction of targeted therapy based on tumor
subtypes, an increasing number of patients that receive neoadjuvant
chemotherapy achieve a pathologic complete response (pCR). Previous
studies have shown that the accuracy of MRI is poor at predicting the
response to neoadjuvant chemotherapy in locally advanced and often
non-resectable breast cancers, where the rate of pCR is low. The purpose of
this study is to evaluate MRI’s ability to predict a pCR in operable breast
cancers after neoadjuvant therapy. Methods: All patients enrolled in the
NSABP B-40, B-41, FB-5 and FB-6 protocols in a single tertiary care
centre, that had an MRI done before and after neoadjuvant therapy were
reviewed. A radiologist, blinded to the pathology results, interpreted the
pre- and post- treatment MRI’s and made a prediction as to whether or not
patients would have a pCR. In this study, a true negative was defined as a
reading of a complete response on MRI that was confirmed as a pCR on final
pathology. pCR was defined as having no residual invasive or in situ disease
in the breast. Results: 129 women with a median age of 51 years were
identified. 90% had invasive ductal carcinoma; 8% had invasive lobular.
58% were ER�, 21% were triple negative and 21% were Her2�. 16% of
patients had a pCR. 25% of patients had no residual invasive cancer in the
breast. pCR rates for ER� tumors was 5%, triple negative 37%, and Her2�
26%. 19% of patients that had a pCR had a total mastectomy. The
sensitivity and specificity of MRI for predicting residual disease were 88%
and 52% respectively. The positive predictive value was 90% and the
negative predictive value was 46% with an accuracy of 82%. Conclusions:
MRI has limited value for determining which patients had a pCR after
neoadjuvant chemotherapy, even in operable breast cancers. When residual
disease is suspected on MRI, it is unlikely that a pCR has been
achieved. Surgical excision following neoadjuvant therapy remains the gold
standard to identify which patients have achieved a pCR. Other modalities
will need to be used in order to accurately determine which patients would
be eligible for studies evaluating non operative management following
neoadjuvant therapy.
Breast Cancer—HER2/ER
35s
615 General Poster Session (Board #11B), Sat, 8:00 AM-12:00 PM
Long-term follow-up (FU) of patients with durable complete response
(DCR) after chemotherapy (CT) and trastuzumab (T) for HER2-altered
(HER2�) metastatic breast cancer (MBC). Presenting Author: Giuseppe
Gullo, St Vincent’s University Hospital, Dublin, Ireland
Background: Durable CT-induced complete response (DCR) of MBC is
reported only anecdotally. The addition of monoclonal antibody T to
conventional CT results in significant increase in response rates and
prolonged survival for patients (pts) with HER2� MBC. Though CRs are
reported in most phase II and III clinical trials of CT�T, there is lack of data
about clinical features and long-term outcome of these pts. Methods: We
retrospectively reviewed all pts with HER2� MBC treated at our Institutions
with any CT�T and identified cases who achieved DCR. HER2
positivity was defined as 3�score at immunohistochemistry and/or amplification
at FISH test. DCR was defined as a CR according to RECIST 1.0
criteria lasting �36 months. Results: We identified 13 pts with HER2�
MBC who met criteria for DCR following CT�T. Their characteristics are:
median age: 56yrs (range 30-65), stage at diagnosis: M0 54%/M1 46%,
histology: ductal 84%/mixed ductal-lobular 8%/unknown 8%, tumour
grade: G3 54%/G2 23%/unknown 23%, oestrogen receptors (ER): neg
62%/pos 38%, sites of metastatic disease: liver only 54%/other visceral
15%/bone and soft tissues only 31%, chemotherapy regimen:
taxane�carboplatin 54%/single-agent taxane 23%/docetaxel�lapatinib
15%, capecitabine 8%. Metastatic disease was biopsy-proven in 54% of
pts. All pts received maintenance T. Median follow-up (FU) is 6.5 yrs (range
3.0-11.6). Median duration of T was 59 months (17-121). At database
cut-off date 12 pts (92%) are alive. Five pts (38%) had disease relapse (3
on maintenance T, 2 after T discontinuation), 2 of whom had radical
surgery. In 8 pts (62%) disease has not relapsed (characteristics: all
1st-line therapy, ER negative 75%, liver only distant disease 75%).
Conclusions: This is the largest series to our knowledge analyzing long-term
FU of HER2� MBC pts with DCR following CT�T. DCR pts appear to have
prolonged survival and a substantial subgroup of them have no further
disease relapse. These pts are usually treated in the 1st-line setting, have
more frequently ER-negative disease and liver-only metastases. We are
currently investigating the molecular profile of this subset of pts.
617 General Poster Session (Board #11D), Sat, 8:00 AM-12:00 PM
Neuromedin U: A potential predictive biomarker for HER2-targeted drugs.
Presenting Author: Sweta Rani, School of Pharmacy and Pharmaceutical
Sciences and Molecular Therapeutics for Cancer Ireland, Trinity College
Dublin, Dublin, Ireland
Background: Not all HER2-positive breast cancer patients respond to
HER2-targeted drugs and some, who initially respond, subsequently
relapse. There is a need to identify biomarkers for improved patient
selection. Here we aimed to identify gene expression changes associated
with response to HER2-targeted drugs. Methods: To identify mRNA changes
associated with resistance, whole genome microarrays were used to profile
conditioned medium (CM) from HER2-positive cell lines (SKBR3,
HCC1954) and their lapatinib (L)-resistant variants (SKBR3-LR , HCC1954-
LR ). Neuromedin U (NmU) over-expression, identified in this way, was
confirmed in the L-resistant cells and corresponding CM by qPCR and
ELISA. Pooled data from 21 published expression datasets (n�3489
patients) was mined to relate NmU mRNA to tumour subtype and patients’
outcome. NmU cDNA and siRNAs were used to over-express and knockdown
expression of NmU, respectively, prior to assessing it’s association
with response to L, Trastuzumab (T) and Neratinib (N). Results: Analysis of
the tumour data showed NmU expression to be particularly associated with
poor outcome for patients with HER2-positive tumours (p�0.000005). In
cell lines, we identified NmU mRNA and protein to be at significantly
higher levels in L-resistant cells and corresponding CM, compared to that of
L-sensitive SKBR3 and HCC1954. This trend was observed for T-resistant
and N-resistant SKBR3 cells (SKBR3T , SKBR3N ) and their CM, compared
to sensitive parent SKBR3 cells and CM. NmU cDNA over-expression in
SKBR3 and HCC1954 increased resistance to L, T and N. Knock-down of
NmU endogenous levels in SKBR3-LR and innately L-resistant MDA-MB-
361 and T47D sensitised these cells to L, T and N. Further analysis of our
NmU over-expressing and knock-down cells indicated NmU to also be
associated with increased motility, invasion and anoikis resistance.
Conclusions: NmU expression is prognostic of poor outcome in HER2positive
breast tumours. In cell line models, NmU over-expression is
associated with resistance to L, T and N. Taken together, we propose NmU
as a possible prognostic and predictive biomarker for HER2-positive
cancers, with potential also as a co-target to help circumvent resistance to
these drugs. [SFI: 08/SRC/B1410]
Visit abstract.asco.org and search by abstract for the full list of abstract authors and their disclosure information.