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

10618 General Poster Session (Board #52C), Mon, 1:15 PM-5:15 PM
Biomarkers that predict sensitivity to heat shock protein 90 inhibitors
(HSP90i). Presenting Author: Komal L. Jhaveri, Memorial Sloan-Kettering
Cancer Center, New York, NY
Background: HSP90 is an attractive target in tumors as it mediates the
maturation and stabilization of client oncoproteins. HSP90i are potentially
active in a variety of tumors, but therapeutic benefit is confirmed in only a
small subset to date. Predictive biomarkers could identify patients (pts)
and/or tumors with sensitivity to HSP90i improving the therapeutic index.
We explored potential biomarkers across multiple studies of HSP90i.
Methods: Archived pre-treatment tumor specimens from pts treated with
any of several HSP90i (17-AAG, 17-DMAG, CNF2024, retaspimycin &
ganetespib) on 8 phase I/II trials at MSKCC from 1999 to 2011 were
identified. The following antibodies were validated at MSKCC and tumor
tissue was tested by immunohistochemistry (IHC) with results defined for
each as: ER, PR & AR: �1% pos & �1% neg; HSP90 & HSP70: 0, 1� neg
&2�, 3� pos; PTEN: 0, 1� neg&2�pos; HER2: 0, 1� neg, 2�
equivocal, 3� pos; EGFR: 0 neg & 1�,2�,3�pos. Clinical response was
correlated with IHC using Fisher’s exact test. Results: Of the 164 pts
identified, adequate tissue was available for 51, including 11 different
solid tumors & 1 CML. Breast (N�31), melanoma & prostate (N�4 each)
were most frequent. The mean age at primary diagnosis was 50 yrs (24-81).
The median no. of lines of prior chemotherapy for metastatic disease was 2
(0-8). The mean duration of HSP90i therapy was 79 days (0-411). There
were 19 responses (5 PR; 14 SD); 14/19 responses strongly correlated with
HER2� status (p � 0.001); 1 pt with ER�/HER2-/EGFR- & 4 with EGFR
�/HER2-/ER- disease also responded. A correlation for ER & EGFR with
response was not excluded (p�0.07 & p�0.086 respectively). Conclusions:
Our findings confirm HER2 as a sensitive client and an effective biomarker
for sensitivity to HSP90 inhibitors. ER & EGFR did not meet statistical
significance but may warrant further exploration in prospective settings.
10620 General Poster Session (Board #52E), Mon, 1:15 PM-5:15 PM
Serum activinA and TGF-� as biomarkers of breast cancer bone metastasis
behavior. Presenting Author: Christina L. Addison, Ottawa Hospital Research
Institute, University of Ottawa, Ottawa, ON, Canada
Background: Bisphosphonates (BP) are prescribed to pts with metastatic
bone disease every 3-4 weeks regardless of individual risk for skeletal
related events (SREs). In an era of personalized medicine this “one size fits
all” approach is not appropriate and novel markers of SRE risk are required.
TRIUMPH is an ongoing clinical trial evaluating 12 weekly IV BP therapy
for 1 year in women with low risk bone metastases from breast cancer (BC)
as defined by the bone resorption marker C-telopeptide (CTx,) levels �600
ng/L. This sub-study evaluated the utility of novel biomarkers in better
predicting the risk of developing SREs. Methods: Serum obtained from pts
at baseline and 6 weeks post-entry were analyzed for tumor growth factor-�
(TGF-�) and activinA levels by ELISA (sensitivity ~15-30 pg/ml). Levels
were correlated with pt parameters including time to development of bone
metastasis, and number of previous SREs using linear regression analysis.
Changes in levels of biomarkers from baseline to 6 weeks were used to
calculate odds ratios using logistic regression analysis. Results: Baseline
activinA correlated with baseline CTx and bone specific alkaline phosphatase
(p�0.004 and p�0.0001 respectively). Baseline activinA also correlated
with weight (p�0.02), BMI (p�0.007) and trended towards total
number of prior SREs (p�0.07). Baseline TGF-� correlated with pt age
(p�0.02), weight (0.006), BMI (p�0.0005) and duration of metastatic
bone disease (p�0.004), but did not correlate with any other biomarker.
Change in activinA (baseline to week 6) was the only biomarker that
trended to predict coming off study early (p�0.053) as per protocol (i.e.
CTx�600 ng/ml, SREs or pt/physician choice). Conclusions: Baseline levels
of activinA trended to predict incidence of SREs in patients with bone
metastases, and changes in levels from baseline to 6 weeks trended to
predict coming off study early. These findings warrant future studies in BC
pts assessing activinA as a predictor of risk associated with breast cancer
bone metastases. This study was conducted with the support of the Ontario
Institute for Cancer Research through funding provided by the Government
of Ontario, and with funding from the Ontario Chapter of the Canadian
Breast Cancer Foundation.
Tumor Biology
685s
10619 General Poster Session (Board #52D), Mon, 1:15 PM-5:15 PM
Effect of everolimus on angiogenic biomarkers in patients with tuberous
sclerosis complex (TSC): Results from EXIST-1 and EXIST-2. Presenting
Author: David Neal Franz, Cincinnati Children’s Hospital Medical Center,
Cincinnati, OH
Background: The efficacy and safety of everolimus, an oral mTOR inhibitor,
was assessed in two randomized, double-blind, placebo-controlled, phase
3 trials: EXIST-1 (NCT00789828) and EXIST-2 (NCT00790400). EX-
IST-1 examined everolimus for the treatment of subependymal giant cell
astrocytoma (SEGA) associated with TSC and EXIST-2 for the treatment of
renal angiomyolipoma (AML) associated with either TSC or sporadic
lymphangioleiomyomatosis. In each instance, everolimus was superior to
placebo for the primary endpoints, SEGA and renal AML response rates.
Inhibitors of mTOR have antiangiogenic effects on tumor growth in vitro
and in vivo. Methods: Patients were randomized to receive everolimus
(n�78) starting at 4.5 mg/m2 /day (target trough, 5-15 ng/mL) or placebo
(n�39) in EXIST-1 and 10 mg/day everolimus (n�79) or placebo (n�39)
in EXIST-2. Plasma samples were taken at baseline and pre-dosing on day
1 of weeks 4, 12, 24, 36, and 48 of treatment. Angiogenic markers of
interest were vascular endothelial growth factor (VEGF)-A and -D, placental
growth factor (PlGF), soluble VEGF receptor-1 (sVEGFR1), soluble VEGF
receptor 2 (sVEGFR2), c-Kit, and collagen type IV. Results: Compared with
placebo, a sustained ~30% and ~60% increase in VEGF-A was observed in
the everolimus arm of EXIST-1 and EXIST-2, respectively. A concomitant
decrease in collagen type IV (~25% EXIST-1; ~45% EXIST-2) and
sVEGFR2 (~25% both trials) was also observed in the everolimus arm. A
sustained decrease (~60%) in VEGF-D was observed in the everolimus arm
of EXIST-2, but not EXIST-1. In both studies, no change was observed in
PlGF, sVEGFR1, or c-Kit plasma concentrations in the everolimus arm or
any biomarkers evaluated in the placebo arm. Baseline sVEGFR2 and
VEGF-D were ~40% and ~4-fold higher, respectively, while VEGF-A was
~50% lower in EXIST-2 compared with EXIST-1. A similar baseline plasma
concentration for the other biomarkers was noted in both studies.
Conclusions: Patients presenting with SEGA or renal AML associated with
TSC had a reduction in plasma concentrations of sVEGFR2, collagen type
IV, and VEGF-D (AML only) and an increase in VEGF-A. Everolimus may
have antiangiogenic properties in TSC patients.
10621 General Poster Session (Board #52F), Mon, 1:15 PM-5:15 PM
Identification and validation of plasma protein biomarker panels for breast
cancer diagnosis by using multiple reaction monitoring-based mass spectrometry.
Presenting Author: Hyeong-Gon Moon, Department of Surgery,
Seoul National University Hospital, Seoul, South Korea
Background: Multiple reaction monitoring-based mass spectrometry (MRM-
MS) has the ability to perform a wide range of proteome analysis in a single
experiment using a small volume of specimen. We aimed to develop a
plasma protein signature for breast cancer diagnosis using the MRM-MS
technology. Methods: Previously, we have identified lists of breast cancerrelated
proteins from various models of proteomic discovery including
cancer plasma vs healthy plasma, cancer cell line secretome vs nontumorigenic
cell line secretome, cancer tissue vs normal tissue, and
literature search. Based on these protein panels, total of 29 proteins were
selected for further experiments. We verified and validated the protein
signature in two independent cohorts of breast cancer patients and healthy
women. Results: In the verification cohort of 80 breast cancer patients and
80 healthy women, MRM-MS showed significant differences in plasma
concentration for 11 proteins. Among them, the difference was not
significant for 4 proteins when the cases were limited to stage I and II
patients. Based on p values and consistent expression level along the AJCC
stages, we have created a plasma protein signature comprised of 3 plasma
proteins. The 3 plasma protein signature effectively discriminated cancer
and healthy cases with the AUC of 0.831 (sensitivity 78.7%, specificity
78.7%). The performance of the 3 plasma protein signature was validated
in the cohort of 100 cancer patients and 100 healthy women. The accuracy
of the 3 protein signature was still meaningful with the AUC of 0.746 and
0.797 for all stages and stage I or II patients, respectively. Conclusions: The
3 plasma protein signature for breast cancer diagnosis, developed by the
MRM-MS technology, showed promising results in the present study.
Visit abstract.asco.org and search by abstract for the full list of abstract authors and their disclosure information.