Download the ESMO 2012 Abstract Book - Oxford Journals

Annals of Oncology
229P RELATIONSHIP BETWEEN INTESTINAL FUNCTION AND
EXPOSURE TO SORAFENIB AND SUNITINIB IN CANCER
PATIENTS
J. Durand 1 , B. Blanchet 2 , M. Buyse 3 , N. Neveux 4 , P. Boudou-Rouquette 1 ,
O. Mir 1 , M. Vidal 2 , L. Cynober 5 , F. Goldwasser 1
1 Medical Oncology, Cochin Teaching Hospital, Paris, FRANCE,
2 Pharmacology-Toxicology, Cochin Teaching Hospital, Paris, FRANCE,
3 Pharmacy, St Antoine Teaching Hospital, Paris, FRANCE, 4 Clinical Chemistry
Laboratory, Cochin Teaching Hospital, Paris, FRENCH GUYANA, 5 Laboratory of
Biological Nutrition Ea 4466, Université Paris Descartes, Paris, FRANCE
Background: Variability in exposure to sorafenib (SO) and sunitinib (SU), two oral
tyrosine kinase inhibitors (TKI) approved for the treatment of various solid tumours,
is large. Plasma citrulline (CIT) levels reflect the functional small bowel cell mass.
We studied the relationship between TKI exposure and intestinal function in adult
cancer patients (pts).
Methods: CIT concentration and drug exposure were determined in 56 pts under SO
(n = 38) or SU (n = 18) on day (D) 0 (baseline), then D8 and D22 after treatment
initiation. The clinical results led to in vitro studies. In vitro concentration
dependent-cytotoxicity of SO and SU was evaluated in Caco-2 cell lines using
mitochondrial toxicity test (MTT) assay. Under clinically relevant concentrations of
SO (2-10 µg/mL) or SU (0.05-0.10 µg/mL), transepithelial electrical resistance
(TEER) and CIT levels at the basolateral side of Caco-2 cells were determined.
Results: In SO-treated pts, mean citrullinemia on D8 was lower than at D0 (26.2 ±
10.9 mmol/L vs 35.2 ± 13.4, p < 0.0001). Baseline citrullinemia correlated with SO
Area Under Curve (AUC) (r = 0.59, p =0.0001). Median SO AUC on D22 was
1.23-fold lower than SO AUC on D8 (67.0 vs 82.4 mg/L.h respectively, p = 0.033).
Magnitude of decrease in citrullinemia between D0 and D8 correlated with
subsequent decrease in SO AUC (r = 0.61, p < 0.001). In SU-treated pts, baseline
citrullinemia was not correlated with drug exposure (r = -0.29, p = 0.26). Neither
citrullinemia nor SU AUC varied over the study period (p = 0.55 and p = 0.23
respectively). In vitro lethal concentration 50 (LC50) of SO (9.64 + 0.44 µg/mL) was
close to therapeutic concentration unlike SU LC 50 (2.07 + 0.11 µg/mL) which was
40-fold higher. In SO-treated Caco-2 cells, TEER and CIT levels decreased compared
to those of control group (p < 0.0001 and p = 0.017 respectively). SU did not cause
any change of these 2 parameters.
Conclusions: Contrary to SU, SO caused in vitro a cytotoxic effect on Caco-2 cells at
therapeutic concentrations. This likely explains the different kinetics of CIT
concentrations between SO- and SU-treated pts. Correlation between SO AUC and
citrullinemia suggests strongly that functional enterocytic mass contributes to the
intra- and inter-individual variability in sorafenib exposure in cancer pts.
Disclosure: O. Mir: BAYER, PFIZER, ROCHE. F. Goldwasser: BAYER, PFIZER,
ROCHE. All other authors have declared no conflicts of interest.
230P FEASIBILITY OF GENETIC ABERRATIONS ANALYSIS IN THE
CIRCULATING TUMOR CELLS (CTCS)
J. Antonello 1 , E. Rossi 2 , U. Basso 3 , A. Facchinetti 2 , V. Zagonel 3 ,J.
F. Swennenhuis 4 , L.W. Terstappen 4 , A. Amadori 1 , R. Zamarchi 1
1 Immunology and Molecular Oncology, IOV-IRCCS, Padova, ITALY, 2 Department
of Surgery, Oncology and Gastroenterology, University of Padova, Padova,
ITALY, 3 Medical Oncology, IOV-IRCCS, Padova, ITALY, 4 Medical Cell Biophysics,
University of Twente, Twente, NETHERLANDS
Background: In current practice cancer tissue is taken at diagnosis to assess the
presence of treatment targets. This however is suboptimal since tumor cells evolve
due to genomic instability. Assessment of the genotype and phenotype of the CTCs
will provide insights into which treatments would be most beneficial for the
individual patient. Feasibility to detect treatment targets in CTCs has been
demonstrated (Meng, Tripathy et al. 2004; de Bono, Attard et al. 2007; Rossi, Basso
et al. 2010; Wang, Pfister et al. 2010). In this context, the development of a
cytogenetic assay for CTCs will be crucial for successful molecular targeted therapy
in cancer patients. Genetic characterization of CTCs is expected to gather new
knowledge on the mechanism of metastasis and on potential targets of novel
therapeutic strategies.
Patients and methods: Assay optimization for analysis of genetic aberrations of
CTCs was performed with cells from tumor cell lines. Tumor cell lines with known
chromosomal aberrations and mosaicism were spiked into 7.5mL whole blood
samples, at numbers similar to those observed in-vivo in cancer patients. Tumor cells
were enriched by CellSearch System and off-line purified and individually analyzed
for chromosomal alterations. The procedure was next validated by using blood
samples collected from cancer patients.
Results/Conclusions: A robust protocol for the isolation of individual CTCs followed
by DNA extraction and amplification after CellSearch enrichment was established.
The number and the quality of CTCs needed to obtain an informative analysis of
chromosomal aberrations were set up. Accrual of cancer patients for individual CTC
isolation and molecular characterization is ongoing. Updated data including
evaluations on prognostic value of the genetic aberrations analysis of the CTCs and
correlation with clinical stage, tumor burden, metastatic sites and survival will be
available and presented at the meeting.
Disclosure: All authors have declared no conflicts of interest.
231P META-ANALYSIS OF HER3 EXPRESSION AND PROGNOSIS
IN SOLID TUMORS
A. Ocana 1 , F. Vera Badillo 2 , B. Seruga 3 , A. Pandiella 4 ,E.Amir 2
1 Medical Oncology, Albacete University Hospital, Albacete, SPAIN, 2 Medical
Oncology, Princess Margaret Hospital, Toronto, CANADA, 3 Sector of Medical
Oncology, Institute of Oncology, Ljubljana, SLOVENIA, 4 Salamanca Cancer
Research Center, CIC-CSIC, Salamanca, SPAIN
Introduction: Aberrant activation of various ErbB receptors has been linked with
malignant transformation. HER3 is an ErbB family member that can dimerize with
other ErbB receptors such as HER2 and can modulate the transmission of oncogenic
stimuli. The prognostic role of HER3 is not clear, but numerous agents against HER3
are currently in clinical development. Here we present a meta-analysis of studies
evaluating the prognostic impact of HER3 expression.
Material and methods: Pubmed was searched for studies evaluating expression of
HER3 (as measured by immunohistochemistry-IHC) and overall survival (OS) in
solid tumors. Published data were extracted and computed into odds ratios (ORs) for
death at 3 and 5 years. Subgroup analyses were performed to evaluate the association
of HER3 with overall survival in different tumor types. Data were pooled using
generic inverse variance and random-effect modeling.
Results: Eleven studies were eligible for analysis. Median sample size was 126.
Colorectal cancers were evaluated in three studies and gastric in two, while breast,
melanoma, pancreas, ovary, head and neck and cervix cancers were assessed in one
study respectively. The median percentage of cases with HER3 expression was 43%.
HER3 expression was associated with worse OS at both 3 and 5 years (OR: 2.23; 95%
confidence interval [CI], 1.74 to 2.86, p < 0.001 and OR: 2.25; 95% CI, 1.77 to 2.86,
p < 0.001, respectively). Among studies with known HER2 over-expression (breast,
gastric and ovary cancers), the magnitude of effect of HER3 on OS was significantly
greater (3-years OS OR: 3.33, 95% CI, 2.30 to 4.83, p < 0.001 and 5-year OS OR:
3.17, 95% CI, 2.23 to 4.49, p < 0.001).
Conclusion: Expression of HER3 is associated with worse survival in solid tumors.
The effect of HER3 may be greater in those tumors where HER2 is also
over-expressed. A validated method for HER3 assessment is required, but the
potential clinical benefit from targeting HER3 appears favorable.
Disclosure: All authors have declared no conflicts of interest.
232 NESTED CASE CONTROL STUDY OF PROTEOMIC
BIOMARKERS FOR INTERSTITIAL LUNG DISEASE IN
JAPANESE PATIENTS WITH NON-SMALL CELL LUNG CANCER
TREATED WITH ERLOTINIB
N. Katakami 1 , S. Atagi 2 , H. Yoshioka 3 , M. Fukuoka 4 , A. Ogiwara 5 , M. Imai 6 ,
M. Ueda 7 , S. Matsui 8
1 Division of Integrated Oncology, Institute of Biomedical Research & Innovation
Hospital, Kobe, JAPAN, 2 Internal Medicine, National Hospital Organization
Kinki-Chuo Chest Medical Center, Sakai, JAPAN, 3 Respiratory Internal Medicine,
Kurashiki Central Hospital, Kurashiki, JAPAN, 4 Medical Oncology, Izumi
Municipal HospitalCancer Center, Izumi City, JAPAN, 5 Research and
Development, Medical ProteoScope Co., Ltd, Yokohama, JAPAN, 6 Primary
Lifecycle Management, Chugai Pharmaceutical Co., Ltd, Tokyo, JAPAN, 7 Clinical
Research Planning, Chugai Pharmaceutical Co., Ltd, Tokyo, JAPAN, 8 Data
Science, The Institute of Statistical Mathematics, Tachikawa, JAPAN
Background: Interstitial lung disease (ILD) is a serious adverse drug reaction
associated with EGFR tyrosine kinase inhibitors, but its risk factors are yet to be
elucidated. We sought to identify the proteomic biomarkers associated with ILD in
Japanese patients with non-small cell lung cancer treated with erlotinib, and to build
predictive models for development of ILD using the proteomic biomarkers.
Methods: We conducted a nested case control study. The cases were subjects in
whom ILD developed within 120 days after the administration of erlotinib following
enrolment in the cohort, and the controls were randomly selected from patients
without ILD who were treated with erlotinib. For the proteomics analysis, albumin
and IgG were removed from serum samples obtained before the first administration
of erlotinib, then the samples were digested with protease and the resultant peptide
fragments were separated, identified and assayed by LC–MS/MS. Logistic regression
analysis was used for the identification and examination of predictability for ILD.
Results: A total of 645 patients were enrolled in the cohort, 15 cases and 64 controls
were analysed. Logistic regression analysis was performed to identify the peptide
peaks and proteins assossiated with ILD. When multiplicity was taken into account,
we were unable to statistically verify any genuine association between individual
markers and ILD. Investigation of the predictive power based on leave-one-out
cross-validation showed that the area under the ROC curve was 0.73 at a maximum.
Volume 23 | Supplement 9 | September 2012 doi:10.1093/annonc/mds391 | ix91