96. Jahrestagung der Deutschen Gesellschaft für Pathologie e. V ...

Abstracts
FR-029
Whole exome sequencing identifies potential therapeutic
targets for castration resistant prostate cancer
R . Menon 1 , M . Deng 1 , D . Boehm 1 , F . Fend 2 , D . Boehm 3 , S . Biskup 3 , S . Perner 1
1 Institute of Prostate Cancer Research, Institute of Pathology, University Hospital
of Bonn, Bonn, 2 Institute of Pathology, University Hospital Tübingen,
Tübingen, 3 Center for Genomics and Transcriptomics, Tübingen
Aims. Castration resistant prostate cancer (CR-PCa) is the most aggressive
form of prostate cancer (PCa) having a poor prognosis, and is a
significant therapeutic challenge. The key to the development of novel
therapeutic targets for CR-PCa is to decipher the molecular alterations
underlying this lethal disease. The aim of our study was to perform whole
exome sequencing and gene copy number analysis on 5 CR-PCa/normal
paired formalin fixed paraffin embedded (FFPE) samples using the
SOLiD4 next generation sequencing platform.
Methods. Genomic DNA was extracted from 5 CR-PCa/normal paired
FFPE samples. The DNA was subjected to targeted exon capture using
the Agilent Sure Select kit. The captured DNA was sequenced using the
SOLiD4 next generation sequencing platform. The sequencing output
was mapped, sorted, filtered and annotated using well-known human
genome databases. The results were further analyzed for SNPs and copy
number variations. A set of amplified/deleted genes were validated using
fluorescence in-situ hybridization (FISH) assays with a PCa progression
cohort. The cohort consisted of 138 cases for localized cancer, 105 patients
with primary PCa and corresponding LN metastasis, and 39 samples for
castration resistant tumors.
Results. Whole exome sequencing analysis identified focal regions of deletion,
which included well-known tumor suppressors such as NKX3.1
and PTEN. Focal regions of amplification included well-known genes
such as cmYC and AR that are known to play a role in PCa. Furthermore,
we identified several amplified genes as druggable targets e.g. HDAC6,
NTRK1, PLD1, SPHK1, and SIRT7. NTRK1 is a kinase that plays an active
role in cell proliferation. HDAC6, PLD1, SPHK1 and SIRT7 regulate numerous
complex cellular processes including signal transduction, transcription
and apoptosis.
Conclusions. This is the first study to use whole exome sequencing approaches
on FFPE CR-PCa material to identity novel therapeutic targets.
Validation studies would further shed light into the biological understanding
of the disease and its plausible treatment options.
FR-030
Cut-offFinder – a web application for cut-off optimization for
molecular markers
J . Budczies1 , F . Klauschen1 , W . Schmitt1 , C . Denkert1 1Charité Hospital, Institute of Pathology, Berlin
Aims. In order to translate a continuous diagnostic variable into a clinical
decision, it is necessary to determine a cut-off point and to stratify
patients into two groups, each of which requires a different kind of treatment.
Methods. Cut-offFinder is implemented as Java Server Pages (JSPs) that
connect to the statistical engine R. Using a web browser, the user can
upload a molecular data set, assign biomarker and outcome variables
and determine an optimal cut-off point for the biomarker. The web application
offers three different methods for cut-off determination: The
first method fits a mixture model of two Gaussian distribution to the
distribution of the variable. The optimal cut-off is determined as the value
where both probability density functions coincide. For the two other
methods, all possible cut-off points are scanned. The second method
correlates the dichotomized biomarker with a binary outcome variable
using logistic regression. The optimal cut-off is defined as the point with
the most significant (Fisher’s exact test) split. The third method fits Cox
proportional hazard models to the dichotomized variable and the sur-
56 | Der Pathologe · Supplement 1 · 2012
vival variable. Then, the optimal cut-off is defined as the point with the
most significant (log rank test) split.
Results. As example, we have analyzed gene expression data of estrogen
receptor (ESR1) and progesterone receptor (PGR) from a publicly available
microarray data set of 286 breast cancer samples (GSE2034 at www.
ncbi.nlm.nih.gov/geo) using Cut-offFinder. Histograms of the distribution
of ESR1 and PGR showed a clear bimodal shape. Distribution derived
cut-offs were located at 10.6 (ESR1) and 5.0 (PGR). The dependence
on the cut-off of the odds ratio (OR) for correlation ERS1 expression with
ER status (determined by immunohistochemistry) was analyzed and
visualized. The optimal cut-off was determined as 10.1 with OR=67.8
(30.2–152.1). Using ERS1 expression measured by the microarray, determination
of ER status was feasible with a sensitivity of 85.7% and a specificity
of 91.9%. The dependence on the cut-off of the hazard ratio (HR)
for correlation of PGR expression with distance-metastasis-free survival
was analyzed and visualized. The optimal cut-off was determined as 2.5
with HR=0.46 (0.30–0.71). Kaplan Meier analysis showed a significantly
better outcome for patients with high PGR expression (p=0.00028).
Conclusions. In summary, Cut-offFinder is a comprehensive and easyto-use
web application for cut-off determination for molecular markers.
FR-031
BRAF-testing with pyrosequencing: a reliable alternative for the
analysis of highly pigmented and degraded FFPE material
A . Lehmann1 , C . Schewe1 , K . Jöhrens1 , C . Denkert1 , J . Budczies1 , M . Dietel1 1Charité Universitätsmedizin Berlin, Institute of Pathology, Berlin
Aims. The approval of new BRAF inhibitors for the treatment of metastasized
melanoma has led to a great demand for BRAF testing in molecular
pathology laboratories. However, molecular analysis of formalin-fixed,
paraffin-embedded (FFPE) melanoma tissue is challenging. Sanger sequencing
often fails due to high amplification lengths and the influence
of melanin. In this study we tested the applicability of a new pyrosequencing
assay for BRAF analysis on 118 FFPE melanoma tissues and compared
the results with those of Sanger sequencing.
Methods. The study comprised 118 formalin-fixed, paraffin-embedded
tissues of malignant melanoma which were referred to our department
of Molecular Pathology for routine BRAF testing. DNA was extracted
(QIAamp® DNA Mini Kit, Qiagen) and samples were subjected to both,
Sanger sequencing (in-house method) and pyrosequencing (therascreen
BRAF Pyro Kit®, Qiagen) of BRAF exon 15, codons 599 and 600.
Results. BRAF sequences of 102 of 118 samples (86.4%) were evaluable
by both methods pyrosequencing and Sanger sequencing. Mutational
status of these samples was consistent in 98.0% (Cohen’s kappa coefficient
=0.96, p=0.000). Sanger sequencing failed for 15 samples, which
were mostly highly pigmented. Interestingly, 11 of these samples were
still evaluable with pyrosequencing. With a success rate of 95.8% (CI95
[90.5–98.2%]), significantly more cases could be evaluated by pyrosequencing
than by Sanger sequencing [success rate Sanger =87.3%, CI95
(80.1–92.1%); p=0.035].
Conclusions. Pyrosequencing requires comparably short target sequences
for amplification which makes this method feasible even for problematic
material for which standard methods like Sanger sequencing
often fail. Particularly with regard to the high impact of BRAF-testing
for therapy decision, pyrosequencing is a fast and reliable alternative for
BRAF-testing.