Transcript Do we have the technology and availability needed for clinical

Do we have the technology and
availability needed for clinical
implementation?
The 11th NRI conference
Oslo, May 9th 2016
Hege G. Russnes, MD, PhD
Head at Molecular diagnostics, Dept. of Pathology
and
Researcher, Institute for cancer research
Oslo University Hospital
Increased
knowledge
DNA, Epigenetics,
RNA, protein
TCGA: http://cancergenome.nih.gov
ICGC: https://icgc.org
2014
Technological developments
- extreme increase in knowledge
Why can we not get all this in a routine
diagnostic setting?
Example #1, Personalized Medicine:
challenge when changing classification
and thus treatment stratification of
breast cancer patients
Breast cancer, the diagnosis
•Tissue based diagnostics needed for:
• Prognosis
• Prediction of therapy response
• Selection of patients to clinical trials
•17 different types of breast cancer is recognized by histologic
appearance (WHO)
• No major importance for clinical decisions
Subgrouping breast cancer
Histologic grade
Size, nodal
Age
involvement,
metastases
TNM
ER, PgR, Ki67, HER2
Grade 1
Grade 2
Grade 3Estrogen receptor, ER
HER2/erbB2
HER2/erbB2
The Intrinsic classification
Five subtypes by gene
expression:
• Luminal A
• Luminal B
NON-LUMINAL
ER-
LUMINAL,
ER+
• HER2-enriched
• Basal-like
• Normal-like
Perou et al. PNAS 2001
Molecular subtypes
A relationship between phenotypic and genomic subtypes
Russnes et al. JCI 2011
St. Gallen 2013 recommendations
Treatment by MOLECULAR SUBTYPE
”LumA-like”
”LumB-like”
St. Gallen 2015 recommendations
Treatment by MOLECULAR SUBTYPE
”LumA-like”
Multiparameter molecular marker
”intermediate group”
”LumB-like”
Coates et al, Ann Oncol 2015
Ki67 immunohistochemical staining
~2 % positive (ca 2/100 positive nuclei)
~45% positive (ca 21/47 positive nuclei)
Courtesy, Elin Borgen
~17% positive (ca 9/52 positive nuclei)
~ 93% positive (ca 70/75 positive)
A multigene test
•
•
•
•
•
•
•
•
High reproducibility
Single-sample test
Fast (days-week)
Tissue demands?
Tissue fixation?
Send or in-house?
Cost per sample?
What about the “grey zones”?
EMIT EBC
Establishment of Molecular profiling for
Individual Treatment decisions in Early
Breast Cancer
PI: Bjørn Naume
Co-PIs: Elin Borgen, Vessela N. Kristensen, Hege G. Russnes, Therese Sørlie
Oslo University Hospital
Phase 0:
Retrospective cohort:
Five subgroups
EMITEBC, phase 1:
Standard
histopatology
Treatment according
to guidelines
“Parallel” testing of study-related analyses, comparison of treatment
decision according to molecular analyses vs standard
Who will “change” treatment category?
PAM50 ROR What would be the consequences?
analysis
Intervention study needed?
Analysis at
OUS in
”Clinical mode”
ROR-score
Intrinsic subtype
Feasibility of
PAM50 in routine setting
Other analyses of
primary tumor in
parallel
Test logistics, questionnaires,
health economy
parameters/analyses
Example #2: Personalized Medicine:
Challenge making therapy decision
based on mutation status
MetAction: Actionable Targets in
Cancer Metastasis
- From Bed to Bench to Byte to Bedside
Preliminar
cytology report
IonTorrent
FISH
Molecular Board
MolPat (Validation)
(week 1.5)
Comparison with primary
tumor histopathology
Final pathology report
Final molecular report
PI: G. Mælandsmo, A-L Børresen-Dale, OUH
Tumor Board
(week 2 - 2.5)
Challenging case
•
Female, born 196~
Diagnosed with a large tumor in the ovary July 2014. Both ovaries were involved in the tumor,
forming a mass at the back of the uterus.
•
Surgical treatment:
Hysterectomy and bilateral salpingo-oophorectomi, omentectomi, lymfadenectomi.
•
Histopathology:
Small cell malignant tumor, can represent an undifferentiated small cell sarcoma located
in both ovaries and in pelvis, growing into the parametries bilaterally and in serosa of uterus
including cervix.
•
Tumor phenotype (IHC):
Diffuse positivity for CD10, CD99 and vimentin, focally for inhibin. Tumor was negative for pancytokeratin (AE1/AE3), ER, CD45, myeloperoxidase, HMB-45, Melan A, chromogranin,
synaptophysin, desmin, Myf-4, calretinin og EMA.
=
Targeted
sequencing:
four mutations
Activating mutation
=
Wnt signalling
=
Potentially stabilazing
mutation
HER2
CTNNB1
KRAS
Potentially activating
mutation
RAF
=
Activating mutation
PIK3CA
AKT
nucleus
TCF/c-myc
proliferation
MEK/ERK MEK inhibitors?
proliferation
mTOR mTOR inhibitors?
survival
Challenges:
• Several driver genes affected, but do not know if
all are functional?
• What kind of validation would be needed?
Customize for every patient?
• Not eligible for any trials
• The mutation pattern reflected an ovarian epithelial
carcinoma and not a sarcoma
• Histology and protein analyses (IHC) showed a
dedifferentiated tumor (…sarcomatous)
• Important to have knowledge of genomic alterations as
well as phenotypic information
Example #3: Personalized Medicine:
Challenge dealing with small biopsies
Tissue sampling
Bias in sampling
Not representative!
Representative!
• Methodology in demand of “Fresh” tissue piece; tissue selected
prior to microscopic examination (“blinded”)
• Methodology using FFPE tissue will secure selection of
representative part of tumor by dissection
NB: Intra tumor heterogeneity, subpopulations
NB: Immune response varies within tumors
NB: Small biopsies, representativity
Some challenges for implementation of
Personalized Medicine
• Technology not easily available, too expensive or not
robust/standardized
• Demands large tissue pieces and/or special preservation
• Algorithms/data analyses suited for collections of samples, few are
for single-sample prediction
• Data analyses challenging, time consuming
• Data handling, data storage
• Research cohorts can have different types of selection bias, need
validation in clinical trials
• Treatment regimen changes, a predictor might “loose” its power
Logistics must be suited for “on demand” situations, a test
needs to be performed whenever a patient needs it
Technology and availability…
• Changing classification is demanding
• Predicting therapy response based on mutation
status is challenging
• Diagnostics must be performed in an integrative
setting:
– Clinical information
– Imaging
– Morphology
– DNA alteration
– Phenotype alteration (RNA, protein)
A huge demand for many types of technology and a diversity in competence
– Immune response is needed
Shared resources between “routine” laboratories and research laboratories
Anne-Lise
Børresen-Dale
Hege Russnes’ project
group:
Inga H. Rye
Bente Risberg
Helen Vålerhaugen
Arne V. Pladsen
Jiqiu Cheng
Veronica O. Wang
Geir A. Kongelf
Universitetet i Oslo: Ole CHr. Lingjærde, Arnoldo Frigessi
OSBREAC: Kristine Kleivi Sahlberg, Rolf Kåresen, Bjørn Naume, Anne-Lise Børresen-Dale, Vessela N. Kristensen, Øystein
Fodstad, Jahn M. Nesland, Torill Sauer, Jon Lømo, Øystein Garred, Gunhild Mælandsmo, Tone Baaten, Helle Skjerven,
Jurgen Geissler, Britt Fritzmann, Ellen Schlichting, Olav Engebråten, Solveig Hofvind, Elin Borgen, Gry Geitvik
Hans Kristian M. Vollan, Åslaug Helland, Anna Sætersdal, Therese Sørlie
Kornelia Polyak
Vanessa Almendro
Michael Stratton
Peter Campbell
David Wedge
Anders Zetterberg
Michael Wigler
James Hicks
Carlos Caldas
Sarah-Jane Dawson
Florian Markowetz
Per Eystein Lønning
Stian Knappskog
Peter Van Loo
IHC
Protein
Few proteins
Visual
interpretation,
-simple
-correlation with
type of cell,
histology
Variation
between
interpretators
Easy to
integrate
FISH
DNA
Few genes
Visual
interpretation,
-simple
-correlation with
type of cell,
histology
PCR
Sequencing
RNA
DNA, RNA
Single
genes/subsets
Automatized
interpretation,
-dependent on
software
-No correlation
with type of cell,
Variation between histology
interpretators
Company
Easy to integrate based, difficult to
integrate
Single geneswhole genome
Automatized
interpretation,
-dependent on
software
-No correlation
with type of cell,
histology
From easy to
difficult to
integrate