Transcript EGFR mutations

Diagnostic Molecular Pathology
Update
Christopher D. Gocke, M.D.
Associate Professor of Pathology & Oncology
Johns Hopkins Medical Institutions
Baltimore, MD
April 9, 2011
Disclosures
Christopher D. Gocke, M.D. has no relevant disclosures
Recent advances
•
•
•
•
•
•
•
EGFR testing in lung cancer
New HPV testing recommendations
Expression analysis of breast cancer
Warfarin side effects prediction
HCV genotyping
Molecular screening for MRSA
MPDs defined by JAK2 mutation
EGFR TESTING IN LUNG CANCER
The EGFR tyrosine kinase pathway
ErbB receptor family:
ERBB1 (EGFR)
ERBB2 (HER2)
ERBB3
ERBB4
These form homoonly heterodimers
or heterodimers
Mod Path (2008) 21:S16
EGFR in lung cancer
• NSCLC (80% of all lung cancers) is a
clinical, morphologic and genetic mixed
bag
• Historically, EGFR is over-expressed in
62% and correlates with poor prognosis
• Ligands are also over-expressed 
autocrine loop hyperactivity
• Small molecule inhibitors and antibodies
Hydrophobic pocket
Erlotinib
EGFR therapy in lung cancer
gefitinib and erlotinib reversibly block the kinase
region of the ERBB proteins
Nature Rev Drug Disc (2004) 3:1001
EGFR inhibitor therapy
• Early (phase II) trials
– Response in previously treated patients:
9-19%, median survival 6-8.4 mo (no benefit
over chemo alone)
• But, some subgroups did better
–
–
–
–
Women (19% response vs 3% men)
Adenocarcinoma histology
Asian ethnicity
Never smoker (36% vs 8%)
Why the better response?
• Acquired, somatic mutations in EGFR
– Most responding patients had
heterozygous mutations
– Mutations result in increased receptor
activation (phosphorylation)
NEJM (2004) 350:2129
Mutations are clinically
significant
• Gain-of-function mutations in 77% of
responders vs 7% of refractory patients
EGFR status
(N= 487)
Response rates
(%)
PFS (mo)
OS (mo)
WT
5 – 14
1.7 – 3.6
4.9 – 8.4
Mutant (34%)
16 – 84
9.9 – 21.7
13 – 30.5
• 10-20% responders without EGFR mutations
indicate other molecular causes (e.g., EGFR
amplification or mutations in other ERBB family
members
EGFR mutations
Nat Rev Cancer (2007)7:169
Resistance to EGFR inhibitors
• Develops after 6-12 months in most
• About ½ acquire a T790 exon 20
mutation
• Altered EGFR trafficking, active drug
excretion are possible alternatives
• 2nd generation drugs are targeting
irreversible EGFR binding and other
ERBB family members
Current recommendations for
molecular assays
• FFPE tissue of adenocarcinoma only
(but not mucinous BAC); may need
microdissection
• Direct sequencing is gold standard, but
mutation-specific methods may be OK
• Several replicates should be run
• Heterogeneity exists in tumor and
metastasis—be aware
JCO (2008)26:983
Direct sequencing of EGFR
mutations
NEJM (2004)350:2129
Other considerations for
EGFR testing
• EGFR mutant pts treated with placebo do
better than controls—a prognostic marker
• IHC for protein positive in 50-90% of NSCLC—
1 study (BR.21) suggests survival benefit—very
inconsistent scoring
• Amplification by FISH/RT-PCR positive in 3145%—significant overlap with mutation
positive—unclear how important
• KRAS and EGFR mutations mutually
exclusive—KRAS positive are inhibitor resistant
NEW HPV RECOMMENDATIONS
Infectious disease testing:
Improving the Pap smear
• Relative risk of developing high grade dysplasia
(premalignancy) if infected with high risk HPV is
increased 76-fold
Univ Utah
Guidelines for management of
abnormal cervical screening tests
• Consensus statement guided by
American Society for Colposcopy and
Cervical Pathology
• Am J Obstet Gynecol (2007) 197:346
• Only testing for high-risk (oncogenic)
types of HPV is indicated
Infectious disease testing:
Improving the Pap smear
• Hybrid capture HPV detection (Digene)
1. Lyse cervical cells
2. Hybridize with cRNA
5. Detect label
3. Bind with antiDNA/RNA antibodies
4. Add labeled antiDNA/RNA antibodies
Management of women with
ASC-US
• Reflex testing (of the ASC-US
specimen) for HPV (preferred on cost
basis) OR repeat cytology OR colpo
• Except: adolescents (≤20 yo), no HPV
Management of women with
ASC-H
• Colposcopy
– If negative (no CIN2,3 or greater):
• HPV testing at 12 mo OR
• Cytology at 6 and 12 mo
• Repeat colposcopy if these don’t
normalize
Management of women with
LSIL
• LSIL is a good indicator of high-risk HPV
infection (76.6%)
– Prevalence of CIN2+ in this group is 12-16%
• Colpo with biopsy
– If negative, manage as for ASC (HPV testing at
12 mo OR cytology at 6 and 12 mo)
• Except:
– Adolescents: no HPV testing
– Postmenopausal: monitor with HPV test or
cytology, or go direct to colpo
Management of women with
HSIL
• High rate of HPV positivity and CIN2 or
greater (84-97% by LEEP)
• Colpo with biopsy—HPV testing plays
no role in these patients
Management of women with
atypical glandular cells (AGC)
• High rate of CIN2+ and invasive cancer
(3-17%)
• HPV testing should be done at colpo,
but not as a part of triage
– If these women are biopsy negative:
• And HPV+, repeat HPV/cytology at 6 mo
• And HPV-, repeat HPV/cytology at 12 mo
• And HPV unknown, repeat HPV/cyto at 6 mo
Screening
• Most HPV infections clear spontaneously,
therefore only women ≥30 yr should
undergo HPV testing
Detection of CIN2+ in 35+ y.o. women
Pooled test parameters (Int J Cancer (2006)119:1095)
Sensitivity (%)
Specificity (%)
HPV testing
94
93
Cytology
60
97
Screening (2)
• Final guidance on 2 controversial areas:
– What is the appropriate interval for
rescreening a cyto-/HPV- patient?
3 years (compare to annual Paps); <2%
develop CIN3+ over next 10 years (JNCI
(2005) 97:1072)
Screening (3)
– What to do with cyto-/HPV+ patients?
• A common phenomenon (58% HPV+ women
were cyto- in Kaiser study)
• Most (60%) of HPV+ women become HPVspontaneously over 6 months (BJC (2001)
89:1616)
• Risk of CIN2+ in them is 2.4-5.1%
Follow-up testing in 12 mo, with repeat positives
going to colpo
EXPRESSION ANALYSIS OF
BREAST CANCER
Current management guidelines
for breast cancer
Not uniform consensus
NCCN BINV-6
Gene expression array
• Quick way of examining simultaneous
expression of thousands of genes
• Several uses:
– Class finding (unsupervised clustering):
•
•
•
•
Basal-like (ER/PR/HER2 negative)
Luminal A & B
Clinically aggressive
HER2 over-expressing
Normal-like
(These classes correlate with outcome and
response to therapy, but not beyond standard
clinicopathologic markers)
Gene expression array (2)
– Identification of molecular targets for
specific therapy (e.g., androgen receptor
pathway?)
– Prognostication (supervised analysis)
• Is there a pattern that separates 2 different preidentified populations, e.g., metastasis-free
survival vs not, or cancer recurrence vs not?
Gene expression array (3)
Harvest RNA, label
Read intensity
Hybridize
Wash
Aggregate data
Gene expression array (4)
• Problems
– Tissue requirements: prefer frozen tissue
– Small sample size for validation (hundreds)
– Population bias
– Gene list instability
• E.g., 2 different signatures (Mammaprint and
Rotterdam) share only 3 genes, yet identify the
same groups of low- and high-risk patients
21 gene RT-PCR assay
(Oncotype Dx)
• Not an array, but a specific gene panel
• FDA exempt (CLIA) for node negative,
ER +, tamoxifen-treated
• Provides a “recurrence score” from 1-100
• Requires FFPE tissue
• Performed at company lab in California
21 gene RT-PCR assay
RS <18
RS ≥31
Prognostic
Predictive
JCO (2008) 26:721
70-gene signature
(MammaPrint)
• FDA approved for Stage I-II, nodenegative, ≤5 cm invasive tumor, ≤60 yr
• Predicts risk of distant metastasis in next
5-10 yr
• First IVDMIA cleared by FDA (2/07)
• Requires fresh preserved tumor with a
minimum % tumor
• Performed at company lab in Amsterdam
“Good” = 87% @ 10 yr
“Bad” = 44% @ 10 yr
MammaPrint signature outcome
(without adjuvant therapy)
NEJM (2002) 347:1999
Comparison of commercially available tests
Oncotype DX
MammaPrint
Starting material
FFPE
Fresh mRNA
Number of genes
21
70
Rank of importance
of pathways
1) Proliferation
2) ER
3) HER2
1) Proliferation
2) ER
3) HER2
Current indication
Node -, ER +
Node +/-, ER +/-
Eligible patients
Older
Young and old
Prognostic vs
predictive
Prognostic and
predictive
Prognostic
Outcome prediction
Continuous
Dichotomous
Cost effectiveness?
In 1 study
?
US FDA status
Exempt
Approved
Clinical trial
TAILORx
MINDACT
Study question
Who with intermed.
risk will respond to
chemo?
Who will have good
outcome without
chemo?
Cost
~$3600
~$3200
Oncologist (2008) 13:477
WARFARIN MANAGEMENT
Maintenance dose variability
Molec Interv (2006) 6:223
Warfarin metabolism
Molec Interv (2006) 6:223
Genetic variation in warfarin
related enzymes
• CYP2C9 (hepatic metabolizer):
– *1 (wild-type), rapid metabolizer, ~80%
Caucasian population
– *2 and *3, slow metabolizers, 12.2% and
7.9% allele frequencies—heterozygotes
take twice or three times as long to achieve
steady-state when their daily dose is
adjusted down
Genetic variation in warfarin
related enzymes
• VKORC1 (Vit K production)
(transcriptional variants)
– B (wild-type), high-dose phenotype, 60%
allele frequency in Caucasian (wide
variation reported)
– A, low-dose, 40% Caucasian (most Asians)
Clinical variability in warfarin
dosing
Genet Med (2008) 10:89
FDA action on warfarin
• 2006: adds a “black box” warning for
Coumadin about risk of severe or fatal
bleeding
• 8/2007: adds notice that genetic
variation in CYP2C9 and VKOR1
influence phenotype
• 9/2007: clears Nanosphere’s Verigene
test for identification of some alleles
Nanogold DNA genotyping
technology
Science (2000) 289:1757
Problems with genotyping for
warfarin dosing
• Most studies use INR as endpoint; little
data from randomized studies on direct
clinical utility in predicting/preventing
bleeding or thrombosis (underanticoagulation)
• Little data in non-Caucasians, children
• How to present dosing advice to clinicians
• Ethical, legal, social implications