Hereditary Non-Polyposis Colorectal Cancer (HNPCC) Advisory Work Group Update

1
HEREDITARY NON-POLYPOSIS COLORECTAL CANCER
(HNPCC) ADVISORY WORK GROUP UPDATE by HENRY T.
LYNCH, M.D. With VA Advisory Group
Members Creighton University School of
Medicine Omaha, Nebraska
2
Family History

• Must be comprehensive
• Cancer of all anatomic sites, verification
• whenever possible.

3
Patients Modified Nuclear Pedigree
4
Family History, Cancer Risk,and Diagnosis
Factors obfuscating pedigree interpretation

• incomplete gene penetrance in key
• relatives other than cancer affecteds
• early phenotype (cancer) expression
• incomplete family history
• false paternity
• cancer occurrence may be sporadic
• limited patient and/or M.D. cooperation
• records lost or destroyed.

5
Familial
Hereditary
Lynch Syndrome
AC-1 without MMR
Sporadic
FAP AFAP Mixed Polyposis Syndrome Ashkenazi
I1307K CHEK2 (HBCC) MYH PJSFJP CD BRRS
as yet undiscovered hereditary cancer variants
Hamartomatous Polyposis Syndromes
6
Role of Genetics in Pathogenesis of CRC

• Most cases of LS are not recognized.
• This is partially attributed to poor patient
  recollection of family history and inadequate
  direction by physicians.
• There is a role of the pathologist in selecting
  patients for MSI testing.
• Scand J Gastroenterol 41(Suppl 243)146-152,
  2006.

7
Genetic Events Associated with Colorectal
Tumorigenesis
Loss of DNA mismatch repair
KRAS2 or BRAF mutation
18q loss (DCC)
17p loss (TP53)
DNA Hypomethylation
Normal Epithelium
Proliferative Epithelium
Early Adenoma
Intermediate Adenoma
Late Adenoma
Carcinoma
8

• Pathology
• Carcinoma of Colon
• mucinoous carcinomas
• signet cell carcinomas
• diploid tumors (on flow cytometry)
• TILs
• Adenoma
• 1) Found in 20 of colons with CRC
• 2) Jass and Stewart (Gut 33783-786, 1992)
  adenomas in HNPCC were larger, more often
  villous, and had more high grade dysplasia
• 3) Consistent with our hypothesis that adenomas
  in HNPCC have a greater procilvity for malignant
  degeneration than sporadic adenomas.

HNPCC (Lynch Syndrome I II)
9
(No Transcript)
10
Genetic Heterogeneity in HNPCC
MSH6
MLH1
MSH2
PMS2
PMS1
Chr 7
Chr 3
Chr 2
HNPCC is associated with germline mutations in
any one of at least five genes
11
Lynch Syndrome Genotypic Heterogeneity
Clinical cancer phenotypes differ with each of
the MMR mutations 1. MSH2 has ? extracolonic
cancer types and ? Muir-Torre syndrome 2.
MLH1 may have ? CRC expression. 3. MSH6 may be
more benign with ? CRC but ? endometrial
cancer.
12
Molecular Genetics and Lynch Syndrome

• ?15 of all CRCs show MSI
• Most not caused by Lynch syndrome.
• Acquired silencing (epigenetic inactivation)
• of MLH1 gene by methylation of promoter.
• Immunohistochemistry (IHC) useful for
• identifying protein loss (MSH2, MLH1).

13
Molecular Diagnosis of LS Toward a Consensus

• Molecular diagnostics changed the landscape of LS
  and FAP considerably in the past decade.
• Advantage of mutational testing Identify risk
  status - carrier vs. non-carrier - and therein
  clinical surveillance and management implications
  abound.
• Data show that MSI alone should not be used as a
  basis for selecting patients for mutational
  testing, for LS given the modest but real
  fraction of patients with MSI-negative tumors in
  which mutations were found.
• Lynch et al. J Natl Cancer Inst 99261-263,
  2007.

14
Role of Genetics in the Pathogenesis of
Colorectal Cancer

• Boland notes that MSI is the mutational
  signature in CRCs that evolve as a result of
  inactivation of the DNA MMR system.
• MSI found in ?15 of all CRCs
• ?3 of all CRCs are LS nearly all LS CRCs
  have MSI
• ?12 of CRCs represent the non-inherited form
  of
• DNA MMR inactivation induced by methylation
  of the
• promotor of the MLH1 gene, which silences
  gene
• expression.
• J Clin Oncol 25754-756, 2007.

15
Role of Genetics in the Pathogenesis of
Colorectal Cancer

• MSI is a premier molecular marker in CRC which
  indicates the pathophysiologic genesis of the
  tumor.
• It provides clinical information to use in caring
  for this subset of CRC patients.
• Testing for MSI is within the grasp of molecular
  dx labs.
• Utility in identifying LS families.
• Boland. J Clin Oncol 25754-756, 2007.

16
Molecular Diagnosis of LS Toward a Consensus

• If tumor is MSI-positive, IHC is then done to
  direct mutational testing to a specific MMR gene,
  which MSI alone cannot do.
• If tumor is MSS, must weigh low probability of an
  informative IHC test and cost of performing it.
• Engel et al. Int J Cancer 118115-122, 2006.
• Lynch et al. J Natl Cancer Inst 99261-263,
  2007.

17
Molecular Diagnosis of LS Toward a Consensus

• Decision to perform MSI, some routinely do IHC
  staining concurrently
• Mutational testing if either test is informative
• Performing both tests serves a quality assurance
  role.
• Discrepancies between MSI and IHC occur in up to
  10 of cases further assessment of technical
  issues accounting for the discrepancy can lead to
  performance improvement.
• Lindor et al. J Clin Oncol 201043-1048, 2002.

18
Molecular Diagnosis of LS Toward a Consensus

• In clinically marginal cases, if MSI is present
  and accompanied by loss of MLH1 protein, an
  argument may be made for proceeding with BRAF
  mutation testing and methylation assay, rather
  than expensive MLH1 testing.
• BRAF mutation pretty much excludes LS.
• Domingo et al. J Med Genet 41664-668, 2004.

19
VA Hereditary Non-Polyposis Colorectal Cancer
Advisory Working Group

• Results from meeting on 9/27/07

20
Members

• Henry T. Lynch, MD (Chair)
• John M. Carethers, MD
• Albert de la Chapelle, MD, PhD
• James Eshleman, MD, PhD
• Stephen N. Thibodeau, PhD

21
Charge

• To provide insight whether genetic/genomic
  analysis should become a routine element in
  management of HNPCC and sporadic colorectal
  cancer
• To provide advice on use of immunohistochemical
  and molecular assessment of MSI/mismatch repair
  enzyme defects in HNPCC and sporadic colon cancer

22
Charge (cont)

• To provide advice on therapeutic decision making,
  especially adjuvant chemotherapy in MSI-H
  patients
• The charge was extended to provide advice on a
  research agenda pertaining to the impact of
  genotypic and phenotypic heterogeneity on
  different racial groups.

23
Patient Screening

• Lynch Syndrome is the single condition in which
  real patient benefit can be obtained through
  genetic testing.
• Non uniform approach for diagnostics
• At the university hospital in San Diego, patients
  get IHC testing.
• At the San Diego VAMC, genetic tests are done on
  an individual basis and need approval from the
  local VA administrators.

24
HNPCC and Lynch Syndrome are NOT Synonymous
25

• Albert de la Chapelle
• The Ohio State University

26
Is Lynch syndrome frequent enough to be screened
for? Proportion of unselected cases having LS
(Columbus study 1999-2005)

• Colorectal cancer 44/1566 (2.8)
• Endometrial cancer 13/562 (2.3)
• Hampel et al. NEJM 2005
• Hampel et al. Cancer Res 2006
• Hampel et al. unpublished

27
How many family members of a newly diagnosed
proband with LS will be studied for the mutation,
and how many mutation carriers will be found?
Total No. of new probands 44 No. counseled so
far 33 No. of family members tested 246
(7.5/proband) No. of family members with
mutation 108 (3.3/proband) No. of family members
without mutation 138 (4.2/proband)
Hampel et al NEJM 2005 and unpublished
28
Idealized national figures, USA
Newly diagnosed CRC 150,000
Thereof LS mutation carriers (2.8) 4,200
Family members tested (7.5/proband) 31,500
Family members mutation-positive (3.3/proband) 13,860
Family members mutation-negative (4.2/proband) 17,640

Total
New Lynch syndrome 4,200 13,860 18,060
High-risk individuals returned to average risk 17,640
Hampel et al. NEJM 2005 and unpublished
29
Use of high-risk criteria as a prescreen before
molecular screening
Criteria Met Criteria Did not meet criteria
lt 50 years of age 22 22
Amsterdam II 4 40
Bethesda guidelines 29 15
The Bethesda guidelines have the highest
sensitivity (66) but the of false positives is
high.
Hampel et al. NEJM 2005 and unpublished
30
Choice of molecular prescreenData from Columbus
study N500Mutation positive N18
Microsatellite instability (5 markers) Microsatellite instability (5 markers) Microsatellite instability (5 markers) Microsatellite instability (5 markers)
Positive Mutation positive Sensitivity
High 64 18 18/18 (100)
Low 34 0 0
Immunohistochemistry (4 genes) Immunohistochemistry (4 genes) Immunohistochemistry (4 genes)
Abnormal Mutation positive Sensitivity
71 17 17/18 (94)
Hampel et al. NEJM 2005 and unpublished
31
Recommendations for screening
1. MSI and IHC have similarly high sensitivities
to detect LS. IHC has numerous advantages and
is cheaper 2. Bethesda guidelines as pre-screen
leads to loss of 1/3 of cases but is
significantly cheaper than IHC only 3. IHC only
as prescreen has high sensitivity and reasonable
cost and is therefore preferred
32
Algorithm prepared by our VA Working Group
33
Algorithm for Population Screening
Colorectal Cancer Family History
STOP (sporadic)
IHC

for MLH1
for MSH2 or MSH6
Sequencing (directed by IHC) MLPA (sequential
or concurrent?)
Hypermethylation, BRAF (V600E)

mutation -
mutation

STOP (Lynch)
conversion other tumor testing if available
STOP (sporadic)
34
Algorithm for Population Screening(v.2, if
sequencing cost drops)
Colorectal Cancer Family History
STOP (sporadic)

IHC

Sequencing MLPA
For MLH1 cases

STOP (Lynch)
Hypermethylation, BRAF (V600E)
For MSH2 or MSH6 cases


conversion other tumor testing if available
STOP (sporadic)
35
Algorithm for Moderate/High Risk Patients
(ModerateBethesda HighAmsterdam II)
Colorectal Cancer Family History, Age lt 50,
etc.
Syndrome X ???
STOP (sporadic)
MSI/IHC

for MLH1
for MSH2 or MSH6
Sequencing (directed by IHC) MLPA (sequential
or concurrent?)
Hypermethylation, BRAF (V600E)

mutation -
mutation

Stop or Test blood for germline methylation ???
STOP (Lynch)
Conversion Or Linkage, Or Testing of other
tumor if available

-
STOP (Lynch)
STOP (other)
36
Conclusions Research Still Needed

• Financial studies comparing population and
  moderate/high risk screening are needed.
• A multi-center study comparing a
  non-5-Fluorouracil regimen to 5-Flurouracil in
  MSI-H patients is needed.
• Research is needed about MSI in the
  African-American community.

37
Conclusions (contd)

• Family history is not taken properly.
• There are quality control issues with some
  laboratories.
• Guidelines on the standard of care are needed.
• Educational tools for healthcare workers and
  patients are needed.

38
Molecular Diagnosis of LS Toward a Consensus

• MMR germline mutations achieved the LS diagnostic
  pinnacle!
• However, they were helpful in only 60 of
  Amsterdam-positive cases remainder (40) due to
  not-yet-identified mutations
• Now we have a new phenomenon of LS-like (syndrome
  X) that lack MMR and MSI.
• Lynch et al. J Natl Cancer Inst 99261-263, 2007.

39
AC-I LS Families without MSI

• Results
• AC-I families
• 59.4 with tumor MSI
• 40.6 tumor MSS.
• Valle et al. J Clin Oncol 25781-786, 2007.

40
AC-I LS Families without MSI

• Comparing MSI and MSS groups
• Age of onset (MSI, 41 yrs MSS,53 yrs)
• CRC more frequently proximal in MSI cases
• Fewer mucinous tumors in MSS
• Loss of MMR protein expression in MSI tumors
• Predominance of individuals with multiple
  primary tumors in MSI pedigrees
• ? CRC and endometrial carcinoma in MSS.
• Valle et al. J Clin Oncol 25781-786, 2007.

41
AC-I LS Families without MSI

• Some distinctive clinical pathologic features of
  LS without MSI have been observed
• MSS LS patients older than those with MSI
• Tumors less commonly proximal
• Less often clearly differentiated and
  mucinous
• More often DNA aneuploidy
• Higher proportion of polyps in MSS but
  differences
• do not reach statistical significance.
• Lindor et al. JAMA 2931979-1985, 2005.
• Jass et al. Lancet 3461200-1201, 1995.

42
AC-I LS Families without MSI

• Features of LS without MSI (continued)
• Did not present multiple cancers
• Lesser penetrance than LS with MMR
  deficiency
• Lower incidence of CRC and endometrial cancer
  
• observed in MSS group.
• In the MSS group, more than half the families
  
• showed only CRCs in their pedigrees.
• Lindor et al. JAMA 2931979-1985, 2005.
• Renkonen et al. J Clin Oncol 213629-3637, 2003.

43
AC-I LS Families without MSI

• What do we know about AC-1 families with MSS?
• No genes have been identified as altered
• MSS tumors develop at later age (mean age of
• onset 53 years)
• Most (79) localized in the distal colon and
  rectum
• do not normally produce mucin and have higher
  
• association with polyps (but nonsignificant).
• Valle et al. J Clin Oncol 25781-786, 2007.

44
AC-I LS Families without MSI

• Cancer screening guidelines may differ from those
  of classical LS for familial CRC of undetermined
  type and may not include annual colonoscopy,
  endometrial or ovarian cancer screening.
• Mechanisms of carcinogenesis implicated in
  MSS-hereditary form of CRC remain elusive.
• Is this an entity by itself or does it comprise
  several MSS LS subclasses?
• Too early to exclude conventional
  screening/management.
• Valle et al. J Clin Oncol 25781-786, 2007

45
AC-I LS Families without MSI

• Clinicopathologic differences in MSS AC-1
  families support different entity for syndrome
  within LS.
• Lindor et al. coined the term familial
  colorectal cancer type X
• The term familial CRC of undetermined type has
  recently been proposed.
• Valle et al. J Clin Oncol 25781-786, 2007,
  citing
• Lindor et al. JAMA 2931979-1985, 2005.
• Lynch et al. Eur J Hum Genet 14390-402,2006.

46
Lynch syndrome (LS) and Mismatch Repair (MMR)
GenesWhat Are the Risks?
A new era for genetic counseling evolved thanks
to the cloning of MSH2, MLH1 and MSH6.
Prior to MMR mutation discovery we had to rely
solely on an individuals family history.
Now we can determine lifetime risks for CRC and
extra colonic cancers by MMR testing.
This enables patients to make decisions about
screening, prophylactic surgery, or
chemoprevention strategies.
47
Extracolonic malignancies in HNPCC(Lynch Syndrome)
lt1 Brain 3.7
lt1 Stomach 13
lt1 Biliary Tract 2
lt1 Small Intestine 5
lt1 Urinary Tract 4
1 Ovary 12
1.5 Endometrium 60
2 Colon/Rectum 82
Woman with HNPCC (Age 70)
Average risk woman (Age 70)
48
Cardinal Features of Lynch Syndrome

• Accelerated carcinogenesis

49
(No Transcript)
50
Molecular Diagnosis of LS Toward a Consensus

• Reliance on so-called cardinal principles,
  including
• a. early age of onset (? 44),
• ? synchronous/metachronous CRCs,
• proximal predilection
• b. accelerated CRC carcinogenesis
• c. litany of integral extracolonic cancers
• d. cutaneous stigmata (sebaceous tumors) in the
  
• Muir-Torre syndrome variant
• e. brain tumors (glioblastomas) in Turcots
  syndrome
• variant.
• Lynch et al. J Natl Cancer Inst 99261-263, 2007.

51
Cardinal Features of Lynch Syndrome

• Early onset
• CRC predominance, right sided proclivity, and ?
  synchronous and metachronous occurrences
• Multiple primary cancer excess (pattern
  significant)
• Mismatch repair mutations (MSH2, MLH1, MSH6,
  MLH3, PMS2)

52
CRC Screening in HNPCC
Järvinen et al. evaluated the efficacy of CRC
screening in a controlled HNPCC trial extending
over 15 years. CRC incidence and survival were
compared in 2 cohorts of at-risk members of 22
HNPCC families. 133 subjects had colonic
screening at 3-year intervals, while 119 controls
had no screening.
Gastroenterology 118829-834, 2000.
53
CRC Screening in HNPCC
Results All CRCs in the study group (8, or 6),
compared with 19 controls (16 p 0.014), were
local, causing no deaths, compared with 9 deaths
caused by CRC in the controls. The overall death
rates were 10 vs. 26 subjects in the study and
control groups (p 0.003) in mutation-positive
subjects, the rates were 4 vs. 12 (p 0.05).
Järvinen et al. Gastroenterology 118829-834,
2000.
54
CRC Screening in HNPCC
Conclusion CRC screening at 3-year intervals
more than halves the risk of CRC, prevents CRC
deaths, and decreases overall mortality by about
65 in HNPCC families.
Järvinen et al. Gastroenterology 118829-834,
2000.
55
Colonoscopy
We initiate at age 25 (because of proximal
colonic CRC predilection) and repeat annually
because of accelerated carcinogenesis.
56
J Clin Oncol253534-3542, 2007.
57
Who Should Be Tested?
The molecular genetic revolution of the past
decade has enabled we physicians, and genetic
counselors, to have a level of certainty
regarding patients cancer destiny! However we
need to determine who needs to be tested.
58
Who Should Be Tested?
1. Pedigree consistent with hereditary colorectal
cancer (CRC) syndrome 2. Known germline
mutation predisposing to cancer 3. Patients
at acceptable high cancer risk status 4.
Presence of cancer syndrome stigmata
(phenotype) e.g., polyposis in FAP 5. Genetic
counseling, risks/benefits understood 6. Consent
given 7. Results full explanation of
surveillance/management advice.
59
Who Should Be Tested?

• How do you determine high cancer risk status?
  What to look for
• Cancer-causing mutation known to be
• segregating in family.
• 2. Patient in direct line of descent (affected
  parent,
• sibling, progeny).
• 3. Search for cancer risk stigmata, e.g.,
  multiple
• colonic polyps in FAP, perioral pigmentation
  in
• Peutz-Jeghers syndrome, other.

60
Who Should Be Tested?
How do you determine acceptable high cancer risk
status? Procedure 1. Develop pedigree. 2.
Patient consents, receives genetic counseling. 3.
In absence of existence of known cancer-causing
mutation, may participate in research program
with IRB approval, full consenting process,
patient understands and signs a consent form
following full explanation of risk-benefit.
61
Who Should Be Tested?
At-risk relatives of highly-informative
patient(s) 1. Ideally, inform primary/secondary
relatives, particularly those in clinical
catchment area. 2. Encourage patients to take
responsibility for informing those
relatives. 3. When possible, distribute
educational brochures. 4. Consider FIS.
62
Who Should Be Tested?
Referral to medical geneticist/center of
hereditary cancer expertise 1. When physician
lacks expertise, referral should be made to a
cancer genetics center. 2. Key medical/genetic/gen
ealogic findings made available to referral
center. 3. Patient/family acceptance. 4.
Candidate for DNA testing accepts, is tested,
and results are fully explained in genetic
counseling setting.
63
Who Should Be Tested?
a) Mutation confirms precise dx in proband and
allows testing in relatives Mutation-specific
testing b) Inexpensive compared with
initial test, accuracy near 100 since it
examines presence or absence of the mutation
found in index case c) Screening and
management can be based on these results.
64
Molecular Diagnosis of LS

• Therapeutics
• Patients with CRC with MSI-H did not gain a
  survival benefit with 5-FU, as compared with
  patients with MSS tumors.
• Support has been found in MSI-H CRC cell lines,
  which are more resistant to 5-FU compared with
  MSS cell lines.
• Jo Carethers. Cancer Biomark 251-60,2006.

65
Molecular Diagnosis of LS

• Conclusion
• a. tumor cytotoxicity of 5-FU mediated by DNA
• mechanisms
• b. patients with MSI-H tumors may not benefit
  from
• 5-FU therapy
• c. Future research required for greater
  elucidation of
• cellular mechanisms of the DNA recognition
  of
• 5-FU.
• Jo Carethers. Cancer Biomark 251-60,2006.

66
Family Information Service (FIS)
Cost-effective and highly efficient way of
educating and counseling all available family
members from a geographic catchment area during a
single setting. Makes best use of physicians
time and effort, has group therapy potential and
patients welcome it.
67
(No Transcript)
68
(No Transcript)
69
Role of Genetics in Pathogenesis of CRC

• Kwak Chung estimate that at least 5 of the
  140,000 cases of CRC dx annually are attributable
  to an underlying hereditary CRC syndrome they
  and we believe this is an underestimation.
• These high-penetrance syndromes would include
  Lynch Syndrome (LS), FAP, and MYH-associated
  polyposis.
• They have a high risk for extra colonic-CRCs.
• It is crucial to recognize their unique features,
  molecular genetics, and natural history, for dx
  and management.
• Clin Colorectal Cancer 6340-344, 2007.