Stratifying cancerrisk using family history

1
Stratifying cancer-risk using family history
molecular data

• Tyler Wish, PhD Candidate, Memorial University
• Supervisors Dr. Parfrey, Dr. J. Green, Dr. R.
  Green

2
Outline

• Defining FCC-X
• NL Cancer Risk Data
• Identifying and stratifying high risk subgroups
  with the use of
• Family history criteria
• Molecular data
• Family history scores

3
Autosomal dominant CRC kindredsEtiology unknown

• Familial Colorectal Cancer Type X (FCC-X)
• Defined as Amsterdam I families without MMR
  deficiencies - (Lindor et al, JAMA, 2005)
• 30-60 of Amsterdam Cases fulfill FCC-X
• Compared to Lynch Syndrome
• Later CRC onset (55-60 years)
• Lack of extracolonic cancers Primarily CRC
• Lack of synchronous / metachronous cancer
• More often Left-sided
• Less often mucinous

4
NL Population-based Data

• Microsatellite stable, autosomal dominant, late
  onset colorectal cancer may occur frequently in
  the population.
• Wish T, Green J, Green RC, Woods MO, Younghusband
  HB, Dicks EL, Gallinger S, McLaughlin J,
  Stuckless S, Parfrey PS

5
Research Purpose

• To investigate the relative size of kindreds
  defined by family history criteria and genetic
  definitions of risk
• Focusing on microsatellite stable families
• Characterize their cancer-risk profiles
• To further identify subgroups of high risk cancer
  kindreds based on family histories scores and
  pathology
• Aims
• More precisely identify subgroups of high risk
  familial CRC kindreds
• Identification of high risk kindreds leads to
  proper screening and improved outcomes

6
Familial Risk
7
CRC risk stratified by familial and genetic
definitions
8
Lynch syndrome-Cancer Risk
9
(No Transcript)
10
Conclusions

• FCC-X accounts for a significant proportion of
  Amsterdam kindreds (67)
• Large FCC-X population
• NL FCC-X kindreds have similar cancer risk
  profiles as other populations
• Later age onset
• Primarily CRC
• Absence of synchronous / metachronous cancers

11
What about ACMAC and Bethesda Kindreds?

• Heterogeneity
• Significant heterogeneity amongst kindreds
  fulfilling ACMAC and Bethesda criteria although
  some similarities to FCC-X
• Cancer risk profiles
• Clinical characteristics
• Age of onset
• Types of Cancer CRC versus extracolonic
• Explanations..
• Autosomal dominant disease
• Multiple allele, Low penetrant disease
• Sporadic clustering
• Family History Score
• A tool to independently evaluate familial risk
  within these broad family history criteria

12
Family History Score (FHS)- Yang. Am. J. Epi,
1998.

• Family specific score
• Calculates deviation from expected cancer
  incidence for each individual
• Expected incidence
• age, race, gender specific U.S cancer incidence
  from the Surveillance, Epidemiology, and End
  Results (SEER) database

13
FHS Purpose?

• To identify subgroups of high-risk familial CRC
  kindreds
• Allows a more precise estimation of familial
  cancer risk
• Increases specificity of identifying high risk
  families
• Able to differentiate between CRC kindreds and
  kindreds presenting with multiple malignancy types

14
FCC-ACMAC kindreds (n41)
15
FCC-ACMAC (n27) versus FCC-X (n18)CRC Risk
16
Utility of Family History Score

• FHS enabled us to more precisely identify
  kindreds within ACMAC which are representative of
  FCC-X
• Remaining kindreds appear to be
• other disease?
• sporadic clustering?
• Demonstrated that the FCC-X population is
  potentially twice as large as identified by the
  Amsterdam I definition only

17
What about remaining MSS Bethesda Kindreds?

• Non-ACMAC Bethesda
• Expect a 2-4X increase in risk from population
  risk
• Lifetime CRC risk 15

18
Family History ScoreNon-ACMAC Bethesda MSS
kindreds (intermediate risk families)
19
Non-ACMAC Bethesda MSS kindredsCRC Risk
20
Microsatellite Stable High Risk Kindreds
identified by family history criteria and
FHSFCC-X (n18) / ACMAC (n27) / Bethesda
(n50)
21
Conclusion

• The combined use of family history criteria,
  molecular data, and family history score enables
  a much more precise and informative analysis of
  familial risk
• Our data suggest that potentially 15 of
  families in this study cohort represent
  microsatellite stable autosomal dominant disease
  (FCC-X)
• Sensitivity of identifying FCC-X kindreds with
  Amsterdam I criteria is low
• Size of the FCC-X is potentially significantly
  larger than currently recognized
• The specificity of identifying high cancer risk
  kindreds with current criteria (Bethesda) is low

22
Future Work

• Model to predict cancer risk using clinical
  criteria, FHS, and pathology
• FHS may increase specificity of identifying high
  risk families
• Use available pathology data to further identify
  and stratify cancer risk

23
Many Thanks

• Funding
• CIHR
• Supervisors Dr Parfrey, Dr. Green, Dr. Green
• Co-authors
• Newfoundland Research Team

24
FCC-X FHS
25
Lynch syndrome kindreds - FHS
26
Non-ACMAC Bethesda MSS kindredsLynch syndrome
cancer risk