Predicting Individual Breast Cancer Risk Peter Devilee Leiden - - PowerPoint PPT Presentation

Predicting Individual Breast Cancer Risk

Peter Devilee Leiden University Medical Center

Br46 44 Br63 Br45

“Am I at risk?”

10-Year risk to develop breast cancer in the Netherlands

0,00% 0,50% 1,00% 1,50% 2,00% 2,50% 3,00% 3,50% 4,00% 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90

Data: Netherlands Cancer Registry, 1999-2003

Absolute risk Age

Breast cancer incidence in NL

You?

Lifetime breast cancer risk

0%

Risk stratification

100% Communicating risk:

• Relative risk
• Absolute risk
• Lifetime risk
• Cumulative risk
• 10-year risk

Br46 44 Br63 Br45

“How high is my risk?”

Breast cancer prevention

Options

 Chemoprevention  Prophylactic surgery  Screening

Disadvantages

 Side-effects  Over-diagnosis  Increased cost Target those most likely to benefit Identify women at greatest disease risk

Br46 44 Br63 Br45

“Is all this risk genetic?”

Risk factors for breast cancer

 Biology

 Age  Hormonal factors  Breast density on mammogram

 Lifestyle

 Reproductive factors  Alcohol / smoking / physical exercise / obesity  Use of HRT, radiation exposure

 Family history

 Familial relative risk (FRR)

Twin studies

Cancer Proportion of variance due to Heritability Shared environment Non-shared environment Breast 0.31 0.16 0.53 Ovary 0.39 0.61 Lung 0.18 0.24 0.58 Colon 0.15 0.16 0.69 Prostate 0.57 0.43

Cancer incidence among 203.691 twin pairs from Scandinavia

Mucci et al. (2016) JAMA 315:68.

FRR for breast cancer is ~2-fold

Number of Affected First Degree Relatives Relative Risk (99% CI) 1.00

(0.97 – 1.03)

1 1.80

(1.70 – 1.91)

2

2.93

(2.37 – 3.63)

≥3

3.90

(2.03 – 7.49)

Collaborative Group on Hormonal Factors in Breast Cancer (2001)

~12% of all patients and 7% of controls

Prevalence of familial cancer

Cancer Genetic* Familial** Sporadic Breast ~5 ~15 ~85 Colon ~5 ~25 ~75 Prostate ~3 ~15 ~85

* % of cases with a germline mutation in a known susceptibility gene ** % of cases with one or more family members with same disease

1 10 0,000001 0,00001 0,0001 0,001 0,01 0,1 1

Relative Risk Allele frequency

BRCA1 BRCA2 TP53 PALB2 CHEK2 ATM CDH1 STK11 PTEN

Risk SNPs

Genetic architecture of breast cancer

1. BRCA1, BRCA2 2. Non-B1/2 genes 3. Risk SNPs

FRR explained by “genetics”

(25 years of genetic research)

Unexplained: 41%*

BRCA1 BRCA2 CHEK2 ATM PALB2 TP53 PTEN LKB1 27 pre-iCOGS SNPs (9%) ~80 new SNPs

• n iCOGS (5%)

~65 new SNPs Oncoarray (4%) Estimated

• n chip (18%)

* For overall breast cancer in Europeans (Lower for ER-negative disease, early onset disease, and breast cancer in non-Europeans)

Genetic testing in NL

Linkage analysis

1991

BRCA1 BRCA2

1994 1995

CHEK2

2014

Gene panels (Research) PALB2

2018

Specific Cancer Syndromes: TP53, PTEN, CDH1, NF1, . . . .

The result of a genetic test

 A pathogenic variant detected

 What is the breast cancer risk?  Risk to other cancers?

 A pathogenic variant excluded  An unclassified variant detected

Known breast cancer genes: BRCA1, BRCA1, PALB2, ATM, CHEK2 “syndromic” genes: PTEN, TP53, CDH1, NF1, STK11

Uncertainties in risk level

Easton et al. 2015

Low Moderate High

PALB2 CHEK2 2x lifetime risk 4x lifetime risk

Genetic risk Population risk 95% CI

Risk management in NL

Low (RR <2) Moderate (RR: 2-3) High (RR >3) Life time risk <20% 20-30% >30% Start screening 50 yr 40 yr 35 yr Physical examination

• +

Mammography population screening <50 yr: annually >50 yr: population screening <60 yr: annually >60 yr: population screening MRI

• BRCA1/2 mutation carriers:

Panel testing: which genes are relevant?

Gene Breast Ovary BRCA1 BRCA2 TP53 PTEN CDH1 STK11 PALB2 ATM CHEK2 NF1 BRIP1 RAD51C RAD51D

Yes > 2x risk? Probably Unlikely No Unknown

Douglas Easton, unpublished

BRIDGES Gene Panel Sequencing

Studies Controls Cases ER-pos ER-neg Triple negative Population- based 30 41,797 33,538 20,174 5,428 2,044 “Familial/ Other”

(clinic-based, selected for FH/bilaterality)

14 (2,156) 11,411 5,099 1,681 469 TOTAL 44 43,953 44,949 25,273 7,109 2,513

35 Genes from commercial panels

Truncating variants – known genes

BRIDGES consortium, preliminary data

Population-based BRCA1 BRCA2 PALB2 CHEK2 ATM

0.1 1 10 100

Odds Ratio

0.1 1 10 100

Odds Ratio

Familial enriched 1.90 3.34 2.45 5.13 5.32 8.37 5.63 4.95 11.60 7.14

Truncating variants – other DSB

BRIDGES consortium, preliminary data

0.1 1 10 100

Odds Ratio

0.1 1 10 100

Odds Ratio

Population-based Familial enriched BARD1 BRIP1 FANCC FANCM MRE11 NBN RAD50 RAD51C RAD51D XRCC2 (P= .0018) (P= .00011) 2.86 7.15 1.87 2.81

SNP profiles

> 15 million in human genome

Breast cancer associated SNPs

 ~180 identified today  Minor allele frequency in general

population: 5-40%

 Risk per allele: 1.03 – 1.26  Each individual will have an almost unique

pattern of homozygosity/heterozygosity at each of these loci

All breast cancer SNPs

77 SNPs (max 154 risk alleles)

Mavaddat et al (2015), JNCI 107, djv036

(N = 33,381) (N = 33,673)

Relative risk RR = 1

Two-way SNP interactions

Mavaddat et al (2015), JNCI 107: djv036

A polygenic risk score (PRS)

5% of women 5% of women

Under a polygenic multiplicative model Mavaddat et al (2015), JNCI 107: djv036

For women in the highest 1% of the PRS, the estimated OR compared with women in the middle quintile was 3.36 (95%CI: 2.95-3.83, p= 7.5x10-74)

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 20 25 30 35 40 45 50 55 60 65 70 75 Absolute risk Age (years)

All Breast Cancers

>99% 95-99% 90-95% 80-90% 60-80% 40-60% 20-40% 10-20% 5-10% 1-5% <1%

Risk of developing breast cancer

80% of women

All Breast Cancers – Absolute lifetime risk

Mavaddat et al (2015), JNCI 107: djv036

PRS and family history

Mavaddat et al (2015), JNCI 107: djv036

Combining PRS with family history

Lakeman et al., Submitted

Don’t just do panel testing!

0.1 1 10

CHEK2 ATM PALB2 BRCA2 BRCA1

• Approx. Centile

10% <0.1% 1%

Relative Risk

Towards better risk prediction

Genetics Family history Lifestyle Hormonal factors Breast density

Risk Calculator Personal risk estimate

Avoidable risks

Breast cancer risk prediction tools

 Absolute Risk Prediction Models  Gene Carrier Status Risk Prediction

Models

 Risk Prediction Models of Women at High

Risk

Risk stratification: BOADICEA

Lee and Antoniou, submitted

Female, age 20, moderate-risk gene carrier (CHEK2* 1100delC)

Br46 44 Br63 Br45

“Am I at risk?”

BRCA1/2, PALB2, CHEK2 = Negative SNPs + Gene Panel Testing

+ Other risk factors:

• Family history
• Breast density etc.

Examples

Hilbers et al., submitted

* * * *

CHEK2* 1100delC 1.24 1.62 1.02 1.50 1.68 1.55 1.62 PRS-160 Odds Ratio

Outstanding issues

 Which genes confer risk?  Allelic diversity and risk*  Imprecise risk estimates  How do risk factors combine?*

* Journal Club

DNA variation and breast cancer

Long-range enhancers Transcription factor binding sites

5’

Aminoacid substitutions UTR Splicing signals UTR

3’

miRNA binding sites Polyadenylation signals

• Real breast cancer genes
• Candidate breast cancer genes
• Common SNPs

Complementation and HR efficiency

Mesman et al. (2018), Genet Med, in press

Translating functional results into cancer risk

OR 1 OR 1.5 OR 2.0 OR 2.5 OR 3.0

0,2 0,4 0,6 0,8 1 1,2

Relative HR efficiency

Continuous Model

OR > 3

OR 8.0

Shimelis et al. (2017) Cancer Res. 77:2789-2799.

Acknowlegdements

LUMC

 Romy Mesman, Inge

Lakeman, Rick Boonen, Harry Vrieling, Haico van Attikum, Maaike Vreeswijk, Mar Rodriguez

 Clinical Genetics: Christi

van Asperen, Setareh Moghadasi, Juul Wijnen (Inter)national

 Douglas Easton  Jacques Simard  Fergus Couch  Matti Rookus (for Hebon)