Etiologic Heterogeneity Etiologic Heterogeneity In Endometrial - - PowerPoint PPT Presentation

Etiologic Heterogeneity Etiologic Heterogeneity In Endometrial Cancer Advances in Endometrial Cancer E id i l d Bi l Epidemiology and Biology March 17-18, 2014

Louise A. Brinton, Ph.D. Di i i f C E id i l d G ti Division of Cancer Epidemiology and Genetics National Cancer Institute h d Bethesda, MD

Background

• Bokhman in 1983 proposed two endometrial

cancer groups based on endocrine/metabolic cancer groups based on endocrine/metabolic functioning and risk factor differences

• Type I cancers correspond pathologically to

Type I cancers correspond pathologically to endometrioid adenocarcinomas, whereas Type II cancers encompass most non- Type II cancers encompass most non endometrioid types (serous as prototype)

• Epidemiologic studies suggest Type II cancers

Epidemiologic studies suggest Type II cancers less strongly linked to classic risk factors, albeit based on small numbers, incomplete , p risk factors, non-standardized pathology

Endometrial Cancer Progression Model g

PTEN mutations Endometrioid cancer (EC) Atypical h l i (EH) Hyperplasia ith t t i PTEN mutations MMR defects ( ) (>90%) hyperplasia (EH) Obesity Normal without atypia

28% risk of invasion

• ver 12 years

5% risk of invasion

• ver 14 years

Hormone exposure Normal endometrium Atrophy Serous cancer (5 10%) Endometrial intraepithelial Atrophy (5-10%) p carcinoma (EIC) TP53 mutations Sherman 2000 Mod Pathol; Lacey 2011 JCO

Etiologic Heterogeneity of d i l i hi h Endometrial Cancer within the NIH-AARP Diet & Health Study

• Cohort was established in 1995-1996 in 6

Cohort was established in 1995 1996 in 6 states and 2 metropolitan areas, by inviting 3.5 million AARP members aged 50-71 years 3.5 million AARP members aged 50 7 years to complete a questionnaire

• Follow-up of 114 409 members through 2006
• Follow-up of 114,409 members through 2006
• 1,491 incident endometrial cancers
• 1,312 Type I and 138 Type II cancers

Definition of Type I and II Tumors

• Type I

Endometrioid, mucinous, tubular, Endometrioid, mucinous, tubular, adenocarcinoma with squamous differentiation, and adenocarcinoma NOS

• Type II

Serous, clear cell, mixed cell, small cell, Serous, clear cell, mixed cell, small cell, squamous cell

Significant Risk Factor Relationships for Type I d II E d i l C NIH AARP S d and II Endometrial Cancers: NIH-AARP Study

Risk Type I (N=1,312) Type II (N=138) P-het factor yp ( , ) yp ( ) N RR 95% CI N RR 95% CI Race Race White 1,228 1.00 referent 115 1.00 referent Black 47 0.66 0.49-0.88 14 2.18 1.24-3.84 Oth 37 0 57 0 41 0 79 9 1 35 0 68 2 67 0 0004 Other 37 0.57 0.41-0.79 9 1.35 0.68-2.67 0.0004 BMI <30 708 1.00 referent 86 1.00 referent >30 570 2.93 2.62-3.2 47 1.83 1.27-2.63 0.001 MHT use N 787 1 00 f 93 1 00 f Never 787 1.00 referent 93 1.00 referent Ever 525 1.18 1.05-1.32 45 0.84 0.57-1.22 0.01

Study Population: GOG-210

• Endometrial cancer patients, recruited at 53

i b 2003 2007 i sites between 2003-2007 into a molecular/surgico-pathological staging trial

• Questionnaires obtained from 91% of trial

participants, enabling risk factor evaluation for disease subtypes among 3,434 patients

• Specialized review conducted for select

p histologies (grade 3 endometrioid, serous, carcinosarcomas, mucinous, clear cell , , tumors, mixed epithelial)

Statistical Analyses

• Logistic regression used to estimate ORs

and 95% CIs with referent group and 95% CIs, with referent group comprised of patients with grades 1-2 endometriod (EM) tumors endometriod (EM) tumors

• ORs >1.0 indicated more common risk

f t ti t ith ith T factors among patients with either Type II or grade 3 EM

• Linear hypothesis testing compared

equality of ORs across specific tumor subgroups (e.g., Type II vs. grades 1-2 EM; Type II vs. grade 3 EM)

Patient Characteristics

• 56% of patients were >60 years of age

d 11% hi and 11% were non-white

• Tumor classification

75% Type I (2,244 grades 1-2 EM, 354 grade 3 EM 17% Type II (321 serous, 141 carcinosarcomas, 77 clear cell, 42 mixed i h l ll ) with serous or clear cell components)

• Type II patients older, more often non-

whites as compared with grades 1-2 EM

ORs for Type II and Grade 3 EM (vs. Grades 1-2 EM) for BMI

BMI (kg/m

2)

n OR (95% CI) % n OR (95% CI)

phet*<0.001 phet**=0.57 Type II, n=581

Normal 119 20.5 Reference

%

Overweight Obese 244 42.0 149 25.6 0.88 (0.66, 1.18) 0.58 (0.45, 0.76)

Grade 3 Endometrioid, n=354

Normal 76 21.5 Reference Overweight Obese 90 25.4 160 45.2 0.62 (0.46, 0.84) 0.91 (0.65, 1.28) 0.1 1.0 10.0 Comparison Group: Grade 1-2 Endometrioid Cancers, n=2,244 Normal: 16.6% Overweight: 19.7% Obese: 51.8%

ORs for Type II and Grade 3 EM (vs. Grades 1-2 EM) for Number of Livebirths

Total number of live births n OR (95% CI) % n

Type II, n=581

1 52 74 12.7 9.0

phet**=0.92 phet*<0.001

OR (95% CI)

Reference 0.91 (0.61, 1.35)

%

2 3 > 4 115 171 29.4 19.8 150 25.8 ( , ) 1.19 (0.86, 1.63) 1.26 (0.90, 1.77) 1.83 (1.31, 2.55)

Grade 3 Endometrioid, n=354

1 54 15.3 39 11.0 Reference 1.04 (0.67, 1.62) 2 3 101 28.5 75 21.2 > 4 75 21.2 1.26 (0.88, 1.80) 1.40 (0.96, 2.06) 1.65 (1.11, 2.46) 0.1 1.0 10.0 Comparison Group: Grade 1-2 Endometrioid Cancers, n=2,244 Nulliparous: 21.4% 1: 13.8% 2: 29.3% 3: 18.4% >4: 13.6%

ORs for Type II and Grade 3 EM (vs. Grades 1-2 EM) for Smoking Status

Smoking status OR (95% CI) n OR (95% CI)

phet*=0.02 phet**=0.40

Type II, n=581

Non-smoker 363 62.5 Reference

%

Current smoker Former smoker 41 7.1 156 26.9 1.80 (1.22, 2.67) 1.09 (0.87, 1.36)

Grade 3 Endometrioid, n=354

Non-smoker 220 62.1 Reference Current smoker Former smoker 31 98 8.8 27.7 1.80 (1.18, 2.75) 1.06 (0.82, 1.38) 0.1 1.0 10.0 Comparison Group: Grade 1-2 Endometrioid Cancers, n=2,244 Non-smoker: 64.9% Current smoker: 6.0% Former smoker: 26.8%

ORs for Type II and Grade 3 EM (vs. Grades 1-2 EM) f B C & T if EM) for Breast Cancer & Tamoxifen use

Breast cancer and tamoxifen use (Br/tam) n OR (95% CI) %

3 3 phet**=0.004 19 466 Br+/tam

( )

Type II, n=581

Reference Br-/tam- phet*<0.001 80.2 7.9 0.3 3.3 3.02 (1.95, 4.67) 1.04 (0.16, 6.55) 2 19 Br+/tam+ Br-/tam+ Br+/tam- 46 1.29 (0.74, 2.27)

Grade 3 Endometrioid, n=354

5.4 87.6 2.25 (1.30, 3.92) Reference Br-/tam- Br+/tam- 19 310 2.5 0.3 0.98 (0.48, 2.02) 1.08 (0.12, 9.55) Br-/tam+ Br+/tam+ 9 1 0.1 1.0 10.0 Comparison Group: Grade 1-2 Endometrioid Cancers, n=2,244 Br-/tam- : 89.4% Br+/tam- : 2.1% Br-/tam+ : 0.2% Br+/tam+ 2.4%

Summary of Findings: Type II vs Grades 1 2 EM Type II vs. Grades 1-2 EM

• Obesity less frequent

– Provides support for a less hormonally- dependent etiology of Type II cancers, although t l t ff t f b it we cannot rule out some effect of obesity

• Multiparity and current smoking more

frequent

– Less protective effect – Hormonal factors likely driving smoking association, but mechanism for parity less clear

• More frequent treatment with Tamoxifen

for breast cancer

Summary of Findings: y g Grades 3 vs. 1-2 EM

• Risk factors for grade 3 EM and Type II

cancers were generally similar

• Patients with grade 3 EM tumors had

more non-Tamoxifen treated breast

• cancer. Possible explanations include:

– Similar risk profiles as for breast cancer p – Radiation treatment of proximate organs – Rare inherited cancer syndromes y – Mutations in cancer predisposing genes

Classification of Type I and II Cancers

• Type I
• 2,244 grades 1-2 EM
• 354 grade 3 EM
• Type II
• 321 serous
• 141 carcinosarcomas

77 l ll

• 77 clear cell
• 42 mixed with serous or clear cell

components) components)

Type II Histologies vs Grade 1 2 EM Type II Histologies vs. Grade 1-2 EM

M lti it d t ki

• Multiparity and current smoking were

more common among women with i d l ll carcinosarcomas and clear cell cancers

– Implies less protective effects for these tumors

• Obesity was less common among those

with serous or clear cell cancers

– Less of a risk factor for these cancers

ORs for Type II Histologic Subgroups (vs. Grades 1 2 EM) f B t C /T if 1-2 EM) for Breast Cancer/Tamoxifen

Breast cancer and tamoxifen use (Br/tam) n OR (95% CI) %

phet<0.001

2 12 Br-/tam- Br+/tam- Br-/tam+

Serous, n=321

Reference 1.53 (0.78, 3.01) 1.52 (0.21, 10.86)

252 78.5 3.7 0.6 25 Br+/tam+

Carcinosarcoma, n=141

3.24 (1.92, 5.46)

7.8 108 76.6 5 3.5

Reference 1.48 (0.56, 3.94)

Br-/tam- Br+/tam-

not estimable

Clear cell, n=77

0.0 11.3 16 69 89.6

5.37 (2.88, 10.03) Reference

Br-/tam+ Br+/tam+ Br-/tam-

not estimable

1 1.3 0.0 3.9 3

0.44 (0.06, 3.30) 1.40 (0.42, 4.66)

Br+/tam- Br-/tam+ Br+/tam+ 0.1 1.0 10.0 Comparison Group: Grade 1-2 Endometrioid Cancers, n=2,244 Br-/tam- : 89.4% Br+/tam- : 2.1% Br-/tam+ : 0.2% Br+/tam+ : 2.4%

Tamoxifen Exposure Among Patients with p g Serous Cancers and Carcinosarcomas

• Possibly unexpected given that

Tamoxifen is a weak estrogen; can also Tamoxifen is a weak estrogen; can also form DNA adducts which could produce cancers via non-hormonal effects cancers via non hormonal effects

• Tamoxifen also associated with benign

endometrial polyps which may undergo endometrial polyps, which may undergo malignant transformation

Etiologic Heterogeneity of Endometrial g g y Cancer: Summary

• Type I tumors largely due to an

imbalance in endogenous hormones g

• Type II patients more likely to be black

and have a poor prognosis and have a poor prognosis.

• Type II tumors more often develop in a

less estrogen dependent manner and are less estrogen-dependent manner and are weakly or unrelated to classic risk factors factors.

Endometrial Cancer Heterogeneity: Endometrial Cancer Heterogeneity: Is Dualistic Model Oversimplified?

• Morphologic, gene expression and

genomic data indicate that endometrial g carcinomas display more diversity than captured by binary classification scheme p y y

• Recent TCGA publication focusing on

endometrioid and serous carcinomas endometrioid and serous carcinomas suggests 4 broad groups based on copy number and exome sequence analysis number and exome sequence analysis

Unsupervised Hierarchical Clustering of Somatic Copy Number Alterations in Endometrial Carcinomas Number Alterations in Endometrial Carcinomas

50/53 serous tumors (94%) 8/13 mixed tumors (62%) 36/289 endometrioid tumors (12%) Nature 497, 67-73 (2013) doi:10.1038/nature12113

Frequency of Molecular Alterations in Type I and II Endometrial Cancers yp

Pathway Type I (%) Type II (%) ER overexpression 47-79 22-30 ER overexpression 47-79 22-30 PR overexpression 74 29 PTEN mutation 37-83 5 PTEN loss of expression 45-61 6-25 Exon 9 16 3 Exon 20 16 12-21 P53 overexpression 16-29 Serous 75-86, Clear cell 23, Gr 3 35 HER2/neu overexpression 9 42 HER2/neu amplification 1-2 17-28 HER2/neu amplification 1-2 17-28 P16 10-30 30-90 ARIDIA 50-70 Gr 3 >60, Clear cell >50, Serous 10- 20 20 Beta-catenin 31-47 0-3 HNF-1B 10-20 Clear cell >40

Endometrial Cancer Heterogeneity: Next Steps with GOG 210 Next Steps with GOG-210

• Create TMAs for 1,000 cases, with
• ver-representation of rare histologies
• Conduct supervised and unsupervised
• Conduct supervised and unsupervised

analyses of data from the TMAs D l l l ifi ti h b d

• Develop novel classification scheme based
• n expression and somatic copy number

lt ti alterations

• Identify antibodies and probes that most

robustly discriminate risk factors, outcomes, and responses to therapy

ORs for Type II Histologic Subgroups (vs. G d 1 2 EM) f BMI Grades 1-2 EM) for BMI

BMI (kg/m

2)

n OR (95% CI) % n

Serous, n=321

Normal Overweight 72 80 22.4 24 9

phet<0.001

OR (95% CI)

Reference 0.76 (0.53, 1.10)

%

Overweight Obese 80 127 24.9 39.6

0.76 (0.53, 1.10) 0.49 (0.35, 0.68)

Carcinosarcoma, n=141

Normal 22 15.6

Reference

Overweight Obese

Clear cell n=77

31 71 22.0 50.4

1.00 (0.56, 1.78) 0.91 (0.55, 1.53)

Clear cell, n=77

Normal Overweight Obese 16 20.8 30 39.0 26 33.8

0.50 (0.26, 0.97) 1.40 (0.74, 2.63) Reference

0.1 1.0 10.0 Comparison Group: Grade 1-2 Endometrioid Cancers, n=2,244 Normal: 16.6% Overweight: 19.7% Obese: 51.8%

ORs for Type II Histologic Subgroups (vs. G d 1 2 EM) f N b f Li bi h Grades 1-2 EM) for Number of Livebirths

Total number of live births n OR (95% CI) %

Serous, n=321

1 2 46 30 77 14.3 9.3 24.0 20 2

phet=0.01

Reference 0.78 (0.47, 1.29) 0.95 (0.64, 1.41) 1 11 (0 73 1 69)

3 > 4 65 90 20.2 28.0

Carcinosarcoma, n=141

1 17 11 12.1 7.8

1.11 (0.73, 1.69) 1.44 (0.95, 2.17) Reference 0.79 (0.36, 1.73)

2 3 > 4 32 30 48 22.7 21.3 34.0

Clear cell, n=77

1 7 6 9.1 7 8

1.09 (0.59, 2.01) 1.35 (0.72, 2.54) 2.12 (1.16, 3.88) Reference 1.14 (0.38, 3.44)

2 3 > 4 6 23 13 26 7.8 29.9 16.9 33.8

( , ) 1.95 (0.82, 4.61) 1.50 (0.59, 3.86) 3.25 (1.35, 7.84)

0.1 1.0 10.0 Comparison Group: Grade 1-2 Endometrioid Cancers, n=2,244 Nulliparous: 21.4% 1: 13.8% 2: 29.3% 3: 18.4% >4: 13.6%

ORs for Type II Histologic Subgroups (vs. G d 1 2 EM) f S ki S Grades 1-2 EM) for Smoking Status

Smoking status n OR (95% CI) %

phet=0.05 Serous, n=321

Non-smoker Current smoker 202 62.9 18 5.6

Reference 1.53 (0.90, 2.62)

Former smoker 91 28.3

1.09 (0.82, 1.44)

Carcinosarcoma, n=141

Non-smoker 59.6 84

Reference

Current smoker Former smoker

Clear cell, n=77

12 36 8.5 25.5

2.32 (1.20, 4.50) 1.07 (0.71, 1.62)

,

Non-smoker Current smoker Former smoker 49 8 18 63.6 10.4 23.4

Reference 2.53 (1.15, 5.58) 0.91 (0.52, 1.58)

0.1 1.0 10.0 Comparison Group: Grade 1-2 Endometrioid Cancers, n=2,244 Non-smoker: 64.9% Current smoker: 6.0% Former smoker: 26.8%

Study Strengths and Limitations

• Strengths

– Large numbers of carefully clinico- Large numbers of carefully clinico pathologically characterized rare cancers – Extensive epidemiologic information p g

• Limitations

– Lack of a control group – Lack of a control group – Risk factor questionnaire self-completed Still limited numbers of rare cancers – Still limited numbers of rare cancers