Roadmap Complexities of screening Illustrative example: breast - - PowerPoint PPT Presentation

6/23/2016 1

The WISDOM of Screening

May: Laura Esserman, MD MBA Professor, UCSF Departments of Surgery and Radiology June: Yiwey Shieh, MD Instructor, UCSF Division of General Internal Medicine

Roadmap

• Complexities of screening
• Illustrative example: breast cancer

screening

• The current state of risk-based

(precision) screening

• Adopting a risk-based practice in a

time of evidence flux

Available screening tests

Cancer site Test(s) Breast Mammography Prostate PSA Colorectal Colonoscopy, FOBT, sigmoidoscopy Cervical Cytology (Pap smear) Lung Low-dose CT

Complexities of screening

6/23/2016 2

Detectable Metastases Normal Cell Atypical Cell Carcinoma In Situ Stage 1 Cancer Stage 2-3 Cancer Cancer death

Old Paradigm: inexorable progression

Early Detection Will Reduce Mortality

Esserman Lancet Onc 2014

“cancer” is one disease . . .

Detectable Metastasis

New Paradigm: variable progression

Normal Cell Atypical Cell/CIS Stage 1 Cancer Stage 2-3 Cancer Cancer death Detectable Metastasis Normal Cell Stage 1-3 Cancer Cancer death Normal Cell Atypical Cell/CIS Stage 1 Cancer Systemic Therapy Key to Reducing Mortality Early Detection Will Not Impact Mortality Early Detection Reduce Mortality INDOLENT LESIONS RAPID PROGRESSION SLOW PROGRESSION

IDLE condition : Indolent lesions

• f epithelial origin

Indolent Tumors: Rare metastases, course Indolent

The good: mortality has declined

U.S. Men, 1975-2013 U.S. Women, 1975-2013 Colorectal cancer Prostate cancer Colorectal cancer Cervical cancer Breast cancer

seer.cancer.gov

The good: incidence has declined (for some cancer sites)

U.S. Men, 1975-2013 U.S. Women, 1975-2013 Colorectal cancer Colorectal cancer Cervical cancer

seer.cancer.gov

6/23/2016 3

Evidence for benefit in cervical and colorectal cancer screening

• Cervical cancer

– Incidence rate decline in multiple cohort studies – RCTs in India showing mortality benefit of both HPV testing and visual acetic acid exam – Most new cases occur in unscreened women

• Colorectal cancer

– RCTs and cohort studies on sigmoidoscopy

• Drop in incidence rate 18-23%
• Drop in mortality 22-31%

– Modeling: 50% of decline in incidence/mortality attributable to screening

Detectable Metastasis

New Paradigm: variable progression

Normal Cell Atypical Cell/CIS Stage 1 Cancer Stage 2-3 Cancer Cancer death Detectable Metastasis Normal Cell Stage 1-3 Cancer Cancer death Normal Cell Atypical Cell/CIS Stage 1 Cancer Systemic Therapy Key to Reducing Mortality Early Detection Will Not Impact Mortality Early Detection Reduce Mortality INDOLENT LESIONS RAPID PROGRESSION SLOW PROGRESSION

IDLE condition : Indolent lesions

• f epithelial origin

Indolent Tumors: Rare metastases, course Indolent

The bad: incidence has increased (for certain cancers)

U.S. Men, 1975-2013 U.S. Women, 1975-2013

seer.cancer.gov

Colorectal cancer Prostate cancer Colorectal cancer Cervical cancer Breast cancer

Early stage cancers driving increase in incidence rates

Esserman JAMA 2009

Breast Prostate

6/23/2016 4

Overdiagnosis occurs when screening picks up IDLE or indolent disease

IDLE INDOLENT

An example of overdiagnosis:

Thyroid cancer screening in Korea

Ahn NEJM 2014

A thought experiment…

10 20 30 40 50 Fast Slow Very Slow IDLE % of overall cancers detected

Breast

10 20 30 40 50 Fast Slow Very Slow IDLE % of overall cancers detected

Colorectal & Cervical

10 20 30 40 50 Fast Slow Very Slow IDLE % of overall cancers detected

Prostate

Screening does not always prevent metastatic disease at time of first presentation

Welch NEJM 2015

6/23/2016 5

What makes screening so complex?

• Benefits of screening are proportionate to

distribution of biologic tumor types

• Distribution of tumor types is changed by

screening

• Perceptions of screening benefit are also

proportionate to distribution of biologic tumor types

• Impact of screening changes with advances in

treatment

The case for “IDLE” or indolent tumors in breast cancer Defining indolent breast cancers using gene expression profiling

70 significant prognosis genes

Vant Veer Nature 2002

Ultralow Threshold

Ultra Low risk: Threshold determination and locking

Ultra low group (yellow curve):

• Threshold at 0.7
• 100% overall survival @ 25 years (n=8)

Ultra low group (yellow curve):

• Threshold at 0.7
• 100% overall survival @ 25 years (n=8)

Threshold different and refined from Esserman et al BCRT: Nature paper 5 yrs FU

Ultra Low Threshold determined using NEJM publication with 25 years FU data (van de Vijver et al, NEJM 2002; Drukker et al, BCRT 2014)

• Node negative at time of diagnosis
• 100% overall survival at 25 years
• Ultra low risk threshold locked at MP-score 0.7

6/23/2016 6

Indolence = excellent survival regardless of treatment

94% Survival NO Systemic Tx 20 years 97% Survival 20 years

Evidence for indolent prostate cancers

Prostate cancer mortality in low risk (Gleason ≤ 6, PSA ≤ 10) lesions followed with active surveillance

Klotz JCO 2014

30% of Screen Detected Are Categorized as “Ultralow Risk” Cancers

Women aged 49-60

Esserman, Shieh, vant Veer BCRT 2011

70-gene prognosis signature index score distribution

Women aged 49-60

High Unscreened, symptomatic Screened, asymptomatic

Esserman, Shieh, vant Veer BCRT 2011

6/23/2016 7

DCIS Increased 500% after the Advent of Mammographic Screening . . .

10 20 30 40 50 60 70 80 90 100 1975 1980 1985 1990 1995 2000 2005 Incidence rate (per 100,000) Year of diagnosis

Figure 2. SEER9 Age-adjusted incidence rate of breast cancer by stage (1973-2005)

In situ Rate Localized Rate Regional Rate Distant Rate

Localized Regional In Situ Metastatic

Li CEBP 2005

Consequences: Treatment of DCIS

Wells J Am Coll Radiol 2013

Questionable benefit of radiation or mastectomy in DCIS

• Observation of >100,000

women with DCIS

• 20-year breast cancer

mortality = 3.3%

• Mortality rates equivalent

among women treated with lumpectomy, lumpectomy + radiation, and mastectomy

• >50% of women who died
• f breast cancer did not

have in-breast recurrence

Narod JAMA Oncology 2015

For low-grade DCIS, no benefit to surgery

• 57,222 women (SEER)
• 2% (1169) of women

had observation only

• Survival in low grade

DCIS IDENTICAL (98.6 vs 98.8%) for surgery vs. not

6/23/2016 8

Precursor of Indolent tumor fits definition of IDLE

DCIS Dx IDLE Condition

Recap

• Screening will uncover indolent tumors, or

precursors to indolent tumors.

• We must be prepared to deal with them.

PATIENTS ASSUME THAT CANCER, LEFT UNTREATED , WILL KILL YOU

Physicians too

31

Some women screened too much, others too little Women are caught in the middle... and some are choosing not to screen at all

Unintended Consequences of the Screening Controversy

6/23/2016 9

Mammography guidelines, 2016 USA

Organization Start screening at age Terminate screening at age Frequency of assessment United States Preventive Services Task Force (USPSTF) 50 74 Every 2 years (for women at average- risk of breast cancer) American Cancer Society (ACS) 45 As appropriate based on life expectancy Annually then biennially at 55 years of age and

• lder

American College of Obstetricians and Gynecologists (ACOG) 40 As appropriate based on life expectancy Annually American College of Radiology (ACR)/Society

• f Breast Imaging (SBI)

40 As appropriate based on life expectancy Annually

Mammography guidelines, 2016 Rest of world

Country Start screening at age Terminate screening at age Frequency of assessment Canada 50 74 Every 2-3 years United Kingdom 50 (47) 70 (73) Every 3 years The Netherlands 50 75 Every 2 years Australia 50 74 Every 2 years Switzerland recently considering ending mammography screening altogether because

• f lack of evidence that the benefits outweighs the harms.

Biller-Andorno and Jüni, NEJM, 2014.

RISK-BASED SCREENING

A proposed approach

Shieh Nature Rev Clin Onc 2016

6/23/2016 10

Screening cascade, part 1

Shieh Nature Rev Clin Onc 2016

Screening cascade, part 2

Shieh Nature Rev Clin Onc 2016

Choosing whom to screen

• Lung cancer screening

– Age > 55 years – 30+ pack-year history of smoking – If former smoker, quit within 15 years

• BRCA mutation carriers

in breast cancer

– Mammogram + MRI every 6 months (ACS guidelines)

Deciding how often to test

• HPV-negative after age

30 retest in 5 years

• Colonic polyps

– Few small, hyperplastic polyps 10 years – Tubular adenomas 5- 10 years (depending on size, number) – Sessile serrated polyps 3-5 years

6/23/2016 11

Defining the target

Fintelmann RSNA Radiographics 2015

Women Informed to Screen Depending On Measures of risk WISDOM Study Design: Pragmatic Trial

44

Eligible Patients Consent

Randomized Cohort

Randomize

Annual Screening Personalized Screening Observational Cohort Annual Screening Personalized Screening

adapts over time

Risk Calculator

USPSTF

BCSC 9 genes Polygenic risk

Mammogram

• breast density

Athena Health Questionnaire

• family history,

comorbidities, previous biopsies, age, race/ethnicity

Personalized Risk Profile: Risk Assignment notification, assigned screening frequency Genomic profiling

• BRCA, comprehensive

hereditary breast cancer risk gene panel, SNPs

• saliva collection

RBS trial consent RBS Consumer Engagement

Lowest risk Average risk Elevated risk Highest risk

BCSC

Follow-Up:

• Mammography Frequency Assigned by Risk Profile
• Annual Athena Questionnaire to re-assess risk

Cancer detected: Molecular profiling

Breast Health Specialist counseling

No cancer: repeat

45

Personalized Screening Arm

Shared Decision-Making Report

6/23/2016 12

The Personalized Arm: Better Buckets

Age 40-49 with a <1.3% 5-year risk Repeat MMG at age 50 Age ≥50 *OR* >1.3% 5-year risk Biennial MMG Age 40-49 with

• ext. dense breasts

OR BCSC/SNPs 5year risk 4.5-6% Annual MMG Positive Genetic test OR >6% 5-year risk Annual MMG + MRI

All: yearly update of risk

Risk Based Screening Will Teach Us

Who is At Risk for What Kind of Cancer

LEARN

who gets what kind of cancer

CONTINUOUS IMPROVEMENT ADAPT/TAILOR

Prevention Biopsy Treatment Screening

PRACTICE GENERATING EVIDENCE ADAPT

Determine if personalized screening (as compared to annual screening):

• 1. Is as safe
• 2. Is less morbid
• 3. Is more accepted by women
• 4. Enables prevention
• 5. Has greater health care value

WISDOM Study Aims Integrating prevention into risk assessment

Risk category Average Above Average Moderate High Very High Genetic Information <20th percentile based on PRS 20-40th PRS 40-60th PRS CHEK2, ATM, 60-80th PRS PALB2, CDH1, STK11 >80th PRS BRCA1/2, TP53, PTEN Screening Mammogram ≥50 years Every 2 years Mammogram ≥40 years Every 2 years Mammogram Yearly Mammogram MRI Yearly Mammogram MRI Yearly Medical & Lifestyle- based Risk Reduction Lifestyle Tamoxifen, Aromatase inhibitors, Lifestyle Tamoxifen, Aromatase inhibitors, Lifestyle Surgical Risk Reduction BSO BPM

6/23/2016 13

IBIS 2: Anastrazole

Cuzick Lancet 2014

Prevention Trials: Endocrine Risk Reduction

• Investment:

– 100’s of millions spent testing endocrine therapy in the setting of elevated risk in the U.S.; Italy; U.K.

• Impact:

– 50% risk reduction in population w/average Gail Risk (~ 3) – 85% reduction for women with atypia

• Agents

– Tamoxifen, Raloxifen, Exemestane, Arimidex

• Current Uptake:

– Likely less than 5% of eligible women

Lifestyle Modifications

• Lifestyle modifications are recommended as

potential risk reduction strategies for women

• ther than BRCA carriers and include:

– Weight control – No cigarette smoking – Decreased alcohol consumption – Exercise – Discontinue hormone therapy if appropriate

• Lifestyle modifications (especially optimal

weight maintenance after menopause and decreased animal fat) can decrease breast cancer risk by 30-45%

Ross D. 2000 San Antonio Breast Cancer Symposium…need abstract number of updated reference Vogel V. CA Cancer J Clin. 2000;50:156-170. Holmes MD et al. Breast Cancer Res 2004;6:170-178..

How much of cancer risk is modifiable?

• Cohort of ~90,000 F and

~45,000 M

• Healthy lifestyle defined

as:

– 0-5 pack-year smoking – ≤1 or ≤2 alcoholic drinks/day (for F/M) – BMI 18.5-27.5 – weekly aerobic physical activity

Female Male Lung 85% 90% Colorectal 60% 59% Breast 15% Prostate 40%

Population attributable risk (PAR): proportion of cases that would not

• ccur if all individuals adopted the

lifestyle of the low-risk population

Song JAMA Onc 2016

6/23/2016 14

NOMENCLATURE CHANGE

Dictionary.com Definition

can·cer noun

• 1. Pathology
• a. a malignant and invasive growth or tumor, especially one
• riginating in epithelium, tending to recur after excision

and to metastasize to other sites.

• b. any disease characterized by such growths.
• 2. any evil condition or thing that spreads destructively; blight.

“Cancer” today encompasses many diseases with distinct trajectories:

Which should still be called “cancer”?

When Nomenclature Changes . . . Treatment changes

• Cervical Cancer

– CIS CIN 1, 2, & 3 (Bethesda System 1998) – CIN 1 now followed, 50% disappear by 1 year without treatment

• Bladder Cancer

– Superficial bladder cancer papillary urothelial neoplasm of low malignant potential (PUNLMP)

• Thyroid Cancer

– encapsulated follicular variant of papillary thyroid carcinoma noninvasive follicular thyroid neoplasm with papillary-like nuclear features

Course correction

JAMA Oncology April 14, 2016

6/23/2016 15

Other Indolent or IDLE conditions

Cancer site Corresponding IDLE condition Prostate Gleason 3+3 disease Breast Indolent invasive cancers Breast Low-grade DCIS

• Opportunity for watchful waiting or

prevention

• IDLE conditions:

– Should not be targets of screening

INTEGRATING A RISK-BASED MINDSET INTO PRIMARY CARE

“Population-level early diagnosis and screening strategies need to fully engage primary care to maximise their potential.”

Rubin Lancet Onc 2015

“Primary care needs to move beyond its focus on smoking and alcohol use in primary prevention, and engage effectively in initiatives to promote physical activity and reduce obesity.”

health.harvard.edu

Focus on prevention

6/23/2016 16

“PCPs have several roles in cancer screening, including promotion of uptake and informed choice, [and] information provision”

Fintelmann RSNA Radiographics 2015

Risk stratification

• Risk prediction models

– Breast: BCSC risk model (https://tools.bcsc- scc.org/bc5yearrisk/calculator.htm) – Prostate: PCPT risk calculator

• Consider comorbidities: ePrognosis.ucsf.edu

Join the search for a solution

• “Primary care should prepare itself for growth

in genomic information and how this information can be incorporated with lifestyle and other factors to develop individualised preventive strategies.”

• Wisdom study enrolling Fall 2016

http://wisdomstudy.org