The role of MRI in Screening for Volpara Solutions Breast Cancer in - - PowerPoint PPT Presentation

6/7/2017 1

The role of MRI in Screening for Breast Cancer in Dense Breasts: the DENSE trial

Carla H. van Gils, PhD

Disclosure

• Research funding or support

Industry: – Bayer AG – Volpara Solutions Governmental and charity funding:

• ZonMW
• Dutch Cancer Society
• Pink Ribbon / A Sisters Hope
• Breast screening in The Netherlands
• MRI in dense breasts
• DENSE trial

Outline

• Nationwide since 1990; fully digital since 2010
• Biennial mammography
• 50-75 years
• Participation: 79%
• >1.2 mln examinations per year
• No breast density laws
• Breast density not routinely assessed in screening

Netherlands population-based breast screening program

6/7/2017 2

• ~112,000 examinations, with Volpara measurements
• 2003-2011
• 50-75 years

Breast density distribution in NL

22% 41% 29% 8%

Fatty – category 1 Scattered – category 2 Heterogeneous – category 3 Extremely dense – category 4 Wanders et al. Breast Cancer Res Treat 2017;162:95-103

Breast density, risk & masking:

Biennial digital population- based screening program

VDG 1 n=24,210 VDG2 N=46,426 VDG3 N=32,330 VDG 4 N=8,932 p-trend Total BC /1000 exam. 4.6 8.3 9.4 11.2 <0.001 Screen-det BC /1000 exam. 4.0 6.4 6.6 6.8 <0.001 Interval BC/1000 exam. 0.7 1.9 2.9 4.4 <0.001 Sensitivity 85.7% 77.6% 69.5% 61.0% <0.001 VDG = ‘volpara density grade’ measured with Volpara software BC/1000 = screen-detected tumors + interval tumors/1000 Sensitivity = # screen-detected tumors/ (# screen-detected tumors + # interval tumors) Wanders et al. Breast Cancer Res Treat 2017;162:95-103

Is MRI an option? What is known on use of MRI in dense breasts?

• 612 women with ACR density 3/4 and ≥ 1other risk factor
• MRI after 3 annual rounds of Mx+US

Study No. Exam. Extra cancer detection rate 95% CI Recall rate 95% CI PPV 95% CI Berg et al. 612 18/1000 (6-30) 23/100 (19-26) 9% (5-14%)

6/7/2017 3

What is known on use of MRI in dense breasts?

• Subgroup: excluding those with BRCA1/2 mutations, chest

radiation or personal history of breast cancer

Study No. Exam. Extra cancer detection rate 95% CI Recall rate 95% CI PPV 95% CI Berg et al. 612 18/1000 (6-30) 23/100 (19-26) 9% (5-14%) Melnikow et al. subgroup 334 21/1000 (9-43) 23/100 (19-28) 9% (4-17%)

What is known on use of MRI in dense breasts?

Study No. Exam. Extra cancer detection rate 95% CI Recall rate 95% CI PPV 95% CI Kuhl et al. (1st round) 443 25/1000 (13-45) 7/100 (5-10) 33% (18-52%)

• 443 women with a mildly to moderately increased risk
• Of which 427 with ACR density 3 or 4 and negative mx+US

What is known on use of MRI in dense breasts?

• Subgroup: excluding those with BRCA1/2 mutations, chest

radiation or personal history of breast cancer

Study No. Exam. Extra cancer detection rate 95% CI Recall rate 95% CI PPV 95% CI Kuhl et al. (1st round) 443 25/1000 (13-45) 7/100 (5-10) 33% (18-52%) Melnikow et al. subgroup 105 29/1000 (6-81) 9/100 (4-16) 32% (12-66%)

Current guidelines on MRI for dense breasts

For example:

6/7/2017 4

Current guidelines on MRI for dense breasts

Or:

What is unknown, and what are the hurdles?

• Berg et al. and Kuhl et al.: Promising in relation to extra

detection Cave: small numbers of cancers But:

• What if density only selection criterium?
• What happens after a ‘prevalent’ screening round?
• Does extra cancer detection improve health outcomes?

(not solely overdiagnosis?) (N.B. not clear for other ‘adjunct’ modalities either)

What is known on use of MRI in dense breasts?

Study No. Exam. Extra cancer detection rate 95% CI Recall rate 95% CI PPV 95% CI Total Prevalent Incident 3861 2120 1741 15.5/1000 22.6/1000 6.9/1000 (12-20) (17-30) (4-13) 9/100* 12/100* 5/100* (8-10) (11-13) (4-7) 17%* 19%* 13%* (14-22) (15-24) (7-21)

• 2120 women at average risk, prevalent and incident rounds
• Calculated Gail lifetime risk 6-12%, no significant risk factors
• Of which 60% with dense breasts (C or D) and negative mx+US

February 2017

* BIRADS 3,4,5 are considered as positive 55% of 48 cancers were in dense breasts 85% of 12 cancers were in dense breasts

From the USPSTF:

6/7/2017 5

• Paired design:

– Extra detection of cancers

• Parallel design: randomized (or comparison cohort)

– Extra detection of cancers, but also: – To compare interval cancer rates – To compare advanced breast cancer rates – Mortality rates

• Long-term:

– Not only prevalent, but also incident rounds:

• What happens to detection in ongoing screening

program?

• What happens to recall/false-positives?

Why important? Set-up of DENSE trial

Dense tissue and Early breast Neoplasm ScrEening) trial

ClinicalTrials.gov Identifier: NCT01315015

DENSE trial

• RCT with ~7,000 MRI and

28,000 ‘usual care’ after a negative DM and extremely dense breasts

• 4 Dutch screening organizations,

8 hospitals in various parts of the Netherlands

• Target: contrast-enhanced MRI

in 4,776 breast screening participants (to be compared to a 4-fold number of controls)

DENSE: MRI screening for women with extremely dense breasts

Domain:

• Current screening population: 50-75 years
• With ‘extremely dense’ breasts (Volpara in mammography units)
• With a ‘negative’ screening mammography

Design:

• RCT with interval cancer rate as main outcome
• Other studies: observational paired design (everyone

mammography + MRI) – higher detection rate due to

• verdiagnosis!?

Intervention:

• ‘Additional contrast-enhanced MRI’ vs ‘no further examination

after negative screening mammography’

• 3 screening rounds (prevalent + 2 incident rounds)

Emaus et al. Radiology 2015;277:527-37

6/7/2017 6

Single-consent prerandomization design

• Breast density is not routinely assessed in NL (no awareness)
• Disadvantages of classical RCT in case of DENSE:

– Anxiety – Self-referral induced contamination

• Therefore: prerandomization + invitation of MRI group only

(control arm not contacted; its evaluation is registry-based.)

• Approval Minister of Health, Welfare and Sports (2011)
• Comes with a price:

– Exclusion and non-participation occur after randomization – No other information on control group than age, cancer occurrence, tumor characteristics, treatment, mortality

Participant selection and recruitment

• Dutch breast cancer screening organization
• Living within 60 km of ‘DENSE’ hospital
• Extremely dense breasts (Volpara in

>40 mammography units) and negative mx

• Randomization 1:4
• Mail-invitation for MRI by screening organization
• Expression of interest to researchers through website or reply card
• If positive, contact potential participant by phone to:

– Check eligibility – Provide further information – Schedule MRI appointment

• Written informed consent just preceding MRI examination

Outcomes DENSE trial

• How many tumors detected extra with MRI?
• More important, primary outcome: Is the

interval cancer rate reduced compared to usual screening? (and the rate of advanced cancers?)

• Simulation models for breast cancer mortality

reduction

• Costs
• % false-positives and quality of life

In 2018 all participants will have undergone their 2-year follow-up

6/7/2017 7

Experience with 1st round MRI and re-invitation for the 2nd round Participation DENSE trial

Not interested N=1,872 (23%) Nonresponders N=913 (11%) Control group N≈32,312 Intervention group (N=8,061) Interested in participation N=5,276 (66%)

Interested women:

• Were slightly younger (median 54 yrs vs 55 yrs)
• Lived slightly closer to MRI center (median 24.5 vs 25.5 km)
• Were of higher SES:
• Q4 SES: 68% interested
• Q1 SES: 59% interested

Most frequently stated reasons for nonparticipation

Not mutually exclusive

• MRI-related inconveniences or contraindications: 27%

E.g. claustrophobia (12%), refusing contrast (7%), intracorporeal metals (3%)

• Anxiety regarding result of supplemental screening: 21%

E.g. emotional burden too high (15%), concerns about false-postives or

• verdiagnosis (2%)
• Personal health: 21%

E.g. other diseases (9%), or low estimated risk (7%)

• Practical reasons 14%

E.g. time/travel related (12%), financial (<1%)

• Burden too high – no further explanation 12%

Participation DENSE trial

Not interested N=1,872 (23%) Nonresponders N=913 (11%) Control group N≈32,312 Intervention group (N=8,061) Interested in participation N=5,276 (66%) Agreed to participate Completed MRI N = 4,783 (59%) Incomplete MRI N=71 (1%) Ineligible N=103 (1%) Withdrawal after more extensive information N=156 (2%) Cancelled MRI / no show N=163 (2%)

6/7/2017 8

Who are the participants?

• 4783 participants in MRI group
• From the lower age range of screening participants
• Less than 1% with personal history of breast cancer in the

past

• 15% with family history of breast cancer in 1st degree

relative 70+ yrs 65-69 yrs 60-64 yrs 55-59 yrs 50-54 yrs

3% 9% 12% 21% 55%

What is their risk profile?

BCSC-BDD 5-year risk Extremely dense Low (<1.00%) 0% Average (1.00-1.66%) 28% Intermediate (1.67-2.49%) 48% High (2.50-3.99%) 21% Very high (≥4.00%) 2%

Cave:

• DENSE population older than study Kerlikowske
• DENSE: BCSC-BDD vs Kerlikowske study: BCSC

See also: poster no. 23 S.V. (Stéphanie) de Lange, on BCSC and Gail in DENSE women

How did they experience MRI?

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Very Quite A little Not

Pain Discomfort Anxiety Measured with screen-specific items questionnaire

– Essink-Bot et al. JNCI 1998;90:925-931

How did they experience MRI?

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Very Quite A little Not

Pain Discomfort Anxiety

* * *

* Compared to MRISC – Rijnsburger et al. Br J Cancer 2004; 91:69-76

6/7/2017 9

How did they experience MRI?

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Very Quite A little Not

Pain Discomfort Anxiety

* Compared to MRISC – Rijnsburger et al. Br J Cancer 2004; 91:69-76

* * * * * *

DENSE in progress: > 8,000 MRI scans

• Inclusion (1st round) completed, 2nd round close to end
• Re-invitation for women with again negative Mx two years

later, and who are still eligible

Round 3 Round 2 Round 1 122 3121 4783

Re-invitation for 2nd MRI two years later

Participants approximately 2 years after 1st MRI Underwent Mx (negative) and still eligible Undergo 2nd MRI: 74 % of total no. participants (81% of eligible participants) Active drop-out (1%) No longer eligible (9%) (No Mx, or positive Mx

• r contra-indications)

Decline invitation 2nd round (16%)

See also: poster no. 24 M.F. (Marije) Bakker, on continued participation in 2nd DENSE round

• Ca. 1% because of

possibility of Gd retention

Experience with information on possibility of Gd deposition

• In August 2016 we informed all participants, with an extra

patient information leaflet, on the possibility of Gd deposition in the brain

• Because DENSE study is:

1. Research 2. Screening

• We explained:

– Gd based contrast used for many years in very large numbers for diagnosis or treatment effects – No health effects known in relation to Gd deposition “Despite absence of health effects we inform you because it may influence your willingness to participate”

6/7/2017 10

In summary

• In current guidelines: MRI not recommended for

screening women selected on the basis of breast density alone

• MRI may be of value for supplemental screening but

needs further proof in large randomized trials with several screening rounds and relevant health outcomes

• DENSE trial fills this gap
• Primary outcome: effect on interval cancer rate
• In 2018 all participants will have undergone their 2-year

follow-up

In summary

• Non-participation : conservative estimate

– MRI and contrast-related – Related to anxiety of supplemental screening – improvement possible!

• Risk profile:

– Majority at increased risk, based on risk models

• Little pain and anxiety due to MRI, somewhat discomfort
• >80% of eligible women participates in 2nd round
• Validating abbreviated

protocol (Kuhl JCO 2014)

• Minimizing false-positives:

– Automated pattern recognition techniques (ARGOS CAD system – Gilhuijs et al.)

• Studying effectiveness in

higher-risk subgroups

• Estimating lead-time +

modelling screening intervals

……..working towards a future-proof screening tool And in between….

Thanks to all study partners

6/7/2017 11

Financial and in kind support DENSE: Experience with information on possibility of Gd deposition

• After informing 4783 women in August 2016:

– 29 immediate drop-outs – Upon re-invitation for 2nd (or 3rd) round ~3% extra drop-out

• In emails/telephone calls received after the letter,

participants indicate that they appreciate to have been informed