Breast Screening Remarks about mammography screening trials

1
Breast Screening Remarks about mammography
screening trials

• Anthony B. Miller
• Professor Emeritus, Department of Public Health
  Sciences, University of Toronto,
• Head, Division of Clinical Epidemiology
• Deutsches Krebsforschungszentrum,
• Heidelberg, Germany

2
Cluster randomization in the Swedish Two-County
trial

• Number of Units Women ASP PSP
  ASP PSP
• Kopparberg 14 7 39051 18846
• Ostergotlund 12 12 39034 37036

3
Cluster randomization in the Swedish Two-County
trial

• Nixon et al, 2000
• Fixed effects and a variety of random effect
  models show a strong degree of agreement and
  yield a significant 29 or 30 reduction in
  breast cancer mortality
• The heterogeneity among clusters and strata was
  relatively small.

4
Updated Swedish overview analysis (Nystrom et al,
2002)

• Kopparberg not included
• Ostergotlund - data cited to show that the
  screened and control areas had similar breast
  cancer incidence and mortality prior to
  randomisation

5
Ostergotlund breast cancer data (Nystrom et al,
2002)

• Pre-trial period (1968-77) Intervention period
  (1978-82)
• Incidence
• Invited clusters 162.4 257.9
• Control clusters 162.0 185.8
• Mortality
• Invited clusters 60.6 62.7
• Control clusters 63.4 57.7

6
Treatment in the Two-county trial

• Holmberg et al, 1986
• Standard treatment according to stage of disease
  was applied to study and control groups
• Tabar et al, 1999
• Hardly any prophylactic chemotherapy or hormone
  therapy given to women with node positive disease

7
Women with breast cancer age 50-59 Survival
during 13 years follow-up

• Trial and group N Alive ()
• Swedish Two county (Late 1970s)
• ASP 349 290 (83)
• PSP 290 221 (75)
• CNBSS-2 (1980s)
• MP 622 515 (83)
• PO 610 505 (83)

8
Conclusion on the Two-county trial

• The survival experienced by the women with breast
  cancer in the controls, is not the current
  expectation.
• This must have some impact, perhaps a major
  impact, on the estimated benefits that are likely
  to be derived from breast screening.

9
Cancers found by physical examination alone, when
mammography also used, according to age

• Trial 40-49 50-59
• HIP (1960s) 58 40
• BCDDP (1970s) 8 7

10
Cancers found by mammography alone, when physical
examination also used, according to age

• Trial 40-49 50-59
• HIP (1960s) 25 39
• BCDDP (1970s) 45 47

11
Recommendation of the Working Group that reviewed
the US BCDDP (1979)

• A trial to evaluate the magnitude of benefit and
  net benefit-risk in the use of mammography
  screening should be conducted.
• CNBSS-2 is the only trial designed to meet this
  need.

12
Canadian National Breast Screening Study (CNBSS)-2

• 39,405 volunteers age 50-59 randomized with
  informed consent to
• Annual two-view mammography physical
  examination BSE (MP)
• Annual physical examination BSE only (PO)
• 5 or 4 screens and 11-16 years follow-up

13
Occurrence of Invasive Breast Cancers in CNBSS-2

• MP PO
• Screen detected 267 148
• Interval cancers 50 88
• Incident cancers 305 374
• Total 622 610
• Total in situ 71 16

14
Characteristics of screen-detected invasive
breast cancers in CNBSS-2

• MP PO
• Detected by Ma alone PE PE
• Number 126 141 148
• Node positive 20 33 36
• 15mm or more 38 67 72
• with or without mammography

15
CNBSS-2 Deaths from breast cancer, 11-16 years
follow-up

• MP PO
• Women years (103) 216 216
• Breast cancer deaths 107 105
• Rate/10,000 4.95 4.86
• Rate ratio (95 CI) 1.02 (0.78, 1.33)

16
Performance indicators for women age 50-69, first
screen

• Programme Detection lt15mm
• (/1000) (/1000)
• CNBSS-2 (50-59) 7.2 1.8
• Canada (50-69) 6.9 1.5
• UK NHS (50-64) 6.0 1.3
• Netherlands (50-64) 6.5 1.5

17
Performance indicators for women age 50-69,
rescreens

• Programme Detection lt15mm
• (/1000) (/1000)
• CNBSS-2 (annual) 3.0 1.1
• Canada (2-yrly) 3.8 1.1
• UK NHS (3-yrly) 3.8 0.8
• Netherlands (2-yrly) 4.3 1.0

18
Conclusion on CNBSS-2

• The benefit from screening derives from the
  earlier detection of advanced breast cancers,
  coupled with good therapy, not from the early
  detection of impalpable cancers.
• This is accomplished both by good BPE and by
  mammography

19
(No Transcript)
20
Updated Swedish overview analysis Follow-up
model, (Nystrom et al, 2002)

• Age RR (95 CI)
• 40-49 0.91 (0.76-1.09)
• 50-59 0.93 (0.78-1.11)
• 60-69 0.73 (0.61-0.87)
• 70-74 1.12 (0.73-1.72)

21
IARC Working Group, 2002 (press release, 20
Mar., 2002)

• The group.concluded that the trials have
  provided sufficient evidence for the efficacy of
  mammography screening of women between 50 and 69
  years.
• The reduction in mortality from breast cancer
  among women who chose to participate in screening
  programmes was estimated to be about 35.

22
IARC Working Group, 2002 (press release, 20
Mar., 2002)

• For women aged 40-49 years, there is only limited
  evidence for a reduction.

23
IARC Working Group, 2002 (press release, 20
Mar, 2002)

• The working group also concluded that there is
  insufficient evidence that clinical breast
  examination or self-examination reduce mortality
  from breast cancer.

24
(No Transcript)
25
(No Transcript)
26
Effect of NHS programme on reduction in breast
cancer mortality, England Wales (Blanks et
al, 2000)

• Effect of Estimate 1990-98
• Screening 6.4 (range 5.4-11.8)
• Treatment 14.9 (range 12.2-14.9)

27
Conclusions

• The benefit from breast cancer screening derives
  from the earlier detection of more advanced
  disease, not the early detection of impalpable
  cancers, i.e. from a reduction in average tumor
  size from gt 20 mm to lt 20mm, providing modern
  therapy is used.

28
Conclusions

• The meta-analyses to date have over-estimated the
  benefit likely to be achieved by mammography
  screening in the era of adjuvant chemotherapy and
  hormone therapy
• In many countries, mortality from breast cancer
  is falling, but the contribution of screening is
  small

29
Conclusions

• Screening is an expensive use of health care
  resources
• Screening can not abolish mortality from cancer,
  and people who accept screening should not be
  deceived that it will
• As treatment improves, the benefit from screening
  will fall
• As prevention improves, the value of screening
  will diminish