Original Article

J Med Screen
0(0) 1–9
The effect of mammography screening ! The Author(s) 2018
Reprints and permissions:
regimen on incidence-based breast sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0969141318780152
cancer mortality journals.sagepub.com/home/msc

Martin J Yaffe1,2,3, Nicole Mittmann4,5,6, Oguzhan Alagoz7,8,

Amy Trentham-Dietz7, Anna NA Tosteson9 and
Natasha K Stout10

Abstract
Objectives: Incidence-based mortality quantifies the distribution of cancer deaths and life-years lost, according to age
at detection. We investigated the temporal distribution of the disease burden, and the effect of starting and stopping
ages and interval between screening mammography examinations, on incidence-based mortality.
Methods: Incidence-based mortality was estimated using an established breast cancer simulation model, adapted and
validated to simulate breast cancer incidence, screening performance, and delivery of therapies in Canada. Ten strategies
were examined, with varying starting age (40 or 50), stopping age (69 or 74), and interval (1, 2, 3 years), and
“No Screening.” Life-years lost were computed as the difference between model predicted time of breast cancer
death and that estimated from life tables.
Results: Without screening, 70% of the burden in terms of breast cancer deaths extends between ages 45 and 75.
The mean of the distribution of ages of detection of breast cancers that will be fatal in an unscreened population is
61.8 years, while the mean age of detection weighted by the number of life-years lost is 55, a downward shift of
6.8 years. Similarly, the mean age of detection for the distribution of life-years gained through screening is lower than
that for breast cancer deaths averted.
Conclusion: Incidence-based mortality predictions from modeling elucidate the age dependence of the breast cancer
burden and can provide guidance for optimizing the timing of screening regimens to achieve maximal impact. Of the
regimens studied, the greatest lifesaving effect was achieved with annual screening beginning at age 40.

Keywords
Breast cancer screening, mammography, incidence-based mortality, mammography screening regimens, quality-adjusted
life-years, age to begin screening, screening interval
Date received: 14 September 2017; accepted: 10 May 2018

Introduction
In developing strategies for reducing the burden of
death and morbidity associated with breast cancer, it
is useful to understand the origin of that burden,
namely: At what age are breast cancers that ultimately
1 10
Physical Sciences Program, Sunnybrook Research Institute,
Toronto, Canada
2
Departments of Medical Biophysics and Medical Imaging, University of
Toronto, Toronto, Canada
3
Ontario Institute for Cancer Research, Toronto, Canada
4
Health Outcomes and PharmacoEconomic (HOPE) Research Centre,
Sunnybrook Research Institute, Toronto, Canada
5
Applied Research in Cancer Control, Department of Pharmacology,
University of Toronto, Toronto, Canada
6
Office of Research, Cancer Care Ontario, Toronto, Canada
7
Department of Population Health Sciences and Carbone Cancer Center,
University of Wisconsin-Madison, Madison, USA
8
Department of Industrial and Systems Engineering, University of
Wisconsin-Madison, Madison, USA
9
The Dartmouth Institute for Health Policy and Clinical Practice, Geisel
School of Medicine at Dartmouth, Lebanon, USA
Department of Population Medicine, Harvard Medical School and
Harvard Pilgrim Health Care Institute, Boston, USA
Corresponding author:
Martin J Yaffe, Sunnybrook Research Institute, Room S657, 2075 Bayview
Ave, Toronto, ON M4N 3M5, Canada.
Email: martin.yaffe@sri.utoronto.ca
2 Journal of Medical Screening 0(0)
lead to death, detected? and How can we intervene most
effectively to reduce that burden? One metric for this
burden is incidence-based mortality (IBM),1,2 where
the number of deaths due to breast cancer or
the number of life-years lost (a measure of premature
mortality) is expressed versus the age at which the dis-
ease surfaces clinically or is detected by screening.
We have recently completed an evaluation of the
effects of different screening scenarios, i.e. the ages to
begin and discontinue mammography screening and
the interval between screens, on health outcomes,
cost-effectiveness, and cost–utility.3–5 We used a well-
established breast cancer model, developed using
US data and adapted to the Canadian context. The
model allows the impact of screening to be quantified
in terms of lives saved (premature breast cancer deaths
averted) as well as life-years and quality-adjusted life-
years (QALY) saved.
The model also allows the prediction, within a birth
cohort, of the age at which each breast cancer would be
detected by screening or through clinical finding, and
the age at which each woman in the cohort would die
from breast cancer or some other cause. In designing
an intervention program, it is helpful to understand the
distribution of disease burden, i.e. IBM, and how an
intervention may affect that distribution. In an analo-
gous manner, the reduction of breast cancer-related
deaths or the increase in life-years attributable to
screening can be assessed in relation to the age at
which cancer was or could have been detected.
Methods
We utilized a version of the Wisconsin Cancer
Intervention and Surveillance Modeling Network
(CISNET) breast cancer model6–10 which we had pre-
viously tailored to the Canadian population.3
The model is a discrete event simulation where a
large birth cohort (in this case two million women)
moves through life. Events are triggered by random
number selections with weightings for initiation of a
breast cancer, its growth rate, hormone and HER2
receptor status, detection by screening or clinically,
etc., all calibrated such that average values match
empirical data for these variables. Once a breast
cancer is initiated in the model, its growth follows a
Gompertz-type function, with various sizes used as sur-
rogates for progression of the cancers from in situ, to
locally invasive, regional, or distant metastases. For
example, for treatment, the model uses a cure/no cure
approach, with a cure rate that has been calibrated to
match 10-year survival. The model also incorporates
components reflecting cancers with limited malignant
potential.8 The model begins with baseline survival
data in the absence of treatment and modifies survival
on the basis of data derived from the meta-analysis of
the chemotherapy trials conducted by the Early Breast
Cancer Trialists’ Collaborative Group.11–15 Patients are
either cured or not cured (die from breast cancer) and, in
the model, are also exposed to competing causes of
death. For those not cured, the cancers continue to
grow until they reach a specified size that is used as a
surrogate for the point of death. Women without breast
cancer eventually die from other causes, again according
to life-table-weighted random number selection. The
model reflects the performance of modern digital mam-
mography, and includes use of trastuzumab as part of
the therapeutic regimen,3 with survival improvements
based on the report by Howlader et al.16 In addition,
we updated the sensitivities and specificities for screen-
ing mammography, based on the data from the NCI’s
Breast Cancer Surveillance Consortium (BCSC) in the
US6 and from the Screening Mammography Program of
British Columbia (SMPBC).17
The model incorporates four major components: (1)
initiation, growth, and progression of a breast cancer,
including its characteristics; (2) detection clinically or
through screening; (3) response to therapy; and (4)
death of cohort members by causes other than breast
cancer. These are more fully described in publications
on the University of Wisconsin CISNET Model.6–10
Inputs are either applied directly in the model, or are
used indirectly to calibrate model outputs, such as stage-
specific incidence and mortality, to empirical data. Some
key aspects of the model and input variables, as well as
identification of the source of data are summarized in
Table 1. Data are either applied directly in the model
(D in Table 1), or used indirectly (I) in an iterative cal-
ibration to match model outputs such as incidence, stage
distribution, or mortality to empirical data.6–10,19
We validated the model’s fit to the Canadian popu-
lation in three ways. First, we demonstrated the ability
of the modified model to predict age-specific breast
cancer incidence in Canada in the absence of screening.
Second, we estimated the ratios of detection of ductal
carcinoma in situ (DCIS) to invasive cancer and found
that these ratios are comparable with those observed in
the SMPBC.17 Finally, we compared the overall stage
distribution of detected cancers predicted by the model
with that observed in the SMPBC and found that the
agreement was good.3
In this study, we applied the model to a hypothetical
1960 birth cohort of two million women, to estimate
for each woman who will develop breast cancer the
age at which the cancer will either surface clinically
or be detected by screening, and the age of death if a
breast cancer death ultimately occurs. Utilizing
Canadian life table data,23 we have also estimated the
age at which each woman would be expected to die if
she did not die of breast cancer. The difference between
Yaffe et al. 3

Table 1. Summary of inputs to the model. Sources of the input data are also indicated.

Direct or
indirect
Category Characteristic input Description Source (ref number)

Model type Discrete event

Population Birth cohort—1960
Update interval Six months
Tumor growth model Gompertz equation Alagoz et al.9
Onset proportion I For calibrating to incidence Statistics Canada18
Annual % incidence D To reflect temporal patterns of University of Wisconsin19
change incidence/age–period– and Holford20
cohort model
Onset lag I Delay in tumor growth Alagoz et al.9
Mean gamma I Growth rate Alagoz et al.9
Variance gamma I Distribution of growth rates Alagoz et al.9
In situ boundary I Calibrating to stage distribution Alagoz et al.9
Asymptotic (max) D—8 cm Alagoz et al.9
diameter
Tumor characteristics Receptor status D Population data Howlader et al.16 and Ries
(ER, PR, HER2) and Eisner21
Cancer detec- Occult onset size D Alagoz et al.9
tion/diagnosis Sensitivity I Versus age, breast density, prev- Stout et al.,6 Screening
alence or recurring screen Mammography Program,17 and
Breast Cancer Surveillance
Consortium22—also Table 1 in
Yaffe et al.3
Specificity of screening D Versus age, breast density, prev- Stout et al.,6 Screening
alence or recurring screen Mammography Program,17 and
Breast Cancer Surveillance
Consortium22—also Table 1 in
Yaffe et al.3
Specificity of D Affects benign biopsy Screening
noninvasive workup rates/BCSC Mammography Program17
Treatment use D Adherence assumed 100% Mittmann et al.5 (Appendix
and adherence Table A), and personal com-
munication with oncologists
Response to Cure fraction I Stage dependent
treatment Hazard reduction factor D Specified versus tumor size, Early Breast Cancer Trialists’
of therapies stage, age, receptor status, Collaborative Group,11–13
type of therapy, and agent Clarke et al.,14 Peto
et al.,15 and Howlader et al.16
Deaths from D Life tables Statistics Canada23
other causes
BCSC: Breast Cancer Surveillance Consortium; HER 2: human epidermal growth factor receptor 2; ER: estrogen receptor; PR: progesterone receptor.

these ages gives the number of years lost due to prema-
ture breast cancer death and is a conventional measure
of premature mortality.24 The model was run for 10
active screening regimens and a no screening scenario
(Table 2). These regimens represent screening strategies
currently being used in different jurisdictions within
Canada, as well as recommendations from the US
Preventive Services Task Force and The Canadian
Task Force on Preventive Health Care.25–27
The model recorded the history of each woman by
indicating the age at detection of a breast cancer if this
occurred, its “stage” at diagnosis (in situ, local inva-
sive, regional involvement, or distant metastasis), the
age at death, and whether the death was attributable to
breast cancer or another cause. Results were expressed
as age- and stage-specific incidence and mortality per
100,000 women alive at age 40. In some examples,
results were expressed per 100,000 in the initial
cohort when so indicated in the text. The breast
cancer mortality reductions were calculated as illustrat-
ed in Figure 1. Screened women were “followed” from
the age of the first screen until 15 years after the
4 Journal of Medical Screening 0(0)

termination of screening for each regimen. Unscreened due to breast cancer. Here, follow-up on deaths is from
women were followed over the same time period as for age 40 to age 89 (i.e. 15 years after the age at which
that regimen. screening would end) for both the screened and
In the example in Figure 1, the arrow at left symbol- unscreened women. When this difference is summed
izes detection of cancer at age Ab, at a screen for a over the cohort for all ages at death in the follow-up
group of women screened annually from age 40 to period, the number of life-years lost to breast cancer,
74. In this case, the cancer was successfully treated detectable at age Ab with screening or Ac without, can
and the woman lives to her normal life expectancy, be estimated.
i.e. the expected time of death from all causes other To quantify the limitation or harms associated with
than breast cancer, at age Ae using Canadian life mammography screening we also estimated the number
tables.23 The arrow at age, Ac illustrates a more of abnormal recalls following screening, and the number
advanced cancer that surfaces clinically and results in of biopsy procedures that were negative for breast cancer,
a premature breast cancer death at age Ad. The differ- as well as the numbers of screens and women required to
ence between the actual and expected times of death, be screened per life saved and per life-year saved. These
i.e. (Ae
Ad), provides an estimate of the life-years lost have been reported previously.4 In the model, benefits
and harms were quantified in a common “currency” by
Table 2. Screening scenarios modeled, predicted breast cancer expressing both in terms of quality-adjusted years of life
deaths per 100,000 women alive at age 40 and percent mortality lost or gained. Age-specific utility and disutility weighting
reduction relative to No Screening. Full adherence with screen- factors were obtained from published work,28–30 with
ing is assumed. adjustment for health states following the method of
Breast cancer Mortality Tosteson et al.31
Scenario deaths/100,000 women reduction (%) It is recognized that some cancers will be overde-
tected, i.e. detected at screening when they otherwise
No Screening 1940 0
would not have become clinically apparent before the
Annual 40–74 930 53.4
individual had died due to some cause other than breast
Annual 40–69 1030 50.2
Annual 50–74 1100 49.2 cancer. The probability of overdetection is greatest for
Annual 50–69 1200 45.5 screening of older women and for DCIS. Treatment of
Annual 40–49, 1150 42.0 these overdetected (often referred to as overdiagnosed)
Biennial 50–74 cancers can be considered a harm. This effect will be
Annual 40–49, 1240 38.7 included in future analysis because of the very large dis-
Biennial 50–69 parity in estimates of this phenomenon at this time.
Biennial 50–74 1330 35.9 Finally, because the model generates histories
Biennial 50–69 1420 32.3 including cancer stages for individual women, the
Triennial 50–74 1460 27.9
incidence-based paradigm can also be used to consider
Triennial 50–69 1540 24.6
the characteristics and outcomes associated with

premature
br. cancer br ca death
detected expected
through screening no screening life-years lost time
of death

life expectancy
Ab A c Ad Ae

screening (annual) follow-up

Age Birth 40 50 60 70 80 90
Calendar Year 1960 2000 2050 2060

Figure 1. Theoretical effect of earlier detection through screening on years of life saved. In this example breast cancer is detected
clinically at age Ac and breast cancer death occurs at age Ad. Screening provides earlier detection at age Ab and woman lives to her
normal life expectancy at Ae with difference in life-years of Ae
Ad. In our model, effects of screening are followed to 15 years past
time at which screening is terminated.
Yaffe et al. 5

cancers detected at specific times. We examined the
predictions at three time points: that of the first
screen, one year prior to that, and one year later.
Results
The model predicts that mortality begins to drop short-
ly after the onset of screening at age 40 or 50, with the
number of breast cancer deaths per 100,000 women
alive at age 40 dropping from 1940 for unscreened
women to 1100 for a cohort screened annually between
ages 50 and 74 and to 940 for annual screening from
ages 40 to 74 (Table 2). Mortality increases relative to
annual screening about two years after a transition
from annual to biennial screening at age 50, but
remains well below that for an unscreened population.4
Presumably these fairly rapid changes are due to earlier
or later treatment, respectively, of more advanced or
rapidly growing cancers occurring because of changes
in detectability related to screening. The mortality
reduction begins to diminish shortly after the cessation
of screening, but persists through the remainder of life,
becoming minimal by approximately age 89, 15 years
the IBM demonstrates a prevalence “spike” at the first
year of screening (ages 40 or 50). This phenomenon is
unrelated to the time at which a woman will die of
breast cancer, and reflects the fact that those deaths
are attributed to the pool of cancers that had been
previously undetected in the cohort, but which are
found earlier due to the lead time provided by screen-
ing. This spike is followed by a sharp drop in IBM.
An example of the life-years lost, versus the age at
which the cancer surfaces or is detected by screening, is
shown in Figure 3 for different screening regimens. For
the no screening case, these curves express the burden
of breast cancer in terms of loss of life. The distribu-
tions of breast cancer deaths according to the age at
which cancer is detected (regardless of the age at which
the death actually occurs) and life-years lost (which
does depend on when death occurs) are summarized
in Table 3. Also shown in Table 3 are the number of
1200
1000
800
QALY lost

after screening has ceased. For the two screening 600

scenarios mentioned above, the predicted mortality
400
reductions are 49.2 and 53.4%, respectively (Table 2).
It must be noted, however, that these numbers corre- 200

spond to 100% of the cohort participating in screening 0

and adhering to the regimen, and will be lower if there 20 30 40 50 60 70 80
is only partial compliance. Partial compliance can be Age at diagnosis

estimated by a weighted combination of results from

Figure 3. Number of QALYs lost due to breast cancer death
women screened in a particular regimen and women (from initial birth cohort of 100,000) versus age at which cancer
who are unscreened. was detected for No Screening , for annual screening from
When the predicted breast cancer deaths are plotted 40 to 74 and for biennial screening from 50 to 74 .
versus the age at which cancer is detected (Figure 2), QALY: quality-adjusted life-year.

60

50
Number of deaths

40

30

20

10

0
20 30 40 50 60 70 80 90 100
Age at diagnosis
Figure 2. IBM. Number of breast cancer deaths (from initial birth cohort of 100,000) versus age at which cancer was detected for
No Screening for annual screening from 40 to 74 and for annual and biennial screening from 50 to 74 , .
6 Journal of Medical Screening 0(0)
0.05
Rel. no. of lives or QALY
0.04
0.03
gained
0.02
0.01
0.00
30 40 50 60
-0.01
Age at diagnosis
Figure 4. Relative number of breast cancer deaths averted
and QALYs gained versus age at breast cancer detection for
annual screening from 40 to 74. QALY: quality-adjusted life-year.
lives saved and the number of women-years of life
saved versus the age of detection, which, for each
screening regimen, are obtained by subtracting the
curve for the deaths and years of life lost, respectively,
from that for the unscreened group. While the largest
percent reduction in mortality occurs in women aged
65–69, the most life-years saved is observed in women
aged 55–59. Absolute and relative distributions of lives
and life-years saved versus age at cancer detection
are given for annual screening from age 40 to 74 in
Figure 4 and the lower part of Table 3.
We also examined model predictions for the stage
distribution of cancers at the time of detection for three
points: the year before screening begins, the year of the
prevalence screen at either age 40 or 50, and the year
after (assuming annual screening). For ages 39, 40, and
41, the proportions of all detected invasive cancers that
are predicted by the model to be either regional or dis-
tant are 45, 24, and 19%, respectively. For the ages 49,
50, and 51, the values are similar at 46, 24, and 16%.
For an annual screening scenario that begins at age
40, the model predicts that 7% of the women with
prevalent invasive cancers detected at age 40 will die
due to breast cancer, while 5.2% of those women
whose cancers were detected in the next screen at age
41 will die of breast cancer. Although far fewer cancers
would surface in the prescreening year at age 39, 21%
of them would result in breast cancer death. Similarly,
if routine screening is started at age 50, 7.1% of those
with breast cancers detected in the prevalence screen at
age 50 will die of breast cancer, whereas 4.2% of those
with incident cancers at age 51 and 20.4% of those with
cancers surfacing at age 49 are predicted to die.
Discussion
Most (70%) of the breast cancer deaths in an
unscreened population arise from cancers that appear
between age 45 and 74 (Figures 2 and 3). These results,
in terms of the shape of the distributions, are in general
agreement with those published by the American
Cancer Society;2 however, their calculations are based
on SEER data, which cover an undefined mix of
screened and unscreened women, while here we have
considered a birth cohort with well-defined screening
patterns. Additionally, the American Cancer Society
data are presented in five-year age bands, which sup-
presses the peak that we observe at the onset
70 80 90 of screening.
Shortly following the point at which screening
begins, mortality attributable to cancers detected at
that age drops, as evidenced by the difference in area
under the curves for screened versus unscreened women
(Figure 2). This is preceded by a pronounced spike in
mortality attributable to cancers detected in this previ-
ously unscreened cohort, at the first round of screening,
at ages 40 and 50 in this example. This suggests that
many of the cancers detected in the prevalence screen
are progressive rather than indolent cancers, and while
some of the prevalent cancers are detected sufficiently
early to allow successful treatment, others, having
passed the period for which a lead time advantage of
screening was available, are now destined to be lethal.
Comparison of the curves for starting screening at ages
40 and 50 suggests that some of these deaths may be
preventable if women begin screening earlier.
Considering only the invasive cancers, the model
shows a sharp downward stage shift from the year
prior to onset of screening, presumably where the can-
cers detected come from a pool of cancers that have
developed at different rates over time, to the initial
screen, where there is a mixture of these cancers and
earlier-stage or smaller lesions that, at this point in
time, can be detected only by imaging. There is a fur-
ther, smaller, but still marked shift between the initial
and second screens, where the cancers detected are
mainly those that can be found by mammography.
The same phenomenon occurs both for beginning
screening at age 40 or 50.
The prediction of breast cancer deaths is consistent
with the staging predictions of the model shown in our
previous publication.4 Although the actual number of
breast cancers detected in the year preceding screening
is lower, the IBM drops dramatically from that year to
the time of the prevalence screen, and then further for
the second screen. These modeling results imply that
the pool of more advanced cancers in the population
would be cleared at the prevalence screen and, thereaf-
ter, there would be a subsequent downward stage shift
(or downward shift of size within a stage) of detected
cancers, and an improvement in outcome.
For all the screening strategies studied here, a mor-
tality reduction is associated with detection occurring
as early as one to two years after the onset of screening.
Yaffe et al. 7

Given the natural history of breast cancer, this suggests

Table 3. Absolute number and age distribution of eventual breast cancer deaths versus age at diagnosis. Also shown are deaths averted through screening, life-years (LY) lost and

85–89

64.7

287

190
3.7

0.8

0.7

0.1
that there is some mortality advantage of earlier detec-

10
3
0
tion even for cancers that are more advanced.
The model gives a measure of premature deaths,
80–84

97.3

649

436
5.5 where the mean age of detection with respect to the
1.8

1.6

0.3
10

43
1
number of eventual breast cancer deaths in unscreened
women is 61.8 years. When the mortality burden is
75–79

1170

considered in terms of life-years lost, rather than pre-

125

807

275
7.1

3.3

3.0
40
4

2
mature deaths, Table 3 and a comparison of Figures 2
and 3 show that the distribution shifts to lower ages,
70–74

1849

1330

approximately 40–69 years, with a mean age of 55,

saved through screening for a birth cohort of 100,000 women. Scenarios compared are No Screening and Annual screening from ages 40 to 74.

148

104

958
8.4

5.2

4.9

11

6 reflecting the increased number of years lost when

younger women die from breast cancer. Similarly,
65–69

when the benefits attributable to screening are exam-

3852

2838

2016
14.2

10.8

10.4
250

176
19

13

ined (Figure 4), it is seen that the relative benefit is

shifted to earlier age of detection (mean age reduces
60–64

from 60.5 to 57.3 for annual screening from 40 to

4557

3446

2439
13.6

12.8

12.7
240

167
18

16

74), when life-years saved, rather than deaths averted,

are considered.
We have shown in previous work4 that more fre-
55–59

5605

4320

3022
14.0

15.8

15.9
246

170

quent screening detects more breast cancer cases than

18

20

No Screening and results in fewer breast cancer deaths.

In addition, more frequent screening is associated with
50–54

5389

4165

2789
11.4

15.2

15.3

14.2
201

133

more recalls of women, who, upon further noninvasive

19

imaging, are found not to have cancer, and more neg-

ative biopsies as well as an increased ratio of in situ to
45–49

4792

3764

2232
13.5

13.8

invasive cancers detected. In prior modeling analyses,

154

Values do not add to 100% because of deaths due to cancer diagnosed before age 35 and after age 89.
8.7

9.6
90

15

under idealized situations, it was estimated that two

negative biopsies would be performed for each life-
40–44

3397

2717

1212

year gained due to annual screening between ages 40

5.4

9.6

4.4
96

10
42

and 74.4
8

Our work has both strengths and limitations. The

35–39

University of Wisconsin model, as modified for use in

2210

1778

NA
NA
NA
NA
3.2

6.2

6.5
55

the Canadian context, performs quite well in predicting

breast cancer incidence in the absence of screening.3
Age distribution of Br Ca Deaths averted (%)a

The model was calibrated using modern empirical

data on the sensitivity and specificity of screening
QALY saved Compared to No Screening

mammography versus age and breast density to

describe the screening process. Canadian data on utili-
Age distribution of QALY saved (%)

zation of therapies and their efficacy were used wher-

ever available, and no assumptions on the mortality
Percent of all Br Ca Deaths

reduction of screening were applied explicitly in the

Percent of all QALY lost

model. In this work we found it informative to analyze

QALY: quality-adjusted life-year; LY: life-year.
Br Ca deaths averted
Percent of all LY lost
Breast cancer deaths

birth cohorts. It is straightforward to combine calcula-

tions from multiple cohorts to make predictions on
Age at diagnosis

populations with arbitrary age distribution. Of neces-

QALY lost

sity, there are limitations in modeling the natural his-

LY lost

tory of breast cancer. Some of the sources of data on

therapies are quite old. A balance must be struck
between the currency of the data (over time, some ther-
apies are being applied in different ways, and to
Annual 40–74
No Screening

expanded subsets of patients or a wider range of can-

cers) and the duration of follow-up available from the
trials. Aside from application of probabilities for can-
cers expressing estrogen, progesterone, and HER2/neu
a
8 Journal of Medical Screening 0(0)
receptors, no attempt has been made to apply further
characterization, e.g. by molecular subtype.
Furthermore, the progression from in situ to invasive
disease has been based strictly on the modeled diameter
of the tumor, and this has resulted in some discrepan-
cies between modeled and actual ratios of DCIS and
invasive cancers detected by screening. Finally, in the
model, the growth rate of cancers was not treated as
being age dependent, although there is some evidence
that premenopausal cancers may grow more quickly.32
Conclusion
This work illustrates how presentation of the results
from simulation in an incidence-based manner can be
helpful in understanding the age distribution of the dis-
ease burden of breast cancer. This approach provides
insights that may be useful in optimizing the design of
an intervention program aimed at reducing that
burden. Consideration of the burden and benefit in
terms of QALYs rather than lives lost or saved incor-
porates both benefits and harms into the analysis. The
use of QALYs or life-years shifts both the burden of
disease and the impact of screening downward mark-
edly in terms of the age at diagnosis. The model esti-
mates that approximately 24% of the QALYs lost to
breast cancer (compared with 14% of deaths) are a
result of cancers that are detectable between ages 40
and 49, and 23% of the QALYS gained by annual
screening between ages 40 and 74 are attributable to
cancers detected in women in their 40s. As models
become more sophisticated, incorporating such fea-
tures as cancer subtypes and more detailed data on
response to therapy, the incidence-based approach
may provide guidance for the design of stratified inter-
ventions by identifying opportunities for maxi-
mal impact.
Acknowledgments
We are grateful to Statistics Canada and to Dr Michael Wolfson at the
University of Ottawa for supplying key incidence data. The content is solely
the responsibility of the authors and does not necessarily represent the official
views of Statistics Canada, the National Cancer Institute or the National
Institutes of Health. We thank the participating women, mammography facil-
ities, and radiologists for the data they have provided for this study. A list of
the BCSC investigators and procedures for requesting BCSC data for research
purposes are provided at: http://breastscreening.cancer.gov/.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the
research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the
research, authorship and/or publication of this article: This work was sup-
ported by a contract from The Canadian Breast Cancer Foundation. The
University of Wisconsin breast cancer simulation model used in this analysis
was also supported by grant number U01 CA152958 from the National Cancer
Institute as part of the CISNET. Model input data on the performance of
screening mammography was also supported by the National Cancer
Institute-funded BCSC grant number UC2CA148577 and contract number
HHSN261201100031C. For a full description of U.S. data sources, please
see: http://www.breastscreening.cancer.gov/work/acknowledgement.html. To
discuss use of model run data, please contact Dr Yaffe.
References
1. Chu KC, Miller BA, Feuer EJ, et al. A method for partitioning cancer
mortality trends by factors associated with diagnosis: an application to
female breast cancer. J Clin Epidemiol 1994; 47: 1451–1461.
2. Oeffinger KC, Fontham ETH, Etzioni R, et al. Breast cancer screening for
women at average risk 2015 Guideline update from the American Cancer
Society. JAMA 2015; 314: 1599–1614.
3. Yaffe MJ, Mittmann N, Lee P, et al. Modelling mammography screening
for breast cancer in the Canadian context: modification and testing of a
microsimulation model. Health Rep 2015; 26: 3–8.
4. Yaffe MJ, Mittmann N, Lee P, et al. Clinical outcomes of modelling
mammography screening strategies. Health Rep 2015; 26: 9–15.
5. Mittmann NM, Stout NK, Lee P, et al. Total cost-effectiveness of mam-
mography screening strategies. Health Rep 2015; 26: 16–25.
6. Stout NK, Lee SJ, Schechter CB, et al. Benefits, harms and costs for breast
cancer screening following US implementation of digital mammography.
J Natl Cancer Inst 2014; 106: dju092.
7. Batina NG, Trentham-Dietz A, Gangnon RE, et al. Variation in tumor
natural history contributes to racial disparities in breast cancer stage at
diagnosis. Breast Cancer Res Treat 2013; 138: 519–528.
8. Fryback DG, Stout NK, Rosenberg MA, et al. The Wisconsin Breast
Cancer Epidemiology Simulation Model. J Natl Cancer Inst Monogr
2006; 36: 37–47.
9. Alagoz O, Ergun MA, Cevik M, et al. The University of Wisconsin Breast
Cancer Epidemiology Simulation Model: an update. Med Decis Making
2018; 38: 99S–111S.
10. National Cancer Institute, Cancer Intervention and Surveillance Modeling
Network. www.cisnet.gov (accessed 26 May 2018).
11. Early Breast Cancer Trialists’ Collaborative Group. Systemic treatment of
early breast cancer by hormonal, cytotoxic, or immune therapy. 133 rand-
omised trials involving 31,000 recurrences and 24,000 deaths among 75,000
women. Part I. Lancet 1992; 339: 1–15.
12. Early Breast Cancer Trialists’ Collaborative Group. Systemic treatment of
early breast cancer by hormonal, cytotoxic, or immune therapy. 133 rand-
omised trials involving 31,000 recurrences and 24,000 deaths among 75,000
women. Part II. Lancet 1992; 339: 71–85.
13. Early Breast Cancer Trialists’ Collaborative Group (EBCTCG). Effects of
chemotherapy and hormonal therapy for early breast cancer on recurrence
and 15-year survival: an overview of the randomised trials. Lancet 2005;
365: 1687–1717.
14. Clarke M, Coates AS, Darby SC, et al. Early Breast Cancer Trialists’
Collaborative Group (EBCTCG). Adjuvant chemotherapy in oestrogen-
receptor-poor breast cancer: patient-level meta-analysis of randomised
trials. Lancet 2008; 371: 29–40.
15. Peto R, Davies C, Godwin J, et al. Early Breast Cancer Trialists’
Collaborative Group (EBCTCG). Comparisons between different poly-
chemotherapy regimens for early breast cancer: meta-analyses of long-
term outcome among 100,000 women in 123 randomised trials. Lancet
2012; 379: 432–444.
16. Howlader N, Altekruse SF, Li CI, et al. US incidence of breast cancer sub-
types defined by joint hormone receptor and HER2 status. J Natl Cancer
Inst 2014; 106: dju055.
17. BC Cancer Agency. Screening Mammography Program 2012 Annual
report, Vancouver, BC, Canada, www.bccancer.bc.ca/screening/
Documents/SMP_Report-Annual Report2012.pdf (2012, accessed 9
February 2017).
18. Statistics Canada. Table 103-0550 New cases of primary cancer (based on
the February 2014 CCR tabulation file), by cancer type, age group and sex,
Canada, provinces and territories, annual. CANSIM (database), http://
www5.statcan.gc.ca/cansim/a26?lang¼eng&id¼1030550 (2015, accessed
31 July 2015).
19. University of Wisconsin. CISNET readers guide. Core profile documenta-
tion. Madison: University of Wisconsin, https://cisnet.flexkb.net/mp/pub/
CISNET_ModelProfile_BREAST_UWISC_001_07232013_58567.pdf
(2013, accessed 27 April 2018).
Yaffe et al.
20. Holford TR. Analysing the temporal effects of age, period and cohort.
Stat Methods Med Res 1992; 1: 317–337.
21. Ries LAG and Eisner MP. Cancer of the female breast. In: Ries LAG,
Young JL, Keel GE, et al. (eds) Cancer survival among adults: U.S. SEER
program, 1988–2001, patient and tumor characteristics (NIH pub. No. 07-
6215). Bethesda, MD: National Cancer Institute, 2007.
22. Breast Cancer Surveillance Consortium. BCSC Mammography Data.
National Cancer Institute, http://breastscreening.cancer.gov/statistics/
mammography_data.html (2014, accessed 27 April 2018).
23. Statistics Canada. Life tables, Canada, provinces and territories, https://
www.statcan.gc.ca/pub/84-537-x/84-537-x2017001-eng.html (2011,
accessed 28 May 2018).
24. Ries LAG, Melbert D, Krapcho M, et al. (eds) SEER cancer statistics
review, 1975–2005. Bethesda, MD: National Cancer Institute, http://seer.
cancer.gov/csr/1975_2005/ (2008, accessed 28 May 2018).
25. US Preventive Services Task Force. Screening for breast cancer: U.S.
Preventive Services Task Force recommendation statement. Ann Intern
Med 2009; 151: 716–726.
26. Nelson HD, Tyne K, Naik A, et al. Screening for breast cancer: an update
for the U.S. Preventive Services Task Force. Ann Intern Med 2009;
151: 727–737.
9
27. Canadian Task Force on Preventive Health Care. Recommendations on
screening for breast cancer in average-risk women aged 40–74 years.
CMAJ 2011; 183: 1991–2001.
28. Fleishman JA. Demographic and clinical variations in health status. Report
#15. Agency for healthcare research and quality, http://meps.ahrq.gov/
mepsweb/data_files/publications/mr15/mr15.shtml (accessed 28
May 2018).
29. de Haes JCJM. The impact of a breast cancer screening programme on
quality-adjusted life-years. Int J Cancer 1991; 49: 538–544.
30. Hanmer J, Lawrence WF, Anderson JP, et al. Report of nationally repre-
sentative values for the non-institutionalized U.S. adult population for
seven health-related quality of life scores. Med Decis Making 2006;
24: 391–400.
31. Tosteson ANA, Stout NK, Fryback DG, et al. Cost-effectiveness of digital
mammography breast cancer screening: results from ACRIN DMIST. Ann
Intern Med 2008; 148: 1–10.
32. Peer PG, van Dijck JA, Hendriks JH, et al. Age-dependent growth rate of
primary breast cancer. Cancer 1993; 71: 3547–3551.