Mammography Versus Clinical Examination of

the Breasts
Cornelia J. Baines, Anthony B. Miller*

commented that ‘‘in an era when modern mammography is

Using published data from screening trials, this article com- available, the use of clinical breast examination in screening is
pares two-modality (mammography and clinical examina- unethical and irrational.’’ He raised an important issue. What is
tion) and single-modality (clinical examination alone) the role of clinical breast examination in screening for breast
screening by evaluating cancer detection rates, program sen- cancer? Is it dispensable? Answering these questions is difficult
sitivities, mode of cancer detection in two-modality screen- because there are few opportunities allowing valid comparisons
ing, nodal status at time of detection, survival 10 years post- of two-modality screening (mammography and clinical breast
diagnosis, and breast cancer mortality 10 years after entry. examination) with single-modality screening (clinical breast ex-
Consistently, two-modality screening achieved higher cancer amination alone). The question may be particularly important for
detection rates and program sensitivity estimates than either women age 40 to 49, for whom there is widespread controversy
modality alone; mammography alone achieved higher rates about the efficacy of mammography screening.
than clinical examination alone; interval cancer detection
rates between screening examinations were higher following Data Sources
clinical examination alone than mammography alone; Of eight randomized controlled trials (RCTs) of breast cancer
single-modality screening with mammography failed to de- screening reported to date (3), the four Swedish RCTs used only
tect breast cancers identified by clinical examination alone; mammography, leaving four that incorporated clinical breast ex-
the sensitivity of mammography was lower in younger than amination in their protocol, namely the New York Health Insur-
older women, while the reverse was true for clinical exami- ance Plan (HIP) Study (4), the Edinburgh RCT (5), and the
nation; and mammography identified a higher proportion of Canadian National Breast Screening Studies (CNBSS) I (6) and
node-negative breast cancer than clinical examination. We II (7) (Table 1). The manner in which these RCTs differed from
conclude that combining clinical breast examination with each other must be understood. Respective ages at entry were
mammography is desirable for women age 40–49 because 40–64 years, 45–64 years, 40–49 years, and 50–59 years. The
mammography is less sensitive in younger than older intervention group in the HIP study and both CNBSS trials
women. Careful training and monitoring are, however, as received annual two-view mammography and clinical breast ex-
essential with clinical examiners as with mammographers. amination. In the Edinburgh trial, the intervention group re-
[Monogr Natl Cancer Inst 1997:22:125–129] ceived two-view mammography with clinical breast examina-
tion at the first screen visit, one-view mammography with
clinical breast examination at the third, fifth, and seventh
In countries where breast cancer is not a major priority and screens, and clinical breast examination alone at the second,
where funding for and expertise in screening mammography fourth, and sixth screens. Control groups in the HIP and Edin-
are scarce, clinical examination of the breasts as a single screen- burgh trials received no screening at all. In CNBSS-I, the control
ing modality unquestionably deserves consideration. However, group received a single clinical breast examination and thereaf-
in North America, where breast cancer is a priority and mam- ter depended on ‘‘usual care’’ in the community. They were
mography is relatively accessible, the real issue is not ‘‘screen- followed annually by mailed questionnaire. In CNBSS-II, the
ing mammography versus clinical examination,’’ but rather control group received annual clinical breast examinations.
‘‘screening mammography with clinical examination versus In addition to results from screening trials, clinical breast
screening mammography without clinical examination.’’ Unfor- examination has been evaluated in case series (8) and in screen-
tunately, this issue is rarely addressed, probably due to two ing projects (9,10), all disadvantaged by the lack of an appro-
major factors: pervasive confidence in technology as a solution priate comparison group. In contrast to the case series, two
for most problems facing society; and the population’s generally screening projects—the Breast Cancer Detection Demonstration
inflated view of the risks of getting breast cancer, of dying from Project (BCDDP) (9) in the United States and the DOM Project
breast cancer and, in particular, of benefiting from mammo- in Utrecht (10), both of which used two-modality screening—do
graphic screening (1,2). Furthermore, when clinical breast ex- provide useful data on clinical breast examination for compari-
amination is considered for inclusion in a mammography screen- son purposes with the RCTs.
ing program, it is often dismissed for economic reasons. In
general, mammography gets much attention and clinical breast
*Affiliation of authors: University of Toronto, ON, Canada.
examination is usually given short shrift, just as chest x-rays and Correspondence to: Cornelia J. Baines, Associate Professor, Department of
electrocardiograms have diminished reliance on percussion and Public Health Sciences, University of Toronto, 12 Queen’s Park Crescent West,
auscultation of the chest. 3rd Floor, Toronto, ON M5S 1A8, Canada.
Several years ago, at a meeting on breast cancer, a speaker © Oxford University Press

Journal of the National Cancer Institute Monographs No. 22, 1997 125
Downloaded from https://academic.oup.com/jncimono/article-abstract/1997/22/125/2952583
by guest
on 12 May 2018
 Table 1. Available data sources for evaluating clinical breast examination*

Source (start date) Design Age at entry (year) Study intervention Frequency Control intervention

HIP (1963) (4) RCT 40–64 MA + CBE qly No screening

Edinburgh (1979) (5) RCT 45–64 MA + CBE q2y No screening
(Rounds 1, 3, 5, 7)
CBE q2y No screening
(Rounds 2, 4, 6)
CNBSS-I (1980) (6) RCT 40–49 MA + CBE q1y Single CBE
CNBSS-II (1980) (7) RCT 50–59 MA + CBE qly Annual CBE
BCDDP (1972) (9) Project 37–74 MA + CBE q1y NA
Utrecht (1975) (10) Project 50–64 MA + CBE Variable NA

*MA ⳱ mammography; CBE ⳱ clinical breast examination; q 1 y ⳱ annually; q 2 y ⳱ every two years.

The quality of the breast examination is also an important tion alone for four or five successive annual screening exami-
issue. Clearly, high performance standards are as important for nations.
clinical examination as for mammography. The CNBSS has es- Not only can breast cancer be detected as a direct consequence
tablished what competent clinical breast examination alone can of a screening examination, it can also be detected in the interval
achieve in terms of cancer detection (11). Recent research on the between screening visits. Such ‘‘interval cancers’’ may only
efficacy of breast self-examination (BSE) also reinforces the become detectable after the screening examination, or they
importance of high standards: benefit from BSE seems to be may be missed by the screening process (in both cases, the
restricted to competent practitioners (12,13). We are not aware screens are said to be ‘‘false negative’’). Any evaluation of
of any published document describing training, monitoring, or clinical breast examination must consider interval cancer
routine evaluation of clinical examiners in the HIP study, the rates.
BCDDP, or the Edinburgh trial. In the Utrecht Project, the clini- b) Program sensitivities (detection method): Only the
cal examination was performed by the radiological technologist CNBSS studies yield sensitivity estimates for a screening pro-
at the time of mammography, and she used a cupped hand to tocol that includes clinical examination alone.
palpate four quadrants of each breast (personal communication). c) Mode of cancer detection: For women receiving two-
This is in marked contrast to the method applied in the CNBSS, modality screening, the proportions of breast cancer detected by
where the examiners were trained to visually examine the mammography alone, by clinical examination alone, and by both
breasts, to palpate the whole breast (not just the cone), to use a simultaneously can be documented. This is possible within the
systematic search pattern, and to apply the pads of their fingers. intervention arm in the four RCTs and in the two screening
Women were examined both sitting up and lying down (11). projects.
Overall, the standards achieved by CNBSS clinical breast ex- d) Nodal status at time of detection: This offers yet another
amination were high. way to evaluate clinical examination. With the data available,
comparisons of nodal status are possible according to both mode
Approaches to Analysis of detection within the intervention arms in all four RCTs and
The constraints on evaluation of clinical breast examination intervention versus control status in the two CNBSS trials. The
arising from the various study designs are apparent in Table 1. latter is of greater relevance in evaluating clinical breast exami-
Even so, the four RCTs and the two screening projects that nation.
combined clinical breast examination with mammography allow e) Survival postdiagnosis: For this approach, data from the
several approaches to evaluating the role of the former: cancer three North American studies—CNBSS, HIP, BCDDP—relate
detection rates, program sensitivities, mode of cancer detection, survival to mode of detection for two-modality screening. Ad-
nodal status at time of cancer detection, survival 10 years post- ditionally, in the CNBSS, survival associated with single-
diagnosis, and breast cancer mortality 10 years after entry. modality screening can be reported.
a) Cancer detection: Only the CNBSS allows detection rates f) Mortality from breast cancer 10 years after entry:
to be compared for combined versus single-modality screening, CNBSS-I breast cancer mortality in cases with year-one screen
since in the other two RCTs the comparison was screening ver- and interval detections will be described because it offers the
sus no screening, and in the screening projects there were no only opportunity to compare mortality following a single epi-
control groups. sode of two-modality screening to mortality following a single
In CNBSS-I, for women 40–49 years on entry, two-modality episode of single-modality nonmammographic screening in
screening can be compared with clinical examination alone, but women age 40–49.
only for cancers detected at the first screening visit and up to 12 What Can Screening With Clinical Breast
months thereafter. (Because intervention women received their
second screening examination 12 months after the first, subse-
Examination Achieve?
quent comparisons on the role of clinical breast examination are a) Cancer detection: Unfortunately, there are only three trials
not possible; one can only compare program outcomes with in which clinical breast examination was conducted in the
respect to breast cancer incidence and breast cancer mortality.) absence of mammography: the Edinburgh trial (5) and CNBSS
In contrast, in CNBSS-II, for women 50–59 years on entry, I and II (6,7). It is clear from Table 2 that two-modality screen-
two-modality screening can be compared with clinical examina- ing will detect more breast cancer than clinical examination

126 Journal of the National Cancer Institute Monographs No. 22, 1997
Downloaded from https://academic.oup.com/jncimono/article-abstract/1997/22/125/2952583
by guest
on 12 May 2018
 alone. Edinburgh detection rates for clinical examination at cal examination alone were slightly higher in younger women
screening rounds 2, 4, and 6 were lower than the rates observed (68% vs. 63%). Thirdly, two-modality screening achieved a
in the CNBSS for women age 50–59 (7). However, the Edin- higher sensitivity in both older and younger women than sin-
burgh women received mammography in rounds 1, 3, 5, and 7, gle-modality screening with clinical breast examination. Fourth,
and this may have depleted the breast cancers available for di- in the CNBSS, the observed sensitivity for clinical examina-
agnosis by clinical examination in the following years. The rates tion alone in women age 40–49 is of the same order of mag-
might also be lower because the quality of the clinical exami- nitude (68%) as that observed in the two-county and Stock-
nation did not match that in the CNBSS. A straightforward com- holm mammography-alone trials (62% and 53%, respectively)
parison of the Edinburgh and Canadian trials is clearly impos- (3). However, the fact remains that mammography achieves
sible. higher detection rates than clinical examination alone (Ta-
The CNBSS-I detection rate for women age 40–49 screened ble 3).
with clinical examination alone at the first screening round was It may be advisable to use 1) two-modality screening for
2.46/1,000 (6), a rate exceeding the rates reported for the Swed- women age 40–49, based on lower mammography and higher
ish two-county and Stockholm mammography-alone trials, clinical examination sensitivity estimates for this age group
which were 2.09/1,000 and 2.06/1,000, respectively, in women compared to older women and 2) single-modality, two-view
age 40–49 at entry (3). mammographic screening for women 50 years and over, based
Interval cancer rates can be expected to be higher following on higher mammography sensitivity estimates and lower esti-
screening with clinical breast examination alone compared to mates for clinical examination for older women relative to
screening with mammography. At the first screen, for those age younger women.
40–49 at entry, CNBSS interval cancer rates were higher in c) Mode of cancer detection: Table 3 displays the mode of
women allocated to receive clinical breast examination only cancer detection (for screen-detected tumors) in the intervention
(1.11/1,000 women) than in those receiving two-modality arms of the four RCTs and the two screening projects. The
screening (0.75/1,000 women). Given that CNBSS mammogra- proportions detected by clinical breast examination alone vary
phy achieves detection rates and sensitivity estimates that match from 3.3% in Edinburgh (5) to 44.7% in the HIP study (15). If
other trials (3), it cannot be suggested that the comparison of one looks at the proportion of all clinically positive screening
CNBSS interval rates is unduly favorable to clinical examina- examinations, it varies from 44% for Utrecht to 74% for Edin-
tion. For women age 40–49, this raises the question, How im- burgh. The usefulness of clinical breast examination is demon-
portant is the difference between interval rates in the two groups strated by the fact that there is no trial in which mammography
being compared? identified all breast cancers. Although Edinburgh comes close at
b) Program sensitivity (detection method): Sensitivity es- 96%, its provision of mammography every second year pre-
timates are higher for two-modality screening than for single- cludes results that can truly evaluate the role of clinical breast
modality with mammography. For women age 40–49 on entry, examination.
the sensitivity of two-modality screening with two-view mam- The data in Table 3 reinforce the advisability of adding clini-
mography was much higher in the CNBSS (81%) compared to cal breast examination to mammography screening in younger
single-modality with single-view mammography in the two- women. In the CNBSS, for women age 40–49 allocated to mam-
county study (62%) and in the first screen of the Stockholm mography, clinical examination alone was the mode of cancer
study (53%) (3). detection in 23.5%, compared to only 12% for women age 50–
Only the CNBSS trials can compare the sensitivity of mam- 59 allocated to mammography.
mographic screening with screening by clinical breast exam- d) Nodal status: Mammographically detected cancers are
ination alone (11). Comparing CNBSS sensitivity rates for two- more likely to be node negative than those detected by clinical
modality screening to single modality with clinical breast examination. Table 4 reveals not only that single-modality
examination in the two age groups, 40–49 at entry and 50–59 at screening in CNBSS-I detected fewer cancers (55) at the first
entry, four observations can be made. First, two-modality screening round than two-modality (86), but also that the latter
screening achieved a higher sensitivity in older (88%) than is associated with a higher proportion of node-negative invasive
younger (81%) women. Secondly, sensitivity estimates for clini- tumors and a marginally higher proportion of node-positive tu-

Table 2. Screen detection rates/1000: two- versus single-modality screening*

CNBSS Edinburgh (5)

Trial age 40–49 (6) 50–59 (7) 45–64
intervention
rounds MA + CBE CBE MA + CBE CBE MA + CBE CBE

1 3.89 2.46 7.20 3.45 6.15 —

2 1.74 — 3.74 1.95 — 1.75
3 1.99 — 2.48 1.28 3.15 —
4 2.38 — 3.14 0.89 — 0.85
5 1.84 — 2.84 1.64 3.33 —
6 — — — — — 1.03
7 — — — — 3.08 —

*MA ⳱ mammography; CBE ⳱ clinical breast examination.

Journal of the National Cancer Institute Monographs No. 22, 1997 127
Downloaded from https://academic.oup.com/jncimono/article-abstract/1997/22/125/2952583
by guest
on 12 May 2018
 Table 3. Mode of cancer detection with two-modality screening*

Percent detected at screening

Study Age (year) n MA only CBE only MA + CBE

HIP (1988) (14) 40–64 132 33.3 44.7 22.0

BCDDP (1987)† (9) 37–74 3548 35.5 7.9 53.3
Edinburgh (1990)‡ (5) 45–64 88 22.7 3.4 73.9
CNBSS† (6) 40–49 255 40.4 23.5 36.1
CNBS† (7) 50–59 325 53.2 12.0 34.8
Utrecht (1984)‡ (10) 50–69 196 55.6 9.7 34.6

*MA ⳱ mammography; CBE ⳱ clinical breast examination.

†All cancers.
‡Invasive cancers.

Table 4. Screen-1 nodal status of invasive cancers in patients age 40–49 observed for women in the mammography arm who were de-
from CNBSS* tected by clinical examination only. Although the survival rates
Annual MA + CBE Single CBE associated with detection by mammography alone in all three
Allocation studies exceed those for the other two modes of detection, this is
nodal status n (%) n (%)
insufficient to prove that mortality from breast cancer has been
Node-negative 52 (60) 30 (54) reduced.
Node-positive 33 (38) 20 (36)
Status unknown 1 (2) 5 (10)
The major conclusion of the Boston case series (8), implau-
Total 86 (100) 55 (100) sibly endorsed in the journal Science (15), was that five-year
survival postdiagnosis was excellent at 95% for women whose
*MA ⳱ mammography; CBE ⳱ clinical breast examination. breast cancers were detected by mammography alone, while that
for women whose breast cancer was physically palpable was
much lower at 74%. The usefulness of such case series, how-
mors. Since equal numbers of women were assigned to the in-
ever, is limited by the lack of an appropriate comparison group,
tervention and control arms, 25,214 and 25,216, respectively, it
lead-time bias, and selection bias. Indeed, in CNBSS-1, seven-
is appropriate to show frequencies rather than rates.
year survival for breast cancer patients age 40–49 detected by
e) Survival postdiagnosis: Table 5 compares survival at 10
mammography alone was 95%, while that for women who did
years postdiagnosis according to mode of detection for women
not receive mammography was 91% (6).
with screen-detected breast cancer. Lead-time bias is not an
f) Deaths 10 years after entry: Much concern has been
issue here because what is being described are the survival rates
expressed about the asymmetric distribution of advanced breast
associated with the three modes of detection possible in two-
cancer in CNBSS-I at the first screening round in women age
modality screening. The comparisons across these North Ameri-
40–49 on entry (16,17), namely an excess of advanced breast
can screening studies are impeded by unmatched age groupings
cancer detected in the two-modality arm of the trial compared
for the cohorts, with younger women having a lower risk of
to the control arm. Table 6 displays the distribution of deaths
breast cancer than older. CNBSS-I has the youngest cohort (age
that have occurred approximately 10 years after entry, in
40–49) (6) compared with the HIP Study (age 40–64) (14) and
CNBSS-I women who had breast cancer detected either at the
the BCDDP (age 34–74) (9). The highest survival rates are ob-
first screening round or in the first 12 months thereafter. For
served in the CNBSS for every mode of detection in women who
two-modality and single-modality groups, the distribution of
received combined mammography and clinical examination. Be-
breast cancer deaths in cases detected in the first year is now
cause the CNBSS is the most recently conducted study of the
21 versus 19, respectively, compared to 16 versus 10 at the
three displayed, one contributing element may be better mam-
seven-year follow-up (6). Including deaths in women with breast
mographic technology in the CNBSS compared to the two older
cancer due to other causes [all causes of death in breast can-
trials. CNBSS 10-year survival postdiagnosis for women as-
cer patients are verified by external panel review (18)], the totals
signed to receive a single clinical examination matched that
are 22 and 21 for the two groups, respectively. This near equal-
ization in distribution should lessen the persuasiveness of criti-
Table 5. Survival at 10 years by mode of detection—screen cancers
only (%)*
Table 6. Deaths due to invasive breast cancer 10 years* after entry among
Study (age, year)
CNBSS subjects aged 40–49†
HIP BCDDP CNBSS
(40–64) (4) (34–74) (9) (40–49) (6) Allocation MA + CBE CBE only

Mode of detection MA + CBE MA + CBE MA + CBE Single CBE Screen-1 15 9

Interval-1 6 10
MA only 77 85 93 NA Deaths due to other causes 1 2
CBE only 59 76 84 86 Total 22 21
MA + CBE 55 77 81 NA
*One CBE screen death occurred at 10 y 3 m and one at 10 y 16 d.
*MA ⳱ mammography; CBE ⳱ clinical breast examination. †MA ⳱ mammography; CBE ⳱ clinical breast examination.

128 Journal of the National Cancer Institute Monographs No. 22, 1997
Downloaded from https://academic.oup.com/jncimono/article-abstract/1997/22/125/2952583
by guest
on 12 May 2018
 cism directed at the CNBSS (16,17), especially in light of be employed in screening, examiners will need to be carefully
similar patterns of mortality observed in other trials (18). The trained and monitored. If the costs of a screening program must
CNBSS mortality results are compatible with conclusions be limited, one could recommend that clinical breast examina-
reached from meta-analyses, namely that benefit from screening tion should, at the very least, be part of the screening protocol for
women age 40–49 is slow to appear. The recently published women under age 50 because, at that age, the sensitivity of
external review of CNBSS randomization (19) by forensic ex- mammography is lower than in later years.
perts found no evidence of subversion in CNBSS randomiza-
References
tion procedures. An accompanying editorial (20) includes fac-
tual inaccuracies that have recently been corrected (21). This is not (1) Black WC, Nease RF Jr, Tosteson AN. Perceptions of breast cancer risk
and screening effectiveness in women younger than 50 years of age. J Natl
the first time that factual inaccuracies have been published (18). Cancer Inst 1995;87:720–31.
(2) Baines CJ. Women and breast cancer: is it really possible for the public to
Conclusions be well informed? [editorial]. Can Med Assoc J 1992;146:2147–8.
(3) Fletcher SW, Black W, Harris R, Rimer BK, Shapiro S. Report of the
Proponents of mammography screening in women age 40–49 International Workshop on Screening for Breast Cancer. J Natl Cancer Inst
have rightly said it is inappropriate to recommend clinical breast 1993;85:1644–56.
examination for screening in the absence of evidence. Certainly (4) Shapiro S, Venet W, Strax P, Venet L, Roescer R. Ten- to fourteen-year effect
of screening on breast cancer mortality. J Natl Cancer Inst 1982;69:349–55.
evidence from an RCT in North America comparing screening (5) Roberts MM, Alexander FE, Anderson I, Chetly U, Donnan PT, Forrest P,
with clinical breast examination to no screening will never be et al. Edinburgh trial of screening for breast cancer: mortality at seven
available. Therefore, evidence on clinical breast examination years. Lancet 1990;335:241–6.
(6) Miller AB, Baines CJ, To T, Wall C. Canadian National Breast Screening
from existing trials and projects must be examined. In fact, only Study: 1. Breast cancer detection and death rates among women aged 40 to
the CNBSS allows comparative evaluation of clinical breast ex- 49 years [published erratum appears in Can Med Assoc J 1993;148:718].
amination, and the comparison is with two-modality screening, Can Med Assoc J 1992;147:1459–76.
(7) Miller AB, Baines CJ, To T, Wall C. Canadian National Breast Screening
not ‘‘no screening.’’ Study: II. Breast cancer detection and death rates among women aged
Because proponents of mammography have repeatedly called 50–59 years [published erratum appears in Can Med Assoc J 1993;148:
the CNBSS mammography ‘‘flawed’’ (16,17), the question 718]. Can Med Assoc J 1992;147:1477–88.
(8) Stacey-Clear A, McCarthy KA, Hall DA, Pile-Spellman E, White G, Hulka
arises, Are the achievements of clinical breast examination in the C, et al. Breast cancer survival among women under age 50: is mammog-
CNBSS enhanced because of ‘‘flawed mammography’’? As has raphy detrimental? Lancet 1992;340:991–4.
been reported before (18), there is much evidence to answer (9) Seidman H, Gelb SK, Silverberg E, LaVerda N, Lubera JA. Survival experi-
ence in the Breast Cancer Detection Demonstration Project. CA 1987;37:
‘‘no.’’ The CNBSS has achieved results equal to or better than 258–90.
other RCTs with respect to successful randomization (which (10) DeWaard EF, Collette HJA, Rombach JJ, Banders-van Halewijn EA, Hon-
cannot be similarly documented for any other trial), cancer de- ing C. The DOM project for the early detection of breast cancer, Utrecht,
The Netherlands. J Chron Dis 1984;37:1–44.
tection rates, prevalence/incidence ratio, and survival. In short, (11) Baines CJ, Miller AB, Bassett AA. Physical examination. Its role as a
there is no persuasive evidence that ‘‘flawed’’ mammography single screening modality in the Canadian National Breast Screening Study.
enhanced the achievements of clinical breast examination ob- Cancer 1989;63:1816–22.
(12) Newcomb PA, Weiss NS, Storer BE, Scholes D, Young BE, Voigt LF.
served in the CNBSS. Breast self examination in relation to the occurrence of advanced breast
Unquestionably, by any of the parameters examined, screen- cancer. J Natl Cancer Inst 1991;83:2605.
ing with mammography, alone or in combination with clinical (13) Harvey BJ, Miller AB, Baines CJ, Corey PN. A nested case-control study
of breast self examination practice. Can Med Assoc J. In press.
exam, performs better than clinical breast examination alone. (14) Shapiro S, Venet W, Strax P, Venet L. Current results of the breast cancer
The differences described may be smaller and possibly less im- screening randomized trial: the Health Insurance Plan (HIP) of Greater
portant than many would predict, with one major exception: New York Study. In: Day NE, Miller AB, editors. Screening for Breast
Cancer. Toronto: Hans Huber, 1988:3–15.
cancer detection rates. These are always considerably higher (15) Good news on mammograms. Science 1992;258:739.
when mammography is used. Nevertheless, this may not be (16) Tarone E. The excess of patients with advanced breast cancer in young
an unqualified benefit given the likelihood of overdiagnosis women screened with mammography in the Canadian National Breast
Screening Study. Cancer 1995;75:997–1003.
(22,23). (17) Kopans DB, Feig SA. The Canadian National Breast Screening Study: a
Because two-modality screening out-performs mammography critical review. AJR Am J Roentgenol 1993;161:755–60.
alone, there is a role for clinical breast examination in breast (18) Baines CJ. The Canadian National Breast Screening Study: a perspective
on criticisms. Ann Int Med 1994;120:326–34.
screening if women are to gain the most benefit from screening. (19) Bailar JC III, MacMahon B. Randomization in the CNBSS: a review for
It has long been known that biopsy of a palpable mass should not evidence of subversion. Can Med Assoc J 1997;156:193–9.
be deferred because of negative mammograms (24). With mam- (20) Boyd NF. The review of randomization in the Canadian National Breast
Screening Study. Is the debate over? Can Med Assoc J 1997;156:207–9.
mography alone, lumps will be overlooked, especially in (21) Cohen M, Kaufert P, MacWilliam L, Tate R. Checking random assignment
younger women. with claims data. Can Med Assoc J 1997;156:1269–70.
As with mammography, breast examination technique must (22) Fletcher SW. Breast cancer screening among women in their forties: an
overview of the issues. Monogr Natl Cancer Inst 1997;22:5–9.
be excellent in order to be useful. And excellence can be (23) Andersson I, Janzon L. Reduced breast cancer mortality in women under
achieved. It has been demonstrated that medical school curricula age 50: Updated results from the Malmö Mammographic Screening Pro-
could be revised to enhance clinical breast examination compe- gram. Monogr Natl Cancer Inst 1997;22:63–7.
(24) Winchester DP. Physical examination of the breast. Cancer 1992;69:
tence among medical students (25) and that educational pro- 1947–9.
grams can effectively improve examination competence among (25) Campbell HS, McBean M, Mandin H, Bryant H. Teaching medical students
health professionals (26). The need to achieve excellence should how to perform a clinical breast examination. Acad Med 1994;69:993–5.
(26) Campbell HS, Pilgrin CA, Fletcher SE, Morgan TM, Lin S. Improving
not be a deterrent to clinical breast examination any more than physicians’ and nurses’ clinical breast examination. A randomized con-
it has been to mammography. If clinical breast examination is to trolled trial. Am J Prev Med 1991;7:1–8.

Journal of the National Cancer Institute Monographs No. 22, 1997 129
Downloaded from https://academic.oup.com/jncimono/article-abstract/1997/22/125/2952583
by guest
on 12 May 2018