ArticlePDF Available

Overdiagnosis in Breast Cancer Screening: Time to Tackle an Underappreciated Harm

Overdiagnosis in Breast Cancer Screening: Time to Tackle an
Underappreciated Harm
The earlier that cancer is found, the better. This under-
lying tenet of cancer screening has saved many lives.
However, studies are beginning to challenge the certainty
that finding cancer early is always better. One stark exam-
ple was the practice of screening infants for neuroblastoma.
After the introduction of a simple urine test in Japan led to
nationwide screening, the incidence of neuroblastoma ap-
proximately doubled, whereas that of mortality and late-
stage disease remained unaffected. Widespread urine test-
ing of infants was subsequently abandoned (1). The excess
cases of cancer found on screening were examples of over-
diagnosis, defined as occurring when “a condition is diag-
nosed that would otherwise not go on to cause symptoms or
death” (2). Overdiagnosis need not imply that a given screen-
ing effort is ineffective or ill-advised. Indeed, overdiagnosis has
been documented in effective screening programs for several
types of cancer, including breast cancer.
Pathologically diagnosed breast cancer is heteroge-
neous; whereas some tumors grow rapidly, others grow
slowly, and still others may never grow. Tumors that grow
slowly or not at all can lead to overdiagnosis. Unfortu-
nately, mammography screening programs cannot distin-
guish between fatal and harmless breast cancer. Breast can-
cer overdiagnosis can only harm the affected woman.
Whereas other harms of breast cancer screening, such as
pain from compression of the breasts during examination
or anxiety due to false-positive results, are transitory, the
impact of a cancer diagnosis lasts a lifetime.
Reported estimates of breast cancer overdiagnosis
range from 0% to 54%, highlighting the complexity of this
topic (3–8). Long-term follow-up of women in random-
ized trials provide some estimates. For example, in the
Malmo¨ randomized, controlled trial (3), the total number
of diagnosed invasive breast cancer remained higher in the
screened group than in the control group after 15 years of
follow-up, a persistent excess of 115 cases. A spike in breast
cancer incidence would be expected early in the screening
group, but the number of cases in the control group should
“catch up” over time if no overdiagnosis occurred.
Population-level cancer registries offer another approach,
tracking the incidence of breast cancer before and after
screening is introduced. However, changes other than the
introduction of screening might influence breast cancer in-
cidence. Estimates of overdiagnosis tend to be higher when
studies use a denominator that is restricted to only
screening-detected cancer (because overdiagnosis can occur
only in this subset), include ductal carcinoma in situ cases
as well as invasive cancer, involve shorter follow-up, or
include control groups that have little or no access to
Kalager and colleagues’ study (9) in this issue adds
valuable data by presenting a population-based trial that
describes the incidence of invasive breast cancer in Nor-
way; the authors took advantage of the fact that breast
cancer screening was gradually introduced in different geo-
graphic areas of the country between 1996 and 2005. Stag-
gered introduction enabled comparison of both concurrent
and historical trends in breast cancer incidence and allowed
comparison of regions with and without mammography
screening. The study included 39 888 patients with inva-
sive breast cancer, of whom almost 8000 were diagnosed
after the introduction of routine screening. Overall, Kal-
ager and colleagues estimated that 15% to 25% of the cases
diagnosed in the screening areas represented overdiagnosis
(9). Their estimates varied depending on the length of
follow-up and were calculated from a denominator that
included all cases of cancer, not just those detected by
Strengths of the study include the high attendance rate
(77%) of women in the screening program and the re-
ported low rate of mammography in areas where the
screening program had not yet been introduced. However,
because the study was not a randomized trial, women in
the various regions may have differed in ways other than
breast cancer screening; the authors point out that different
regions of Norway varied in breast cancer incidence rates
and temporal trends. The study is also limited in that the
follow-up after introduction of screening in some areas
may not have been long enough to allow stable estimates of
the degree of overdiagnosis.
Two factors suggest that estimates of overdiagnosis
from Norway may not generalize to the United States.
First, U.S. radiologists are more likely than their European
counterparts to report abnormalities found on mammo-
grams (10, 11). Second, U.S. women often start annual
mammography screening at age 40 years, whereas Norwe-
gian women start biennial screening at age 50 years. Given
more frequent screening over a longer time, overdiagnosis
probably occurs more often in the United States than in
Instead of focusing on the exact extent of overdiagno-
sis, it is time to agree that any amount of overdiagnosis is
serious and to start dealing with this issue now. Ultimately,
better tools are needed to reliably identify which breast
cancer will be fatal without treatment and which can be
safely observed over time without intervention, but we can-
not wait for these tools to be developed. Mammographers,
especially those in the United States, could help by consid-
ering changes in the threshold for calling a mammographic
feature abnormal. Evaluating strategies for observing
change in some lesions over time instead of recommending
Annals of Internal MedicineEditorial
536 © 2012 American College of Physicians
Downloaded From: by a Universiy of Washington User on 12/19/2014
an immediate biopsy has been suggested (10, 12, 13). This
may be a tough sell for women with anxiety as a result of
the “watch-and-wait” approach, as well as for radiologists
who do not want to miss any sign of disease and fear
malpractice lawsuits. Nevertheless, unless serious efforts are
made to reduce the frequency of overdiagnosis, the prob-
lem will probably increase as newer imaging modalities,
such as breast magnetic resonance imaging, are introduced.
Finally, we have an ethical responsibility to alert
women to this phenomenon. Most patient-education aids
do not even mention overdiagnosis (14), and most women
are not aware of its possibility (15). Effective communica-
tion about overdiagnosis of breast cancer will require great
care—and evaluation to determine how best to do it; oth-
erwise, women may become fearful or angry. Just because
communicating with patients will be difficult does not mean
that we should not tackle this problem. Informed women
deserve no less when deciding about breast cancer screening.
Joann G. Elmore, MD, MPH
University of Washington School of Medicine
Seattle, WA 98104-2499
Suzanne W. Fletcher, MD
Harvard Medical School and Harvard Pilgrim Health Care
Boston, MA 02215
Potential Conflicts of Interest: Disclosures can be viewed at www
Requests for Single Reprints: Joann G. Elmore, MD, MPH, 325
Ninth Avenue, Room 10EH03, Box 359780, Seattle, WA 98104-2499.
Current author addresses are available at
Ann Intern Med. 2012;156:536-537.
1. Sawada T. Past and future of neuroblastoma screening in Japan. Am J Pediatr
Hematol Oncol. 1992;14:320-6. [PMID: 1456397]
2. Welch HG, Black WC. Overdiagnosis in cancer. J Natl Cancer Inst. 2010;
102:605-13. [PMID: 20413742]
3. Zackrisson S, Andersson I, Janzon L, Manjer J, Garne JP. Rate of over-
diagnosis of breast cancer 15 years after end of Malmo¨ mammographic screening
trial: follow-up study. BMJ. 2006;332:689-92. [PMID: 16517548]
4. Duffy SW, Agbaje O, Tabar L, Vitak B, Bjurstam N, Bjo¨rneld L, et al.
Overdiagnosis and overtreatment of breast cancer: estimates of overdiagnosis
from two trials of mammographic screening for breast cancer. Breast Cancer Res.
2005;7:258-65. [PMID: 16457701]
5. Jørgensen KJ, Gøtzsche PC. Overdiagnosis in publicly organised mammog-
raphy screening programmes: systematic review of incidence trends. BMJ. 2009;
339:b2587. [PMID: 19589821]
6. Wu D, Pe´rez A. A limited review of over diagnosis methods and long term
effects in breast cancer screening. Oncol Rev. 2011;5:143-147. [PMID:
7. Biesheuvel C, Barratt A, Howard K, Houssami N, Irwig L. Effects of study
methods and biases on estimates of invasive breast cancer overdetection with
mammography screening: a systematic review. Lancet Oncol. 2007;8:1129-38.
[PMID: 18054882]
8. Gøtzsche PC, Nielsen M. Screening for breast cancer with mammography.
Cochrane Database Syst Rev. 2009:CD001877. [PMID: 19821284]
9. Kalager M, Adami HO, Bretthauer M, Tamimi RM. Overdiagnosis of inva-
sive breast cancer due to mammography screening: results from the Norwegian
screening program. Ann Intern Med. 2012;156:491-9.
10. Elmore JG, Nakano CY, Koepsell TD, Desnick LM, D’Orsi CJ, Ransohoff
DF. International variation in screening mammography interpretations in
community-based programs. J Natl Cancer Inst. 2003;95:1384-93. [PMID:
11. Smith-Bindman R, Chu PW, Miglioretti DL, Sickles EA, Blanks R,
Ballard-Barbash R, et al. Comparison of screening mammography in the United
States and the United Kingdom. JAMA. 2003;290:2129-37. [PMID: 14570948]
12. Welch HG, Schwartz LM, Woloshin S. Overdiagnosed: Making People Sick
in the Pursuit of Health. 1st ed. Boston: Beacon Pr; 2012.
13. Esserman L, Thompson I. Solving the overdiagnosis dilemma [Editorial]. J
Natl Cancer Inst. 2010;102:582-3. [PMID: 20413743]
14. Jørgensen KJ, Gøtzsche PC. Content of invitations for publicly funded
screening mammography. BMJ. 2006;332:538-41. [PMID: 16513713]
15. Schwartz LM, Woloshin S, Sox HC, Fischhoff B, Welch HG. US women’s
attitudes to false positive mammography results and detection of ductal carci-
noma in situ: cross sectional survey. BMJ. 2000;320:1635-40. [PMID:
EditorialOverdiagnosis in Breast Cancer Screening 3 April 2012 Annals of Internal Medicine Volume 156 • Number 7 537
Downloaded From: by a Universiy of Washington User on 12/19/2014
Current Author Addresses: Dr. Elmore: 325 Ninth Avenue, Room
10EH03, Box 359780, Seattle, WA 98104-2499.
Dr. Fletcher: 208 Boulder Bluff, Chapel Hill, NC 27516.
Annals of Internal Medicine
W-174 3 April 2012 Annals of Internal Medicine Volume 156 • Number 7
Downloaded From: by a Universiy of Washington User on 12/19/2014
... Inconsistent or inaccurate interpretation of germline TP53 variants may lead to serious harms not only from missing a diagnosis of LFS and the consequent lack of life-saving cancer screening and risk management, but also from overscreening of those inadvertently diagnosed with LFS, causing unnecessary anxiety and possible rare complications from screening modalities. 20,21 Individuals may choose to undergo invasive risk-reducing procedures, such as prophylactic mastectomy, on the basis of having a P and/or LP TP53 variant, a decision that cannot be reversed if a variant is downclassified. 22,23 Patients and families may face significant clinical and psychologic impact when management recommendations are changed because of variant reclassification or uncertainty. ...
PURPOSE The use of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology guidelines has improved germline variant classification concordance, but discrepancies persist, sometimes directly affecting medical management. We evaluated variant discordance between and within families with germline TP53 variants in the National Cancer Institute's Li-Fraumeni syndrome longitudinal cohort study. MATERIALS AND METHODS Germline TP53 genetic testing results were obtained from 421 individuals in 140 families. A discordant test result was defined as a report of pathogenicity that differed between two clinical testing laboratories, between a testing laboratory and the ClinVar database, or between either the laboratory or ClinVar database and variant classification by internal study review. RESULTS There were 141 variants in 140 families (one family had two different TP53 variants). Fifty-four families had discordant interpretations (54 of 140, 39%). Sixteen families had discordant classifications leading to clinically important differences in medical management (16 of 140, 11%). Interfamilial discordance was observed between four families (two different variants). Intrafamilial discordance was observed within six families. One family experienced both intrafamilial and interfamilial discordance. CONCLUSION This large single-gene study found discordant germline TP53 variant interpretations in 39% of families studied; 11% had a variant with the potential to significantly affect medical management. This finding is especially concerning in patients with Li-Fraumeni syndrome because of their exceedingly high risks of multiple cancers and intensive cancer screening and risk-reducing recommendations. Centralized data sharing, gene-specific variant curation guidelines, and provider education for consistent variant interpretation are essential for optimal patient care.
... Many of these tests are too sensitive and less specific, leading to false positive cases as they may suffer from "source uncertainty" which is associated with circulating nucleic acid fragments: circulating mutant fragments of DNA can emanate from diverse sources that may not necessarily represent viable malignancy, for example alterations captured in cfDNA due to clonal hematopoietic mutations of indeterminate potential (5). Also, the test may turn out to be positive too early, making radiologic or clinical verification almost impossible, thereby causing overdiagnosis and anxiety (6,7). For these reasons, blood-based tests have not yet gained wider acceptance or adoption. ...
Full-text available
We have previously shown that circulating ensembles of tumor-associated cells (C-ETACs) are a systemic hallmark of cancer based on analysis of blood samples from 16,134 individuals including 10,625 asymptomatic individuals and 5,509 diagnosed cases of cancer. C-ETACs were ubiquitously (90%) detected across all cancer types and were rare (3.6%) among the asymptomatic population. Consequently, we hypothesized that asymptomatic individuals with detectable C-ETACs would have a definitively elevated risk of developing cancer as compared with individuals without C-ETACs. In the present manuscript we present 1-year follow-up data of the asymptomatic cohort which shows that C-ETAC positive individuals have a 230-fold (P < 0.00001) higher 1-year cancer risk as compared with individuals where C-ETACs were undetectable. Simultaneously, we also expanded the study to include 4,419 symptomatic individuals, suspected of cancer, prior to undergoing an invasive biopsy for diagnosis. C-ETACs were detected in 4,101 (92.8%) of these 4,419 cases where cancer was eventually confirmed. We conclude that detection of C-ETACs can identify patients at risk of cancer and can be reliably used to stratify asymptomatic individuals with an elevated 1-year risk of cancer. Prevention Relevance The study evaluated a blood test that can determine if healthy (‘asymptomatic’) individuals without a history of cancer have an increased risk of developing cancer within the next one year. This test can significantly minimize radiological or invasive screening in the majority individuals who do not have any increased risk.
... In fact, the estimated prevalence of breast cancer overdiagnosis has been up to 54% in some populations, which highlights the extent of the problem. 56 Such overdiagnosis may enhance the problem of overtreatment instead of lowering it, because of the overenthusiastic administration of treatment on demand by patients who are scared by the diagnosis of cancer. 57 Another problem documented with early screening for breast cancer is bias, such as healthy volunteer bias, length-biased sampling, and lead-time bias. ...
Full-text available
The prevalence and mortality of breast cancer is increasing in Asian countries, including India. With advances in medical technology leading to better detection and characterization of the disease, it has been possible to classify breast cancer into various subtypes using markers, which helps predict the risk of distant recurrence, response to therapy, and prognosis using a combination of molecular and clinical parameters. Breast cancer and its therapy, mainly surgery, systemic therapy (anticancer chemotherapy, hormonal therapy, targeted therapy, and immunotherapy), and radiation therapy, are associated with significant adverse influences on physical and mental health, quality of life, and the economic status of the patient and her family. The fear of recurrence and its devastating effects often leads to overtreatment, with a toxic cost to the patient financially and physically in cases in which this is not required. This article discusses some aspects of a breast cancer diagnosis and its impact on the various facets of the life of the patient and her family. It further elucidates the role of prognostic factors, the currently available biomarkers and prognostic signatures, and the importance of ethnically validating biomarkers and prognostic signatures.
... Overdiagnosis is the diagnosis of a medical condition that would never have caused any symptoms or problems. Reported estimates of breast cancer overdiagnosis range from 0% to 54% (Elmore and Fletcher 2012). ...
Full-text available
Classic chronic diseases progression models are built by gauging the movement from the disease free state, to the preclinical (asymptomatic) one, in which the disease is there but has not manifested itself through clinical symptoms, after spending an amount of time the case then progresses to the symptomatic state. The progression is modelled by assuming that the time spent in the disease free and the asymptomatic states are random variables following specified distributions. Estimating the parameters of these random variables leads to better planning of screening programs as well as allowing the correction of the lead time bias (apparent increase in survival observed purely due to early detection). However, as classical approaches have shown to be sensitive to the chosen distributions and the underlying assumptions, we propose a new approach in which we model disease progression as a gamma degradation process with random starting point (onset). We derive the probabilities of cases getting detected by screens and minimize the distance between observed and calculated distributions to get estimates of the parameters of the gamma process, screening sensitivity, sojourn time and lead time. We investigate the properties of the proposed model by simulations.
... A flood of damaging reports amplified perceived risk of "overdiagnosis" and overtreatment. Study after flawed study-like a house of cards-touted mammography's alleged harms and costs [10][11][12][13][14][15][16][17]. ...
It's been ten years since the U.S. Preventive Services Task Force pulled back on recommendations for breast cancer screening in women ages 40 - 49 years. After a decade of negative reports, most physicians are aware of mammography's limits. Today, many women avoid, delay or deliberately skip getting screened. As invasive breast cancer rates have been rising, and breast cancer remains a leading cause of death, truthful information about screening is critical for public health. Unfortunately, many reports about mammography exaggerate its harms and over-estimate overdiagnosis. The public should be aware of current evidence supporting the benefit of breast cancer screening, including a 40% decline in the U.S. mortality rate in the mammography era. Delayed diagnosis has a downside, about which women should be informed. Contrary to popular views, breast cancer stage remains a key determinant of long-term prognosis. For the most common form of breast cancer, small tumor size and lack of lymph node involvement portend significantly better outcomes than larger tumors with positive nodes. Although mammography is not full-proof, the technology continues to improve; it is currently the best tool for finding breast cancer before it is greater than 2 centimeters or has spread. Interdisciplinary discussion of this topic by primary care physicians, oncologists, radiologists, public health experts, pathologists, and patient advocates would serve women's health.
... Physicians have never been obliged to offer nonbeneficial care and they can confidently recommend against marginally beneficial care that they believe is not worthwhile. The principle of nonmaleficence is particularly pertinent in the case of this young woman, inasmuch as many have called attention to the risk of harm from overdiagnosing breast cancer in women her age [8,9]. Since 2009 the United States Preventive Services Task Force guidelines no longer unequivocally recommend mammograms for women younger than 50 but rather defer to shared decision making based on individual risk-to-benefit assessments [10], and in 2015 the American Cancer Society updated its guidelines, recommending that women with average breast cancer risk begin regular screening mammograms at age 45 [11]. ...
If she has private insurance, the ACA mandates that her plan cover a screening mammogram without cost sharing [1]. (This is interesting given the weak evidence supporting mammograms for women 40-50 years old [10] and points to the strong political sensitivities surrounding breast cancer screening.) Refusing to refer an insured patient for mammography will not reliably benefit other patients more in need of services since the money is just as likely to increase the takings of insurance company shareholders. Such savings offer little justification to withhold the service [23]. If all physicians restricted the use of mammograms for this low-risk group, it could possibly decrease the cost of insurance and thus benefit other patients. Given the universally mandated insurance coverage for breast cancer screening and fear of litigation for delayed breast cancer diagnosis, however, there would have to be a paradigm shift in both insurance coverage and tort reform for the practice patterns of physicians to change. Thus Dr. Perry has no ethical obligation based on resource allocation to limit Ms. Hollowell’s access to a mammogram covered by her insurance in accordance with the law. Best interest or nonmaleficence arguments could be used to justify not yielding to Ms. Hollowell’s autonomous request and limiting her access because of the risk of harm from overtreatment as outlined above. However, in the current environment, in which mammograms are considered standard of care, Dr. Perry would be incurring significant personal liability were Ms. Hollowell to be diagnosed with breast cancer at a later stage.
Full-text available
This paper deals with degradation processes whose onset is triggered at a random time and which stay hidden until they are discovered through inspection or when they begin to show symptoms. This is applicable in many healthcare and industrial scenarios, for example, in the modeling of breast cancer or termite infestation. In our model, we assume that symptoms appear after hitting a random critical threshold and that inspections may have a sensitivity less than one as well as a nonzero false positive rate. The expected cost of repair is derived, and the inspection rate is optimized for a cycle (which lasts from degradation‐free to repaired state). This gives results for three cases: the first is for a finite observation period with no degradation recurrence, the second for infinite time horizon allowing recurrence. In the third case, we derive an upper bound for the expected cost in a given constant time period. Finally, the model is applied to determine the optimal strategy for breast cancer screening with regard to the effects of different parametrizations.
Background Little is known about the impact of primary melanoma diagnosis on healthcare utilization and changes in utilization over time. Objectives To evaluate population‐based temporal trends in healthcare utilization following primary melanoma diagnosis. Methods We conducted a before‐and‐after multiple time series study of Medicare beneficiaries aged ≥ 66 years with primary melanoma diagnoses between 2000 and 2009 using the Surveillance, Epidemiology, and End Results Medicare database. Primary exposure was time from primary melanoma diagnosis at 3–6 months and 6–24 months postdiagnosis. Covariates included tumour‐, patient‐ and geographical‐level characteristics and healthcare utilization in the 6 months before diagnosis. Poisson regression was used to estimate population‐based risk‐adjusted utilization rates for skin biopsies, benign skin excisions, internal medicine office visits and dermatology office visits. Results The study population included 56 254 patients with first diagnoses of primary melanoma. Most patients were ≥ 75 years old (56·8%), male (62·1%), and had in situ melanoma (42·4%) or localized invasive melanoma (45·9%). From 2000 to 2009, risk‐adjusted skin biopsy rates 24 months postdiagnosis increased from 358·3 to 541·3 per 1000 person‐years (P < 0·001), and dermatology visits increased from 989·0 to 1535·6 per 1000 person‐years (P < 0·001). Benign excisions and internal medicine visits remained stable. In 2000, risk‐adjusted skin biopsy rates 6 months postdiagnosis increased by 208·5 relative to the 6 months before diagnosis (148·7 vs. 357·2) compared with an observed absolute increase of 272·5 (290·9 vs. 563·1) in 2009. Trends in dermatology visits were similar. Conclusions Utilization of skin biopsies and dermatology office visits following primary melanoma diagnosis has increased substantially over time. These results may inform optimization of care delivery for melanoma within the Medicare population.
Full-text available
The risk of breast cancer (BC) overdiagnosis attributed to mammography screening is an unresolved issue, complicated by heterogeneity in the methodology of quantifying its magnitude, and both political and scientific elements surrounding interpretation of the evidence on this phenomenon. Evidence from randomized trials and also from observational studies shows that mammography screening reduces the risk of BC death; similarly, these studies provide sufficient evidence that overdiagnosis represents a serious harm from population breast screening. For both these outcomes of screening, BC mortality reduction and overdiagnosis, estimates of magnitude vary between studies however overdiagnosis estimates are associated with substantial uncertainty. The trade-off between the benefit and the collective harms of BC screening, including false-positives and overdiagnosis, is more finely balanced than initially recognized, however the snapshot of evidence presented on overdiagnosis does not mean that breast screening is worthless. Future efforts should be directed towards (a) ensuring that any changes in the implementation of BC screening optimize the balance between benefit and harms, including assessing how planned or actual changes modify the risk of overdiagnosis; (b) informing women of all the outcomes that may affect them when they participate in screening using well-crafted and balanced information; and (c) investing in research that will help define and reduce the ensuing overtreatment of screen-detected BC.
Full-text available
Screening mammography differs between the United States and the United Kingdom; a direct comparison may suggest methods to improve the practice. To compare screening mammography performance between the United States and the United Kingdom among similar-aged women. Women aged 50 years or older were identified who underwent 5.5 million mammograms from January 1, 1996, to December 31, 1999, within 3 large-scale mammography registries or screening programs: the Breast Cancer Surveillance Consortium (BCSC, n = 978 591) and National Breast and Cervical Cancer Early Detection Program (NBCCEDP, n = 613 388) in the United States; and the National Health Service Breast Screening Program (NHSBSP, n = 3.94 million) in the United Kingdom. A total of 27 612 women were diagnosed with breast cancer (invasive or ductal carcinoma in situ) within 12 months of screening among the 3 groups. Recall rates (recommendation for further evaluation including diagnostic imaging, ultrasound, clinical examination, or biopsy) and cancer detection rates were calculated for first and subsequent mammograms, and within 5-year age groups. Recall rates were approximately twice as high in the United States than in the United Kingdom for all age groups; however, cancer rates were similar. Among women aged 50 to 54 years who underwent a first screening mammogram, 14.4% in the BCSC and 12.5% in the NBCCEDP were recalled for further evaluation vs only 7.6% in the NHSBSP. Cancer detection rates per 1000 mammogram screens were 5.8, 5.9, and 6.3, in the BCSC, NBCCEDP, and NHSBSP, respectively. Recall rates were lower for subsequent examinations in all 3 settings but remained twice as high in the United States. A similar percentage of women underwent biopsy in each setting, but rates of percutaneous biopsy were lower and open surgical biopsy higher in the United States. Open surgical biopsies not resulting in a diagnosis of cancer (negative biopsies) were twice as high in the United States than in the United Kingdom. Based on a 10-year period of screening 1000 women aged 50 to 59 years, 477, 433, and 175 women in the BCSC, NBCCEDP, and NHSBSP, respectively, would be recalled; and for women aged 60 to 69 years, 396, 334, and 133 women, respectively. The estimated cancer detection rates per 1000 women aged 50 to 59 years were 24.5, 23.8, and 19.4, respectively, and for women aged 60 to 69 years, 31.5, 26.6, and 27.9, respectively. Recall and negative open surgical biopsy rates are twice as high in US settings than in the United Kingdom but cancer detection rates are similar. Efforts to improve US mammographic screening should target lowering the recall rate without reducing the cancer detection rate.
Full-text available
Breast cancer screening programs have been effective in detecting tumors prior to symptoms. Recently, there has been concern over the issue of over-diagnosis, that is, diagnosis of a breast cancer that does not manifest prior to death. Estimates for over-diagnosis vary, ranging from 7% to 52%. This variability may be due partially to issues associated with bias and/or incorrect inferences associated with the lack of probability modeling. A critical issue is how to evaluate the long-term effects due to continued screening. Participants in a periodic screening program can be classified into four mutually exclusive groups depending on whether individuals are diagnosed and whether their symptoms appear prior to death: True-early-detection; No-early-detection; Over-diagnosis; and Not-so-necessary. All initially superficially healthy people will eventually fall into one of these four categories. This manuscript reviews the major methodologies associated with the over-diagnosis and long-term effects of breast cancer screening.
Full-text available
To estimate the extent of overdiagnosis (the detection of cancers that will not cause death or symptoms) in publicly organised screening programmes. Systematic review of published trends in incidence of breast cancer before and after the introduction of mammography screening. PubMed (April 2007), reference lists, and authors. Review methods One author extracted data on incidence of breast cancer (including carcinoma in situ), population size, screening uptake, time periods, and age groups, which were checked independently by the other author. Linear regression was used to estimate trends in incidence before and after the introduction of screening and in older, previously screened women. Meta-analysis was used to estimate the extent of overdiagnosis. Incidence data covering at least seven years before screening and seven years after screening had been fully implemented, and including both screened and non-screened age groups, were available from the United Kingdom; Manitoba, Canada; New South Wales, Australia; Sweden; and parts of Norway. The implementation phase with its prevalence peak was excluded and adjustment made for changing background incidence and compensatory drops in incidence among older, previously screened women. Overdiagnosis was estimated at 52% (95% confidence interval 46% to 58%). Data from three countries showed a drop in incidence as the women exceeded the age limit for screening, but the reduction was small and the estimate of overdiagnosis was compensated for in this review. The increase in incidence of breast cancer was closely related to the introduction of screening and little of this increase was compensated for by a drop in incidence of breast cancer in previously screened women. One in three breast cancers detected in a population offered organised screening is overdiagnosed.
Precise quantification of overdiagnosis of breast cancer (defined as the percentage of cases of cancer that would not have become clinically apparent in a woman's lifetime without screening) due to mammography screening has been hampered by lack of valid comparison groups that identify incidence trends attributable to screening versus those due to temporal trends in incidence. To estimate the percentage of overdiagnosis of breast cancer attributable to mammography screening. Comparison of invasive breast cancer incidence with and without screening. A nationwide mammography screening program in Norway (inviting women aged 50 to 69 years), gradually implemented from 1996 to 2005. The Norwegian female population. Concomitant incidence of invasive breast cancer from 1996 to 2005 in counties where the screening program was implemented compared with that in counties where the program was not yet implemented. To adjust for changes in temporal trends in breast cancer incidence, incidence rates during the preceding decade were also examined. The percentage of overdiagnosis was calculated by accounting for the expected decrease in incidence following cessation of screening after age 69 years (approach 1) and by comparing incidence in the current screening group with incidence among women 2 and 5 years older in the historical screening groups, accounting for average lead time (approach 2). A total of 39,888 patients with invasive breast cancer were included, 7793 of whom were diagnosed after the screening program started. The estimated rate of overdiagnosis attributable to the program was 18% to 25% (P < 0.001) for approach 1 and 15% to 20% (P < 0.001) for approach 2. Thus, 15% to 25% of cases of cancer are overdiagnosed, translating to 6 to 10 women overdiagnosed for every 2500 women invited. The study was registry-based. Mammography screening entails a substantial amount of overdiagnosis. Norwegian Research Council and Frontier Science.
In a review in this issue of the Journal (1), Welch and Black clearly document that surveillance routinely identifies lesions that many patients would not need to know about in their lifetimes. These lesions only become a problem because we feel compelled to diagnose and treat them. What motivates intervention is the opportunity to prevent disease progression, metastasis, and death and the philosophy that “early detection is always better.” The patient’s fear of cancer and clinician’s concern about malpractice are also driving factors. But often overlooked are the profound consequences of treatment and diagnostic interventions. The article by Welch and Black should serve as a clarion call to acknowledge the spectrum of cancer behavior and the need to reclassify “indolent” lesions with a term other than “cancer” and to improve the specificity of our screening tests. This study is not “bad news,” but “good news” because it points a way forward. First, we must accept that population screening and diagnostic scans detect substantial numbers of indolent tumors and benign lesions in addition to potentially lethal disease. Second, we must resolve that we can and must address the problem. The unintended consequence of finding a precancerous lesion is exemplified by the 41-year-old research scientist who called one of us in a panic. Her first mammogram showed a cluster of calcifi cations; magnetic resonance imaging showed another focus of ductal carcinoma in situ, which led to a mastectomy that showed both lobular and ductal carcinoma in situ and an axillary sentinel node biopsy that showed isolated tumor cells. Now she faces a decision about chemotherapy and prophylactic contralateral mastectomy. Have we “helped” this patient in her goal to avoid death from cancer? The answer is “unlikely.” Much of what we call cancer is not destined for an inexorable progression to metastasis and death. We can no longer say that we must intervene because we cannot tell the difference. Raising the fraction of people diagnosed with cancer has grave consequences. It adds the burden of diagnosis to hundreds of thousands and engenders needless fear. It obscures our ability to identify and focus on tumors that need more aggressive or tailored treatment where our current approaches are unsuccessful. Cancer is a serious disease, but we have to redefine what cancer truly is.
This article summarizes the phenomenon of cancer overdiagnosis—the diagnosis of a “cancer” that would otherwise not go on to cause symptoms or death. We describe the two prerequisites for cancer overdiagnosis to occur: the existence of a silent disease reservoir and activities leading to its detection (particularly cancer screening). We estimated the magnitude of overdiagnosis from randomized trials: about 25% of mammographically detected breast cancers, 50% of chest x-ray and/or sputum-detected lung cancers, and 60% of prostate-specific antigen–detected prostate cancers. We also review data from observational studies and population-based cancer statistics suggesting overdiagnosis in computed tomography–detected lung cancer, neuroblastoma, thyroid cancer, melanoma, and kidney cancer. To address the problem, patients must be adequately informed of the nature and the magnitude of the trade-off involved with early cancer detection. Equally important, researchers need to work to develop better estimates of the magnitude of overdiagnosis and develop clinical strategies to help minimize it.