Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Dec 1.
Published in final edited form as: Mayo Clin Proc. 2022 Nov 3;97(12):2316–2323. doi: 10.1016/j.mayocp.2022.08.006

Relative Survival With Early-Stage Breast Cancer in Screened and Unscreened Populations

Andrea R Marcadis 1, Luc GT Morris 1,*, Jennifer L Marti 2,*
PMCID: PMC10314986  NIHMSID: NIHMS1898903  PMID: 36336518

Abstract

Relative survival and disease-specific survival are two statistics that measure net survival from a cancer diagnosis, excluding other causes of death. In most cases, these two rates are comparable. However, in some cancer types for which cancer screening is performed, relative survival is often greater than disease-specific survival. This divergence has been attributed to mechanisms such as the “healthy user effect” and overdiagnosis of indolent tumors detected by screening. Using relative survival rate as a marker of these mechanisms, we examined the association of breast cancer screening with relative survival rates for women diagnosed with early stage breast cancer.

In population-based data from the National Cancer Institute’s Surveillance, Epidemiology and End Results registry, we examined relative survival rates in women diagnosed with stage I breast cancer or ductal carcinoma in situ, who were in highly-screened versus less highly-screened groups, based on time period, age group, and insurance status.

In this analysis, relative survival rates for early stage breast cancer were higher than disease specific survival, even exceeding 100%, in populations experiencing higher rates of screening: women diagnosed during the era of widespread uptake of mammography, age above 40, and women with health insurance coverage.

The favorable outcomes observed in screen-detected breast cancers are at least in part attributable to the healthy user effect and overdiagnosis of indolent tumors. Therefore, survival rates may not accurately reflect the effectiveness of cancer screening. These findings have implications for counseling of patients and future clinical studies of active monitoring approaches in breast cancer.

Keywords: breast cancer, DCIS, screening, survival, overdiagnosis

Introduction

Survival in oncology is commonly measured by statistics such as overall or disease-specific survival. Another measure for estimating cancer-specific survival is relative survival, which compares the probability of dying in a population with cancer compared to a matched population without cancer. Relative survival is the ratio of observed to expected survival, matching patients with cancer to the general population, based on age, race, sex, and time-period. In most cancers, relative and disease-specific survival rates are similar, and relative survival statistics have the potential advantage of not being affected by misattributed cause of death.1

However, relative survival can differ from disease-specific survival in certain scenarios, where baseline characteristics of the persons diagnosed with cancer diverge from those of persons without cancer. For example, patients with certain cancer diagnoses (e.g. lung or head and neck cancers) tend to have lower life expectancy compared to the general population, due to factors such as a higher prevalence of smoking. Therefore, relative survival rates for patients with these cancers will tend to be lower than disease-specific survival.2,3

The opposite finding is also possible: in certain cancer types, relative survival rates could be greater than disease-specific survival, if the cancer population is healthier than the general population. This can be observed in incidentally or screen-detected cancers, as patients who seek out or receive preventive care are known to more often engage in other healthy behaviors, a phenomenon known as the “healthy user effect” or “healthy screenee bias.”4,5 If some of these cancers have disease-specific survival rates of nearly 100%6, this effect could elevate relative survival rates to over 100%. Indeed, this phenomenon has been described in early-stage cancers of the breast, prostate, thyroid, and skin (melanoma).7 Therefore, relative survival rates over 100% are characteristic of cancer types in which screening or incidental detection in healthy persons lead to overdiagnosis – defined as the diagnosis of indolent tumors that would not cause harm if they remained undetected.

Here, we use relative survival rates in patients with breast cancer to better understand the association between cancer screening and diagnosis of indolent disease in healthy or health-conscious persons. We examined differences in relative survival with invasive breast cancer and ductal carcinoma in situ (DCIS) among women with higher and lower exposure to screening mammography based on different time periods, age groups and insurance status.

Methods

The National Cancer Institute’s Surveillance, Epidemiology and End Results (SEER) program is the source of United States population-based cancer incidence and survival data. We accessed the data for the SEER 9 Research Data set, April 2020 release, based on data from the November 2019 submission. Data were accessed using SeerStat (version 8.3.6, National Cancer Institute, Bethesda, MD). The Memorial Sloan Kettering Cancer Center Institutional Review Board determined that this research using SEER cancer registry data, which is deidentified, was exempt from review. The investigators signed a data use agreement with the SEER program prior to accessing and analyzing data.

Disease-specific survival and relative survival were analyzed for women diagnosed with early-stage invasive carcinoma of the breast (categorized as stage I cancers based on the AJCC 6th edition, the staging system contemporaneous with diagnosis in this cohort) and ductal carcinoma in situ (using histology codes for DCIS, ‘8230/2: Ductal carcinoma in situ, solid type’, ‘8453/2: Intraductal papillary-mucinous carcinoma, in situ’,’8500/2: Intraductal carcinoma, noninfiltrating, NOS’, ‘8507/2: Intraductal micropapillary carcinoma’). American Joint Committee on Cancer (AJCC) 6th edition staging is available in SEER for cancers diagnosed in years 2004–2016. For the analysis of trends in survival rates over time, the cohort was restricted to patients diagnosed from 1980–2006, so that patients in all diagnosis years had complete 10 year survival data. Rates of mammographic screening among United States women were obtained from published reports from Centers for Disease Control National Center for Health Statistics.9

Relative survival in SEERStat was calculated using the default Ederer II method, which calculates the expected survival rates for patients under observation at each point of follow-up, so the matched individuals are considered to be at risk until the corresponding cancer patient dies or is censored (National Cancer Institute Surveillance Research Program Technical Report, https://surveillance.cancer.gov/reports/tech2011.01.pdf). Expected survival during the time period of the study comes from US annual life tables from the National Center for Health Statistics (NCHS, https://seer.cancer.gov/expsurvival/documentation.html). Confidence intervals for relative survival rates were calculated using the linear method.

Results

Trends in relative survival across different time periods

Recommendations for screening mammography began in the early 1980s, when the American Cancer Society began recommending mammographic screening every 1–2 years for women starting at age 40.8 Since then, the rates of screening mammography in the US among women age 40 and older have increased markedly, more than doubling between 1985–2005 (Figure 1A); in contrast, fewer than 10% of women under 40 in the US have received breast cancer screening9 10. We analyzed trends in 10 year relative survival among women age ≥ 40 and age < 40 diagnosed with ductal carcinoma in situ (DCIS) and stage I breast cancer during this time period, to determine whether increasing uptake in mammographic screening has been associated with changes in relative survival.

Figure 1. Trends in 10-year relative survival (1980–2006) for women with ductal carcinoma in situ and stage I breast cancer, compared to rates of screening mammography, stratified by age.

Figure 1.

Open in a new tab

A) Rates of mammographic screening among United States women 40 and older. Data obtained from published reports from CDC National Center for Health Statistics9 B) Relative survival rates for women aged 40 and older with ductal carcinoma in situ or stage I breast cancer has been over 100% since the late 1980s, when screening became widespread. American Joint Committee on Cancer 6th edition stages are used (system in use at the time). Data are from the National Cancer Institute’s Surveillance, Epidemiology and End Results 9 Research Data set, April 2020 release, based on data form the November 2019 submission (SeerStat version 8.3.6).

During the period when rates of mammographic screening in the US increased, there was a concomitant rise in relative survival rates among women 40 and older with DCIS and stage I breast cancer (Figure 1B). Relative survival rates for DCIS and stage I breast cancer in women age 40 and older have exceeded 100% since the late 1980s, when screening became widespread (Figure 1B). In contrast, relative survival rates in women under age 40 have remained below 100%. These data indicate that, among women in the age group for which breast cancer screening has been widely recommended, relative survival with DCIS and breast cancer has steadily risen to exceed 100%.

Relative survival by age group

In the United States between 2004–2016, 10-year relative survival for women aged 40 and older with stage I breast cancer was 101.4% (95% CI, 100.9%–101.8%), meaning that women with this diagnosis lived longer compared to age, race and sex-matched peers. This rate was higher than the 10-year disease-specific survival of 95.9% (95% CI 95.7–96.1%) (Figure 2A). During the same time period, among women in the United States younger than 40, in whom routine breast cancer screening is not recommended and rarely performed, 10-year relative survival for stage I breast cancer was below 100% (95.2% [95% CI, 93.9–96.4%]), and closely mirrored the disease-specific survival rate of 94.9% (95% CI, 93.7–96.0%) (Figure 2B).

Figure 2. Relative and disease-specific survival for women with stage I breast cancer, stratified by age and health insurance status.

Figure 2.

Open in a new tab

A) Women 40 and older with stage I breast cancer had 10-year relative survival of greater than 100% when compared with age- and race-matched counterparts without cancer, and tracked higher than their disease-specific survival B) In women younger than 40 with stage I breast cancer, 10-year relative survival was less than 100%, and tracked closely with disease-specific survival. C) Women age ≥40 with stage I breast cancer and private or Medicare-based insurance had a relative survival of greater than 100% when compared with age- and race-matched counterparts without cancer D) In women aged ≥40 with no insurance or Medicare-only insurance, relative survival was less than 100%, and tracked more closely with disease-specific survival. American Joint Committee on Cancer 6th edition stages are used (system in use at the time). Data are from the National Cancer Institute’s Surveillance, Epidemiology and End Results 9 Research Data set, April 2020 release, based on data form the November 2019 submission (SeerStat version 8.3.6), using years 2004–2016 for A-B, and 2007–2016 for C-D (years insurance data was available).

For DCIS, relative survival trends for women were similar (10-year RS for women 40 and older, 104.6%; younger than 40, 99.1%), as were disease-specific survival trends (10-year DSS for women 40 and older, 98.2%; younger than 40, 99.0%) (Figure 3A-B).

Figure 3. Relative and disease-specific survival for women with ductal carcinoma in situ (DCIS), stratified by age and health insurance status.

Figure 3.

Open in a new tab

A) Women 40 and older with DCIS had 10-year relative survival of greater than 100% when compared with age- and race-matched counterparts without cancer, and tracked much higher than their disease-specific survival. B) In women younger than 40 with DCIS, 10-year relative survival was less than 100%, and tracked closely with disease-specific survival. C) Women age ≥40 with DCIS and private or Medicare-based insurance had a relative survival of greater than 100% when compared with age- and race-matched counterparts without cancer D) In women aged ≥40 with no insurance or Medicare-only insurance, relative survival was less than 100%, and tracked closely with disease-specific survival. American Joint Committee on Cancer 6th edition stages are used (system in use at the time). Data are from the National Cancer Institute’s Surveillance, Epidemiology and End Results 9 Research Data set, April 2020 release, based on data form the November 2019 submission (SeerStat version 8.3.6), using years 2004–2016 for A-B, and 2007–2016 for C-D (years insurance data was available).

These numbers indicate that, in women in the age group undergoing breast cancer screening, relative survival rates are greater than disease-specific survival and greater than 100%. In contrast, in women under 40, in whom screening is not commonly performed, relative and disease-specific survival rates were identical.

Relative survival and insurance status

The association of breast cancer screening with relative survival can also be observed by comparing patients with differing health insurance status. Women with private or Medicare-based insurance are more likely to undergo screening mammography than women who are uninsured or have Medicaid-only insurance (76–80% vs. 50–68%).11,12 Among women with private or Medicare-based insurance, relative survival with stage I breast cancer exceeded 100% (103.0% at 9 years), and far exceeded disease-specific survival (Figure 2C). In contrast, among women over age 40 who were uninsured or insured with Medicaid (between 2007–2016), the relative survival with stage I breast cancer was <100% (89.9% at 9 years), similar to women under age 40 not undergoing routine cancer screening, and similar to disease-specific survival (Figure 2D).

Relative survival rates for women with DCIS were similar (105.5% for women with private or Medicare-based insurance; 99.3% at 9 years for women with Medicaid or no insurance). Disease specific survival rates for women with DCIS also similar (99.0% for women with private or Medicare-based insurance; 97.7% for women with Medicaid or no insurance) (Figure 3C-D). The gap between relative and disease-specific survival rates was observed in women with private or Medicare-based insurance, but not in women with no insurance or Medicaid insurance.

Discussion

We report 3 comparisons of women undergoing higher versus lower rates of screening mammography (based on time period, patient age, and patient insurance status). All 3 comparisons show consistent associations between increasing use of screening and relative survival rates that exceed 100%. These data indicate that heavily screened women diagnosed with early-stage breast cancers (stage I invasive breast cancer or DCIS) are living longer than their age, gender, and race-matched counterparts.

These findings illustrate how survival statistics are poor indicators of the effectiveness of cancer screening. Relative and disease-specific survival statistics each have strengths and weaknesses, but comparison of the two rates allows identification of populations in which the healthy user effect and overdiagnosis are prevalent. The healthy user effect will produce relative survival rates that are higher than disease-specific survival. If disease-specific survival rates are close to 100%, this will result in relative survival rates that may exceed 100%. This phenomenon is observed with some screen-detected cancers, where disease-specific survival rates are close to 100%, due to lead-time bias (screening brings forward the date of diagnosis), and length-time bias (screening preferentially detects slowly-progressing or non-progressing cancers). These are characteristics of cancer types that are prone to overdiagnosis – the detection of cancers with indolent biology that would not go on to cause symptoms, death or other harm in the person’s lifetime. Therefore, relative survival rates over 100% indicate that the population being studied is healthier than the comparison non-cancer population, and that many of the screen-detected tumors are very low-risk tumors with indolent behavior.

In each case, comparing women undergoing higher levels of mammographic screening – in more recent time periods, above age 40, or with health insurance – relative survival rates were higher than disease-specific survival rates, and above 100%. These findings are consistent with the presence of healthy screenee bias and overdiagnosis. It is important to note here that insurance status may also be associated with other factors that affect overall life expectancy and therefore relative survival rates.

There are several implications to these findings. First, these data indicate that the favorable outcomes that have been observed in screen-detected breast cancers are, in part, attributable to the healthy user effect and overdiagnosis of biologically indolent tumors, and cannot be solely attributed to screening practices intercepting aggressive cancers at an early stage.13,14,15 Second, acknowledgement of these data might mitigate some of the anxiety that a woman with a new breast cancer diagnosis faces, and reduce the tendency of patients and physicians to favor aggressive treatment of small, low-risk, screen-detected breast cancers. Third, these data support investigation of active surveillance for breast cancer, because they show that the diagnosis of an early-stage breast cancer does not necessarily mean that a woman’s survival has been compromised in comparison to the general population. Active surveillance is a cancer management approach that is accepted in other early-stage cancers with high disease-specific survival rates, and relative survival rates that exceed 100%, including certain prostate and thyroid cancers. Active surveillance is being investigated for DCIS with clinical trials in the United States (COMET), United Kingdom (LORIS), Europe (LORD), and Japan (LORETTA),16,17 and may also be worthy of study in early-stage hormone receptor-positive breast cancer.

Conclusion

Our observation that women with early-stage breast cancer have relative survival rates of over 100% illustrates how survival rates in general are poor indicators of the effectiveness of cancer screening. In particular, while the divergence between relative survival and disease-specific survival rates indicates the weaknesses of each of these statistics, this observation does help to identify factors such as healthy user, lead time and length time biases, that affect survival rates in screen-detected cancers. Our data suggest that the presence of healthy user bias and overdiagnosis are relevant to stage I invasive breast cancer. Investigation of active surveillance for low-risk, screening-detected invasive breast cancers should also be considered.

Acknowledgements and competing interests

This work was supported by the National Cancer Institute at the National Institutes of Health Cancer Center Support Grant (P30 CA008748), the National Institutes of Health (grant numbers T32 CA009685 to A.R.M, K08 DE024774 and R01 DE027738 to L.G.T.M), and the Frederick Adler Chair Fund (to L.G.T.M). The authors have declared no conflicts of interest related to this work. L.G.T.M. is an inventor on intellectual property owned by Memorial Sloan Kettering Cancer Center outside of this work.

Abbreviations:

DCIS

ductal carcinoma in situ

SEER

Surveillance, Epidemiology and End Results Prorgam

NCHS

National Center for Health Statistics

AJCC

American Joint Committee on Cancer

Footnotes

Conflicts of interest: all authors have no disclosures

CRediT author statement

Andrea Marcadis: Methodology, software, formal analysis, data curation, writing – original draft preparation, visualization. Luc Morris: Conceptualization, methodology, software, formal analysis, investigation, writing – review and editing, supervision, funding. Jennifer Marti: Conceptualization, methodology, software, formal analysis, investigation, writing – review and editing, resources, supervision, project administration.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Data availability

All data are publicly available through the SEER program, with details of data query structure needed to replicate analyses provided above in the Methods section.

References:

  • 1.Howlader N, Ries LAG, Mariotto AB, Reichman ME, Ruhl J, Cronin KA. Improved Estimates of Cancer-Specific Survival Rates From Population-Based Data. JNCI: Journal of the National Cancer Institute. 2010;102(20):1584–1598. doi: 10.1093/jnci/djq366 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Makkar N, Ostrom QT, Kruchko C, Barnholtz-Sloan JS. A comparison of relative survival and cause-specific survival methods to measure net survival in cancer populations. Cancer Med. 2018;7(9):4773–4780. doi: 10.1002/cam4.1706 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Henson DE, Ries LA. The relative survival rate. Cancer. 1995;76(10):1687–1688. [DOI] [PubMed] [Google Scholar]
  • 4.Shrank WH, Patrick AR, Alan Brookhart M. Healthy User and Related Biases in Observational Studies of Preventive Interventions: A Primer for Physicians. Journal of General Internal Medicine. 2011;26(5):546–550. doi: 10.1007/s11606-010-1609-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Davies L, Petitti DB, Martin L, Woo M, Lin JS. Defining, Estimating, and Communicating Overdiagnosis in Cancer Screening. Ann Intern Med. 2018;169(1):36–43. doi: 10.7326/M18-0694 [DOI] [PubMed] [Google Scholar]
  • 6.Welch HG, Kramer BS, Black WC. Epidemiologic Signatures in Cancer. N Engl J Med. 2019;381(14):1378–1386. doi: 10.1056/NEJMsr1905447 [DOI] [PubMed] [Google Scholar]
  • 7.Marcadis AR, Marti JL, Ehdaie B, Hakimi AA, Davies L, Morris LGT. Characterizing Relative and Disease-Specific Survival in Early-Stage Cancers. JAMA Intern Med. December 2019. doi: 10.1001/jamainternmed.2019.6120 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.History of ACS Recommendations for the Early Detection of Cancer in People Without Symptoms. https://www.cancer.org/health-care-professionals/american-cancer-society-prevention-early-detection-guidelines/overview/chronological-history-of-acs-recommendations.html. Published July 30, 2020. Accessed January 4, 2021.
  • 9.Table 33. Use of mammography among women aged 40 and over, by selected characteristics: United States, selected years 1987–2015. 2015. https://www.cdc.gov/nchs/fastats/mammography.htm. Accessed December 30, 2020.
  • 10.Qin J, White MC, Sabatino SA, Febo-Vázquez I. Mammography use among women aged 18–39 years in the United States. Breast Cancer Res Treat. 2018;168(3):687–693. doi: 10.1007/s10549-017-4625-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Sabatino SA, Coates RJ, Uhler RJ, Breen N, Tangka F, Shaw KM. Disparities in Mammography Use Among US Women Aged 40???64 Years, by Race, Ethnicity, Income, and Health Insurance Status, 1993 and 2005: Medical Care. 2008;46(7):692–700. doi: 10.1097/MLR.0b013e31817893b1 [DOI] [PubMed] [Google Scholar]
  • 12.Sabatino SA, Thompson TD, Richardson LC, Miller J. Health Insurance and Other Factors Associated With Mammography Surveillance Among Breast Cancer Survivors: Results From a National Survey. Medical Care. 2012;50(3):270–276. doi: 10.1097/MLR.0b013e318244d294 [DOI] [PubMed] [Google Scholar]
  • 13.Welch HG, Black WC. Overdiagnosis in cancer. J Natl Cancer Inst. 2010;102(9):605–613. doi: 10.1093/jnci/djq099 [DOI] [PubMed] [Google Scholar]
  • 14.Bleyer A, Welch HG. Effect of three decades of screening mammography on breast-cancer incidence. N Engl J Med. 2012;367(21):1998–2005. doi: 10.1056/NEJMoa1206809 [DOI] [PubMed] [Google Scholar]
  • 15.Marcadis AR, Davies L, Marti JL, Morris LGT. Racial Disparities in Cancer Presentation and Outcomes: The Contribution of Overdiagnosis. JNCI Cancer Spectrum. 2020;4(2):pkaa001. doi: 10.1093/jncics/pkaa001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Hwang ES, Hyslop T, Lynch T, et al. The COMET (Comparison of Operative versus Monitoring and Endocrine Therapy) trial: a phase III randomised controlled clinical trial for low-risk ductal carcinoma in situ (DCIS). BMJ Open. 2019;9(3):e026797. doi: 10.1136/bmjopen-2018-026797 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Kanbayashi C, Thompson AM, Hwang ESS, et al. The international collaboration of active surveillance trials for low-risk DCIS (LORIS, LORD, COMET, LORETTA). JCO. 2019;37(15_suppl):TPS603–TPS603. doi: 10.1200/JCO.2019.37.15_suppl.TPS603 [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

All data are publicly available through the SEER program, with details of data query structure needed to replicate analyses provided above in the Methods section.

RESOURCES