Andrey Nailor, M.D.·Published 11 June 2008·17 min read
Key points
Following are brief summaries of the major randomized trials of breast cancer screening.
Smith: Cancer Screening, Department of Cancer Control, American Cancer Society, Atlanta, Georgia C.
Each followed a somewhat different protocol, and the outcome in each has been influenced by a number of factors that have important implications for the interpretation of study results.
Health Insurance Plan of New York Randomized Clinical Trial The HIP study was initiated in December 1963.
Following are brief summaries of the major randomized trials of breast cancer screening. They represent study designs in which a large group of women was randomly assigned to a group that either would or would not receive an invitation to be screened for breast cancer. As can be seen, the first of these studies was initiated in the early 1960s, and for some, follow-up still continues.
Each followed a somewhat different protocol, and the outcome in each has been influenced by a number of factors that have important implications for the interpretation of study results. These factors include the study methodology, the clinical protocol, participation rates in the study group (compliance), screening rates in the control group (contamination), and the number of screening rounds before an invitation was extended to the control group. Other factors that likely influenced end results include the quality of the screening process, thresholds for diagnosis, and follow-up mechanisms for women with an abnormality. For these factors, we have few insights into individual trials.
It is also the case that, with the exception of the randomized trials conducted in Canada (NBSS-I and -II), which, as noted in the description of the Canadian trials below, have been controversial, none was specifically designed to evaluate the efficacy of breast cancer screening in age-specific subgroups (e.g., women aged 40 to 49). The availability of data in each of the trials has provided the opportunity for retrospective analysis of screening efficacy by age; however, limitations in statistical power and questions regarding the appropriateness of applying the same screening protocol to pre- and postmenopausal women have been the focus of debate over whether health policy could reliably be informed by the existing data.
Health Insurance Plan of New York Randomized Clinical Trial The HIP study was initiated in December 1963. It was the first randomized, controlled trial to evaluate the efficacy of breast cancer screening with CBE and mammography. Approximately 62,000 women aged 40 to 64 were randomly assigned to two groups: The study group would be offered annual CBE and two-view mammography (craniocaudal and mediolateral views) for 4 years, and the control group would receive usual care.
After 10 years from entry into the study, approximately 30% fewer breast cancer deaths occurred in the study group compared with the control group. At 18 years of follow-up, there were 23% fewer deaths. Mammography and CBE contributed independently to breast cancer detection, with 33% of cases detected by mammography alone, 45% with CBE alone, and 22% with combined modalities.
Swedish Two-County Study The largest randomized trial of breast cancer screening is the Swedish Two-County study, so-called because women were randomized into study and control groups within two counties in Sweden, Kopparberg and Ostergotland. The trial consisted of approximately 133,000 women aged 40 to 74 years, with approximately 77,000 women invited to the screening.
Written and revised by
Andrey Nailor, M.D.
Our doctors write and review articles for accuracy and keep them up to date.
health.am provides general information, not medical advice. Always consult a qualified health professional about your own health.
The screening intervals differed by age, with women aged 40 to 49 invited every 24 months and women aged 50 and older invited every 33 months. The screening examination included only single-view mammography (mediolateral oblique view) and did not include CBE. After 8, 11, and 14 years of follow-up, approximately 30% fewer breast cancer deaths occurred in the study group invited to the screening compared with the control group.{pagebreak}
Malmö In Malmö, Sweden, a randomized trial to evaluate the efficacy of breast cancer screening was initiated in 1976, designated the Malmö Mammographic Screening Trial I. Approximately 42,000 women aged 45 to 69 were randomized on the basis of birth cohort (women born between 1927 and 1932) into study and control groups. The study group was invited to receive two-view mammography (craniocaudal and oblique views) at 18- to 24-month intervals for five rounds. CBE was not part of the screening regimen.
No overall reduction in deaths was observed in the study group compared with the control group at 8 and 11 years of follow-up, although a nonsignificant 20% lower breast cancer death rate was observed for women aged 55 and older. Among the explanations that have been offered for these results is the finding that approximately 25% of the control group was screened with mammography during the study period, and a similar percentage were nonattenders in the study group, thus diluting the differences between the original allocation. A second cohort of women born between 1933 and 1945, designated Malmö Mammographic Screening Trial II, were invited to be screened between 1978 and 1990.
The most recent update reports only results of the combined cohorts for women younger than age 50 at the time of randomization. The investigators observed a statistically significant 36% reduction in breast cancer mortality among the group invited to the screening.
{embed="ref/stat-menu-br-cancer1"}
Stockholm The Stockholm randomized trial of breast cancer screening was initiated in 1981. Sixty thousand women aged 40 to 64 were randomized on the basis of birth date into a study group of approximately 40,000 women who would receive invitations to screening and a control group of 20,000 women. The study consisted of only two rounds of screening with single-view mammography (oblique view) without CBE at a 28-month interval and observed a nonsignificant 29% reduction in breast cancer mortality in the group invited to the screening.
Gothenburg The Gothenburg randomized trial of breast cancer screening was initiated in 1982. Approximately 52,000 women aged 39 to 59 years were randomized into groups to be invited and not to be invited to five screening rounds. The screening interval was 18 months, and usually women were screened with two-view mammography (craniocaudal and mediolateral oblique views), unless prior screening examinations indicated such low breast density that single-view mammography was justified (approximately 30% of mammograms). CBE was not part of the screening regimen. The investigators observed a 44% statistically significant reduction in breast cancer mortality among women aged 39 to 49 after 11 years of follow-up.
Edinburgh The Edinburgh trial was initiated in 1979 and was an evaluation of the efficacy of CBE and, initially, two-view mammography (depending on initial findings, single-view mammography was often performed in subsequent screening rounds). CBE was done annually and mammography every 2 years. Eighty-seven general practices representing approximately 45,000 women were randomized into study and control groups, which led to unanticipated confounding by socioeconomic status. Women aged 45 to 64 in the study group received invitations to four rounds of screening with mammography. At 7 years of follow-up, 17% fewer breast cancer deaths had occurred in the study group compared with the control group, a difference that was not statistically significant.{pagebreak}
Canada–National Breast Screening Study 1 The Canadian randomized trial of the efficacy of breast cancer screening in women aged 40 to 49 was initiated in 1980. It was designed to specifically test the efficacy of breast cancer screening in women in their 40s and consisted of four to five rounds of annual CBE and mammography, depending on time of entry into the study. After a physical examination that included CBE, women were invited to participate in the study, and volunteers were then randomized into a study group or control group.
Approximately 50,000 women participated in the study, with virtually equal numbers randomized to study and control groups. At 7 years of follow-up, approximately 36% more breast cancer deaths had occurred in the study group than in the control group. At 10.5 years of follow-up, the rate of excess mortality had declined to 14%. This study and its conclusions have been controversial for a number of reasons, including study design (all study participants received a high-quality CBE before randomization), questions about the randomization process, the quality of mammography, and the observation that there was a significant excess of patients with advanced tumors in the group invited to screening compared with the usual care group.
More specifically, Tarone has shown that the excess of patients in the National Breast Screening Study 1 (NBSS-1) who had four or more positive nodes in the screening arm is statistically significant and inconsistent with the stage distribution of disease in other trials, both in the first round and at the conclusion of the study. There is no known explanation for the excess of advanced tumors in the group invited to the screening, and a review of the randomization process did not uncover any irregularities. However, it is apparent that this yet unexplained excess of advanced cancers accounts for the higher breast cancer death rate in the screening arm.
Canada–National Breast Screening Study 2 The National Breast Screening Study 2 (NBSS-2) was also initiated in 1980 and was a trial of the efficacy of breast cancer screening in women aged 50 to 59. It was designed to test specifically the efficacy of breast cancer screening in this age group with mammography and CBE versus CBE alone and consisted of four rounds of annual examinations in the study and control groups.
{embed="ref/stat-menu-br-cancer1"}
As with the NBSS-1, after a physical examination that included CBE, women were invited to participate in the study. Volunteers were then randomized into a study group or control group. Approximately 39,000 women participated in the study, with virtually equal numbers randomized to study and control groups. At 7 years of follow-up, only 3% fewer deaths had occurred from breast cancer in the study group than in the control group. This particular study has received less attention than the study of women aged 40 to 49, but many of the same critiques apply.
Metaanalyses Before 1997 both the HIP study and the Swedish Two-County study provided clear and early evidence that mammography could reduce breast cancer mortality in women aged 50 and older, providing a sound basis for the promotion of screening for this age group. Evidence supporting the value of screening for women in their 40s was less clear. Although there was indirect evidence of a benefit both from the trials and observational studies, until recently no single trial had shown a statistically significant reduction in breast cancer deaths among women randomized to an invitation to screening in their 40s.
Whereas the absence of definitive evidence of a benefit was troubling to some, especially in light of early findings of an excess of breast cancer deaths in the NBSS-1, others were persuaded that the absence of a benefit was an artifact of methodologic shortcomings, in particular low statistical power in the trials for subgroup analysis.
To overcome the limits of small sample sizes, investigators began to conduct meta-analyses of trial data, combining age-specific results from the various studies to overcome the limits of small sample sizes. There are two important observations about these meta-analyses. First, as was the case with individual trials, benefits for women aged 50 and older appear early in the follow-up period, whereas they occur later for women aged 40 to 49. Second, with accumulating years of follow-up in the 40 to 49 group, the relative risk of mortality steadily improved. {pagebreak}
As shown in recent studies, with an average follow-up of 12.7 years, the meta-analysis using the most current data shows a relative risk of 0.82 (18% fewer breast cancer deaths) when all trials are combined, a relative risk of 0.74 (26% fewer deaths) when only population-based trials are combined (not shown), and a relative risk of 0.71 (29% fewer deaths) for all five Swedish randomized controlled trials. Each point estimate is statistically significant at the 95% confidence level, although the all-trial meta-analysis has the lowest relative risk of breast cancer mortality because of the excess rate of breast cancer deaths in the NBSS-1 in the group invited to screening. The meta-analysis of Swedish trials is included in this comparison because they represent a more homogeneous group and because they include the two second-generation trials (i.e., Gothenburg and Malmö), which applied more advanced screening protocols and observed 44% and 36% fewer breast cancer deaths in the invited groups compared with the control groups.
What is evident in these meta-analyses is that they show very similar mortality reductions among women younger and older than age 50 who were invited to the screening.
The one analysis that stands apart from the others is the all-trial meta-analysis of women in their 40s, in which the size of the NBSS-1 and the excess rate of deaths in the group invited to screening measurably reduce the estimated benefit. Because irregularities in the NBSS-1 have yet to be explained, the consistency of results in the other meta-analyses and the recent results from Gothenburg and Malmö suggest that it is reasonable to regard the potential benefit of screening in younger versus older women as more similar than different, especially if recommended screening intervals are tailored to the age-specific estimated sojourn time.
As mentioned at the beginning of this section, there are notable differences in the cumulative trial mortality trends for women younger and older than age 50, and reasons for this difference have been an ongoing source of debate - long on conjecture but short of a clear explanation. Some have argued that the observed delay in benefit was simply a function of small sample sizes and overall better survival in younger compared with older women.
{embed="ref/stat-menu-br-cancer1"}
By this argument, it doesn't matter when a mortality reduction appears, because screening is beneficial if it averts deaths from breast cancer in the near or long term. However, others have argued that the delayed benefit was a methodologic artifact of trial design, in which mortality reductions attributed to women in their 40s actually were due to diagnoses after age 50 among women randomized in their 40s. Because the analysis of age-specific mortality differences in trials is based on age at randomization (not age at diagnosis), at the time of analysis the group randomized in their 40s includes patients diagnosed after they have had a fiftieth birthday. According to this hypothesis, if a benefit from screening truly begins at age 50, then the observation of benefit for women in their 40s is more apparent than real, because it must be due to the benefit of diagnosing breast cancer after age 50 among women randomized in their 40s.
{embed="ref/stat-menu-br-cancer1"} Thus, a benefit later in the follow-up period would be due to the time required to accumulate the age migration cases. Likewise, for women randomized in their 50s, the benefit from screening was seen earlier in the follow-up period because they were already at an age when mammography is beneficial. In effect, this argument was a call for analysis of trial data by age at diagnosis.
Although it seems logical, analysis of trial data by age at diagnosis has a built-in bias, because the purpose of screening is to diagnosis breast cancer earlier in its natural history, which means, for any individual who develops breast cancer at a younger age than when symptoms would be expected to appear. Because age at diagnosis is influenced by the study intervention (i.e., mammography), it is regarded as a pseudovariable. By introducing lead-time bias, it complicates comparisons between the invited and not-invited group and is therefore methodologically inappropriate. The question of age-specific benefits is important, but to be properly addressed a trial would need to randomize a narrow age range of women (i.e., women aged 40 to 41) to an invited and not-invited group and follow them for the duration of interest (i.e., until age 50), thus avoiding "age-creep." Still, this question has been pursued.
In 1995, de Koning and colleagues published data from the Swedish Two-County trial arguing that age migration in the group randomized in their 40s could account for 70% of the observed benefit and thus explain the late onset of benefit in this age group. In an accompanying editorial, it was argued that this kind of analysis actually violates the underlying logic of trial data analysis as noted in the previous paragraph, but if it were to be pursued, analysis of actual trial data rather than modeling would be the logical approach. Tabar and colleagues subsequently evaluated Swedish Two-County data by age at randomization and age at diagnosis and observed no support for this hypothesis. In fact, the relative mortality reduction among women randomized in their 40s was greater among women diagnosed before their fiftieth birthday than in women diagnosed after age 50. No support for this hypothesis was observed in a similar analysis of Gothenburg trial data.
Recent analysis of Swedish Two-County data provides a clearer and more clinically intuitive explanation for the delay in benefit observed in the individual trials and meta-analyses based on an assessment of the effect of the trial screening protocols on mortality trends by tumor histology. As described in the section Sojourn Time and the Influence of Early Intervention, the mean breast cancer sojourn time (i.e., potential lead time) is shorter in women younger than age 50 than in women older than age 50. {pagebreak}
Because the majority of the world's trials screened women aged 40 to 49 at randomization at an interval of 24 months, faster tumor progression in women in this age group meant that they were less likely to benefit from mammography than would women aged 50 and older. The consequence of screening all women at the same 24-month interval becomes further evident when cumulative mortality trends are examined by age and tumor histology. When cumulative mortality trends are compared for ductal grade 3 tumors, a benefit begins to appear after 5 years among women aged 50 to 74 but is not evident in the follow-up period among women aged 40 to 49. These higher malignancy grade tumors are more aggressive, have a worse prognosis, and, compared with ductal grade 2, medullary, and invasive lobular cancers, account for a higher proportion of breast cancer deaths that occur sooner after diagnosis. When comparing trends in the invited versus control group in the trial for ductal, grade 2, lobular, and medullary carcinoma, women aged 40 to 49 and 50 to 74 showed similar trends in cumulative mortality.
Furthermore, for both women younger and older than age 50 who were diagnosed with these less aggressive cancers, the mortality benefit does not appear until after 7 to 8 years of follow-up, consistent with the observation that survival is better for these tumor types and that associated mortality occurs later. When examining the overall trend in cumulative mortality by age, these same trends are evident; in other words, for the groups invited to screening, a mortality benefit begins to emerge after approximately 5 years for women aged 50 to 74 and after 8 years for women aged 40 to 49.
{embed="ref/stat-menu-br-cancer1"}
The organizers of the Falun meeting concluded that the screening interval of 24 or more months was differentially effective in reducing mortality among women aged 40 to 49 versus women aged 50 and older. Whereas this interval was equally effective in both age groups for grade 2, medullary, and invasive lobular tumors, and effective in reducing deaths among grade 3 tumors diagnosed in women aged 50 and older, it was entirely ineffective for grade 3 tumors diagnosed in women aged 40 to 49. New results from Gothenburg, which screened women aged 39 to 49 at 18-month intervals, show that the timing of the benefit, which appears at 6 to 8 years, is similar to that observed for women aged 50 and older.
Additional insights into these observations come from earlier analysis of Swedish Two-County data comparing long-term survival by tumor size, extent of disease, and histologic grade. When tumors are grouped by size, grade, and nodal involvement, there is little difference in observed survival by age. However, Tabar and colleagues have also observed that the likelihood that a tumor will have progressed to grade 3 at the time of diagnosis is higher in women younger than age 50 than in women aged 50 and older, indicating that one important difference between breast cancer in premenopausal women compared with postmenopausal women is the greater likelihood of tumor dedifferentiation to less favorable histology.
As shown in earlier work, the influence of tumor grade on survival becomes increasingly pronounced once tumors are greater than 10 mm in size. As a general rule, it is important to diagnose tumors at the earliest opportunity, but these data suggest that this goal is especially important in premenopausal women.
The fundamental conclusion that can be drawn from these data is that the common screening interval of 2 or more years in the trials contributed to sufficiently favorable lead times for most tumor types in women aged 50 and older and failed to provide that same benefit to the more aggressive tumors in women younger than age 50.
More than any other explanation, a mismatch between the estimated mean sojourn time and the screening interval contributed to poorer performance in the trials among women aged 40 to 49 and also to the longer period of observation required to see a benefit. The observation that some eventual benefit accrued to women aged 40 to 49 who were invited to the screening is a function of the heterogeneity of breast cancer. Although the wide screening interval provided little benefit to women in their 40s in whom aggressive tumors developed, the detection of less aggressive tumors was still beneficial and appeared later in the follow-up period.
If the screening interval is greater than the mean sojourn time, the potential for the program to reduce the rate of advanced disease is compromised, because a higher proportion of cancers will progress undetected to the point at which they become clinically evident and appear as interval cancers. In Gothenburg, which screened women aged 40 to 49 every 18 months and showed a 44% reduction in breast cancer mortality, the proportional interval cancer incidence was only 18% in the first 12 months after a negative screen but increased to more than 50% in the period from 12 to 18 months.
In the University of California at San Francisco screening program, sensitivity declined at twice the rate in the interval between 1 and 2 years for women in their 40s compared with women aged 50 and older, reflecting faster growth rates for women in their 40s. According to Sickles, interval cancer rates for women aged 40 to 49 with annual screening are approximately equivalent to interval cancer rates in women aged 50 years and older who are screened every 2 years.
Robert A. Smith and Carl J. D'Orsi
R. A. Smith: Cancer Screening, Department of Cancer Control, American Cancer Society, Atlanta, Georgia C. J. D'Orsi: Diagnostic Radiology, University of Massachusetts Memorial Medical Center, Worchester, Massachusetts
News
News
News
