RT indicates radiation therapy.
Oeffinger KC, Ford JS, Moskowitz CS, Diller LR, Hudson MM, Chou JF, Smith SM, Mertens AC, Henderson TO, Friedman DL, Leisenring WM, Robison LL. Breast Cancer Surveillance Practices Among Women Previously Treated With Chest Radiation for a Childhood Cancer. JAMA. 2009;301(4):404–414. doi:10.1001/jama.2008.1039
Author Affiliations: Departments of Pediatrics (Drs Oeffinger and Ford), Medicine (Dr Oeffinger), Psychiatry and Behavioral Sciences (Dr Ford), and Epidemiology and Biostatistics (Dr Moskowitz and Mss Chou and Smith), Memorial Sloan-Kettering Cancer Center, New York, New York; Pediatric Oncology, Dana-Farber Cancer Institute, Boston, Massachusetts (Dr Diller); Oncology (Dr Hudson) and Epidemiology and Cancer Control (Drs Hudson and Robison), St Jude Children's Research Hospital, Memphis, Tennessee; Pediatrics, Emory University, Atlanta, Georgia (Dr Mertens); Pediatrics, University of Chicago, Chicago, Illinois (Dr Henderson); Pediatrics, Vanderbilt-Ingram Cancer Center, Nashville, Tennessee (Dr Friedman); and the Clinical Research and the Public Health Sciences Divisions of the Fred Hutchinson Cancer Research Center, Seattle, Washington (Dr Leisenring).
Context Women treated with chest radiation for a pediatric malignancy have a significantly increased risk of breast cancer at a young age and are recommended to have an annual screening mammogram starting at age 25 years or 8 years after radiation, whichever occurs last.
Objective To characterize the breast cancer surveillance practices among female pediatric cancer survivors who were treated with chest radiation and identify correlates of screening.
Design, Setting, and Participants Between June 2005 and August 2006, a 114-item questionnaire was administered to a random sample of 625 women aged 25 through 50 years who had survived pediatric cancer, who had been treated with chest radiation, and who were participating in the Childhood Cancer Survivor Study (CCSS), a North American cohort of long-term survivors diagnosed from 1970-1986. Comparisons were made with similarly aged pediatric cancer survivors not treated with chest radiation (n = 639) and the CCSS siblings cohort (n = 712).
Main Outcome Measure Screening mammogram within the previous 2 years.
Results Of 1976 cancer survivors and siblings who were contacted, 87.9% participated. Among the 551 women with a history of chest radiation, 55% reported a screening mammogram in the past 2 years (ages 25-39 years, 36.5%; 95% confidence interval [CI], 31.0%-42.0%; ages 40-50 years, 76.5%; 95% CI, 71.3%-81.7%). In comparison, 40.5% of survivors without chest radiation and 37.0% of CCSS siblings reported a screening mammogram in the same time interval. Notably, among women with a history of chest radiation, 47.3% (95% CI; 41.6%-53.0%) of those younger than 40 years had never had a mammogram and only 52.6% (95% CI; 46.4%-58.8%) of women aged 40 through 50 years were being regularly screened (2 mammograms within 4 years). Screening rates were higher among women who reported a physician recommendation than those who did not (ages 25-39 years, 76.0% vs 17.6%; ages 40-50 years, 87.3% vs 58.3%). In multivariate models, the association was particularly strong for younger women (ages 25-39 years, prevalence ratio [PR], 3.0; 95% CI, 2.0-4.0; ages 40-50 years, PR, 1.3; 95% CI, 1.1-1.6).
Conclusions In this cohort of women who had childhood cancer treated with chest radiation, 63.5% of those aged 25 through 39 years and 23.5% of those aged 40 through 50 years had not had mammography screening for breast cancer within the previous 2 years despite a guideline recommendation that survivors of childhood cancer who were treated with chest radiation should undergo annual screening mammography.
Women treated with chest radiation for a pediatric malignancy face a significantly increased risk of breast cancer at a young age.1- 6 The risk of breast cancer begins to increase as early as 8 years after radiation and the median age of breast cancer diagnosis ranges from 32 to 35 years.1,2,4 Risk of breast cancer is greatest among women who were treated for Hodgkin lymphoma with high-dose mantle radiation.2 By age 45 years, it is estimated that from 12% to 20% of women treated with moderate- to high-dose chest radiation will be diagnosed with breast cancer.1,2,4 For perspective, among women with a BRCA gene mutation, the estimated cumulative incidence of breast cancer ranges from 1% to 5% at age 30 years and from 10% to 19% at age 40 years.7- 11 As in the general population, breast cancer outcomes among childhood cancer survivors are strongly associated with stage at diagnosis.12- 14 Notably, treatment options for these women are often limited due to previous chest radiation and possible exposure to anthracycline chemotherapy.
For the past decade, experts have recommended that women in this risk group initiate breast cancer surveillance with annual screening mammography at age 25 years.1- 6,13- 19 In 2003, the Children's Oncology Group (COG) developed the Long-Term Follow-up Guidelines for Survivors of Childhood, Adolescent, and Young Adult Cancer.20,21 At that time, the COG recommended annual screening mammography for women exposed to moderate- to high-dose chest radiation (≥20 Gy), starting at 25 years of age or 8 years after radiation, whichever occurs last. As recently as 2008, the COG updated the guidelines to include an adjunct breast magnetic resonance imaging (MRI) with annual screening mammography. These recommendations are consistent with those of the American Cancer Society and the United Kingdom Department of Health.22,23
To date, there is limited published information describing the breast cancer surveillance practices of women who were treated with chest radiation for a pediatric malignancy.24- 28 The aims of this study were to determine the prevalence of screening mammography and other methods of breast cancer surveillance and to identify predictors of screening in a large and geographically diverse population of young adult women who were treated with chest radiation for a childhood cancer. Breast cancer surveillance practices were compared with 2 groups: female childhood cancer survivors not exposed to chest radiation and siblings of childhood cancer survivors.
The methodology of the Childhood Cancer Survivor Study (CCSS) and a description of the study participants have been previously published.29,30 Briefly, the cohort consists of survivors of specific childhood cancers (leukemia, brain tumors, Hodgkin lymphoma, non-Hodgkin lymphoma, renal tumors, neuroblastoma, soft tissue sarcomas, or bone tumors) who were diagnosed before 21 years of age at 1 of 26 participating centers between 1970 and 1986 and who were alive at least 5 years from their original diagnosis, and a comparison cohort of their siblings. The current analysis is based upon 9307 survivors and 2951 siblings who participated in the 2003 CCSS Follow-up Survey (Figure 1). The study was approved by the institutional review board at each participating institution and written informed consent was obtained from each participant.
From the cohort, a random sample of 625 eligible female survivors who were aged 25 through 50 years and were treated with at least 20 Gy chest radiation therapy (RT) were contacted for this study. The upper age limit of 50 years reflected the oldest women in the cohort with this exposure.
We were interested in comparing the breast cancer surveillance practices of survivors exposed to chest RT with the breast cancer surveillance practices of 2 groups of women matched by age but without this exposure. To assess the influence of having childhood cancer on screening practices, the first comparison group consisted of female survivors who were not treated with chest radiation. The second comparison group represented a noncancer population and consisted of women who were siblings of survivors in the full CCSS cohort.
A 114-item cross-sectional survey was administered by mail and telephone interview between June 6, 2005, and August 24, 2006. The survey can be downloaded from www.stjude.org/ccss.
The primary outcome was a screening mammogram within the previous 2 years. Breast cancer surveillance practices, including screening and diagnostic mammograms, were characterized using 25 questions adapted from the National Health Interview Survey 2000 Cancer Control Module (http://www.cdc.gov/nchs/nhis.htm).
The following independent variables were assessed: race and ethnicity, self-described living area (rural, urban, suburban), having a primary care physician or usual source of care, having a written cancer treatment summary, last Papanicolaou (Pap) smear, general preventive health beliefs, survivor-related health beliefs and psychological factors, breast cancer knowledge, and breast cancer risk perception. Self-reported race and ethnicity was included because mammography screening rates in the general population vary among racial and ethnic groups.31,32
General preventive health beliefs were measured by 4 validated Health Belief Model items with 5-point Likert-style responses. The first 2 items assessed participants' concern with their general health and interest in routine medical check-ups. The second 2 items focused on cancer-related concerns: susceptibility to serious health problems related to cancer therapy and importance of a routine check-up to look for these problems. Each pair of items was averaged to give 2 composite measures (general health and cancer health), ranging from 1 (not at all concerned) to 5 (extremely concerned).33- 35
Participants were considered to have correct breast cancer knowledge if they responded “true” to the statement: “Women who were treated with radiation to the chest or breast area for childhood cancer are more likely to get breast cancer.” Participants who responded “false” or “not sure” were considered to have incorrect knowledge. Perceived risk of breast cancer was assessed using a previously reported 5-point Likert-style item asking participants to estimate their chance of developing breast cancer compared with other women.36,37 Scores ranged from 1 to 5, with high scores indicating higher perceived risk.
Lastly, the survey included the 13-item pros and cons of mammography, a validated instrument developed by Rakowski and colleagues38- 40 that elicits positive opinions or facilitators and negative opinions or barriers of screening mammography. Participants rated their agreement with 6 pro and 7 con items using a 5-point Likert scale; thus, high pros or cons indicated strong agreement with positive or negative perceptions of screening, respectively.
In this particular survey, we did not ask about health insurance or education. However, from the previous 2003 CCSS Follow-Up Survey administered to the entire survivor and sibling cohorts, 85.8% of eligible women in our study had health insurance and 80.5% were high school graduates with some college or vocational training.
In the general population, women aged 40 years or older are recommended to have a screening mammogram every 1 to 2 years.41 For this reason, we analyzed data separately for 2 age groups: 25 through 39 years and 40 through 50 years. Characteristics of chest RT women and the 2 comparison groups were assessed and the frequency of mammography, clinical breast examination, breast self-examination, and Pap smear testing were determined. To explore potential differences between nonparticipants and participants, the Fisher exact test was used.
To evaluate differences in screening mammography between women who received chest RT and the 2 comparison groups, Poisson regression with robust variance estimates was used to directly estimate prevalence ratios (PRs).42,43 Akin to the relative risk, the PR is defined as the ratio of the probability of a screening mammogram in the past 2 years among participants with a particular characteristic relative to the probability of a screening mammogram in the past 2 years among participants without this characteristic. Poisson regression models were used to directly model the PR because the odds ratio from a logistic regression model would not be a good approximation of the PR in this setting with a fairly common outcome. Analyses were adjusted for race/ethnicity and age at study in 5-year increments.
Poisson regression was also used to evaluate associations among each independent variable and having a screening mammogram within the past 2 years. Variables that were potentially associated with the outcome (P < .10) at the univariate level were assessed by multivariate model. In all analyses, PR and 95% confidence intervals (CIs) are reported.
Prior to the study, sample size calculations were carried out to ensure at least 90% power to detect PR of reasonable magnitude for each planned comparison and for a range of assumed outcome prevalences. Corresponding to the reference group prevalences that were subsequently observed in the analyses, we had anticipated sufficient power to detect PR of 1.7 or higher for any proposed comparison.
In this descriptive study, the 2 comparison groups were used primarily to provide a background rate of screening mammography for the 2 age categories. However, the same univariate and multivariate assessment of factors associated with screening mammography was applied to both comparison groups for women aged 40 through 50 years (available at www.stjude.org/ccss).
All statistical analyses were performed with SAS version 9.1 (SAS Institute Inc, Cary, North Carolina), using 2-sided statistical inferences and a significance level of P ≤ .05.
Of 625 women in the chest RT group who were contacted, 551 participated in the study (response rate, 88.2%; Figure 1). Table 1 reports selected characteristics of study participants. Participants and nonparticipants did not differ by age at time of study, age at childhood cancer diagnosis, interval from cancer to time of study, history of Hodgkin lymphoma, or exposure to anthracyclines or alkylating agents. Nonparticipants were more likely to be racial/ethnic minorities than participants (17.8% vs 7.7%; P = .007).
Of 639 women in the no chest RT group who were contacted, 561 participated (response rate 87.8%). Participants and nonparticipants in this comparison group did not differ by age at time of study, age at childhood cancer diagnosis, interval from cancer to time of study, or exposure to anthracyclines or alkylating agents. Nonparticipants were more likely to be racial/ethnic minorities than participants (19.4% vs 6.8%; P < .001). Of 712 women in the CCSS siblings group who were contacted, 622 participated (response rate, 87.4%). As with the other groups, nonparticipants were more likely to be racial/ethnic minorities than participants (13.6% vs 5.4%; P = .008).
Table 2 reports the breast cancer screening practices of women in the study. Among women aged 25 through 39 years with chest RT, only 36.5% (95% CI, 31.0%-42.0%) reported a screening mammogram within the past 2 years. Although the percentage of screening mammograms among women with chest RT was much lower than expected, the rate was still higher than among women in the general population (CCSS siblings, 10.6%; 95% CI, 7.3%-13.9%) and women who had a childhood cancer but were not at increased risk of breast cancer (no chest RT, 15.5%; 95% CI, 11.4%-19.6%). However, 47.3% (95% CI, 41.6%-53.0%) of women in the target population aged 25 through 39 years, had never had a mammogram and only 23.3% (95% CI, 18.5%-28.1%) had a screening or diagnostic mammogram within the previous year.
Women aged 40 through 50 years with chest RT were more likely to report mammography than their younger counterparts, with 76.5% (71.3%-81.7%) reporting a screening mammogram within the past 2 years compared with 70.0% (95% CI, 64.4%-75.6%) for the group without chest RT (P = .10) and 67.0% (95% CI, 61.6%-72.4%) for the CCSS sibling group (P = .02). Importantly, only 52.6% (95% CI, 46.4%-58.8%) of women aged 40 through 50 with chest RT engaged in regular screening (at least 2 mammograms within 4 years). This was not significantly higher than for women with no chest RT (48.8%; 95% CI, 42.5%-55.1%; P = .39) and only modestly higher than the CCSS siblings (41.5%; 95% CI, 39.5%-43.5%; P < .01). In all groups, women who were older were more likely to have been screened in the prior 2 years or to have received regular screening (Figure 2).
In Table 3, the percentage of women who reported a screening mammogram within the past 2 years, by key characteristics, is provided for the 3 study groups. Table 4 provides univariate and multivariate associations between characteristics of women in the chest RT group, ages 25 through 39, and the likelihood of reporting a screening mammogram within the preceding 2 years. As illustrated in Figure 2, age was an important predictor of screening mammography. For each 5-year incremental increase in age, the likelihood of reporting a mammogram increased nearly 2-fold (PR, 1.8; 95% CI, 1.5-2.2). However, the strongest predictor of mammography in women aged 25 through 39 years was having a physician recommend the test. The likelihood of reporting a mammogram was 3.0 (95% CI, 2.0-4.0) times higher among women who reported a physician recommendation than women who did not report a physician recommendation. Perception of breast cancer risk also predicted screening mammography (PR, 1.3; 95% CI, 1.1-1.5). Lastly, women who considered that the pros of mammography outweighed the cons (positive decisional balance) were more likely to report a screening mammogram (PR, 1.1; 95% CI, 1.0-1.2). The 2 most commonly mentioned barriers to screening mammography among women aged 25 through 39 years who did not have a mammogram in the previous 2 years were “doctor didn't order it” (31%) and “I’m too young” (30%).
Table 5 provides univariate and multivariate associations between characteristics of women in the chest RT group, ages 40 through 50 years, and the likelihood of reporting a screening mammogram. In the final multivariate model, the significant predictors of reporting at least 1 screening mammogram in the previous 2 years (vs none), in addition to older age, were having a primary care physician (PR, 1.5; 95% CI, 1.1-2.3), physician recommendation (PR, 1.3; 95% CI, 1.1-1.6), awareness of increased risk of breast cancer associated with chest radiation (PR, 1.2; 95% CI, 1.1-1.4), increased general health concerns (PR, 1.2; 95% CI, 1.1-1.3), and positive decisional balance in pros and cons of mammography (PR, 1.1; 95% CI, 1.0-1.1). The 2 most important barriers ranked by women in this age group who did not have a mammogram in the previous 2 years were “put it off” or “didn't get around to it” (27%) and “too expensive” or “no insurance/cost” (17%).
We estimate that in the United States, there are approximately 20 000 to 25 000 women who are 25 years or older and were treated for a pediatric malignancy with moderate- to high-dose chest radiation.44,45 Worldwide, about 18% to 20% of female adult survivors of childhood cancer have been exposed to chest radiation. In the past 15 years, numerous studies have reported on the risk of breast cancer among young women treated with chest radiation for a pediatric malignancy.1- 6,12- 19,24,25 Based on these studies, breast cancer screening with annual mammography has been recommended, starting around the age of 25 years.1- 6,13- 19 Importantly, most women in this risk group are not followed up at a cancer center, and they and their clinician may not be aware of this risk or the screening recommendations.46
This is the first large study that we are aware of to provide a detailed assessment of the breast cancer surveillance practices of young women, ages 25 through 50 years, in this high-risk population. There were several notable findings. Nearly half of the women younger than 40 years have never had a mammogram. Encouragingly, the likelihood of initiating screening increased with age; more than 75% of women 40 years or older reported a screening mammogram within the previous 2 years. For perspective, data from the 2006 Behavioral Risk Factor Surveillance System survey indicate that 69% of US women between the ages of 40 and 49 years reported having a mammogram within the past 2 years.47 However, only half were established in a regular pattern of screening (at least 2 screening mammograms within the past 4 years). Although they have a significantly higher risk of breast cancer than women in the general population, the cancer survivors in this study had screening rates in this age period (40-50 years) only minimally higher than women in the 2 comparison groups.
Breast cancer screening, beginning at age 30 years, is recommended for women with a familial risk of breast cancer.23 In the 2000 National Health Interview Survey, there were 480 women aged 30 through 49 years with a first-degree relative with either breast or ovarian cancer. Wu and colleagues48 reported that 23.4% of women aged 30 through 39 years and 55.7% of women aged 40 through 49 years reported a mammogram (screening or diagnostic) within the past year. Similarly, among 551 women in our study, 23.3% of those aged 25 through 39 years and 53.6% of those aged 40 through 50 years reported a mammogram (screening or diagnostic) within the past year. Although the mammography rates for women in our study were similar to this other high-risk population, both were markedly lower than recommended.
There was not a lack of medical contact among our study participants; 92% reported a clinical breast examination within the previous 2 years and almost 90% had a recent Pap smear. Despite this high rate of clinical contact focused on women's health issues, only half of participants reported that a health care professional had recommended a mammogram in the past year. This was particularly evident among younger women, aged 25 through 39 years: 72% reported a clinical breast examination within the past year, but only 33% reported a physician recommendation for a mammogram. Women who received a recommendation were 3 times as likely to have a mammogram. The importance of physician recommendation has been similarly reported for women with a familial risk of breast cancer49 and women 40 years or older in the general population.31 Although a variety of factors undoubtedly contribute to a physician recommendation for mammography, 1 of the primary barriers is likely a lack of clinician familiarity with childhood cancer survivors and their risk of breast cancer. This is supported by our finding that women 40 years or older were more likely to report a physician recommendation than women 25 through 39 years. Thus, studies evaluating the effectiveness of clinician-based interventions are warranted.
This study also provides key insights for developing targeted interventions for women at risk of breast cancer following chest radiation. Having a positive view of screening mammography, reflected by the pros of mammography outweighing the cons, was associated with an increased likelihood of screening, regardless of age. Other facilitators and barriers to screening varied by age group and should be considered when developing targeted screening interventions for this population. One issue that may be raised when designing an intervention aimed at increasing the rate of breast cancer surveillance among this population is the potential harm associated with further radiation exposure with mammography. Among younger women with a BRCA gene mutation, it has been suggested that the increased risk of radiation-induced breast cancers may outweigh the benefit of mammography.50 However, this is less likely to be an issue among women treated with moderate- to high-dose therapeutic chest radiation. The estimated dose of radiation with a standard 2-view screening mammogram is about 3.85 mGy.50- 52 In other words, for a woman who has been treated with 3500 cGy mantle irradiation, a single 2-view screening mammogram increases the radiation exposure from 3500 cGy to 3500.385 cGy. Assuming an annual 2-view screening mammogram from age 25 to age 69 years (44 years), the cumulative radiation exposure would be 16.94 cGy, thus increasing the radiation exposure in this woman from 3500 cGy to 3516.94 cGy (an increase of <1% of the net radiation exposure). Although it is not known whether this small incremental increase in radiation exposure could significantly affect risk of breast cancer in someone who has already had moderate- to high-dose therapeutic irradiation, international expert panels and cooperative groups continue to recommend annual screening mammograms with breast magnetic resonance imaging in this population.1- 6,13- 23
When interpreting our study findings, it is important to consider several limitations. History of mammography was provided by self-report without validation of medical records. However, mammography self-report correlates well with confirmed radiology reports53 and has been used in other studies, including the National Health Interview Survey.48,49,54 Because the study cohort is not population-based, selection bias should be considered when assessing whether the findings are generalizable. That being said, the CCSS is the largest cohort of childhood cancer survivors and previous findings regarding morbidity and mortality are similar to those in smaller population-based studies from Europe.4,55- 58 The study sample was predominantly white, non-Hispanic; thus, the findings may underestimate or overestimate the breast cancer surveillance practices of ethnic or minority groups. Lastly, the participants in this study, including women in our target population and the 2 comparison groups, have participated in a longitudinal cohort for more than 10 years and have been receiving regular newsletters highlighting ways to maintain or improve health, including appropriate cancer screening. It is likely that women in this cohort, as evidenced by their high adherence to general women's health recommendations, represent a highly motivated and educated group. Thus, our estimates of breast cancer screening rates, although much lower than recommended, may overestimate the rates among the many women in this risk group who are not participants in the CCSS, highlighting low screening rates nationally for pediatric cancer survivors.
In summary, our study suggests that most young women at risk of breast cancer following chest radiation for a pediatric cancer, including women at highest risk (Hodgkin lymphoma survivors), are not being appropriately screened. Findings from this study should provide the foundation for targeted interventions involving both clinicians and cancer survivors.
Corresponding Author: Kevin C. Oeffinger, MD, Department of Pediatrics, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY 10021(firstname.lastname@example.org).
Author Contributions: Dr Oeffinger had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Oeffinger, Robison, Diller, Hudson, Mertens, Leisenring.
Acquisition of data: Leisenring, Robison, Oeffinger.
Statistical analysis: Moskowitz, Leisenring, Chou.
Analysis and interpretation of data: Moskowitz, Leisenring, Oeffinger, Ford, Robison, Diller, Hudson, Mertens, Henderson, Friedman, Chou, Smith.
Drafting of the manuscript: Oeffinger.
Critical revision of the manuscript for important intellectual content: Oeffinger, Robison, Moskowitz, Leisenring, Ford, Diller, Hudson, Mertens, Henderson, Friedman, Chou, Smith.
Obtained funding: Oeffinger, Robison.
Study supervision: Oeffinger, Robison.
Financial Disclosures: None reported.
Funding/Support: This work was supported by grants R21-CA-106972 (Dr Oeffinger, principal investigator) from the National Cancer Institute, Department of Health and Human Services and the Centers for Disease Control and Prevention; U24-CA-55727 (Dr Robison, principal investigator) from the Department of Health and Human Services; funding to the University of Minnesota from the Children's Cancer Research Fund; and funding to St Jude Children's Research Hospital from the American Lebanese Syrian Associated Charities.
Role of the Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
The CCSS is a collaborative, multi-institutional project, funded as a resource by the National Cancer Institute. The cohort was assembled through the efforts of 26 participating clinical research centers in the United States and Canada. The study is currently funded by a U24 resource grant awarded to St Jude Children's Research Hospital. For information on contributing centers and how to access and use the CCSS resource, visit http://www.stjude.org/ccss.