Diagnostic Chest X-Rays and Breast Cancer Risk before Age 50 Years for BRCA1 and BRCA2 Mutation Carriers

There is extensive epidemiologic evidence that exposure to moderate- and high-dose ionizing radiation, particularly at young ages, is associated with an increase in a woman's risk of developing breast cancer (1, 2). In contrast, relatively few studies have evaluated the association between low-dose diagnostic chest X-rays and breast cancer risk. Positive associations have been reported for patients with tuberculosis monitored with frequent fluoroscopy (3, 4), patients with scoliosis evaluated with repeated X-rays of the spine (5), and women with frequent diagnostic chest X-rays before the age of 20 years (6, 7) or 30 years (8), or diagnostic chest X-rays administered before 1960 (8), when radiation doses were likely to be higher than current levels.

There is evidence from some (5–7), but not all (8), studies that breast cancer risk associated with diagnostic chest X-rays is greater for women with a family history of breast cancer. Positive associations with diagnostic chest radiation have also been reported for women carrying a mutation in BRCA1 or BRCA2 (9–12), suggesting that genetically predisposed women could be sensitive to low-dose diagnostic radiation exposure. For mammography, a common source of low-dose radiation, 3 studies found no significant associations with breast cancer risk for BRCA1 and BRCA2 mutation carriers (12–14).

As BRCA1 and BRCA2 are involved in the detection and cellular repair of DNA double-strand breaks induced by ionizing radiation (15), it is plausible that BRCA1 and BRCA2 mutation carriers may have reduced ability to repair radiation-induced DNA damage and, therefore, be at increased risk of radiation-induced breast cancer. Women who receive therapeutic radiotherapy for their first breast cancer have an elevated risk for contralateral breast cancer (16) which could be even further increased for those who carry a germline mutation in a gene involved in DNA-damage repair (e.g., ATM, CHEK2, BRCA1, BRCA2; refs. 17–19).

To further study the possible association of low-dose radiation from medical diagnostic procedures, we examined breast cancer risk and diagnostic chest X-rays using self-reported data on exposure from 454 BRCA1 and 273 BRCA2 mutation carriers identified by 3 large family registries of breast cancer in the United States, Canada, and Australia. Because the established association between radiation and breast cancer risk is observed primarily for women exposed to radiation at a young age, we analyzed only mutation carriers ages younger than 50 years.

A detailed description of the study design and analytic methods has been provided elsewhere (20). Briefly, women carrying a deleterious mutation in BRCA1 or BRCA2 (see below) were identified from the Breast Cancer Family Registry (BCFR) composed of 6 registries in the United States, Canada, and Australia (21); the Ontario Cancer Genetics Network (OCGN) in Canada; and the Kathleen Cuningham Foundation Consortium for Research into Familial Breast Cancer (kConFab) in Australia and New Zealand (22). Three registries of the BCFR ascertained families through probands with incident breast cancer who were sampled from population-based cancer registries. All other families were identified through cancer clinics or community outreach efforts. Probands were either affected or unaffected with breast or ovarian cancer and included those with multiple or early-onset cases of breast or ovarian cancer among first- and second-degree relatives or relatives known to carry a mutation in BRCA1 or BRCA2.

In this case–control analysis of mutation carriers, we defined cases as women carrying a deleterious mutation in BRCA1 or BRCA2 who had been diagnosed with a first primary invasive breast cancer and controls as mutation carriers without a history of breast cancer. We assigned each carrier a reference age, defined for cases as the age at breast cancer diagnosis, and for controls as the age at the earliest of the following events: questionnaire completion, bilateral mastectomy, bilateral oophorectomy, or diagnosis of in situ breast cancer. We restricted the analysis to carriers with a reference age <50 years. To minimize survival bias, we further restricted the analysis to carriers who completed the questionnaire no more than 5 years after their reference age. Approximately 82% of carriers completed the questionnaire within 3 years of their reference age. We excluded from the analysis a small number of BRCA1 mutation carriers (2 cases and 1 control) and BRCA2 mutation carriers (3 cases and 2 controls) who reported therapeutic radiation exposures to the chest before breast cancer diagnosis (cases) or interview (controls). The final analysis was based on 454 non-Hispanic white BRCA1 mutation carriers (167 cases and 287 controls) and 273 non-Hispanic white BRCA2 mutation carriers (104 cases and 169 controls).

All participants in the BCFR, OCGN, and kConFab provided written informed consent to participate in research programs that were approved by the Institutional Review Board of each center participating in the BCFR, OCGN, and kConFab. This study was approved by the Institutional Review Board of the Stanford University School of Medicine (Stanford, CA).

All study sites used the same structured questionnaire, completed by interview or mail. Information was collected on established and suspected breast cancer risk factors, including chest X-ray examinations (other than mammograms) and radiation treatment to the chest area before breast cancer diagnosis for cases or interview for controls. The questionnaire asked about treatment with radiation for selected conditions (i.e., tuberculosis, cancer, acne, mastitis, enlarged thymus, or hemangioma), and other conditions in the chest area, as well as diagnostic X-ray examinations for heart catheterization or scoliosis, or other intensive X-ray examinations of the chest area. When a participant reported “therapeutic chest X-rays for other conditions” or “other intensive X-ray examinations of the chest area,” information was requested on the type of condition. The term “other intensive chest X-rays” was not further defined in the questionnaire. Although the questionnaire listed tuberculosis as a condition for treatment with radiation, we recognize that fluoroscopy was used to follow the progress of lung inflation to normal after induced pneumothorax treatment for tuberculosis; radiation was not used as a treatment for tuberculosis per se. For both therapeutic and diagnostic radiation exposure, information was collected on age at first exposure and total number of exposures. As we previously found positive associations between breast cancer risk and chest fluoroscopy for tuberculosis and chest X-rays for pneumonia (8), we examined associations with these 2 exposure categories separately. All other diagnostic chest X-rays (e.g., back or shoulder examinations, upper body bone fractures, routine chest X-rays, and chest X-rays for unspecified indications) were pooled together for the analysis. X-rays from mammograms were not considered, as the questionnaire did not assess a lifetime mammography history.

BRCA1 and BRCA2 mutation testing was conducted for incident breast cancer cases recruited from the 3 population-based BCFR registries and the youngest breast cancer case for families from clinics. Women of Ashkenazi Jewish ancestry were tested for the 3 founder mutations. Testing was conducted either by complete gene sequencing at Myriad Genetics, Inc. or by one of several other methods: enzymatic mutation detection, 2-dimensional gene scanning, chemical cleavage of mismatch, the protein truncation test, single-strand conformation polymorphism analysis, heteroduplex analysis, and denaturing high-conductance liquid chromatography (DHPLC; refs. 23–25). Mutations considered included small insertions and deletions, nonsense mutations, and splice site variants that all resulted in truncated and dysfunctional BRCA1 and BRCA2 proteins. All mutations were confirmed by sequencing. The analysis for this study was restricted to carriers of mutations that were classified as deleterious by applying the criteria used by the Breast Information Core and Myriad Genetics Inc., as described elsewhere (24). When a deleterious mutation was identified, all relatives of the carrier who had provided a DNA sample were subsequently tested for the same mutation.

We used unconditional logistic regression implemented in SAS version 9.1 to estimate ORs and 95% confidence intervals (CI). A robust variance estimator was used to account for possible correlation in covariates between family members (26). This approach accommodates the possibility that some carriers were ascertained because of their breast cancer status. Valid inferences for association parameters can be obtained by ignoring other ascertainment issues provided that the exposure–disease OR for ascertained carriers is equal to the corresponding OR for unascertained carriers having the same family history (26). This condition would be violated only if ascertainment modified the association between exposure and breast cancer. All regression models included reference age as a categorical (<40 and 40–49 years) variable and as a continuous variable, dummy variables for country/region (Canada, Australia, Utah, other U.S. sites), family history of breast or ovarian cancer in first-degree relatives (yes, no), and number of full-term pregnancies (0, 1–2, ≥3). On the basis of our previous finding of increased breast cancer risk associated with chest fluoroscopy for tuberculosis and chest X-rays for pneumonia (8), we estimated breast cancer risk in relation to 3 categories of diagnostic X-ray exposure: chest fluoroscopy for tuberculosis, chest X-rays for pneumonia, and chest X-rays for other reasons. Women in these exposure categories were compared with a single referent category that included women without a history of therapeutic and diagnostic chest X-rays.

Characteristics of cases and controls are shown in Table 1. Cases were older than controls and fewer cases than controls had a college or university degree. More controls than cases were ascertained from clinic-based sources and had a family history of breast or ovarian cancer which is the result of the ascertainment and BRCA1 and BRCA2 mutation testing protocols (20).

For BRCA1 mutation carriers, a history of diagnostic chest X-rays was not associated with breast cancer risk (OR, 1.31; 95% CI, 0.75–2.30; Table 2). There was no evidence of increased risk associated with diagnostic chest X-rays (OR, 1.16; 95% CI, 0.64–2.11), excluding those for pneumonia or chest fluoroscopy. Elevated, but statistically not significant, ORs were associated with a history of chest fluoroscopy for tuberculosis (OR, 2.58; 95% CI, 0.22–30) and chest X-rays for pneumonia (OR, 4.54; 95% CI, 0.53–30). For other chest X-rays (excluding those for pneumonia or chest fluoroscopy), risk did not vary by age or year at first chest X-ray or total number of chest X-rays.

Similarly, for BRCA2 mutation carriers, there was no evidence of increased breast cancer risk associated with a history of diagnostic chest X-rays (OR, 1.13; 95% CI, 0.59–2.18; Table 2). Subgroup analyses showed an elevated risk only for carriers with 3 to 5 chest X-rays (OR, 10.6; 95% CI, 1.93–58), excluding those for pneumonia or chest fluoroscopy.

The results from the present study suggest that for BRCA1 and BRCA2 mutation carriers ages younger than 50 years, there is no association between diagnostic chest X-rays and breast cancer risk, neither for exposure at any age nor for exposure before age 20 years when radiosensitivity is highest. For BRCA2 mutation carriers, our finding of a large OR estimate associated with 3 to 5 diagnostic X-rays, bounded by a wide CI, is likely due to chance, given the lack of association with ≥6 X-rays.

Few prior studies have examined diagnostic X-ray exposure and breast cancer risk for BRCA1 and BRCA2 mutation carriers and, in contrast to our findings, strong associations have been reported by some of the European studies (refs. 9, 11, 12; Table 3). There was, however, some overlap in mutation carriers between the 3 European studies (9, 11, 12); the analysis by Andrieu and colleagues (9) included 319 French mutation carriers from the study by Lecarpentier and colleagues (11), and 21% of mutation carriers included in the analysis by Pijpe and colleagues (12) were also included in the study by Andrieu and colleagues (9). In the incident cohort analysis by Andrieu and colleagues (9) of 969 mostly European BRCA1 and BRCA2 mutation carriers, a history of diagnostic chest X-rays (excluding mammograms) was associated with increased breast cancer risk (HR, 1.75; 95% CI, 1.1–2.8). For women ages 40 years or younger, risk was increased more than 2-fold (HR, 2.75; 95% CI, 1.4–5.3). Similarly, in the French study of 990 BRCA1 and BRCA2 mutation carriers, breast cancer risk associated with a history of diagnostic chest X-rays (excluding mammograms) was increased both for BRCA1 (HR, 4.12; 95% CI, 1.82–9.35) and BRCA2 (HR, 5.43; 95% CI, 1.36–21.7) mutation carriers (11). The largest study to date by Pijpe and colleagues (12) assessed associations with several categories of diagnostic radiation exposure, including fluoroscopy, conventional radiography (X-rays) of chest or shoulders, mammography, computed tomography of chest or shoulders, and other diagnostic procedures (e.g., bone scans). Consistent with our findings, there were no statistically significant associations with a history of diagnostic chest X-rays (HR, 1.38; 95% CI, 0.87–2.20) or chest X-rays before the age of 20 years (HR, 1.29; 95% CI, 0.84–1.98) for BRCA1 and BRCA2 mutation carriers combined. In that study, significant associations with diagnostic radiation exposure emerged only when all types of exposures (fluoroscopy, X-rays, mammography, other diagnostic procedures) before the age of 30 years were considered jointly (HR, 1.90; 95% CI, 1.20–3.00). Finally, a Polish study of patients with breast cancer diagnosed at ages 50 years or younger compared 138 affected BRCA1 mutation carriers with 158 affected non-carriers; affected carriers were more likely to report a history of chest X-rays before the age of 30 years than affected non-carriers (OR, 1.8; 95% CI, 1.2–2.9; ref. 10).

Our findings of a relatively large OR estimate associated with chest fluoroscopy for tuberculosis is in agreement with the study by Pijpe and colleagues (12) that found a 2-fold increased risk for exposure before the age of 20 years, but in both studies, the associations were based on small numbers of exposed mutation carriers and were not statistically significant. Chest fluoroscopy for tuberculosis has also been associated with increased breast cancer risk (3, 8) and mortality (4) in the general population. It is uncertain whether in the other European studies (9–11) reports of chest X-rays included fluoroscopy for tuberculosis. The inclusion of fluoroscopy could have contributed to the positive associations with chest X-rays, particularly for exposures before the age of 20 years. In our study, we asked about diagnostic X-ray exposure for heart catheterization and for the monitoring of scoliosis, which has been associated with increased breast cancer risk (5) and mortality (27) in the U.S. Scoliosis Cohort Study. For other intensive X-rays, we requested the reason for the diagnostic procedure. Although our questionnaire did not ask specifically about chest fluoroscopy, some women reported such exposure under “other intensive diagnostic X-rays” or in the questionnaire section on therapeutic radiation.

The most striking difference between the European studies and ours is the prevalence of self-reported diagnostic X-rays (Table 3). For unaffected carriers, the proportion of women who reported a history of diagnostic chest X-rays ranged from 15% (not including those for pneumonia or chest fluoroscopy for tuberculosis) in our study to 45% (not including fluoroscopy) in the study by Pijpe and colleagues (12), 50% in the study by Andrieu and colleagues (9), and 76% in the study by Lecarpentier and colleagues (11). Differences in the prevalence of reported chest X-rays were particularly pronounced for exposures before the age of 20 years. For affected carriers, proportions ranged from 7% in our study to 23% in the Polish study (10), 31% in the study by Pijpe and colleagues (12), 66% in the study by Andrieu and colleagues (9), and 88% in the study by Lecarpentier and colleagues (11). Large differences were also seen in the prevalence of fluoroscopy; a positive history was reported by 2% of affected mutation carriers in our study, compared with 15% in the European study by Andrieu and colleagues (12).

Differences in the frequency of self-reported diagnostic chest radiation could be due to differences between studies in the wording of questions and/or the inclusion of specific diagnostic procedures, differences between countries in the practice of administering diagnostic X-rays or differences between studies in the age distribution, or birth cohorts, of the mutation carriers. The frequency of medical X-rays has been shown to vary between countries (28). Chest fluoroscopy, for example, has been reported to be more common in Sweden than in the United States (29), and in the Netherlands, mass population screening for tuberculosis was conducted in young people from 1940 to 1960 (12). Analytic differences between the European studies may have also contributed to differences in results. As commented by Pijpe and colleagues (12), the analysis by Andrieu and colleagues (9) assessed lifetime X-rays, including exposure after diagnosis of breast cancer, potentially biasing associations away from the null.

As the frequency and spectrum of mutations in BRCA1 and BRCA2 vary by race/ethnicity (30, 31), we limited our analysis to non-Hispanic white mutation carriers. Breast cancer risk in mutation carriers, however, appears to vary by mutation position (32, 33). It is not known whether radiation sensitivity varies by mutation type and position. Lecarpentier and colleagues (11) found no variation in breast cancer risk associated with chest X-rays by mutation position in BRCA1 or BRCA2. Larger studies will be needed to assess whether the differences in associations between our study, and the European studies are due to differences in mutation types and radiosensitivity.

Several study limitations need to be considered when interpreting our results. As in previous studies (9–12), our exposure assessment relied on self-report. We, therefore, cannot rule out the possibility of inaccurate recall, especially of diagnostic procedures in the distant past. In the few studies conducted in the United States (29, 34) and Sweden (29, 35) that compared self-reported diagnostic X-rays and other diagnostic imaging against medical records, reporting errors tended to be nondifferential by disease status (29, 34, 36), although in one of the Swedish studies, there was evidence of differential recall bias for those ages younger than 50 years (35). A study from the Netherlands (36) that assessed the reliability of self-reported chest X-rays among BRCA1 and BRCA2 mutation carriers reported that the agreement between baseline and follow-up reports was nondifferential by disease status. Nondifferential exposure misclassification would bias risk estimates toward the null and might have reduced the possibility of finding a true association in our study. It is also possible that we failed to find statistically significant associations with diagnostic chest X-rays, especially for chest fluoroscopy and X-rays for pneumonia, due to insufficient power stemming from limited sample sizes. Our risk estimates have wide confidence intervals and are, therefore, imprecise. Future studies of BRCA1 and BRCA2 mutation carriers would benefit from improved exposure assessment, including queries about all types of diagnostic imaging procedures in the chest area, including mammograms which we did not assess in the current study.

Strengths of this study include use of the same questionnaire by the 3 family registry studies to assess therapeutic and diagnostic chest X-rays, which allowed us to identify unexposed carriers as the referent, and the focus on younger mutation carriers who need better guidelines for diagnostic procedures involving ionizing radiation at a young age. To minimize potential survival bias, we limited the affected mutation carriers to those who completed the questionnaire within 5 years of diagnosis. Our results remained unchanged when we further restricted the affected mutation carriers to those who completed the questionnaire within 3 years of diagnosis (data not shown).

Future studies should include larger numbers of BRCA1 and BRCA2 mutation carriers, and obtain detailed information on diagnostic imaging of the chest, including type and number of procedures, age at first exposure, and indications that required diagnostic procedures to the chest. Ideally, such reports would be verified by medical records to minimize any recall bias. However, given that ionizing radiation exposure at a young age (e.g., before the age of 20 years) is the most relevant window of susceptibility for radiation-induced breast cancer, retrospective verification may be very difficult, if not impossible. Study designs that include prospective breast cancer cases in a mutation carrier cohort providing baseline information on low-dose radiation exposure before diagnosis would avoid differential recall bias by case status.

Our findings do not support a positive association between diagnostic chest X-rays and breast cancer risk before the age of 50 years for BRCA1 and BRCA2 mutation carriers. Nevertheless, continued surveillance of the possible effects of diagnostic radiation exposure on breast cancer risk is warranted for these women, given the widespread and increasing use of diagnostic imaging involving ionizing radiation. Special attention might need to be paid to computed tomography, an imaging technology that is being used with increasing frequency both for children and adults and delivers much larger doses of ionizing radiation than conventional chest X-rays (37–39).

No potential conflicts of interest were disclosed.

The content of this manuscript does not necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating centers in the BCFR, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government or the BCFR.

Conception and design: E.M. John, D.C. Thomas, R. Haile, R.L. Milne, D.W. West, M. Daly, J.L. Hopper, G.G. Giles, A.S. Whittemore

Development of methodology: E.M. John, V. McGuire, D.C. Thomas, D.W. West, A. Miron, J.L. Hopper, G.G. Giles, A.S. Whittemore

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E.M. John, R. Haile, D.W. West, J.A. Knight, M.B. Terry, M. Daly, S.S. Buys, I.L. Andrulis, J.L. Hopper, M.C. Southey, C. Apicella, H. Thorne

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E.M. John, D.C. Thomas, H. Ozcelik, R.L. Milne, A. Felberg, D.W. West, A. Miron, J.A. Knight, M.B. Terry, J.L. Hopper, M.C. Southey, A.S. Whittemore

Writing, review and/or revision of the manuscript: E.M. John, V. McGuire, D.C. Thomas, R. Haile, H. Ozcelik, R.L. Milne, D.W. West, J.A. Knight, M.B. Terry, M. Daly, S.S. Buys, I.L. Andrulis, J.L. Hopper, M.C. Southey, G.G. Giles, C. Apicella, H. Thorne, A.S. Whittemore

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): E.M. John, V. McGuire, H. Ozcelik, A. Felberg, J.L. Hopper, M.C. Southey, C. Apicella, H. Thorne, A.S. Whittemore

Study supervision: E.M. John, D.W. West, J.L. Hopper

The authors thank Eveline Niedermayr, all the kConFab research nurses and staff, and the heads and staff of the Family Cancer Clinics for their contributions to this resource. The authors also thank the many families who contribute to the BCFR, kConFab, and OCGN.

This study was supported by the National Cancer Institute grant RO1 CA097359 (to A.S. Whittemore). The Breast Cancer Family Registry (BCFR) is supported by National Cancer Institute grant UM1 CA164920 (to I.L. Andrulis, S.S. Buys, M. Daly, J.L. Hopper, E.M. John, M.B. Terry). The Australian site of the BCFR was also supported by the National Health and Medical Research Council of Australia, the New South Wales Cancer Council, the Victorian Health Promotion Foundation, and the Victorian Breast Cancer Research Consortium. kConFab is supported by grants from the National Breast Cancer Foundation, the National Health and Medical Research Council (NHMRC) and by the Queensland Cancer Fund, the Cancer Councils of New South Wales, Victoria, Tasmania and South Australia, and the Cancer Foundation of Western Australia (to J.F. Sambrook). The Clinical Follow-Up Study is funded by NHMRC grants 145684, 288704 and 454508 (to J.F. Sambrook).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.