Radiation Therapy for Breast Cancer and Increased Risk for Esophageal Carcinoma

Methods

For this study, we used data from the Surveillance, Epidemiology and End Results (SEER) Program of the National Cancer Institute [13]. Using a retrospective cohort design, we separately calculated the age-adjusted (every 5 years) and period-adjusted (every 3 calendar years) standardized incidence ratio [14] as an estimate of relative risk for primary esophageal carcinomas after a first case of primary breast cancer (International Classification of Diseases for Oncology [ICD-O] Topography codes C50.0 to C50.9) for women who received and did not receive radiation therapy for breast cancer. Details on calculating age-adjusted and period-adjusted standardized incidence ratios were described in our previous studies [15, 16]. We separately examined the risk for squamous-cell carcinoma (ICD-O Topography codes C15.0 to C15.9 and Morphology codes 8050 to 8081) and adenocarcinoma (ICD-O Topography codes C15.0 to C15.9 and Morphology codes 8140 to 8381) of the esophagus. Because squamous-cell carcinoma occurs in the cervical and thoracic esophagus (the regions more likely to be exposed to radiation therapy given for breast cancer), we expected that the risk for esophageal squamous-cell carcinomas would be more strongly related to radiation therapy than would the risk for adenocarcinoma of the esophagus, which occurs in the lower third of the esophagus (the region less likely to be exposed to radiation).

Because the induction period for radiation-induced carcinogenesis is considered to be at least several years, we estimated relative risks for different durations of follow-up to examine the carcinogenic effect of radiation therapy for breast cancer. Our selection of cut-points for the duration of follow-up (3 months to <5 years, 5 to <10 years, and ≥ 10 years) was based on the time intervals between radiation therapy and the diagnoses of subsequent esophageal cancer that have been reported previously [2, 3]. Although a few cases of radiation-induced esophageal cancer were reported to have occurred before 10 years of follow-up, most of the reported cases occurred after 10 or more years of follow-up. We excluded women whose esophageal carcinoma was diagnosed within 3 months of the date of diagnosis of breast cancer or whose date of or age at diagnosis, death, or last follow-up visit were unknown.

Results

During the 21-year study period (1973 to 1993), 116 primary esophageal carcinomas occurred among 220 806 women who had first had primary breast cancer (1 216 853 person-years of follow-up). On the basis of incidence rates in the general population, 75.5 primary esophageal carcinomas were expected among these women, giving rise to an overall relative risk of 1.54 (95% CI, 1.27 to 1.84). The relative risks for individual types of cancer with respect to radiation therapy for the first primary cancer are shown in the (Table 1). The pattern of occurrence of esophageal squamous-cell carcinoma after breast cancer is consistent with the radiation-induced carcinogenesis model because the relative risk increased from 1.37 (CI, 0.63 to 2.60) with a follow-up period of 3 months to less than 5 years to 5.42 (CI, 2.33 to 10.68) with a follow-up period of 10 years or more (Table 1). This increase in risk with increasing duration of follow-up was not evident in the group that did not have radiation therapy. A less definitive increasing trend was evident for esophageal adenocarcinoma; the relative risk (based on 2 observed and 0.47 expected cases) after 10 years was 4.22 (CI, 0.47 to 15.25).

When the analysis was restricted to the women whose breast cancer was located in the inner half of the breast (ICD-O Topography codes C50.2, C50.3, and C50.8), the relative risk (based on 4 observed and 0.53 expected cases) of esophageal squamous-cell carcinoma after 10 or more years of follow-up increased to 7.59 (CI, 2.04 to 19.43). We could not perform this subgroup analysis for esophageal adenocarcinoma or subgroup analyses for both histologic types by stage of breast cancer because of small numbers.

Discussion

Evidence from previous case reports raised the possibility of a relation between radiation therapy and esophageal cancer [2-7]. Goffman and colleagues [2] reviewed eight cases of esophageal cancer that occurred a median of 8.5 years (range, 3 to 16 years) after radiation therapy for breast cancer. In a recent report, Fekete and colleagues [3] reported three cases of esophageal cancer in women that occurred 8 to 11 years after radiation therapy for breast cancer. Almost all of these reported cases of radiation-induced esophageal cancer were of the squamous-cell type.

Our controlled study is the first to show that radiation therapy for breast cancer in women may increase the risk for subsequent development of esophageal squamous-cell carcinoma 10 or more years after radiation therapy. The risk for esophageal adenocarcinoma after breast cancer also seemed to be elevated, although the result was based on small numbers [2 observed and 0.47 expected cases] and the 95% CI for the effect estimate after 10 years did not exclude the null value. Two findings are consistent with the theory that radiation therapy has a carcinogenic effect on organs within the radiation field: 1) the increased risk for esophageal carcinoma among women with breast cancer who received radiation therapy but not among women who did not receive this therapy and 2) the accentuation of this risk for cancer of the inner half of the breast.

Our results seem convincing, but our study has several limitations. First, our results are based on small numbers. Second, we could not examine the effect of radiotherapy dose on esophageal carcinomas because no information on dose was available.

Third, the increased risk for esophageal carcinoma could have resulted from a differential distribution of smoking or other strong confounding factors other than radiation therapy among patients with breast cancer who did and did not have radiation therapy (for example, a higher proportion of persons who smoked or drank alcohol in the radiation therapy group). We could not assess this bias because of the absence of information on smoking and alcohol intake, but it is unlikely that a woman's smoking or drinking history would be related to the use of radiation therapy.

Fourth, radiation-induced esophagitis is not uncommon [18], and women with symptoms of this condition may be more likely to have undergone upper-gastrointestinal diagnostic procedures (such as upper gastrointestinal series or endoscopy), which would lead to an increased detection of esophageal carcinomas that otherwise would not have been diagnosed. This bias is unlikely to have affected our results after 10 years because radiation-induced esophagitis is usually an early phenomenon (occurring within weeks or months of radiation therapy) [18] and is less likely to affect risk after several years. In fact, diagnostic bias could explain the excess number of cases of primary esophageal adenocarcinoma seen between 3 months and less than 5 years after breast cancer, regardless of radiotherapy.

Fifth, we could not examine the effect of chemotherapy on esophageal cancer or the influence of chemotherapy on the observed effect of radiation therapy because of lack of information on chemotherapy. Although chemotherapy has been shown to be related, either singly or in combination with radiation therapy, to the subsequent occurrence of second primary cancer in other settings [19, 20], no data are available on its effect on esophageal cancer.

Sixth, we used data on women in whom breast cancer was diagnosed between 1973 and 1993, a period during which considerable changes in the use of radiation therapy for breast cancer occurred [21]. Although radiation to the internal mammary nodes (the field more likely to include the esophagus) was common in the 1970s and early 1980s, modern techniques are less likely to include the internal mammary nodes (and thus the esophagus) in the radiation field [22, 23]. Given these changes in radiation techniques, the dose of radiation to the esophagus is probably lower for women currently receiving radiation therapy for breast cancer than it was for those who received radiation therapy in the past. Thus, our findings apply more to women who received radiation in the past. Unfortunately, because of the small number of second primary cancers and a limited duration of available follow-up for the cases of recently diagnosed breast cancer, we could not examine changes in the risk estimates over time.

In conclusion, our findings suggest that radiation therapy for breast cancer may increase the risk for developing esophageal carcinoma several years after radiation therapy. Future studies are needed to determine whether other risk factors for esophageal carcinomas (such as cigarette smoking, alcohol consumption, and body mass index) interact with radiation therapy in the development of esophageal carcinoma after breast cancer. Because survival among women with breast cancer is gradually increasing, we believe that clinicians should be aware of the possibility of a second primary cancer in the esophagus should suspicious symptoms arise among long-term survivors, particularly those who received radiation therapy for breast cancer in the past and who smoke cigarettes or drink alcohol.

Dr. Neugut: Division of Oncology, Department of Medicine, Columbia University College of Physicians and Surgeons, 630 West 168th Street, New York, NY 10032.