Ionizing Radiation and Tobacco Use Increases the Risk of a Subsequent Lung Carcinoma in Women With Breast Cancer: Case-Only Design
http://www.100md.com
《临床肿瘤学》
the Department of Medical Epidemiology and Biostatistics, Karolinska Institutet
Department of Hospital Physics, Radiumhemmet, and Department of Medicine, Clinical Epidemiology Unit, Karolinska University Hospital
Department of Medical Radiation Physics, Stockholm University, Stockholm, Sweden
Lombardi Cancer Center, Georgetown University Medical Center, Washington, DC.
ABSTRACT
PURPOSE: To analyze the risk of lung cancer in women treated with radiotherapy for breast cancer. We accessed the lung dose in relation to different radiotherapy techniques, provided the excess relative risk (ERR) estimate for radiation-associated lung cancer, and evaluated the influence of tobacco use.
PATIENTS AND METHODS: The Swedish Cancer Registry was used to identify 182 women diagnosed with breast and subsequent lung cancers in Stockholm County during 1958 to 2000. Radiotherapy was administered to 116 patients. Radiation dose was estimated from the original treatment charts, and information on smoking history was searched for in case records and among relatives. The risk of lung cancer was assessed in a case-only approach, where each woman contributed a pair of lungs.
RESULTS: The average mean lung dose to the ipsilateral lung was 17.2 Gy (range, 7.1 to 32.0 Gy). A significantly increased relative risk (RR) of a subsequent ipsilateral lung cancer was observed at 10 years of follow-up (RR = 2.04; 95% CI, 1.24 to 3.36). Squamous cell carcinoma (RR = 4.00; 95% CI, 1.50 to 10.66) was the histopathologic subgroup most closely related to ionizing radiation. The effect of radiotherapy was restricted to smokers only (RR = 3.08; 95% CI, 1.61 to 5.91). The ERR/Gy for women with latency 10 years after exposure was 0.11 (95% CI, 0.02 to 0.44).
CONCLUSION: Radiotherapy for breast cancer significantly increases the risk of lung carcinoma more than 10 years after exposure in women who smoked at time of breast cancer.
INTRODUCTION
The incidence of breast cancer is increasing throughout the industrialized world,1 although at the same time, the mortality is decreasing.2 With recent advances in early diagnosis and treatment, breast cancer is becoming an increasingly survivable disease. Women with breast cancer normally receive postsurgical adjuvant therapy, as radiotherapy, chemotherapy, or hormonal therapy or as a combination of any of these modalities. Adjuvant radiation therapy reduces the risk of local recurrence,3 and its use is increasing as more women today choose partial mastectomies as their surgical choice. However, one of the growing concerns is the chronic or late-occurring complications to the normal tissue from treatment of primary malignancies, such as therapy-related secondary cancers.
Women with breast carcinoma are at increased risk of subsequent primary malignancies, specifically lung carcinoma.4-6 During the last 20 years, the absolute number of lung cancers in women previously diagnosed with breast cancer has shown a steady increase, including in Sweden (Fig 1). 7 The reasons for this are probably the increasing number of long-term surviving breast cancer patients and the fact that the birth cohorts with the highest smoking prevalence (born between 1940 and 1959)8 are now moving into an age when lung cancer more commonly occurs. The increased use of radiotherapy during the last 30 years is probably also a contributing factor. Newer radiotherapy modalities using intensity-modulated radiation therapy, while seeking to reduce acute toxicity, might actually be increasing secondary cancer risk.9,10 This methodology, which uses multiple beams to deliver high doses to the tumor bed, aims to decrease doses to surrounding normal tissue with an attendant reduction in acute radiation complications. However, some clinicians worry that the secondary cancer risk persists or is worse because a greater volume of tissue is radiated at doses still considered to be potentially carcinogenic.9,11-13 Thus, it is important to identify the magnitude of radiation-related lung cancer risk in long-term breast cancer survivors. In addition to radiation, tobacco is a known carcinogen. Smoking is the major risk factor of lung cancer, and a multiplicative effect of smoking on radiation-related risk is suspected.14-16
The aim of our study was to analyze the risk of lung cancer in relation to radiotherapy for breast cancer. We assessed how the lung dose, estimated through a retrospective quantification of the dose distribution, varies with breast cancer radiotherapy techniques and provided excess relative risk (ERR) estimates in terms of radiation dose delivered to the lung. In addition, we evaluated the influence of tobacco use.
PATIENTS AND METHODS
Using the Swedish Cancer Registry, 191 women diagnosed with breast and subsequent lung cancers were identified in Stockholm County during the period from 1958 to 2000. Nine patients were excluded because they were diagnosed with lung cancer within 12 months of the initial breast cancer. It is not likely that these lung cancers could be causally related to radiotherapy. In all, 116 women had received radiotherapy, whereas 66 women were treated by other means (Table 1).
Radiotherapy was centralized to three departments of oncology in Stockholm. For each patient, case records and radiotherapy charts were identified and reviewed. Radiotherapy charts were abstracted for detailed information about physical and geometrical irradiation parameters, such as radiation quality, treatment technique, total dose, fractionation schedule, and treatment field size.
Information about smoking history was identified in case records from departments of oncology, surgery, and thoracic medicine. For patients lacking information (approximately 29%), lifetime smoking habits were retrieved by asking spouse or next of kin for information through a mailed questionnaire. The patient was considered a smoker if the case record or relative indicated that the woman was an occasional smoker or smoked one cigarette or more daily. We did not have the possibility to stratify smoking habits further than smoker and nonsmoker. Among 182 women, the smoking information was available for 71% from medical records and for 22% from next of kin. The smoking status was unknown for 7% of women. A total of 114 women were identified as smokers.
The mean age at diagnosis of breast cancer among the 182 women included in the study was 58 years (median, 57 years; first quartile, 48 years; and third quartile, 65 years). The main characteristics of the study group are listed in Table 1. Fifty-one percent of irradiated women were treated before 1970 compared with nonirradiated women, who were mostly diagnosed between 1970 and 1980. Approximately 40% of the patients were treated with cobalt-60 gamma-radiation (60Co). Thirty-nine (34%) of the 116 radiotherapy-treated women had squamous cell carcinoma, and 33 (28%) had adenocarcinoma. In the nonirradiated group, squamous cell carcinoma occurred in 13 women (20%), and adenocarcinoma occurred in 19 women (29%). Radiotherapy charts were available for 103 of 116 patients. For three patients with missing radiotherapy charts, detailed information about their radiotherapy was found in their case records.
The criteria chosen for organizing the radiotherapy information were the target definition and, consequently, the used treatment techniques. In most clinical charts, the exact definition of target was not given. Target was defined in general terms such as breast parenchyma, chest wall, and lymph nodes of the internal mammary chain (IMC), supraclavicular region, and axillary region.
We identified nine different treatment groups (Table 2). Group 1 consisted of patients irradiated to the IMC. The treatment was administered with overlapping rectangular 60Co fields using a short-distance gamma beam unit. The treatment technique is described in detail elsewhere.17 In group 2, patients were treated with orthovoltage x-ray beams. Ten of the patients were treated with two adjacent anterior-posterior fields to the fossa-supraclaviculary lymph nodes and to the chest wall. Additionally, two patients were treated with one field covering the IMC and one field covering the fossa-axillary lymph nodes. Treatments were administered with a source–skin distance of 60 cm, 0.5 mm copper filtration. Group 3 consisted of patients treated with electron beams covering the fossa-supraclaviculary lymph nodes. Patients in group 4 received radiotherapy to the breast parenchyma/chest wall and the IMC lymph nodes; they were treated with extended tangential photon fields. A shield for the apical part of the lung was often used. Group 5 consisted of patients in whom the IMC, fossa, and axilla were treated with an anterior-posterior 60Co beam. Patients in group 6 were treated like patients in group 5, but they also received an oblique 60Co beam to the thorax region. In group 7, the chest wall and the lower part of the IMC were treated with an oblique electron field. The lymph nodes in the lower part of the IMC, fossa, supraclavicular, and axillary regions were treated with an anterior photon field, whereas a posterior photon field was added to the axillary region. In group 8, the chest wall and the caudal part of the IMC lymph nodes were covered by an anterior electron field. An anterior photon field was used for the cranial part of the IMC, axillary, fossa, and supraclavicular lymph nodes.18 Finally, patients in group 9 received radiotherapy to the remaining breast parenchyma after partial mastectomy and were treated with tangential 60Co beam or with x-ray photon fields.
Radiation treatments were generally administered in 5 fractions per week for 3 to 6 weeks. The highest dose per fraction was administered to patients in group 1, and the lowest dose per fraction was administered in group 9. The treatment techniques of groups 4, 7, and 9 have previously been described in detail.19
Lung Dose Calculation
For most of the patients, original individual three-dimensional dose distributions in the lung were not available because patients were treated before the introduction of three-dimensional treatment planning systems (3D TPS). The reconstruction of treatment techniques on 3D TPS has already been used in studies both on radiation-induced secondary cancer5 and on dose-volume complication relationships.19-21 On the basis of the irradiation parameters given in the radiotherapy charts, we were able to quantify the ipsilateral and contralateral absorbed lung dose for all groups except group 2, where only ipsilateral lung dose was estimated.
For patients treated with short-distance 60Co (group 1), the dose was estimated using published isodose distributions.22 The dose to the ipsilateral lung was calculated on two axial computed tomography slices in a patient operated with radical mastectomy, one at the level of the mamillary plane and one located approximately 11 cm below the jugular region. The dose was calculated on a grid in the two slices. There is uncertainty in the estimate of the contralateral lung dose for this group because only a rough estimate of the leakage and scatter in the collimators used in the short distance treatment techniques could be made, and this contribution to the secondary dose may be considerable.23 For patients in group 2, the mean dose to the ipsilateral lung was calculated from the original depth-dose tables and isodose curves specific for the machines used. To correct for the lower density of the lung tissue, a lung density of 0.25 g/cm3 was used. For this group, the uncertainty in the estimation of the contralateral lung dose was regarded as too large to be of interest.
The treatment technique for group 3 was reconstructed on one model patient, and the treatment techniques for groups 5 and 6 were reconstructed on each patient in the group. Techniques for groups 4, 7, 8, and 9 were reconstructed on a computed tomography–based 3D TPS on 10 model patients.
Statistical Methods
The risk of lung cancer was assessed in a case-only approach, where each woman contributed a pair of lungs that can be considered matched for genetic and environmental factors (primarily smoking). The lung on the breast cancer side was considered as exposed, and the contralateral lung was considered as unexposed. The concordance rates with 95% CIs were calculated as the proportion of lung cancer cases on the same (concordant) side as the breast cancer. The relative risks (RR) and 95% CIs were estimated as in twin-study design, by conditional logistic regression.24 Heterogeneity of the radiation effect by latency, smoking, and histopathology were tested by likelihood ratio tests.
The ERR per Gy was calculated for the patient group with latency 10 years after exposure. The dose-response modeling was performed with the PECAN module of the software package EPICURE, applying an additive relative risk model. The results are presented as ERR/Gy together with likelihood-based CIs. Potential deviance from linearity was investigated by fitting a linear-quadratic function in dose. The individual doses to the ipsilateral and contralateral lung were used. Because the absorbed lung doses to the contralateral lung for treatment group 2 could not be reconstructed, patients from this group were excluded from the analysis.
RESULTS
The mean dose to the ipsilateral lung was 17.2 Gy (range, 7.1 to 32.0 Gy; Table 2), depending on the treatment group. In the first part of the study period (1960s), most of our patients received a mean dose to the lung of approximately 15 Gy, whereas in the 1970s, the mean dose increased to approximately 25 Gy (Fig 2). In the latter part of the study period (1990s), the mean dose decreased to approximately 8 Gy.
We compared latency and risk for lung cancer on the breast cancer side (concordant side) with lung cancer on the contralateral side (discordant side). By using the fact that the two organs are paired, these data could be used to assess the lung cancer risk in a similar fashion as in a disease-discordant twin pair study. Among 116 lung cancer patients who received radiotherapy, there were four women with bilateral tumors in the lung and six women in whom the laterality of lung cancer was not specified, leaving 106 women to be included in the statistical analyses.
The mean latency period between breast and lung cancer in women not treated with radiation therapy was 9.7 years, with no difference between the concordant and discordant side. In patients treated with radiotherapy, however, lung cancers diagnosed on the concordant side showed a mean latency of 17.6 years compared with 13.0 years for discordant lung cancers.
In patients treated with radiotherapy, the risk of developing lung cancer on the same side as the breast cancer compared with the contralateral side was significantly increased after more than 10 years from the breast cancer diagnosis (RR = 2.04; 95% CI, 1.24 to 3.36; Table 3). Within 10 years from breast cancer diagnosis, radiotherapy did not seem to influence the risk of lung cancer. In addition, absence of radiotherapy was not associated with an increase in risk on the concordant side of breast and lung regardless of latency period (Table 3). When time between breast and lung cancer in irradiated women was studied in detail (quantiles), an increased risk was present for concordant lung cancer, with an RR of 2.38 (95% CI, 1.04 to 5.43) in the interval of 15 to 23 years after exposure and of 2.25 (95% CI, 1.01 to 5.17) in the interval of 23 years after exposure (Table 3). In the analysis of the latency between breast cancer and lung cancer, evidence for heterogeneity was found (P = .005).
In calculation of RR for women treated with radiotherapy, the contralateral lung was considered as unexposed and serves as the control for the ipsilateral lung. Therefore, to validate our results, we calculated the RR when the contralateral dose to the lung was known and low ( 15% of the ipsilateral dose to the lung; Table 2). The average dose to the lung was 18.8 Gy when treatment technique groups 1, 2, 5, and 8 were excluded. The RR was significantly increased for women with a latency time of 10 years after exposure (RR = 2.09; 95% CI, 1.02 to 4.29); for a latency time of less than 10 years, the RR was 1.09 (95% CI, 0.48 to 2.47). These results are comparable to the RR for the whole radiation-treated group.
Further analysis of the irradiated study group with a latency time of 10 years after breast cancer treatment demonstrated that the increased risk was most evident for squamous cell carcinoma of the lung (RR = 4.00; 95% CI, 1.50 to 10.66; Table 4). The significant association was not seen in any other histopathologic subtype of lung cancer, including adenocarcinoma. Separate analysis by smoking showed that the increased risk of lung cancer was present for smokers (RR = 3.17; 95% CI, 1.66 to 6.06), but there was no observed increase in risk for nonsmokers (Table 4). In these analyses, evidence of heterogeneity was found for smoking (P = .026), whereas heterogeneity for histopathology did not reach statistical significance (P = .07).
The ERR/Gy was calculated for the irradiated patient group with a latency of 10 years after breast cancer treatment (Table 5). Because further analysis of this group revealed a heterogeneity by smoking, separate estimates of ERR/Gy were presented for smokers and nonsmokers. In the analysis, the individual doses to the ipsilateral and contralateral lung were used, and therefore, patients from the groups with treatment techniques with a known dose to the contralateral lung were selected (group 2 was excluded). For group 1, an approximate dose to the contralateral lung was estimated to be between 1.5 and 2.9 Gy; therefore, a mean value of 2.2 Gy for this group was used in the statistical analysis. Using a model assuming that risk increases linearly with radiation, the estimated ERR/Gy was 0.11 (95% CI, 0.02 to 0.44) for all women and 0.23 (95% CI, 0.04 to 2.13) for smokers (Table 5).
DISCUSSION
Our results demonstrate that, in women treated with radiotherapy, the risk of lung cancer increased after a latency period of more than 15 years. Squamous cell carcinoma seemed to be the histopathologic subgroup most closely related to ionizing radiation. In addition, the increased risk was restricted to women who smoked at time of radiotherapy. Notably, nonsmoking women who received radiotherapy were not found to have an increased risk of lung cancer. In our study group, the mean lung dose decreased over the study period, probably as an effect of an increased prevalence of partial mastectomy.
The case-only design applied in this study, where the unaffected lung serves as the control to the affected lung, has seldom been used before.6 The reason is that a situation like this rarely occurs because the design can only be applied in the case of paired organs where the organs, for some reason, receive different exposures. In the present setting, the radiation exposure to the two lungs differs because of radiation treatment for breast cancer. The design resembles an affected-unaffected monozygotic twin design. In contrast to a twin study, all of our radiation-treated lung cancer patients contributed an informative pair because they were exposure discordant. Furthermore, the design also implies inherent confounding control for all genetic and environmental factors, including smoking and chemotherapy. The main disadvantage is that the only estimable main effect is the effect for radiotherapy. However, potential effect modification can be estimated by stratifying the patients by different characteristics.
The interpretations of our results, as well as of previous studies of lung cancer after breast cancer, are limited by low statistical precision because of the small number of patients. In addition, several previous studies do not have detailed information on radiation dose or smoking history. The strength of our study was the ability to classify individual dosimetry estimated from the original radiotherapy treatment and medical records. However, because the exact position of the lung cancers was unknown, it was only possible to calculate the mean dose to the lung. Our estimated mean dose to the ipsilateral lung was between 7.9 and 26.8 Gy, and the mean dose to the contralateral lung was between 0.4 and 8.6 Gy (Table 2), which is in agreement with other previous studies.5,25,26 One exception is a study by Neugut et al,14 where the mean dose to the lung on the same side as the breast cancer was approximately 2 Gy.
It is generally considered that the induction and latent period of a solid radiation-associated cancer is at least 10 years.5,27,28 However, some studies demonstrated increased excess risk caused by radiation in the period 5 to 10 years after exposure.16,29 In the current study, we showed that women who received radiotherapy had an increased risk of developing lung cancer on the same side as breast cancer after at least 15 years of follow-up. This finding is consistent with other studies that showed that radiation-induced lung cancer develops after a substantial latency period.4,6,25,28,30,31 Notably, women who did not receive radiotherapy were not found to have an increased risk of lung cancer.
In a previous study of lung cancer after radiotherapy for Hodgkin's disease, the most common type of lung cancer was squamous cell carcinoma (45%).16 In a study of lung cancer after radiotherapy for breast cancer, Zablotska and Neugut28 found a significantly increased RR of 5.52 in ipsilateral squamous cell subtype of lung carcinoma. These findings are in agreement with our study, where the squamous cell carcinoma was found to be the most common subtype of lung carcinoma (34%) in irradiated women, with a significantly increased RR of 4.00 (95% CI, 1.50 to 10.66; Table 4).
Smoking is the major risk factor of lung cancer. As a consequence of increased prevalence of smoking, there has been a dramatic increase in the number of lung cancer cases in women.7 Radiation exposure and tobacco smoke both cause direct DNA damage and epigenetic effects.32,33 Thus, they share common carcinogenic mechanisms and might act in consort to increase lung cancer risk when the exposures occur in the same person. Both agents cause cancer in a dose-dependent fashion, and the order of exposure might not be important. When evaluating the combined effects of radiation exposure and smoking, it is necessary to take into account the characteristics of the study population, duration and fractionation of exposure, and volume of the irradiated lung. The data for atomic bomb survivors found the effects of radiation exposure and cigarette smoking on lung cancer to be significantly submultiplicative and quite consistent with additivity,34 whereas in a study of Hodgkin's disease,16 the joint effect of radiation and smoking was consistent with a multiplicative relationship but not with an additive relationship. Because an interaction of ionizing radiation and smoking in women with breast cancer has previously been found,14,35 we studied separately the effect of radiotherapy among smokers and nonsmokers. Despite the small number of women in the study group, we have an indication that the adverse effect of radiotherapy is restricted to smokers because no effect was seen in nonsmoking women (Table 4).
In general, women in the past were treated with orthovoltage x-rays or 60Co and were more likely to undergo more extensive nodal irradiation, increasing the dose to the underlying lung and potentially increasing the risk of carcinogenesis. In our study, the modern technique of radiotherapy uses linear accelerators (6 to 8 MV), whereas in the past, 60Co, orthovoltage x-rays, and MeV radiation were used (Table 2). Partial mastectomy became more common at the end of the study period, as reflected by a decrease in lung dose (Fig 2). The lowest mean dose to the lung was seen in women treated with partial mastectomy and radiotherapy to the remaining breast parenchyma. Still, women treated with radical mastectomy were usually treated with the type of radiation technique where radiotherapy to the IMC and to the lymph nodes usually adds the dose to the lung. The small number of women treated with partial mastectomy and radiotherapy limited our ability to analyze how the surgical technique modified the risk of subsequent lung cancer.
Little36 estimated and then compared the ERR/Gy from studies of cancer patients and atomic bomb survivors. This study found that the risks for cancer patients were smaller than for the atomic bomb survivors of a similar age at exposure. Large differences in dose and dose rates make comparisons of atomic bomb survivors and patients receiving therapeutic doses of external radiation difficult. There is some uncertainty concerning the shape of the dose-response relationship for higher doses of radiotherapy. In our study, we used a model where risk increases linearly with radiation dose, and we have seen no indication of variation from linearity (Table 5) for our studied range of 7.1 to 32.0 Gy. These results are in agreement with a previous study by Gilbert et al.16 They showed little evidence of variation from linearity in a study of lung cancer after radiotherapy for Hodgkin's disease, even though lung doses for the majority of patients exceeded 30 Gy. The radiation-related ERR/Gy in that study was 0.15 (95% CI, 0.057 to 0.39) for all patients and 0.044 (95% CI, –0.009 to –0.53) for female patients, respectively. Those results are similar to the ERR/Gy of 0.11 (95% CI, 0.02 to 0.44) for our lung cancer study group. Also, in the same previously published study, the ERR/Gy for current smokers varied between 0.095 and 0.35 compared with our estimate of 0.23 (95% CI, 0.04 to 0.21).
We show that radiotherapy for breast cancer increases the risk of lung cancer of the ipsilateral lung and that the risk seems to be confined to patients who smoked at time of radiotherapy. However, the CI does not exclude the possibility of increased risk (Table 4). The benefits of radiotherapy certainly outweigh the harmful effects of the treatment for the vast majority of women with breast cancer. Several issues have to be taken into consideration in risk-benefit analyses of radiotherapy for breast cancer. Our results indicate that smoking is one of the important issues to be taken into consideration. We do not have the possibility to study whether the risk of lung cancer decreases with termination of tobacco use at time of diagnosis of breast cancer. We are presently conducting a nationwide Swedish case-control study of women with both breast and lung cancer, and the larger sample size will hopefully allow more detailed analyses of the influence of radiation and smoking exposure.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Parkin DM, Whelan SL, Ferlay J, et al: Cancer Incidence in Five Continents: International Agency for Research on Cancer Scientific Publication, Vol II. Lyon, France, International Agency for Research on Cancer, 1997
Peto R, Boreham J, Clarke M, et al: UK and USA breast cancer deaths down 25% in year 2000 at ages 20-69 years. Lancet 355:1822, 2000
Overgaard M, Hansen PS, Overgaard J, et al: Postoperative radiotherapy in high-risk premenopausal women with breast cancer who receive adjuvant chemotherapy. N Engl J Med 337:949-955, 1997
Harvey EB, Brinton LA: Second cancer following cancer of the breast in Connecticut, 1935-1982. Natl Cancer Inst Monogr 68:99-112, 1985
Inskip PD, Stovall M, Flannery JT: Lung cancer risk and radiation dose among women treated for breast cancer. J Natl Cancer Inst 86:983-988, 1994
Prochazka M, Granath F, Ekbom A, et al: Lung cancer risks in women with previous breast cancer. Eur J Cancer 38:1520-1525, 2002
Socialstyrelsen: Official Statistics of Sweden: Cancer Incidence in Sweden 2002. Stockholm, Sweden, Socialstyrelsen, Report 6: 2003
Statistiska Centralbyr?n: Tobacco. Stockholm, Sweden, Statistiska Centralbyr?n, 2000
Krueger EA, Fraass BA, McShan DL, et al: Potential gains for irradiation of chest wall and regional nodes with intensity modulated radiotherapy. Int J Radiat Oncol Biol Phys 56:1023-1037, 2003
Strom EA: Breast IMRT: New tools leading to new vision. Int J Radiat Oncol Biol Phys 54:1297-1298, 2002
Hall EJ, Wuu CS: Radiation-induced second cancers: The impact of 3D-CRT and IMRT. Int J Radiat Oncol Biol Phys 56:83-88, 2003
Lomax AJ, Cella L, Weber D, et al: Potential role of intensity-modulated photons and protons in the treatment of the breast and regional nodes. Int J Radiat Oncol Biol Phys 55:785-792, 2003
Verellen D, Linthout N, Soete G, et al: Considerations on treatment efficiency of different conformal radiation therapy techniques for prostate cancer. Radiother Oncol 63:27-36, 2002
Neugut AI, Murray T, Santos J, et al: Increased risk of lung cancer after breast cancer radiation therapy in cigarette smokers. Cancer 73:1615-1620, 1994
Pershagen G, Akerblom G, Axelson O, et al: Residential radon exposure and lung cancer in Sweden. N Engl J Med 330:159-164, 1994
Gilbert ES, Stovall M, Gospodarowicz M, et al: Lung cancer after treatment for Hodgkin's disease: Focus on radiation effects. Radiat Res 159:161-173, 2003
Edsmyr F, Walstam R: Method for irradiation of parasternal lymph-node metastases. Acta Radiol 51:308-320, 1959
Lind PA, Rosfors S, Wennberg B, et al: Pulmonary function following adjuvant chemotherapy and radiotherapy for breast cancer and the issue of three-dimensional treatment planning. Radiother Oncol 49:245-254, 1998
Gagliardi G, Bj?hle J, Ottolenghi A, et al: Radiation pneumonitis after radiotherapy for breast cancer: Analysis of the complication probability using the relative seriality model. Int J Radiat Oncol Biol Phys 46:373-381, 2000
Gagliardi G, Lax I, Ottolenghi A, et al: Long-term cardiac mortality after radiotherapy of breast cancer: Application of the relative seriality model. Br J Radiol 69:839-846, 1996
Honore HB, Bentzen SM, Moller K, et al: Sensori-neural hearing loss after radiotherapy for nasopharyngeal carcinoma: Individualized risk estimation. Radiother Oncol 65:9-16, 2002
Webster E, Tsien KC: Atlas of Dose Distribution (Vol 1). Vienna, Austria, International Atomic Energy Agency, 1965
Fraass BA, van de Geijn J: Peripheral dose from megavolt beams. Med Phys 10:809-818, 1983
McGue M: When assessing twin concordance, use the probandwise not the pairwise rate. Schizophr Bull 18:171-176, 1992
Neugut AI, Robinson E, Lee WC, et al: Lung cancer after radiation therapy for breast cancer. Cancer 71:3054-3057, 1993
Rubino C, de Vathaire F, Diallo I, et al: Radiation dose, chemotherapy and risk of lung cancer after breast cancer treatment. Breast Cancer Res Treat 75:15-24, 2002
United Nations Scientific Committee on the Effects of Atomic Radiation.Sources and effects of ionizing radiation: UNSCEAR 2000—Report to the General Assembly, with scientific annexes. New York, NY, United Nations, 2000
Zablotska LB, Neugut AI: Lung carcinoma after radiation therapy in women treated with lumpectomy or mastectomy for primary breast carcinoma. Cancer 97:1404-1411, 2003
Weiss HA, Darby SC, Doll R: Cancer mortality following X-ray treatment for ankylosing spondylitis. Int J Cancer 59:327-338, 1994
Travis LB, Curtis RE, Inskip PD, et al: Re: Lung cancer risk and radiation dose among women treated for breast cancer. J Natl Cancer Inst 87:60-61, 1995
Deutsch M, Land SR, Begovic M, et al: The incidence of lung carcinoma after surgery for breast carcinoma with and without postoperative radiotherapy: Results of National Surgical Adjuvant Breast and Bowel Project (NSABP) Clinical Trials B-04 and B-06. Cancer 98:1362-1368, 2003
Yanagawa N, Tamura G, Oizumi H, et al: Frequent epigenetic silencing of the p16 gene in non-small cell lung cancers of tobacco smokers. Jpn J Cancer Res 93:1107-1113, 2002
Kim DH, Kim JS, Ji YI, et al: Hypermethylation of RASSF1A promoter is associated with the age at starting smoking and a poor prognosis in primary non-small cell lung cancer. Cancer Res 63:3743-3746, 2003
Pierce DA, Sharp GB, Mabuchi K: Joint effects of radiation and smoking on lung cancer risk among atomic bomb survivors. Radiat Res 159:511-520, 2003
Ford MB, Sigurdson AJ, Petrulis ES, et al: Effects of smoking and radiotherapy on lung carcinoma in breast carcinoma survivors. Cancer 98:1457-1464, 2003
Little MP: Comparisons of lung tumour mortality risk in the Japanese A-bomb survivors and in the Colorado Plateau uranium miners: Support for the ICRP lung model. Int J Radiat Biol 78:145-163, 2002(Michaela Prochazka, Per H)
Department of Hospital Physics, Radiumhemmet, and Department of Medicine, Clinical Epidemiology Unit, Karolinska University Hospital
Department of Medical Radiation Physics, Stockholm University, Stockholm, Sweden
Lombardi Cancer Center, Georgetown University Medical Center, Washington, DC.
ABSTRACT
PURPOSE: To analyze the risk of lung cancer in women treated with radiotherapy for breast cancer. We accessed the lung dose in relation to different radiotherapy techniques, provided the excess relative risk (ERR) estimate for radiation-associated lung cancer, and evaluated the influence of tobacco use.
PATIENTS AND METHODS: The Swedish Cancer Registry was used to identify 182 women diagnosed with breast and subsequent lung cancers in Stockholm County during 1958 to 2000. Radiotherapy was administered to 116 patients. Radiation dose was estimated from the original treatment charts, and information on smoking history was searched for in case records and among relatives. The risk of lung cancer was assessed in a case-only approach, where each woman contributed a pair of lungs.
RESULTS: The average mean lung dose to the ipsilateral lung was 17.2 Gy (range, 7.1 to 32.0 Gy). A significantly increased relative risk (RR) of a subsequent ipsilateral lung cancer was observed at 10 years of follow-up (RR = 2.04; 95% CI, 1.24 to 3.36). Squamous cell carcinoma (RR = 4.00; 95% CI, 1.50 to 10.66) was the histopathologic subgroup most closely related to ionizing radiation. The effect of radiotherapy was restricted to smokers only (RR = 3.08; 95% CI, 1.61 to 5.91). The ERR/Gy for women with latency 10 years after exposure was 0.11 (95% CI, 0.02 to 0.44).
CONCLUSION: Radiotherapy for breast cancer significantly increases the risk of lung carcinoma more than 10 years after exposure in women who smoked at time of breast cancer.
INTRODUCTION
The incidence of breast cancer is increasing throughout the industrialized world,1 although at the same time, the mortality is decreasing.2 With recent advances in early diagnosis and treatment, breast cancer is becoming an increasingly survivable disease. Women with breast cancer normally receive postsurgical adjuvant therapy, as radiotherapy, chemotherapy, or hormonal therapy or as a combination of any of these modalities. Adjuvant radiation therapy reduces the risk of local recurrence,3 and its use is increasing as more women today choose partial mastectomies as their surgical choice. However, one of the growing concerns is the chronic or late-occurring complications to the normal tissue from treatment of primary malignancies, such as therapy-related secondary cancers.
Women with breast carcinoma are at increased risk of subsequent primary malignancies, specifically lung carcinoma.4-6 During the last 20 years, the absolute number of lung cancers in women previously diagnosed with breast cancer has shown a steady increase, including in Sweden (Fig 1). 7 The reasons for this are probably the increasing number of long-term surviving breast cancer patients and the fact that the birth cohorts with the highest smoking prevalence (born between 1940 and 1959)8 are now moving into an age when lung cancer more commonly occurs. The increased use of radiotherapy during the last 30 years is probably also a contributing factor. Newer radiotherapy modalities using intensity-modulated radiation therapy, while seeking to reduce acute toxicity, might actually be increasing secondary cancer risk.9,10 This methodology, which uses multiple beams to deliver high doses to the tumor bed, aims to decrease doses to surrounding normal tissue with an attendant reduction in acute radiation complications. However, some clinicians worry that the secondary cancer risk persists or is worse because a greater volume of tissue is radiated at doses still considered to be potentially carcinogenic.9,11-13 Thus, it is important to identify the magnitude of radiation-related lung cancer risk in long-term breast cancer survivors. In addition to radiation, tobacco is a known carcinogen. Smoking is the major risk factor of lung cancer, and a multiplicative effect of smoking on radiation-related risk is suspected.14-16
The aim of our study was to analyze the risk of lung cancer in relation to radiotherapy for breast cancer. We assessed how the lung dose, estimated through a retrospective quantification of the dose distribution, varies with breast cancer radiotherapy techniques and provided excess relative risk (ERR) estimates in terms of radiation dose delivered to the lung. In addition, we evaluated the influence of tobacco use.
PATIENTS AND METHODS
Using the Swedish Cancer Registry, 191 women diagnosed with breast and subsequent lung cancers were identified in Stockholm County during the period from 1958 to 2000. Nine patients were excluded because they were diagnosed with lung cancer within 12 months of the initial breast cancer. It is not likely that these lung cancers could be causally related to radiotherapy. In all, 116 women had received radiotherapy, whereas 66 women were treated by other means (Table 1).
Radiotherapy was centralized to three departments of oncology in Stockholm. For each patient, case records and radiotherapy charts were identified and reviewed. Radiotherapy charts were abstracted for detailed information about physical and geometrical irradiation parameters, such as radiation quality, treatment technique, total dose, fractionation schedule, and treatment field size.
Information about smoking history was identified in case records from departments of oncology, surgery, and thoracic medicine. For patients lacking information (approximately 29%), lifetime smoking habits were retrieved by asking spouse or next of kin for information through a mailed questionnaire. The patient was considered a smoker if the case record or relative indicated that the woman was an occasional smoker or smoked one cigarette or more daily. We did not have the possibility to stratify smoking habits further than smoker and nonsmoker. Among 182 women, the smoking information was available for 71% from medical records and for 22% from next of kin. The smoking status was unknown for 7% of women. A total of 114 women were identified as smokers.
The mean age at diagnosis of breast cancer among the 182 women included in the study was 58 years (median, 57 years; first quartile, 48 years; and third quartile, 65 years). The main characteristics of the study group are listed in Table 1. Fifty-one percent of irradiated women were treated before 1970 compared with nonirradiated women, who were mostly diagnosed between 1970 and 1980. Approximately 40% of the patients were treated with cobalt-60 gamma-radiation (60Co). Thirty-nine (34%) of the 116 radiotherapy-treated women had squamous cell carcinoma, and 33 (28%) had adenocarcinoma. In the nonirradiated group, squamous cell carcinoma occurred in 13 women (20%), and adenocarcinoma occurred in 19 women (29%). Radiotherapy charts were available for 103 of 116 patients. For three patients with missing radiotherapy charts, detailed information about their radiotherapy was found in their case records.
The criteria chosen for organizing the radiotherapy information were the target definition and, consequently, the used treatment techniques. In most clinical charts, the exact definition of target was not given. Target was defined in general terms such as breast parenchyma, chest wall, and lymph nodes of the internal mammary chain (IMC), supraclavicular region, and axillary region.
We identified nine different treatment groups (Table 2). Group 1 consisted of patients irradiated to the IMC. The treatment was administered with overlapping rectangular 60Co fields using a short-distance gamma beam unit. The treatment technique is described in detail elsewhere.17 In group 2, patients were treated with orthovoltage x-ray beams. Ten of the patients were treated with two adjacent anterior-posterior fields to the fossa-supraclaviculary lymph nodes and to the chest wall. Additionally, two patients were treated with one field covering the IMC and one field covering the fossa-axillary lymph nodes. Treatments were administered with a source–skin distance of 60 cm, 0.5 mm copper filtration. Group 3 consisted of patients treated with electron beams covering the fossa-supraclaviculary lymph nodes. Patients in group 4 received radiotherapy to the breast parenchyma/chest wall and the IMC lymph nodes; they were treated with extended tangential photon fields. A shield for the apical part of the lung was often used. Group 5 consisted of patients in whom the IMC, fossa, and axilla were treated with an anterior-posterior 60Co beam. Patients in group 6 were treated like patients in group 5, but they also received an oblique 60Co beam to the thorax region. In group 7, the chest wall and the lower part of the IMC were treated with an oblique electron field. The lymph nodes in the lower part of the IMC, fossa, supraclavicular, and axillary regions were treated with an anterior photon field, whereas a posterior photon field was added to the axillary region. In group 8, the chest wall and the caudal part of the IMC lymph nodes were covered by an anterior electron field. An anterior photon field was used for the cranial part of the IMC, axillary, fossa, and supraclavicular lymph nodes.18 Finally, patients in group 9 received radiotherapy to the remaining breast parenchyma after partial mastectomy and were treated with tangential 60Co beam or with x-ray photon fields.
Radiation treatments were generally administered in 5 fractions per week for 3 to 6 weeks. The highest dose per fraction was administered to patients in group 1, and the lowest dose per fraction was administered in group 9. The treatment techniques of groups 4, 7, and 9 have previously been described in detail.19
Lung Dose Calculation
For most of the patients, original individual three-dimensional dose distributions in the lung were not available because patients were treated before the introduction of three-dimensional treatment planning systems (3D TPS). The reconstruction of treatment techniques on 3D TPS has already been used in studies both on radiation-induced secondary cancer5 and on dose-volume complication relationships.19-21 On the basis of the irradiation parameters given in the radiotherapy charts, we were able to quantify the ipsilateral and contralateral absorbed lung dose for all groups except group 2, where only ipsilateral lung dose was estimated.
For patients treated with short-distance 60Co (group 1), the dose was estimated using published isodose distributions.22 The dose to the ipsilateral lung was calculated on two axial computed tomography slices in a patient operated with radical mastectomy, one at the level of the mamillary plane and one located approximately 11 cm below the jugular region. The dose was calculated on a grid in the two slices. There is uncertainty in the estimate of the contralateral lung dose for this group because only a rough estimate of the leakage and scatter in the collimators used in the short distance treatment techniques could be made, and this contribution to the secondary dose may be considerable.23 For patients in group 2, the mean dose to the ipsilateral lung was calculated from the original depth-dose tables and isodose curves specific for the machines used. To correct for the lower density of the lung tissue, a lung density of 0.25 g/cm3 was used. For this group, the uncertainty in the estimation of the contralateral lung dose was regarded as too large to be of interest.
The treatment technique for group 3 was reconstructed on one model patient, and the treatment techniques for groups 5 and 6 were reconstructed on each patient in the group. Techniques for groups 4, 7, 8, and 9 were reconstructed on a computed tomography–based 3D TPS on 10 model patients.
Statistical Methods
The risk of lung cancer was assessed in a case-only approach, where each woman contributed a pair of lungs that can be considered matched for genetic and environmental factors (primarily smoking). The lung on the breast cancer side was considered as exposed, and the contralateral lung was considered as unexposed. The concordance rates with 95% CIs were calculated as the proportion of lung cancer cases on the same (concordant) side as the breast cancer. The relative risks (RR) and 95% CIs were estimated as in twin-study design, by conditional logistic regression.24 Heterogeneity of the radiation effect by latency, smoking, and histopathology were tested by likelihood ratio tests.
The ERR per Gy was calculated for the patient group with latency 10 years after exposure. The dose-response modeling was performed with the PECAN module of the software package EPICURE, applying an additive relative risk model. The results are presented as ERR/Gy together with likelihood-based CIs. Potential deviance from linearity was investigated by fitting a linear-quadratic function in dose. The individual doses to the ipsilateral and contralateral lung were used. Because the absorbed lung doses to the contralateral lung for treatment group 2 could not be reconstructed, patients from this group were excluded from the analysis.
RESULTS
The mean dose to the ipsilateral lung was 17.2 Gy (range, 7.1 to 32.0 Gy; Table 2), depending on the treatment group. In the first part of the study period (1960s), most of our patients received a mean dose to the lung of approximately 15 Gy, whereas in the 1970s, the mean dose increased to approximately 25 Gy (Fig 2). In the latter part of the study period (1990s), the mean dose decreased to approximately 8 Gy.
We compared latency and risk for lung cancer on the breast cancer side (concordant side) with lung cancer on the contralateral side (discordant side). By using the fact that the two organs are paired, these data could be used to assess the lung cancer risk in a similar fashion as in a disease-discordant twin pair study. Among 116 lung cancer patients who received radiotherapy, there were four women with bilateral tumors in the lung and six women in whom the laterality of lung cancer was not specified, leaving 106 women to be included in the statistical analyses.
The mean latency period between breast and lung cancer in women not treated with radiation therapy was 9.7 years, with no difference between the concordant and discordant side. In patients treated with radiotherapy, however, lung cancers diagnosed on the concordant side showed a mean latency of 17.6 years compared with 13.0 years for discordant lung cancers.
In patients treated with radiotherapy, the risk of developing lung cancer on the same side as the breast cancer compared with the contralateral side was significantly increased after more than 10 years from the breast cancer diagnosis (RR = 2.04; 95% CI, 1.24 to 3.36; Table 3). Within 10 years from breast cancer diagnosis, radiotherapy did not seem to influence the risk of lung cancer. In addition, absence of radiotherapy was not associated with an increase in risk on the concordant side of breast and lung regardless of latency period (Table 3). When time between breast and lung cancer in irradiated women was studied in detail (quantiles), an increased risk was present for concordant lung cancer, with an RR of 2.38 (95% CI, 1.04 to 5.43) in the interval of 15 to 23 years after exposure and of 2.25 (95% CI, 1.01 to 5.17) in the interval of 23 years after exposure (Table 3). In the analysis of the latency between breast cancer and lung cancer, evidence for heterogeneity was found (P = .005).
In calculation of RR for women treated with radiotherapy, the contralateral lung was considered as unexposed and serves as the control for the ipsilateral lung. Therefore, to validate our results, we calculated the RR when the contralateral dose to the lung was known and low ( 15% of the ipsilateral dose to the lung; Table 2). The average dose to the lung was 18.8 Gy when treatment technique groups 1, 2, 5, and 8 were excluded. The RR was significantly increased for women with a latency time of 10 years after exposure (RR = 2.09; 95% CI, 1.02 to 4.29); for a latency time of less than 10 years, the RR was 1.09 (95% CI, 0.48 to 2.47). These results are comparable to the RR for the whole radiation-treated group.
Further analysis of the irradiated study group with a latency time of 10 years after breast cancer treatment demonstrated that the increased risk was most evident for squamous cell carcinoma of the lung (RR = 4.00; 95% CI, 1.50 to 10.66; Table 4). The significant association was not seen in any other histopathologic subtype of lung cancer, including adenocarcinoma. Separate analysis by smoking showed that the increased risk of lung cancer was present for smokers (RR = 3.17; 95% CI, 1.66 to 6.06), but there was no observed increase in risk for nonsmokers (Table 4). In these analyses, evidence of heterogeneity was found for smoking (P = .026), whereas heterogeneity for histopathology did not reach statistical significance (P = .07).
The ERR/Gy was calculated for the irradiated patient group with a latency of 10 years after breast cancer treatment (Table 5). Because further analysis of this group revealed a heterogeneity by smoking, separate estimates of ERR/Gy were presented for smokers and nonsmokers. In the analysis, the individual doses to the ipsilateral and contralateral lung were used, and therefore, patients from the groups with treatment techniques with a known dose to the contralateral lung were selected (group 2 was excluded). For group 1, an approximate dose to the contralateral lung was estimated to be between 1.5 and 2.9 Gy; therefore, a mean value of 2.2 Gy for this group was used in the statistical analysis. Using a model assuming that risk increases linearly with radiation, the estimated ERR/Gy was 0.11 (95% CI, 0.02 to 0.44) for all women and 0.23 (95% CI, 0.04 to 2.13) for smokers (Table 5).
DISCUSSION
Our results demonstrate that, in women treated with radiotherapy, the risk of lung cancer increased after a latency period of more than 15 years. Squamous cell carcinoma seemed to be the histopathologic subgroup most closely related to ionizing radiation. In addition, the increased risk was restricted to women who smoked at time of radiotherapy. Notably, nonsmoking women who received radiotherapy were not found to have an increased risk of lung cancer. In our study group, the mean lung dose decreased over the study period, probably as an effect of an increased prevalence of partial mastectomy.
The case-only design applied in this study, where the unaffected lung serves as the control to the affected lung, has seldom been used before.6 The reason is that a situation like this rarely occurs because the design can only be applied in the case of paired organs where the organs, for some reason, receive different exposures. In the present setting, the radiation exposure to the two lungs differs because of radiation treatment for breast cancer. The design resembles an affected-unaffected monozygotic twin design. In contrast to a twin study, all of our radiation-treated lung cancer patients contributed an informative pair because they were exposure discordant. Furthermore, the design also implies inherent confounding control for all genetic and environmental factors, including smoking and chemotherapy. The main disadvantage is that the only estimable main effect is the effect for radiotherapy. However, potential effect modification can be estimated by stratifying the patients by different characteristics.
The interpretations of our results, as well as of previous studies of lung cancer after breast cancer, are limited by low statistical precision because of the small number of patients. In addition, several previous studies do not have detailed information on radiation dose or smoking history. The strength of our study was the ability to classify individual dosimetry estimated from the original radiotherapy treatment and medical records. However, because the exact position of the lung cancers was unknown, it was only possible to calculate the mean dose to the lung. Our estimated mean dose to the ipsilateral lung was between 7.9 and 26.8 Gy, and the mean dose to the contralateral lung was between 0.4 and 8.6 Gy (Table 2), which is in agreement with other previous studies.5,25,26 One exception is a study by Neugut et al,14 where the mean dose to the lung on the same side as the breast cancer was approximately 2 Gy.
It is generally considered that the induction and latent period of a solid radiation-associated cancer is at least 10 years.5,27,28 However, some studies demonstrated increased excess risk caused by radiation in the period 5 to 10 years after exposure.16,29 In the current study, we showed that women who received radiotherapy had an increased risk of developing lung cancer on the same side as breast cancer after at least 15 years of follow-up. This finding is consistent with other studies that showed that radiation-induced lung cancer develops after a substantial latency period.4,6,25,28,30,31 Notably, women who did not receive radiotherapy were not found to have an increased risk of lung cancer.
In a previous study of lung cancer after radiotherapy for Hodgkin's disease, the most common type of lung cancer was squamous cell carcinoma (45%).16 In a study of lung cancer after radiotherapy for breast cancer, Zablotska and Neugut28 found a significantly increased RR of 5.52 in ipsilateral squamous cell subtype of lung carcinoma. These findings are in agreement with our study, where the squamous cell carcinoma was found to be the most common subtype of lung carcinoma (34%) in irradiated women, with a significantly increased RR of 4.00 (95% CI, 1.50 to 10.66; Table 4).
Smoking is the major risk factor of lung cancer. As a consequence of increased prevalence of smoking, there has been a dramatic increase in the number of lung cancer cases in women.7 Radiation exposure and tobacco smoke both cause direct DNA damage and epigenetic effects.32,33 Thus, they share common carcinogenic mechanisms and might act in consort to increase lung cancer risk when the exposures occur in the same person. Both agents cause cancer in a dose-dependent fashion, and the order of exposure might not be important. When evaluating the combined effects of radiation exposure and smoking, it is necessary to take into account the characteristics of the study population, duration and fractionation of exposure, and volume of the irradiated lung. The data for atomic bomb survivors found the effects of radiation exposure and cigarette smoking on lung cancer to be significantly submultiplicative and quite consistent with additivity,34 whereas in a study of Hodgkin's disease,16 the joint effect of radiation and smoking was consistent with a multiplicative relationship but not with an additive relationship. Because an interaction of ionizing radiation and smoking in women with breast cancer has previously been found,14,35 we studied separately the effect of radiotherapy among smokers and nonsmokers. Despite the small number of women in the study group, we have an indication that the adverse effect of radiotherapy is restricted to smokers because no effect was seen in nonsmoking women (Table 4).
In general, women in the past were treated with orthovoltage x-rays or 60Co and were more likely to undergo more extensive nodal irradiation, increasing the dose to the underlying lung and potentially increasing the risk of carcinogenesis. In our study, the modern technique of radiotherapy uses linear accelerators (6 to 8 MV), whereas in the past, 60Co, orthovoltage x-rays, and MeV radiation were used (Table 2). Partial mastectomy became more common at the end of the study period, as reflected by a decrease in lung dose (Fig 2). The lowest mean dose to the lung was seen in women treated with partial mastectomy and radiotherapy to the remaining breast parenchyma. Still, women treated with radical mastectomy were usually treated with the type of radiation technique where radiotherapy to the IMC and to the lymph nodes usually adds the dose to the lung. The small number of women treated with partial mastectomy and radiotherapy limited our ability to analyze how the surgical technique modified the risk of subsequent lung cancer.
Little36 estimated and then compared the ERR/Gy from studies of cancer patients and atomic bomb survivors. This study found that the risks for cancer patients were smaller than for the atomic bomb survivors of a similar age at exposure. Large differences in dose and dose rates make comparisons of atomic bomb survivors and patients receiving therapeutic doses of external radiation difficult. There is some uncertainty concerning the shape of the dose-response relationship for higher doses of radiotherapy. In our study, we used a model where risk increases linearly with radiation dose, and we have seen no indication of variation from linearity (Table 5) for our studied range of 7.1 to 32.0 Gy. These results are in agreement with a previous study by Gilbert et al.16 They showed little evidence of variation from linearity in a study of lung cancer after radiotherapy for Hodgkin's disease, even though lung doses for the majority of patients exceeded 30 Gy. The radiation-related ERR/Gy in that study was 0.15 (95% CI, 0.057 to 0.39) for all patients and 0.044 (95% CI, –0.009 to –0.53) for female patients, respectively. Those results are similar to the ERR/Gy of 0.11 (95% CI, 0.02 to 0.44) for our lung cancer study group. Also, in the same previously published study, the ERR/Gy for current smokers varied between 0.095 and 0.35 compared with our estimate of 0.23 (95% CI, 0.04 to 0.21).
We show that radiotherapy for breast cancer increases the risk of lung cancer of the ipsilateral lung and that the risk seems to be confined to patients who smoked at time of radiotherapy. However, the CI does not exclude the possibility of increased risk (Table 4). The benefits of radiotherapy certainly outweigh the harmful effects of the treatment for the vast majority of women with breast cancer. Several issues have to be taken into consideration in risk-benefit analyses of radiotherapy for breast cancer. Our results indicate that smoking is one of the important issues to be taken into consideration. We do not have the possibility to study whether the risk of lung cancer decreases with termination of tobacco use at time of diagnosis of breast cancer. We are presently conducting a nationwide Swedish case-control study of women with both breast and lung cancer, and the larger sample size will hopefully allow more detailed analyses of the influence of radiation and smoking exposure.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Parkin DM, Whelan SL, Ferlay J, et al: Cancer Incidence in Five Continents: International Agency for Research on Cancer Scientific Publication, Vol II. Lyon, France, International Agency for Research on Cancer, 1997
Peto R, Boreham J, Clarke M, et al: UK and USA breast cancer deaths down 25% in year 2000 at ages 20-69 years. Lancet 355:1822, 2000
Overgaard M, Hansen PS, Overgaard J, et al: Postoperative radiotherapy in high-risk premenopausal women with breast cancer who receive adjuvant chemotherapy. N Engl J Med 337:949-955, 1997
Harvey EB, Brinton LA: Second cancer following cancer of the breast in Connecticut, 1935-1982. Natl Cancer Inst Monogr 68:99-112, 1985
Inskip PD, Stovall M, Flannery JT: Lung cancer risk and radiation dose among women treated for breast cancer. J Natl Cancer Inst 86:983-988, 1994
Prochazka M, Granath F, Ekbom A, et al: Lung cancer risks in women with previous breast cancer. Eur J Cancer 38:1520-1525, 2002
Socialstyrelsen: Official Statistics of Sweden: Cancer Incidence in Sweden 2002. Stockholm, Sweden, Socialstyrelsen, Report 6: 2003
Statistiska Centralbyr?n: Tobacco. Stockholm, Sweden, Statistiska Centralbyr?n, 2000
Krueger EA, Fraass BA, McShan DL, et al: Potential gains for irradiation of chest wall and regional nodes with intensity modulated radiotherapy. Int J Radiat Oncol Biol Phys 56:1023-1037, 2003
Strom EA: Breast IMRT: New tools leading to new vision. Int J Radiat Oncol Biol Phys 54:1297-1298, 2002
Hall EJ, Wuu CS: Radiation-induced second cancers: The impact of 3D-CRT and IMRT. Int J Radiat Oncol Biol Phys 56:83-88, 2003
Lomax AJ, Cella L, Weber D, et al: Potential role of intensity-modulated photons and protons in the treatment of the breast and regional nodes. Int J Radiat Oncol Biol Phys 55:785-792, 2003
Verellen D, Linthout N, Soete G, et al: Considerations on treatment efficiency of different conformal radiation therapy techniques for prostate cancer. Radiother Oncol 63:27-36, 2002
Neugut AI, Murray T, Santos J, et al: Increased risk of lung cancer after breast cancer radiation therapy in cigarette smokers. Cancer 73:1615-1620, 1994
Pershagen G, Akerblom G, Axelson O, et al: Residential radon exposure and lung cancer in Sweden. N Engl J Med 330:159-164, 1994
Gilbert ES, Stovall M, Gospodarowicz M, et al: Lung cancer after treatment for Hodgkin's disease: Focus on radiation effects. Radiat Res 159:161-173, 2003
Edsmyr F, Walstam R: Method for irradiation of parasternal lymph-node metastases. Acta Radiol 51:308-320, 1959
Lind PA, Rosfors S, Wennberg B, et al: Pulmonary function following adjuvant chemotherapy and radiotherapy for breast cancer and the issue of three-dimensional treatment planning. Radiother Oncol 49:245-254, 1998
Gagliardi G, Bj?hle J, Ottolenghi A, et al: Radiation pneumonitis after radiotherapy for breast cancer: Analysis of the complication probability using the relative seriality model. Int J Radiat Oncol Biol Phys 46:373-381, 2000
Gagliardi G, Lax I, Ottolenghi A, et al: Long-term cardiac mortality after radiotherapy of breast cancer: Application of the relative seriality model. Br J Radiol 69:839-846, 1996
Honore HB, Bentzen SM, Moller K, et al: Sensori-neural hearing loss after radiotherapy for nasopharyngeal carcinoma: Individualized risk estimation. Radiother Oncol 65:9-16, 2002
Webster E, Tsien KC: Atlas of Dose Distribution (Vol 1). Vienna, Austria, International Atomic Energy Agency, 1965
Fraass BA, van de Geijn J: Peripheral dose from megavolt beams. Med Phys 10:809-818, 1983
McGue M: When assessing twin concordance, use the probandwise not the pairwise rate. Schizophr Bull 18:171-176, 1992
Neugut AI, Robinson E, Lee WC, et al: Lung cancer after radiation therapy for breast cancer. Cancer 71:3054-3057, 1993
Rubino C, de Vathaire F, Diallo I, et al: Radiation dose, chemotherapy and risk of lung cancer after breast cancer treatment. Breast Cancer Res Treat 75:15-24, 2002
United Nations Scientific Committee on the Effects of Atomic Radiation.Sources and effects of ionizing radiation: UNSCEAR 2000—Report to the General Assembly, with scientific annexes. New York, NY, United Nations, 2000
Zablotska LB, Neugut AI: Lung carcinoma after radiation therapy in women treated with lumpectomy or mastectomy for primary breast carcinoma. Cancer 97:1404-1411, 2003
Weiss HA, Darby SC, Doll R: Cancer mortality following X-ray treatment for ankylosing spondylitis. Int J Cancer 59:327-338, 1994
Travis LB, Curtis RE, Inskip PD, et al: Re: Lung cancer risk and radiation dose among women treated for breast cancer. J Natl Cancer Inst 87:60-61, 1995
Deutsch M, Land SR, Begovic M, et al: The incidence of lung carcinoma after surgery for breast carcinoma with and without postoperative radiotherapy: Results of National Surgical Adjuvant Breast and Bowel Project (NSABP) Clinical Trials B-04 and B-06. Cancer 98:1362-1368, 2003
Yanagawa N, Tamura G, Oizumi H, et al: Frequent epigenetic silencing of the p16 gene in non-small cell lung cancers of tobacco smokers. Jpn J Cancer Res 93:1107-1113, 2002
Kim DH, Kim JS, Ji YI, et al: Hypermethylation of RASSF1A promoter is associated with the age at starting smoking and a poor prognosis in primary non-small cell lung cancer. Cancer Res 63:3743-3746, 2003
Pierce DA, Sharp GB, Mabuchi K: Joint effects of radiation and smoking on lung cancer risk among atomic bomb survivors. Radiat Res 159:511-520, 2003
Ford MB, Sigurdson AJ, Petrulis ES, et al: Effects of smoking and radiotherapy on lung carcinoma in breast carcinoma survivors. Cancer 98:1457-1464, 2003
Little MP: Comparisons of lung tumour mortality risk in the Japanese A-bomb survivors and in the Colorado Plateau uranium miners: Support for the ICRP lung model. Int J Radiat Biol 78:145-163, 2002(Michaela Prochazka, Per H)