Advertisement

Risk of second primary lung cancer in women after radiotherapy for breast cancer

      Abstract

      Background

      Several epidemiological studies have reported increased risks of second lung cancers after breast cancer irradiation. In this study we assessed the effects of the delivered radiation dose to the lung and the risk of second primary lung cancer.

      Methods

      We conducted a nested case–control study of second lung cancer in a population based cohort of 23,627 early breast cancer patients treated with post-operative radiotherapy from 1982 to 2007. The cohort included 151 cases diagnosed with second primary lung cancer and 443 controls. Individual dose-reconstructions were performed and the delivered dose to the center of the second lung tumor and the comparable location for the controls were estimated, based on the patient specific radiotherapy charts.

      Results

      The median age at breast cancer diagnosis was 54 years (range 34–74). The median time from breast cancer treatment to second lung cancer diagnosis was 12 years (range 1–26 years). 91% of the cases were categorized as ever smokers vs. 40% among the controls. For patients diagnosed with a second primary lung cancer five or more years after breast cancer treatment the rate of lung cancer increased linearly with 8.5% per Gray (95% confidence interval = 3.1–23.3%; p < 0.001). This rate was enhanced for ever smokers with an excess rate of 17.3% per Gray (95% CI = 4.5–54%; p < 0.005).

      Conclusions

      Second lung cancer after radiotherapy for early breast cancer is associated with the delivered dose to the lung. Although the absolute risk is relative low, the growing number of long-time survivors after breast cancer treatment highlights the need for advances in normal tissue sparing radiation techniques.

      Keywords

      Breast cancer is the most common cancer among women in the Western world [

      GLOBOCAN. GLOBOCAN. Updated May 2014. Ref Type: Electronic citation http://globocan.iarc.fr/.

      ] and the cumulative probability for breast cancer in Western Europe is currently 10% [
      • Forouzanfar M.H.
      • Foreman K.J.
      • Delossantos A.M.
      • et al.
      Breast and cervical cancer in 187 countries between 1980 and 2010: a systematic analysis.
      ]. The use of postoperative radiotherapy in the adjuvant setting of early breast cancer treatment has increased over the last two decades, and it has been demonstrated to both decrease loco-regional recurrence rates and improve overall survival [
      • Clarke M.
      • Collins R.
      • Darby S.
      • et al.
      Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials.
      ,
      • Darby S.
      • McGale P.
      • Correa C.
      • et al.
      Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomised trials.
      ,
      • Overgaard M.
      • Hansen P.S.
      • Overgaard J.
      • et al.
      Postoperative radiotherapy in high-risk premenopausal women with breast cancer who receive adjuvant chemotherapy. Danish Breast Cancer Cooperative Group 82b Trial.
      ,
      • Overgaard M.
      • Jensen M.B.
      • Overgaard J.
      • et al.
      Postoperative radiotherapy in high-risk postmenopausal breast-cancer patients given adjuvant tamoxifen: Danish Breast Cancer Cooperative Group DBCG 82c randomised trial.
      ,

      Ebctcg Early Breast Cancer Trialists’ Collaborative Group. Effect of radiotherapy after mastectomy and axillary surgery on 10-year recurrence and 20-year breast cancer mortality: meta-analysis of individual patient data for 8135 women in 22 randomised trials. Lancet 2014. http://dx.doi.org/10.1016/S0140-6736(14)60488-8. [Epub ahead of print].

      ]. The progress in breast cancer treatment with prolonged survival-times [
      • Darby S.
      • McGale P.
      • Correa C.
      • et al.
      Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomised trials.
      ,
      • Engholm G.
      • Gislum M.
      • Bray F.
      • Hakulinen T.
      Trends in the survival of patients diagnosed with cancer in the Nordic countries 1964–2003 followed up to the end of 2006. Material and methods.
      ] has resulted in an increased awareness as of treatment induced second cancers. It has been shown in several epidemiological studies that irradiated breast cancer patients have an increased risk of second lung cancer [
      • Roychoudhuri R.
      • Evans H.
      • Robinson D.
      • Moller H.
      Radiation-induced malignancies following radiotherapy for breast cancer.
      ,

      Berrington de Gonzalez A, Curtis RE, Gilbert E, et al. Second solid cancers after radiotherapy for breast cancer in SEER cancer registries. Br J Cancer 2010; 102: 220–6.

      ,
      • Grantzau T.
      • Mellemkjaer L.
      • Overgaard J.
      Second primary cancers after adjuvant radiotherapy in early breast cancer patients: a national population based study under the Danish Breast Cancer Cooperative Group (DBCG).
      ,
      • Schaapveld M.
      • Visser O.
      • Louwman M.J.
      • et al.
      Risk of new primary nonbreast cancers after breast cancer treatment: a Dutch population-based study.
      ]. However, uncertainty still exists about the dose–response relationship for second cancers in the context of high dose radiation therapy.
      In a population-based cohort of breast cancer patients treated according to the guidelines of the Danish Breast Cancer Cooperative Group (DBCG) from 1982 to 2007, we have previously estimated the risk of developing a radiation induced second lung cancer among ⩾1 year survivors of first primary loco-regional breast cancer to approximately 1 in every 200 irradiated women [
      • Grantzau T.
      • Mellemkjaer L.
      • Overgaard J.
      Second primary cancers after adjuvant radiotherapy in early breast cancer patients: a national population based study under the Danish Breast Cancer Cooperative Group (DBCG).
      ]. To test the hypothesis, that there is a dose–response relationship of second primary lung cancer after breast cancer irradiation, we conducted a nested and matched case-control study with individual dose estimations, containing patients from this cohort. We further wished to estimate the excess relative risk per delivered Gray to the lung tumor. As smoking is strongly correlated to lung cancer, we also looked into the association between the effect of radiation and of smoking.
      To date, only two previous epidemiological studies have published dose–response relations for second lung cancer according to the absorbed lung dose after radiation therapy, with only one smaller study referring to breast cancer patients [
      • Inskip P.D.
      • Stovall M.
      • Flannery J.T.
      Lung cancer risk and radiation dose among women treated for breast cancer.
      ,
      • Gilbert E.S.
      • Stovall M.
      • Gospodarowicz M.
      • et al.
      Lung cancer after treatment for Hodgkin’s disease: focus on radiation effects.
      ]. We therefore conducted this large case-control study of second primary lung cancer among breast cancer patients with long-term follow-up after irradiation of the breast. The individual dose–response estimates were based exclusively on patients treated on linear accelerators; thereby obtaining risk estimates reflecting more contemporary treatment techniques.

      Materials and methods

      Data were supplied from the clinical database under the DBCG. Since 1977, all patients in Denmark diagnosed with primary loco-regional breast cancer have been treated according to uniform and national DBCG-guidelines [
      • Moller S.
      • Jensen M.B.
      • Ejlertsen B.
      • et al.
      The clinical database and the treatment guidelines of the Danish Breast Cancer Cooperative Group (DBCG); its 30-years experience and future promise.
      ]. We conducted a matched case-control study within a Danish national and population-based cohort of 23,627 female one-year survivors aged 20 years or older, treated for early breast cancer between 1982 and 2007. Only patients diagnosed with primary loco-regional breast cancer, no previous malignancy, unilateral breast cancer and surgery performed according to DBCG guidelines were included. We have previously quantified the risk of second solid cancers after radiotherapy in this cohort [
      • Grantzau T.
      • Mellemkjaer L.
      • Overgaard J.
      Second primary cancers after adjuvant radiotherapy in early breast cancer patients: a national population based study under the Danish Breast Cancer Cooperative Group (DBCG).
      ]. For the current nested case–control study all women within this cohort treated with postoperative radiotherapy on linear accelerators according to DBCG guidelines were considered eligible for inclusion. Only the first second primary lung cancer was included in the analysis. Further, women were censored from analysis at date of recurrence or diagnosis of a contra-lateral breast cancer in order to avoid influence from additional treatments given for subsequent cancers.
      Second lung cancers were identified through linkage to the national Danish Cancer Registry [
      • Gjerstorff M.L.
      The Danish Cancer Registry.
      ]. Irradiated patients who subsequently developed a second primary lung cancer were considered as cases. A total of 184 incident cases of second lung cancer were identified in the cohort (Fig. 1). 16 cases were excluded due to missing histopathology of the lung cancer diagnosis or the medical records did not indicate a second lung cancer, two cases due to treatment with a mixture of megavoltage and orthovoltage x-ray and finally 15 cases as medical records were missing. The follow-up for each patient began one year after the initial breast cancer diagnosis (1983–2008) and ended at the date of a second primary cancer diagnosis, recurrence of breast cancer, date of death, date of emigration or end of follow-up (December 31, 2008), whichever occurred first.
      Figure thumbnail gr1
      Fig. 1Flowchart of the included patients. Abbreviations: 1DBCG: The Danish Breast Cancer Cooperative Group; 2RT: radiotherapy,3BC: breast cancer.
      For each case, we attempted to select three controls by incidence density sampling among the 23,627 radiotherapy treated patients. Cases and controls were matched according to the exact age at breast cancer diagnosis, year of breast cancer treatment (±5 years) and cancer free survival for at least as long as the index case. The final study included 151 cases and 443 matched controls.

      Data collection

      Information on the initial breast cancer surgery either mastectomy or breast conserving surgery (BCS), tumor characteristics and adjuvant treatment were extracted from the DBCG database and validated through the hospital medical records for each included patient. For each case and control, the hospital medical records and the full radiotherapy charts were carefully reviewed. Information on patient height and weight around breast cancer diagnosis (±5 years) and smoking status approximately 1 year before lung cancer diagnosis for cases (or comparable date for the controls) were abstracted. Patients were assigned to one of the following smoke categories: never, ever-smoker or unknown smoking status. An estimated average amount of cigarette smoking at breast cancer diagnosis was also noted.

      Radiation dosimetry

      The objective was to estimate the dose to the center of the second lung tumor among the cases and the comparable anatomical location for the matched controls. For each patient the full radiotherapy record of the initial breast cancer treatment was requested. Treatment techniques, number of fields, beam energy, fractionation, delivered dose and monitor units and ±boost were extracted and reconstructed individually for each case and control. To locate the anatomical location of the second lung cancer, CT-scans and X-rays taken at the time of lung cancer diagnosis were collected, including the written reports of these examinations, supplemented with medical records, histopathology, bronchoscopy, and autopsy reports. As the majority of patients had been treated before CT-based radiotherapy planning became common, doses were reconstructed on recently scanned breast cancer patients matched according to Body Mass Index (BMI), breast cancer laterality and type of breast cancer surgery. A total of 13 different treatment techniques were each reconstructed, on 10 left-sided and 10 right-sided breast cancer patients, according to the national guidelines of DBCG [
      • Overgaard M.
      • Christensen J.J.
      Postoperative radiotherapy in DBCG during 30 years. Techniques, indications and clinical radiobiological experience.
      ,
      • Thorsen L.B.
      • Thomsen M.S.
      • Overgaard M.
      • Overgaard J.
      • Offersen B.V.
      Quality assurance of conventional non-CT-based internal mammary lymph node irradiation in a prospective Danish Breast Cancer Cooperative Group trial: the DBCG-IMN study.
      ]. Two different boost techniques were used in the period. They were also reconstructed on an individual basis according to the field size and the anatomical location of the boost. (For further information on the radiation dosimetry and treatment techniques see Supplementary Appendix and Supplementary Table 1.)

      Statistical analysis

      Conditional logistic regression analysis was used to obtain the relative risk (RR) of second lung cancer associated with the radiation dose, comparing cases with matched controls. Analyses were carried out using STATA IC 11.2 (Statistical Software, TX, USA) for doses used as a categorical variable. The RR’s were adjusted for smoking status and systemic adjuvant treatment. Estimates were based on odds ratios (OR) that closely approximate the RR. Since all patients had been treated with radiotherapy, estimated radiation doses of <1 Gray were used as a reference in the categorical analysis. The PECAN module of the EPICURE software [
      • Preston D.L.
      • et al.
      Epicure: Users's Guide (manual).
      ] was used for conditional logistic regression analyses of the excess relative risk per absorbed Gray to the lung, with dose used as a continuous variable. Confidence intervals were here based on profile likelihood methodology to improve the reliability of the results. The dose–response relationship was tested both in a linear- and in a linear quadratic model. When comparing the linear to the linear quadratic model (p = 0.36) we found no better fit of the latter, hence calculations are based on linear functions. All tests are two tailed and p-values less than 0.05 considered significant. p-Values were based on likelihood ratio tests.

      Results

      A total of 151 cases diagnosed with first primary lung cancer and 443 controls were eligible for analysis (Fig. 1). The median time from breast cancer treatment to second lung cancer diagnosis was 12 years (range 1–26 years). The median age at breast cancer diagnoses was 54 years (range 34–74) among cases and controls, and the median age at second lung cancer diagnosis was 68 years (range 46–90) (Table 1).
      Table 1Descriptive characteristics of patients treated for first primary breast cancer who subsequently developed a second primary lung cancer ⩾1 year after breast cancer treatment and matched controls.
      CharacteristicsCasesControls
      Cases and controls were matched according to age at breast cancer diagnosis, year of treatment and cancer free survival, for the controls at least as long as the matched cases.
      p-Value
      No.%
      Percentages may not sum to 100 due to rounding.
      No.%
      Radiotherapy treated patients1512544375
      Age at breast cancer diagnosis (mean and range; years)54(34–74)54(34–74)
      Age at second lung cancer diagnosis (mean and range; years)68(46–90)
      BMI at breast cancer diagnosis (kg/m2) (mean and range)24.2(16.5–35.5)24·1(16.7–38.5)
      Time from BC diagnosis to second lung cancer diagnosis (median and range; years)12.0(1.1–25.8)
      Radiation dose to lung tumor/comparable site for controls (Gy) (mean and range Gy)
      Gy: Gray.
      8.7(0.04–52.2)5.6(0.01–52.3)p = 0.01
      Year of breast cancer treatment
       1982–198621145312
       1987–199124167316
       1992–1996412712027
       1997–2001332210023
       2002–200732219722
      Smoke
       Never9619544p < 0.0001
       Ever smoke1379117740
       Unknown537116
      Adjuvant systemic treatment
       None
      Includes 4 cases and 8 controls treated with ovarian castration.
      684519043p = 0.8
       Endocrine therapy
      Includes patients treated with tamoxifen alone (34 cases, 112 controls) and patients treated with tamoxifen and an aromatase inhibitor (11 cases, 34 controls).
      453014633
       Chemotherapy
      Includes patients who were treated with chemotherapy±endocrine therapy. Includes CEF and CMF regimes. CEF: cyclophosphamide, epirubicin and fluorouracil (8 cases, 29 controls (includes 3 cases and 10 controls who received additional tamoxifen)).CMF: cyclophosphamide, methotrexate and fluorouracil (30 cases, 78 controls).
      382510724
      Surgery
       Mastectomy573817239p = 0.8
       Breast conserving surgery946227161
      Laterality of breast cancer
       Right805320145p = 0.1
       Left714724255
      Lung cancer laterality relative to breast cancer
       Ipsilateral8556
       Contra-lateral6342
       Center32
      Histopathology
      Includes the following histological codes according to the International Classification of Diseases for Oncology (ICD-10); squamous cell carcinoma: 8070; adenocarcinoma: 8140, 8250, 8560; small-cell carcinoma: 8041, 8042, 8045, 8246; large-cell carcinoma: 8012, 8013; NSCL (none-small cell lung cancer): 8046; other/carcinoma NOS: 8010, 8240.
       Squamous cell2517
       Small-cell carcinoma4630
       Large-cell carcinoma75
       Adenocarcinoma4932
       NSCL1711
       Other/Carcinoma NOS75
      a Cases and controls were matched according to age at breast cancer diagnosis, year of treatment and cancer free survival, for the controls at least as long as the matched cases.
      b Percentages may not sum to 100 due to rounding.
      c Gy: Gray.
      d Includes 4 cases and 8 controls treated with ovarian castration.
      e Includes patients treated with tamoxifen alone (34 cases, 112 controls) and patients treated with tamoxifen and an aromatase inhibitor (11 cases, 34 controls).
      f Includes patients who were treated with chemotherapy ± endocrine therapy. Includes CEF and CMF regimes. CEF: cyclophosphamide, epirubicin and fluorouracil (8 cases, 29 controls (includes 3 cases and 10 controls who received additional tamoxifen)).CMF: cyclophosphamide, methotrexate and fluorouracil (30 cases, 78 controls).
      g Includes the following histological codes according to the International Classification of Diseases for Oncology (ICD-10); squamous cell carcinoma: 8070; adenocarcinoma: 8140, 8250, 8560; small-cell carcinoma: 8041, 8042, 8045, 8246; large-cell carcinoma: 8012, 8013; NSCL (none-small cell lung cancer): 8046; other/carcinoma NOS: 8010, 8240.
      Overall cases and controls were well balanced according to type of breast surgery, breast cancer laterality and adjuvant systemic treatment (Table 1). However, the distribution according to smoking status showed significantly more smokers among cases compared to controls: 91% vs. 40%, with 3% of the cases and 16% of the controls having missing information regarding smoking. The overall mean absorbed radiation dose to the lung tumor was 8.7 Gray (range 0.04–52.2) for the cases and 5.6 Gray (range 0.01–52.3) at the comparable site for controls (Table 1). A total of 105 (70%) of the lung cancers were diagnosed five or more years after radiation therapy (range 5–26 years) and 61 (40%) ten or more years after treatment. For second lung cancers diagnosed five or more years after breast cancer treatment 59% developed in the ipsilateral, 38% in the contra-lateral lung and 3% in the center, respectively. Ten or more years after breast cancer treatment these numbers were 62%, 36% and 2%, respectively. As it is estimated to take minimum 5–9 years to develop a radiation induced second cancer [

      IARC, IARC. Ionizing radiation, part 1: X- and gamma (g)-radiation, and neutrons. IARC monographs on the evaluation of carcinogenic risks to humans, World Health Organization International Agency for Research on Cancer, Lyon; 2000.

      ] most of the following results are based on patients with a treatment-lag of five or more years between radiation and second lung cancer diagnosis.
      Table 2 shows the odds ratio (OR) according to the absorbed radiation dose alone and adjusted for smoking status and adjuvant systemic treatment, for doses used in a categorical model. For patients diagnosed with a second lung cancer five or more years after breast cancer treatment the adjusted risk of lung cancer was more than 3-fold increased for doses of 15 Gray or more. Among patients with a treatment-lag of ten or more years the adjusted risk was a significantly 6-fold increased for doses of 25 Gray or more (OR 6.27; 95% confidence interval [CI] = 1.1–34.8). Being an ever smoker gave a 25-fold increased risk of lung cancer compared to never smokers. This was even more profound for patients with a history of consuming more than one pack of cigarettes a day (OR 47; 95% CI = 12–184), though estimates are uncertain due to few numbers of non-smokers among the cases.
      Table 2Odds ratio (OR) of lung cancer among patients treated for early breast cancer according to radiation dose, tobacco consumption and adjuvant systemic treatment.
      CasesControlsCrude rates OR (95% CI)
      Abbreviations: OR: Odds Ratio with corresponding 95% CI, confidence interval.
      Adjusted rates OR (95% CI)
      OR’s were adjusted for smoking status and systemic adjuvant treatment.
      Radiation dose to lung tumor and comparable site for controls (Gy)
      Gy: Gray.
      <5 years treatment lag between BC
      BC: breast cancer.
      diagnosis and second lung cancer diagnosis
       <1 (Gy)18621.0 (reference)1.0 (reference)
       1–418471.46 (0.62–3.82)2.47 (0.90–6.74)
       5–146171.29 (0.41–4.09)1.51 (0.43–5.27)
       15–24242.24 (0.27–18.51)7.47 (0.31–183.03)
       ⩾25261.24 (0.16–9.75)2.81 (0.22–35.70)
      p-Value for trend0.5180.214
      ⩾5 years treatment lag between BC diagnosis and second lung cancer diagnosis
       <1 (Gy)361161.0 (reference)1.0 (reference)
       1–4261090.77 (0.44–1.35)0.63 (0.31–1.26)
       5–1415451.06 (0.51–2.17)0.83 (0.32–2.15)
       15–2411152.78 (1.11–7.00)3.83 (1.24–11.85)
       ⩾2517223.76 (1.48–9.51)3.45 (1.02–11.67)
      p-Value for trend0.0030.016
      ⩾10 years treatment lag between BC diagnosis and second lung cancer diagnosis
       <1 (Gy)21681.0 (reference)1.0 (reference)
       1–410600.51 (0.22–1.18)0.34 (0.11–1.06)
       5–148241.02 (0.35–3.00)0.76 (0.18–3.30)
       15–248103.24 (1.02–10.35)3.84 (0.98–14.98)
       ⩾2514144.53 (1.51–13.63)6.27 (1.13–34.80)
      p-Value for trend0.0020.014
      Adjuvant treatment if ⩾5 years treatment lag between BC diagnosis and second lung cancer diagnosis
       None471451.0 (reference)
       Endocrine therapy27741.16 (0.60–2.23)
       Chemotherapy
      Includes patients who were treated with either chemotherapy alone or chemotherapy and endocrine therapy.
      31881.06 (0.55–2.03)
      Smoking status if ⩾5 years treatment lag between BC diagnosis and second lung cancer diagnosis
       Never61351.0 (reference)
       Ever9611425.34 (8.66–74.12)
       Unknown3581.40 (0.30–6.56)
      Smoking status if ⩾5 years treatment lag between BC diagnosis and second lung cancer diagnosis
       Never61351.0 (reference)
       Ever ⩽one pack per day808527.40 (9.12–82.37)
       Ever >one pack per day14946.91 (11.96–184.05)
       Smoker; but unknown amount2203.11 (0.51–19.06)
       Unknown tobacco status3581.32 (0.28–6.14)
      a Abbreviations: OR: Odds Ratio with corresponding 95% CI, confidence interval.
      b OR’s were adjusted for smoking status and systemic adjuvant treatment.
      c Gy: Gray.
      d BC: breast cancer.
      e Includes patients who were treated with either chemotherapy alone or chemotherapy and endocrine therapy.
      The excess relative risk per Gray (ERR per Gy) among patients with a treatment-lag of 5 years or more with dose used as a continuous variable is shown in Table 3. Among these patients the rate of second lung cancer increased overall with 8.5% per Gray (95% CI = 3.1–23.3%; p < 0.001) using a linear model (Fig. 2).
      Table 3Excess relative risk per Gy according to smoking status, adjuvant treatment and breast cancer surgery among patients with 5 or more years between breast cancer treatment and second lung cancer diagnosis.
      Time since treatment ⩾5 years
      Time since radiotherapy and second lung cancer diagnosis for cases; for controls cancer free survival for at least as long as the index case.
      CasesControlsERR/Gy (95% CI)
      ERR/Gy, CI: excess relative risk per gray and corresponding 95% confidence interval. Estimates based on radiation doses as a continuous variable.
      Two sided p-value for testing ERR/Gy = 0
      RT
      RT: radiotherapy.
      treated all patients, linear model
      1053070.085 (0.031–0.233)<0.001
      Smoking status
      The following estimates are based on a linear dose–response model.
       Never/unknown91930.006 (−0.020–0.163)>0.5
       Ever961140.173 (0.045–0.540)<0.005
      p-Value for homogeneity0.08
      Adjuvant treatment
       None471450.003 (−0.017–0.089)>0.5
       Endocrine therapy27740.326 (0.065–2.731)<0.005
      Chemotherapy
      Includes patients who were treated with either chemotherapy alone or chemotherapy and endocrine therapy.
      31880.091 (0.007–0.316)0.02
      p-Value for homogeneity0.04
      Surgery type
       Mastectomy441250.090 (0.018–0.291)<0.005
       Breast conserving surgery611820.100 (0.011–0.337)<0.025
      p-Value for homogeneity>0.5
      a Time since radiotherapy and second lung cancer diagnosis for cases; for controls cancer free survival for at least as long as the index case.
      b ERR/Gy, CI: excess relative risk per gray and corresponding 95% confidence interval. Estimates based on radiation doses as a continuous variable.
      c RT: radiotherapy.
      d The following estimates are based on a linear dose–response model.
      e Includes patients who were treated with either chemotherapy alone or chemotherapy and endocrine therapy.
      Figure thumbnail gr2
      Fig. 2Estimated excess risk and 95% confidence interval of second primary lung cancer according to estimated radiation dose (Gray (Gy)), among patients diagnosed with a second lung cancer five or more years after breast cancer treatment. The value of the solid line was calculated from individual dose estimation among 105 control and 307 cases with dose used as a continuous variable, based on a linear dose–response relationship. Cases and controls were matched according to age at breast cancer diagnosis, year of breast cancer (±5 years) treatment and cancer free survival for at least as long as the index case. The dots are the risk estimates according to the dose categories <1, 1–4, 5–14, 15–24 and ⩾25 Gy, respectively adjusted for smoking status and systemic adjuvant treatment.
      The risk of lung cancer according to smoking status is shown in Table 3. For ever smokers the excess relative risk increased with 17.3% per delivered Gray to the lung (95% CI = 4.5–54%; p < 0.005. Never smokers and patients with an unknown smoking status had to be combined due to small numbers of none smokers among the cases. The excess relative risk for this small group of patients was 0.6% per Gray, but did not differ significantly from zero (p > 0.5).
      There was an increased excess relative risk per Gray both among irradiated patients treated with a mastectomy and patients treated with BCS (p < 0.005 and p < 0.025) and we observed no difference in risk between the two surgery groups (p-value for homogeneity >0.5) (Table 3).
      The crude effect of adjuvant endocrine therapy and adjuvant chemotherapy as to second lung cancer was not significantly increased, OR for endocrine therapy 1.16 (95% CI = 0.60–2.23) and OR for chemotherapy ± endocrine therapy 1.06 (95% CI = 0.55–2.03) (Table 2). In the linear dose response model with dose used as a continuous variable systemic adjuvant treatment was significantly associated with second lung cancer with an excess relative risk per Gray of 9% for chemotherapy ± endocrine therapy (p = 0.02) and an excess relative risk per Gray of 33% for endocrine therapy (p < 0.005) (Table 3).

      Discussion

      To our knowledge, this is the largest case–control study on second lung cancer after breast cancer irradiation with individual dose estimations to the anatomical location of the second lung tumor with included dose–response estimations. Risk estimates were based on patients treated exclusively on linear accelerators. Radiation doses to the lung of 15 Gray or more was associated with a more than 3-fold increased risk among patients with a treatment-lag of 5 and 10 years or more between breast cancer radiation and lung cancer diagnosis. With dose used as a continuous variable the excess relative risk per Gray among patients with a treatment lag of five or more years after breast cancer treatment increased linearly with 8.5% per Gray (95% CI = 3.1–23.3%; p < 0.001). This rate was enhanced for ever smokers with an excess rate of 17.3% per Gray (95% CI = 4.5–54%; p < 0.005).
      The strength of this study is its population based nature, covering up to 26 years of follow-up. All patients in Denmark diagnosed and treated for early breast cancer according to DBCG guidelines during the study period were included and none of the cohort members were lost to follow-up. Further, controls were selected by incidence density sampling; all giving rise to valid and unbiased estimates. Risk estimates were based on individual dose estimations to the center of the second tumor, enabling better risk estimates than mean doses to the lung. Finally, only patients diagnosed with a histologically confirmed first second primary lung cancer were included, eliminating risk estimates based on misclassified breast cancer metastasis and eliminating the influence of other subsequently cancer treatments.
      As the majority of patients had been treated before the area of CT-based radiotherapy planning, it was not possible to account for individual organ anatomy, as the dose reconstructions had to be performed in a series of recently scanned patients. In order to consider some of the inter-patient variability we matched the BMI of the cases and controls to the BMI of the reference scanned patients with the same breast cancer laterality and breast cancer surgery. Despite having CT-scans and X-rays on the majority of the second lung cancers, determining the exact center of the lung tumor for the dose determination, could be challenging, especially when the tumors were located in or near the radiation field. It should be noted that all patients in this study received postoperative radiotherapy and the reference group are patients receiving <1 Gy, hence risk estimates are based on different levels of radiation exposure.
      Data on smoking were limited by several factors. Information was only collected through medical records and did not include information from the patients or their private physicians. Further, smoking habits are often confirmed more thoroughly among patients with smoking-related illnesses, reflected by that 3% of the cases and 16% of controls had missing information on smoking status. The proportion of female smokers in Denmark is high with an estimated prevalence of 45% in 1985 [

      The Danish Health and Medicines Authority. Updated may 2014. Ref Type: Electronic citation http://sundhedsstyrelsen.dk/en.

      ]. This is close to the number of smokers among controls, suggesting that bias by the unknown smoking status among the controls was negligible.
      Patients with a history of cigarette smoking had a significant increased radiation-induced risk of second lung cancer with an estimated excess rate per Gy of 17%. The combined group of non-smokers and patients with an unknown smoking status had no increased risk; however there was no statistically significant difference between the smoking categories (p = 0.08). This was likely due to very few non-smokers among the cases (6%) and it was therefore not possible to look into the interaction of smoking and radiation. However, two previous studies found that the interaction of radiation and smoking was consistent with a multiplicative effect. This included one study on second lung cancer after radiotherapy for Hodgkin’s disease [
      • Gilbert E.S.
      • Stovall M.
      • Gospodarowicz M.
      • et al.
      Lung cancer after treatment for Hodgkin’s disease: focus on radiation effects.
      ] and one on second esophagus cancer after radiotherapy for breast cancer [
      • Morton L.M.
      • Gilbert E.S.
      • Hall P.
      • et al.
      Risk of treatment-related esophageal cancer among breast cancer survivors.
      ]. Furthermore, Kaufman et al. [
      • Kaufman E.L.
      • Jacobson J.S.
      • Hershman D.L.
      • Desai M.
      • Neugut A.I.
      Effect of breast cancer radiotherapy and cigarette smoking on risk of second primary lung cancer.
      ] found that women treated with postoperative radiotherapy after breast cancer who were smokers had a nearly 19-fold increased risk of second lung cancers whereas women treated with radiotherapy who where never smokers had no increased risk. Combined these results emphasize the strong effect of smoking and second lung cancer and tobacco cessation should strongly be advised to all breast cancer patients.
      Two previous case–control studies have presented dose–response estimations of second lung cancer after radiation therapy. One study was based on breast cancer patients; including 61 cases and 120 controls [
      • Inskip P.D.
      • Stovall M.
      • Flannery J.T.
      Lung cancer risk and radiation dose among women treated for breast cancer.
      ]. The majority of these patients received either treatment with cobalt-60 or orthovoltage X-ray. Although none of the dose estimations were statistically significant, likely due to the small study size, a trend toward a linear dose–response relationship was observed with an excess relative risk per Gray of 20% (95% CI = −62–103%), based on the average dose to the cancerous lung. The other study included Hodgkin lymphoma patients; 227 cases of second lung cancer and 455 matched controls. Treatment techniques were primarily large chest fields with lung blocks or smaller fields to lymph node regions above the diaphragm [
      • Gilbert E.S.
      • Stovall M.
      • Gospodarowicz M.
      • et al.
      Lung cancer after treatment for Hodgkin’s disease: focus on radiation effects.
      ]. Doses were based on individual dose estimates to the center of the lung cancer and doses above 30 Gray were associated with an increased risk. The authors found a linear dose–response correlation with an estimated excess relative risk per Gray of 15% (95% CI = 6–39%). Risk estimates from both studies are reasonably in line with our findings, keeping in mind that the majority of Hodgkin’s lymphoma patients received doses to both lungs. Furthermore, both studies also found a linear dose–response relationship even in the high dose areas.
      In 2012 Berrington de Gonzalez et al. published a systematic review of 28 epidemiological studies which included dose–response estimations across 11 second solid cancer sites [

      Berrington de Gonzalez A, Gilbert E, Curtis R, et al. Second solid cancers after radiation therapy: a systematic review of the epidemiologic studies of the radiation dose–response relationship. Int J Radiat Oncol Biol Phys 2013; 86: 224–33.

      ]. For all second cancers except for second thyroid cancer the authors found a linear dose–response correlation, with no reduction or plateau in the risk, even at doses exceeding 40 Gy. Epidemiological studies with individually dose estimations are generally associated with large uncertainties [
      • Inskip P.D.
      • Stovall M.
      • Flannery J.T.
      Lung cancer risk and radiation dose among women treated for breast cancer.
      ,
      • Gilbert E.S.
      • Stovall M.
      • Gospodarowicz M.
      • et al.
      Lung cancer after treatment for Hodgkin’s disease: focus on radiation effects.
      ,

      Berrington de Gonzalez A, Gilbert E, Curtis R, et al. Second solid cancers after radiation therapy: a systematic review of the epidemiologic studies of the radiation dose–response relationship. Int J Radiat Oncol Biol Phys 2013; 86: 224–33.

      ,
      • Stovall M.
      • Weathers R.
      • Kasper C.
      • et al.
      Dose reconstruction for therapeutic and diagnostic radiation exposures: use in epidemiological studies.
      ]. This applies particularly for the highest dose-categories mainly due to the small number of cases. Despite these uncertainties epidemiological studies across different second cancer sites indicate a linear dose–response relationship; emphasizing the clinical importance of continues improvement of treatment techniques to minimize the dose to normal tissue without compromising the target.
      We observed that the use of chemotherapy was associated with an increased risk of second lung cancer in the linear dose–response model (p = 0.02). All patients treated with chemotherapy in this study received the alkylating agent cyclophosphamide in combination with other chemotherapeutics. Cyclophosphamide has been linked to second cancers [
      • Emadi A.
      • Jones R.J.
      • Brodsky R.A.
      Cyclophosphamide and cancer: golden anniversary.
      ] and alkylating agents to second lung cancer in Hodgkin lymphoma patients; where the largest risk was observed among patient treated with both radiotherapy and chemotherapy [
      • Travis L.B.
      • Gospodarowicz M.
      • Curtis R.E.
      • et al.
      Lung cancer following chemotherapy and radiotherapy for Hodgkin’s disease.
      ]. Valagussa et al. [
      • Valagussa P.
      • Moliterni A.
      • Terenziani M.
      • Zambetti M.
      • Bonadonna G.
      Second malignancies following CMF-based adjuvant chemotherapy in resectable breast cancer.
      ] however, found no association between chemotherapy (cyclophosphamide including regimes) for breast cancer and risk of second cancer. Neither did Morton et al. [
      • Morton L.M.
      • Gilbert E.S.
      • Hall P.
      • et al.
      Risk of treatment-related esophageal cancer among breast cancer survivors.
      ] as to second esophagus cancer after breast cancer, however few patients in this study received alkylating agent-containing chemotherapy.
      We further observed that patients with a treatment lag of 5 or more years receiving adjuvant endocrine therapy had an increased risk of second lung cancer (p < 0.005) compared to patients treated with no additional systemic treatment, using the linear dose–response model. It has been well established that radiation can cause both pneumonitis and pulmonary fibrosis and the risk seems related to dose, volume and treatment-specific factors [
      • Huang E.Y.
      • Wang C.J.
      • Chen H.C.
      • et al.
      Multivariate analysis of pulmonary fibrosis after electron beam irradiation for postmastectomy chest wall and regional lymphatics: evidence for non-dosimetric factors.
      ,
      • Mehta V.
      Radiation pneumonitis and pulmonary fibrosis in non-small-cell lung cancer: pulmonary function, prediction, and prevention.
      ]. A number of studies have linked tamoxifen to an enhancement of radiation induced fibrosis [
      • Huang E.Y.
      • Wang C.J.
      • Chen H.C.
      • et al.
      Multivariate analysis of pulmonary fibrosis after electron beam irradiation for postmastectomy chest wall and regional lymphatics: evidence for non-dosimetric factors.
      ,
      • Bentzen S.M.
      • Skoczylas J.Z.
      • Overgaard M.
      • Overgaard J.
      Radiotherapy-related lung fibrosis enhanced by tamoxifen.
      ,
      • Koc M.
      • Polat P.
      • Suma S.
      Effects of tamoxifen on pulmonary fibrosis after cobalt-60 radiotherapy in breast cancer patients.
      ] through the induction of transforming growth factor β (TGF-β) [
      • Colletta A.A.
      • Benson J.R.
      • Baum M.
      Alternative mechanisms of action of anti-oestrogens.
      ,
      • Butta A.
      • MacLennan K.
      • Flanders K.C.
      • et al.
      Induction of transforming growth factor beta 1 in human breast cancer in vivo following tamoxifen treatment.
      ], that has been associated with the pathogenesis of fibrosis [
      • Rubin P.
      • Johnston C.J.
      • Williams J.P.
      • McDonald S.
      • Finkelstein J.N.
      A perpetual cascade of cytokines postirradiation leads to pulmonary fibrosis.
      ]. Notably, an increased incidence of lung cancer has been reported among patients with pulmonary fibrosis in Interstitial Lung Diseases [
      • Ma Y.
      • Seneviratne C.K.
      • Koss M.
      Idiopathic pulmonary fibrosis and malignancy.
      ,
      • Hubbard R.
      • Venn A.
      • Lewis S.
      • Britton J.
      Lung cancer and cryptogenic fibrosing alveolitis. A population-based cohort study.
      ,
      • Zhang J.Q.
      • Wan Y.N.
      • Peng W.J.
      • et al.
      The risk of cancer development in systemic sclerosis: a meta-analysis.
      ]. It is possible that the increased risk seen among irradiated breast cancer patients treated with endocrine therapy could be linked to increased pulmonary fibrosis that again may predispose to second lung cancer.
      The clinical awareness toward radiation induced cardiac morbidity has resulted in new treatment techniques, such as respiratory gated radiotherapy that spare the heart, to enter clinical practice. McGale et al. estimated the risk of radiation-induced ischemic heart disease among +1 month survivors of primary breast cancer, treated according to DBCG-guidelines from 1976 to 2006 [
      • McGale P.
      • Darby S.C.
      • Hall P.
      • et al.
      Incidence of heart disease in 35,000 women treated with radiotherapy for breast cancer in Denmark and Sweden.
      ]. In this cohort the risk was approximately one in every 400 irradiated patients. This is equivalent to half the risk of developing a radiation induced second lung cancer among ⩾1 year survivors of breast cancer also estimated in a DBCG cohort study, treatment years 1982–2007 [
      • Grantzau T.
      • Mellemkjaer L.
      • Overgaard J.
      Second primary cancers after adjuvant radiotherapy in early breast cancer patients: a national population based study under the Danish Breast Cancer Cooperative Group (DBCG).
      ]. Darby et al. further estimated the dose–response correlation of radiation induced coronary events after radiotherapy for breast cancer among 2168 Swedish and Danish breast cancer patients, treated under DBCG guidelines (1958–2001) [
      • Darby S.C.
      • Ewertz M.
      • McGale P.
      • et al.
      Risk of ischemic heart disease in women after radiotherapy for breast cancer.
      ]. The risk increased linearly with the mean dose to the heart by 7.4% per Gray (95% CI = 2.9–14.5%). This indicates that any dose reduction to the heart would result in a decreased risk of radiation induced cardiac toxicity, equivalently to, any dose reduction to the lung would result in a decreased risk of second lung cancer, as shown in this study. Despite this, the clinical awareness toward radiation induced second cancers is surprisingly far from that of cardiac morbidity in breast cancer treatment.
      In conclusion, we found that the risk of second lung cancer after breast cancer radiotherapy increased linearly with 8.5% per Gray. With the growing number of long-time survivors the risk of radiation induced second cancers needs to be taken into account in established and emerging treatment protocols to reduce the incidence of treatment induced malignancies.

      Funding

      This work was supported by CIRRO (The Lundbeck Foundation Centre for Interventional Research in Radiation Oncology), Aarhus University; Faculty of Health Sciences , the Danish Medical Research Council , the Danish Cancer Society and the European Atomic Energy Community’s Seventh Framework Program (7/2007-2013) under Grant agreement n° 231965 ; ALLEGRO.

      Conflicts of interest statements

      No potential conflicts of interest.

      Acknowledgments

      We thank Dr. L Røhl for assistance in reviewing CT scans and X-rays, Dr. L.B.J. Thorsen for assistance with treatment plan reconstructions, PhD. J. Johansen TLD readings and the hospitals throughout Denmark for their support in data collection.

      Appendix A. Supplementary data

      References

      1. GLOBOCAN. GLOBOCAN. Updated May 2014. Ref Type: Electronic citation http://globocan.iarc.fr/.

        • Forouzanfar M.H.
        • Foreman K.J.
        • Delossantos A.M.
        • et al.
        Breast and cervical cancer in 187 countries between 1980 and 2010: a systematic analysis.
        Lancet. 2011; 378: 1461-1484
        • Clarke M.
        • Collins R.
        • Darby S.
        • et al.
        Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials.
        Lancet. 2005; 366: 2087-2106
        • Darby S.
        • McGale P.
        • Correa C.
        • et al.
        Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomised trials.
        Lancet. 2011; 378: 1707-1716
        • Overgaard M.
        • Hansen P.S.
        • Overgaard J.
        • et al.
        Postoperative radiotherapy in high-risk premenopausal women with breast cancer who receive adjuvant chemotherapy. Danish Breast Cancer Cooperative Group 82b Trial.
        N Engl J Med. 1997; 337: 949-955
        • Overgaard M.
        • Jensen M.B.
        • Overgaard J.
        • et al.
        Postoperative radiotherapy in high-risk postmenopausal breast-cancer patients given adjuvant tamoxifen: Danish Breast Cancer Cooperative Group DBCG 82c randomised trial.
        Lancet. 1999; 353: 1641-1648
      2. Ebctcg Early Breast Cancer Trialists’ Collaborative Group. Effect of radiotherapy after mastectomy and axillary surgery on 10-year recurrence and 20-year breast cancer mortality: meta-analysis of individual patient data for 8135 women in 22 randomised trials. Lancet 2014. http://dx.doi.org/10.1016/S0140-6736(14)60488-8. [Epub ahead of print].

        • Engholm G.
        • Gislum M.
        • Bray F.
        • Hakulinen T.
        Trends in the survival of patients diagnosed with cancer in the Nordic countries 1964–2003 followed up to the end of 2006. Material and methods.
        Acta Oncol. 2010; 49: 545-560
        • Roychoudhuri R.
        • Evans H.
        • Robinson D.
        • Moller H.
        Radiation-induced malignancies following radiotherapy for breast cancer.
        Br J Cancer. 2004; 91: 868-872
      3. Berrington de Gonzalez A, Curtis RE, Gilbert E, et al. Second solid cancers after radiotherapy for breast cancer in SEER cancer registries. Br J Cancer 2010; 102: 220–6.

        • Grantzau T.
        • Mellemkjaer L.
        • Overgaard J.
        Second primary cancers after adjuvant radiotherapy in early breast cancer patients: a national population based study under the Danish Breast Cancer Cooperative Group (DBCG).
        Radiother Oncol. 2013; 106: 42-49
        • Schaapveld M.
        • Visser O.
        • Louwman M.J.
        • et al.
        Risk of new primary nonbreast cancers after breast cancer treatment: a Dutch population-based study.
        J Clin Oncol. 2008; 26: 1239-1246
        • Inskip P.D.
        • Stovall M.
        • Flannery J.T.
        Lung cancer risk and radiation dose among women treated for breast cancer.
        J Natl Cancer Inst. 1994; 86: 983-988
        • Gilbert E.S.
        • Stovall M.
        • Gospodarowicz M.
        • et al.
        Lung cancer after treatment for Hodgkin’s disease: focus on radiation effects.
        Radiat Res. 2003; 159: 161-173
        • Moller S.
        • Jensen M.B.
        • Ejlertsen B.
        • et al.
        The clinical database and the treatment guidelines of the Danish Breast Cancer Cooperative Group (DBCG); its 30-years experience and future promise.
        Acta Oncol. 2008; 47: 506-524
        • Gjerstorff M.L.
        The Danish Cancer Registry.
        Scand J Public Health. 2011; 39: 42-45
        • Overgaard M.
        • Christensen J.J.
        Postoperative radiotherapy in DBCG during 30 years. Techniques, indications and clinical radiobiological experience.
        Acta Oncol. 2008; 47: 639-653
        • Thorsen L.B.
        • Thomsen M.S.
        • Overgaard M.
        • Overgaard J.
        • Offersen B.V.
        Quality assurance of conventional non-CT-based internal mammary lymph node irradiation in a prospective Danish Breast Cancer Cooperative Group trial: the DBCG-IMN study.
        Acta Oncol. 2013; 52: 1526-1534
        • Preston D.L.
        • et al.
        Epicure: Users's Guide (manual).
        HiroSoft International, Seattle WA, USA1988
      4. IARC, IARC. Ionizing radiation, part 1: X- and gamma (g)-radiation, and neutrons. IARC monographs on the evaluation of carcinogenic risks to humans, World Health Organization International Agency for Research on Cancer, Lyon; 2000.

      5. The Danish Health and Medicines Authority. Updated may 2014. Ref Type: Electronic citation http://sundhedsstyrelsen.dk/en.

        • Morton L.M.
        • Gilbert E.S.
        • Hall P.
        • et al.
        Risk of treatment-related esophageal cancer among breast cancer survivors.
        Ann Oncol. 2012; 23: 3081-3091
        • Kaufman E.L.
        • Jacobson J.S.
        • Hershman D.L.
        • Desai M.
        • Neugut A.I.
        Effect of breast cancer radiotherapy and cigarette smoking on risk of second primary lung cancer.
        J Clin Oncol. 2008; 26: 392-398
      6. Berrington de Gonzalez A, Gilbert E, Curtis R, et al. Second solid cancers after radiation therapy: a systematic review of the epidemiologic studies of the radiation dose–response relationship. Int J Radiat Oncol Biol Phys 2013; 86: 224–33.

        • Stovall M.
        • Weathers R.
        • Kasper C.
        • et al.
        Dose reconstruction for therapeutic and diagnostic radiation exposures: use in epidemiological studies.
        Radiat Res. 2006; 166: 141-157
        • Emadi A.
        • Jones R.J.
        • Brodsky R.A.
        Cyclophosphamide and cancer: golden anniversary.
        Nat Rev Clin Oncol. 2009; 6: 638-647
        • Travis L.B.
        • Gospodarowicz M.
        • Curtis R.E.
        • et al.
        Lung cancer following chemotherapy and radiotherapy for Hodgkin’s disease.
        J Natl Cancer Inst. 2002; 94: 182-192
        • Valagussa P.
        • Moliterni A.
        • Terenziani M.
        • Zambetti M.
        • Bonadonna G.
        Second malignancies following CMF-based adjuvant chemotherapy in resectable breast cancer.
        Ann Oncol. 1994; 5: 803-808
        • Huang E.Y.
        • Wang C.J.
        • Chen H.C.
        • et al.
        Multivariate analysis of pulmonary fibrosis after electron beam irradiation for postmastectomy chest wall and regional lymphatics: evidence for non-dosimetric factors.
        Radiother Oncol. 2000; 57: 91-96
        • Mehta V.
        Radiation pneumonitis and pulmonary fibrosis in non-small-cell lung cancer: pulmonary function, prediction, and prevention.
        Int J Radiat Oncol Biol Phys. 2005; 63: 5-24
        • Bentzen S.M.
        • Skoczylas J.Z.
        • Overgaard M.
        • Overgaard J.
        Radiotherapy-related lung fibrosis enhanced by tamoxifen.
        J Natl Cancer Inst. 1996; 88: 918-922
        • Koc M.
        • Polat P.
        • Suma S.
        Effects of tamoxifen on pulmonary fibrosis after cobalt-60 radiotherapy in breast cancer patients.
        Radiother Oncol. 2002; 64: 171-175
        • Colletta A.A.
        • Benson J.R.
        • Baum M.
        Alternative mechanisms of action of anti-oestrogens.
        Breast Cancer Res Treat. 1994; 31: 5-9
        • Butta A.
        • MacLennan K.
        • Flanders K.C.
        • et al.
        Induction of transforming growth factor beta 1 in human breast cancer in vivo following tamoxifen treatment.
        Cancer Res. 1992; 52: 4261-4264
        • Rubin P.
        • Johnston C.J.
        • Williams J.P.
        • McDonald S.
        • Finkelstein J.N.
        A perpetual cascade of cytokines postirradiation leads to pulmonary fibrosis.
        Int J Radiat Oncol Biol Phys. 1995; 33: 99-109
        • Ma Y.
        • Seneviratne C.K.
        • Koss M.
        Idiopathic pulmonary fibrosis and malignancy.
        Curr Opin Pulm Med. 2001; 7: 278-282
        • Hubbard R.
        • Venn A.
        • Lewis S.
        • Britton J.
        Lung cancer and cryptogenic fibrosing alveolitis. A population-based cohort study.
        Am J Respir Crit Care Med. 2000; 161: 5-8
        • Zhang J.Q.
        • Wan Y.N.
        • Peng W.J.
        • et al.
        The risk of cancer development in systemic sclerosis: a meta-analysis.
        Cancer Epidemiol. 2013; 37: 523-527
        • McGale P.
        • Darby S.C.
        • Hall P.
        • et al.
        Incidence of heart disease in 35,000 women treated with radiotherapy for breast cancer in Denmark and Sweden.
        Radiother Oncol. 2011; 100: 167-175
        • Darby S.C.
        • Ewertz M.
        • McGale P.
        • et al.
        Risk of ischemic heart disease in women after radiotherapy for breast cancer.
        N Engl J Med. 2013; 368: 987-998