Effects of Smoking and Radiotherapy on Lung Carcinoma in Breast Carcinoma Survivors

Transcription

1 1457 Effects of Smoking and Radiotherapy on Lung Carcinoma in Breast Carcinoma Survivors Melissa B. Ford, Ph.D. 1 Alice J. Sigurdson, Ph.D. 1 Elaine S. Petrulis, M.P.H. 1 Chaan S. Ng, M.D. 2 Bonnie Kemp, M.D. 3 Catherine Cooksley, Ph.D. 4 Marsha McNeese, M.D. 5 Beatrice J. Selwyn, Sc.D. 6 Margaret R. Spitz, M.D. 1 Melissa L. Bondy, Ph.D. 1 1 Department of Epidemiology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas. 2 Department of Diagnostic Radiation, The University of Texas M. D. Anderson Cancer Center, Houston, Texas. 3 Department of Pathology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas. 4 Department of Biostatistics, The University of Texas M. D. Anderson Cancer Center, Houston Texas. 5 Department of Radiation Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas. 6 The University of Texas School of Public Health, Houston, Texas. Melissa J. Ford s current address: Westat, Rockville, Maryland. Alice J. Sigurdson s current address: Division of Cancer Epidemiology and Genetics, Radiation Epidemiology Branch, National Cancer Institute, Bethesda, Maryland. Bonnie Kemp s current address: Memorial-Hermann Hospital System, Houston, Texas. Supported in part by National Cancer Institute Grant R25 CA The authors thank Dr. Karen Goodman and Dr. Robert Hardy (The University of Texas School of Public Health, Houston, Texas) for their assistance on Dr. Melissa Ford s dissertation research. They also thank Simone Morris, Michelle Benderly, and BACKGROUND. The combined effects of thoracic radiotherapy (XRT) and cigarette smoking are not known with certainty, but they have important implications for lung carcinogenesis after cancer therapy in some patients. The authors analyzed smoking, radiation, and both exposures on lung carcinoma development in women who were treated previously for breast carcinoma. METHODS. Case patients (n 280) were female residents of the United States, ages years, with breast carcinoma prior to primary lung carcinoma diagnosed between 1960 and Control patients (n 300) were selected randomly from 37,000 patients with breast carcinoma who were treated at The University of Texas M. D. Anderson Cancer Center and frequency matched women in the case group based on age at diagnosis (5-year strata), ethnicity, year of breast carcinoma diagnosis (5-year strata), and survival from breast carcinoma diagnosis to lung carcinoma diagnosis. Using stratified analysis and unconditional logistic regression, the authors evaluated the main and combined effects of smoking and XRT on lung carcinoma risk. RESULTS. At the time of breast carcinoma diagnosis, 84% of case patients had ever smoked cigarettes, compared with 37% of control patients, whereas 45% of case patients and control patients received XRT for breast carcinoma. Smoking increased the odds of lung carcinoma in women without XRT (odds ratio [OR], 6.0; 95% confidence interval [95% CI], ), but XRT did not increase lung carcinoma risk in nonsmoking women (OR, 0.5; 95% CI, ). Overall, the OR for both XRT and smoking, compared with no XRT or smoking, was 9.0 (95% CI, ). Logistic regression modeling yielded an adjusted OR of 5.6 for the smoking main effect (95% CI, ), 0.6 for the XRT main effect (95% CI, ), and 8.6 (P 0.08) for the combined effect. CONCLUSIONS. Smoking was a significant independent risk factor for lung carcinoma after breast carcinoma, but XRT alone was not. Smoking and XRT combined enhanced the effect of either alone, with marked increased risks of lung carcinoma after XRT for breast carcinoma. Cancer 2003;98: American Cancer Society. KEYWORDS: breast carcinoma, second primary malignancies, radiotherapy, lung carcinoma, cigarette smoking. Scientists have investigated the genesis of multiple primary malignancies for over a century, and such studies provide a needed strategy for exploring the relations of exposures and host susceptibil- Ervin Bhogte for their help with the medical record abstraction and medical record requests to other institutions. Address for reprints: Melissa L. Bondy, Ph.D., Department of Epidemiology, Box 189, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030; Fax: (713) ; mbondy@mdanderson.org Received March 25, 2003; revision received June 23, 2003; accepted June 30, American Cancer Society DOI /cncr.11669

2 1458 CANCER October 1, 2003 / Volume 98 / Number 7 ity factors. 1 Given improving survival rates, interest has turned to whether treatment, especially radiotherapy (XRT), may increase patient risk for a second primary malignancy. 2,3 It has been suspected that patients with breast carcinoma are at increased risk of subsequent primary malignancies, specifically lung carcinoma, especially if the patients smoke cigarettes. 4 8 In addition to smoking, radiation is a known carcinogen. To assess the relative role of each exposure, this research analyzed a large population to determine the risks of developing a second primary lung carcinoma after breast carcinoma and to determine which exposures increased risk: XRT, cigarette smoking, or both. To evaluate the effects of cigarette smoking and XRT, we conducted a retrospective, hospital-based, chart review study of a large series of women who developed lung carcinoma after breast carcinoma. The problem studied is significant for several reasons: 1) radiation, which is used widely in cancer therapy, is both a breast carcinogen and a lung carcinogen; ) some individuals exhibit radiation sensitivity and ill effects from it; and 3) previous data suggest that XRT and cigarette smoking may increase lung tumorigenesis after breast carcinoma. 5 8 MATERIALS AND METHODS Cases We identified 380 potential cases from The University of Texas M. D. Anderson Cancer Center (M. D. Anderson) tumor registry data base from 1960 to Case patients were female United States residents of all ethnicities, ages years at the time they were diagnosed with breast carcinoma, who subsequently were diagnosed with lung carcinoma, with an interval 6 months between breast carcinoma diagnosis and lung carcinoma diagnosis. Clinical criteria for case selection included all lung carcinoma histologies except mesothelioma; all breast carcinoma histologies except Paget disease; and all stages of breast carcinoma, including ductal carcinoma in situ. A pathologist confirmed case status by reviewing reports, tissue specimens, or documented correspondence from an outside physician or medical institution. All tumors were ascertained as primaries and not metastatic disease and were excluded if there was ambiguity in this regard. After medical record review, we excluded 100 cases, including 29 patients with misclassification of breast or lung carcinoma, 22 patients with unretrievable medical records, 18 patients with simultaneous diagnoses (or diagnoses within 6 months of each other) of breast and lung carcinoma, 16 patients due to the inability to rule out lung metastases, 6 patients with breast carcinoma diagnosed before 1960, 3 patients with primary lung carcinoma diagnosed before breast carcinoma, 3 patients because they were male, 2 patients with diagnoses of mesothelioma, and 1 with a diagnosis of Paget disease. We randomly selected a group of 300 control patients from among the 37,000 patients with breast carcinoma who were evaluated at M. D. Anderson during the same period ( ) during which the case patients were evaluated. Control patients were frequency matched to case patients by age and year of diagnosis with breast carcinoma (5-year strata) and ethnicity. Control patients had to be free of lung carcinoma and must have survived at least as long as the interval between the diagnosis of the case s first and second primary tumors. Data Collection Trained abstractors reviewed the medical records and recorded demographic data (age, gender, ethnicity, marital status, education), dates of primary breast and lung carcinoma diagnoses, methods and dates of treatment for breast carcinoma, laterality and histopathology of the breast and lung lesions, development of other primary malignancies, XRT for other primary malignancies or conditions, reproductive and smoking histories, and other pertinent clinical information. Description of Major Variables XRT information included treatment for breast carcinoma and, if used, the number of courses and when and where it was administered (at M. D. Anderson, elsewhere, or both). We did not collect XRT dosimetry data, because specific dose information sometimes was missing or imprecise for non-m. D. Andersontreated patients with breast carcinoma. Smoking information at the time of breast carcinoma diagnosis included the number of packs per day the patient smoked at the time of her breast carcinoma diagnosis and the total number of years she actually smoked. Data on site and histology of breast and lung lesions were based on the ninth revision of the International Classification of Diseases, Injuries and Causes of Death. 17 Tumor staging for breast and lung lesions followed the TNM classification system. 18,19 Statistical Analysis We initially evaluated frequency distributions and simple cross-tabulations using the chi-square statistic to examine the differences between case patients and control patients. We also calculated crude odds ratios (ORs) and 95% confidence intervals (95% CIs) for all covariates and the disease outcome (lung carcinoma after breast carcinoma). Each crude OR was adjusted

3 Lung Carcinoma in Breast Carcinoma Survivors/Ford et al for the matching variables (age, ethnicity, interval between breast carcinoma and lung carcinoma diagnoses). Because no marked differences were seen, only the crude ORs are reported. To examine the independent and combined contributions of smoking and XRT to lung carcinoma risk after breast carcinoma, we conducted stratified analyses with smoking as a categorical variable (never, former, or current smoker at the time of breast carcinoma diagnosis) and XRT as a dichotomized variable (yes or no). Further stratified analyses identified potential confounding variables or effect modifiers associated with smoking or XRT. Variables that were associated significantly with XRT, smoking, or both were then included in stratified cross-tabulations identifying the variables associated with XRT and smoking in the case group and the control groups. We used unconditional logistic regression, first, to estimate ORs for the main effects of XRT and smoking and, second, to estimate the effect of both exposures together, simultaneously adjusting for other covariates. By adding a product term (XRT*smoking) to the regression analysis, the interaction between smoking and XRT was further evaluated. The logistic models included matching variables, although there was little variation in the ORs for the main and combined effects of XRT and smoking. RESULTS Sample Characteristics For ethnicity and education, no statistically significant differences between case patients and control patients were seen (Table 1). Referral patterns at M. D. Anderson revealed a preponderance of white, non-hispanic women. Control patients were slightly older, and case patients were more likely divorced or widowed. The mean ages were 53.5 years and 56.4 years for case patients and control patients, respectively, with an age range of years. Clinical characteristics showed that women with advanced breast carcinoma were less likely to develop a subsequent primary lung carcinoma, and the proportion of women with adenocarcinoma of the breast who developed lung carcinoma (25.8%) was about double that of women who did not develop a subsequent lung carcinoma (13.7%). A significantly lower proportion of case patients had chemotherapy treatment for breast carcinoma compared with control patients, possibly because more of the case patients had lower stage breast carcinoma. Approximately 37% of the lung tumors were adenocarcinomas without further histologic specification, 11% were bronchioloalveolar adenocarcinomas, 20% were squamous cell carcinomas (with lesser proportions of oat cell carcinoma, a type of small cell carcinoma), and small cell carcinoma that was not otherwise specified (6%). Over half (55%) of the subsequent lung carcinomas occurred between 6 months and 10 years after the breast carcinoma diagnosis; 32% occurred between 10 years and 19 years, and 13% occurred 20 years after the breast carcinoma diagnosis. In 55% of cases, the side of the lung carcinoma was ipsilateral to the breast carcinoma. Smoking and XRT Smoking data at the time of breast carcinoma diagnosis were available for 93% of case patients and 84% of control patients and revealed significant differences in developing lung carcinoma. Of these, 70.1% of case patients (vs. 21.8% of control patients) reported that they were current smokers at the time of their breast carcinoma diagnosis (OR, 12.9; 95% CI, ). Univariate analysis of smoking at breast carcinoma diagnosis yielded increased ORs for both current and former smokers or ever smokers compared with never smokers (Table 2). Forty-five percent of case patients and control patients received XRT for breast carcinoma treatment (OR, 1.03; 95% CI, ). Interaction of Smoking and XRT Table 3 is a simple, stratified analysis of XRT and smoking for case patients and control patients. XRT was not associated with increased risk in nonsmokers (OR, 0.54; 95% CI, ), but smoking increased the odds of lung carcinoma in women who did not receive XRT (OR, 6.0; 95% CI, ). The OR for XRT and smoking combined, compared with neither exposure, was 9.0 (95% CI, ), suggesting a greater than multiplicative effect. Multivariate Modeling We constructed multivariate models with the exposure variables, the outcome variable, the product term of the exposure variables, and selected covariates. Because of missing data for some variables, the analytic sample size was reduced to 439. Selected covariates were risk factors for lung carcinoma after breast carcinoma or were identified as effect modifiers or confounding variables through stratified analysis. Table 4 presents the final adjusted ORs for smoking, XRT, and the product term (smoking*xrt). Smoking independently increased the risk of developing lung carcinoma after breast carcinoma more than five-fold, but XRT alone did not. Breast carcinoma survivors with histories of both smoking and XRT showed an eightfold greater risk compared with survivors without either exposure: This result was nearly significant for a greater than multiplicative effect.

4 1460 CANCER October 1, 2003 / Volume 98 / Number 7 TABLE 1 Selected Demographic and Clinical Characteristics for Breast Carcinoma Survivors with Subsequent Primary Lung Carcinoma (Cases) and Breast Carcinoma Survivors without Lung Carcinoma (Controls) Cases (n 280) Controls (n 300) Characteristic No. (%) No. (%) Chi-square test P value Ethnicity White 262 (93.6) 273 (91.0) Black 12 (4.3) 13 (4.3) Hispanic 4 (1.4) 9 (3.0) Asian/other 2 (0.7) 5 (1.7) Education (yrs) (49.6) 46 (37.1) (46.0) 67 (54.0) 17 6 (4.4) 11 (8.9) Marital status Married 171 (61.3) 207 (69.2) Single 5 (1.8) 13 (4.3) Divorce/separated 37 (13.3) 25 (8.4) Widowed 66 (23.7) 54 (18.1) Age at breast carcinoma diagnosis (yrs) (38.6) 88 (29.3) (61.4) 212 (70.7) Breast carcinoma stage Stage 0 I 78 (43.3) 73 (28.7) Stage IIA 52 (28.9) 75 (29.5) Stage IIB 31 (17.2) 47 (18.5) Stage IIIA 9 (5.0) 23 (9.1) Stage IIIB 8 (4.4) 18 (7.1) Stage IV 2 (1.1) 18 (7.1) Missing 100 (35.7) a 46 (15.3) a Breast carcinoma histology Infiltrating ductal carcinoma 121 (52.8) 166 (59.7) Carcinoma NOS 15 (6.6) 39 (14.0) Carcinoma in situ 12 (5.2) 14 (5.0) Adenocarcinoma NOS 59 (25.8) 38 (13.7) Cystic carcinoma 3 (1.3) 3 (1.1) Other ductal carcinoma 19 (8.3) 18 (6.5) Missing 51 (18.2) a 22 (7.3) a Chemotherapy for breast carcinoma No 225 (80.4) 196 (65.5) Yes 55 (19.6) 103 (34.5) Missing 0 (0.0) a 1 (0.3) a Lung carcinoma histology Adenocarcinoma, NOS 103 (36.8) Bronchioloalveolar adenocarcinoma 32 (11.4) Squamous cell carcinoma, NOS 56 (20.0) Small cell carcinoma, NOS 17 (6.1) Oat cell carcinoma 32 (11.4) Large cell carcinoma, NOS 9 (3.2) Other 28 (10.0) Not recorded 3 (1.1) Yrs between breast and lung carcinoma diagnoses (55.4) (31.8) (10.4) (2.5) Lung carcinoma ipsilateral to the breast carcinoma Yes 155 (55.4) No 113 (40.4) Both or bilateral 9 (3.2) Not recorded 3 (1.1) NOS: not otherwise specified. a Percentage of the total number of cases.

5 Lung Carcinoma in Breast Carcinoma Survivors/Ford et al TABLE 2 Smoking History and Radiotherapy among Breast Carcinoma Survivors with Subsequent Primary Lung Carcinoma (Cases) and Breast Carcinoma Survivors without Lung Carcinoma (Controls) Cases (n 280) Controls (n 300) Characteristic No. (%) No. (%) OR 95% CI Smoking at breast carcinoma diagnosis Never 41 (15.7) 159 (62.8) 1.00 Formerly 37 (14.2) 39 (15.4) Currently 183 (70.1) 55 (21.8) or Never 41 (15.7) 159 (62.8) 1.00 Ever 220 (84.3) 94 (37.2) Radiotherapy for breast carcinoma No 153 (54.6) 162 (55.3) 1.00 Yes 127 (45.4) 131 (44.7) OR: odds ratio; 95% CI: 95% confidence interval. TABLE 3 Independent and Combined Effects of Radiotherapy and Smoking on the Risk of Lung Carcinoma After a Diagnosis of Breast Carcinoma Smoking a XRT Percent of patients Cases (n 261) Controls (n 247) OR 95% CI No No No Yes Yes No Yes Yes XRT: radiotherapy; OR: odds ratio; 95% CI: 95% confidence interval. a At the time of breast carcinoma diagnosis. TABLE 4 Adjusted Odds Ratios for Smoking, Radiotherapy, and Both Variables Together among Breast Carcinoma Survivors with Lung Carcinoma (Cases) and Breast Carcinoma Survivors without Lung Carcinoma (Controls) Smoking XRT OR a P value No No 1.00 No Yes Yes No Yes Yes XRT: radiotherapy; OR: odds ratio. a Adjusted for breast carcinoma histology (adenocarcinoma vs. other), type of breast surgery (excisional biopsy vs. other), breast carcinoma diagnosis date (before 1980 vs or later), age, and ethnicity. DISCUSSION The health effects of cigarette smoking and their relation to cancer are well documented; and, unsurprisingly, the current study showed that smoking is an independent risk factor for lung carcinoma in breast carcinoma survivors. Moreover, a recent study utilizing data from the Swedish Cancer Registry underscored the hypothesis that both XRT and smoking increases the risk of lung carcinoma among women with breast carcinoma, especially after 10 years. 8 In a similar study of lung carcinoma after breast carcinoma, the OR associated with smoking was 14.5 (95% CI, ) and increased to 27.6 (95% CI, ) for the lung ipsilateral to the breast tumor. 7 Several other studies focused on lung carcinoma after breast carcinoma. 4 8 However, each of those studies was limited and was missing either smoking data or radiation dose data. A cohort study of breast carcinoma survivors from the Connecticut Tumor Registry revealed significant excesses of lung carcinoma. 4 Those authors, examining data on 177 lung carcinoma cases in 41,109 breast carcinoma survivors, categorized the survivors as XRT ever or XRT never. Lung carcinoma risks were elevated significantly among irradiated women years and years after treatment (standardized mortality ratios, 2.3 and 4.8, respectively) compared with nonirradiated women (standardized mortality ratios, 1.5 and 1.8, respectively). 4 Another cohort study using National Cancer Institute Surveillance, Epidemiology, and End Results Program data from compared the role of XRT (yes or no) in women with breast carcinoma and found a 2-fold risk of lung carcinoma among women who were treated with XRT for breast carcinoma 10 years after diagnosis for the first malignancy (relative risk [RR], 2.0; 95% CI, ). 5 However, the study by Neugut et al. 5 lacked information on smoking and, thus, could not examine smoking as a potential effect

6 1462 CANCER October 1, 2003 / Volume 98 / Number 7 modifier. It is unknown whether smoking data may have explained all or part of the XRT effect. The current study identified no effect of XRT on lung carcinoma after breast carcinoma and runs counter to studies of lung carcinogenesis among patients receiving XRT for Hodgkin disease A possible explanation for the inconsistent findings may be differences in radiation type, energy dose, and dose rate and differing volume and sites of lung irradiated in the different populations. 28 However, additional studies investigating the effect of XRT on subsequent malignancies, including lung carcinoma after breast carcinoma, found no increased risk from XRT overall or for intervals 10 years after treatment, supporting our results. 5,7,29 Two additional case control studies of subsequent lung carcinoma nested within the Connecticut Tumor Registry did consider the smoking status of women. 6,7 Inskip and colleagues compared 61 cases of lung carcinoma after breast carcinoma with 120 matched controls. 6 Although radiation dosimetry was well quantified, smoking data were missing for 77% of case patients and control patients. Using the available smoking information, after 10 years, women smokers who had received XRT for breast carcinoma had a nearly 3-fold risk of developing lung carcinoma (OR, 2.7; 95% CI, ). Neugut et al. compared 121 cases of lung carcinoma after breast carcinoma with 1043 controls who had nonsmoking, non-xrt-related second primary tumors. 7 Smoking data were available for 78% of cases and for 58% of controls, and their results showed a greater than multiplicative effect for developing lung carcinoma, with a 32.7 times greater risk (95% CI, ) among smokers who received XRT. The magnitude and nature of the effects of smoking, however, have been challenged subsequently. 28 Less dramatic but more compelling was the clear interaction between XRT and smoking in the current study. Similar findings were reported in a nested case control study by Travis et al., who assessed the risk of treatment-associated lung carcinoma after Hodgkin disease reported to seven population-based cancer registries. 27 Those authors found a 20-fold increase in lung carcinoma risk associated with moderate-toheavy tobacco use in patients with Hodgkin disease who received XRT (RR, 20.2; 95% CI, ), revealing an effect compatible with multiplicative interaction between the two exposures. In the current study, we had more complete data on smoking compared with other studies of lung carcinoma after breast carcinoma; and, although we collected XRT information (yes or no), we were not able to obtain sufficient detailed records to estimate an individual s lung doses. We demonstrated that smokers who received XRT for breast carcinoma had greater risks than would have been expected from multiplicative combined effects. For example, under the assumption of a multiplicative effect, we would have expected a risk estimate of 3.0 for smoking and XRT combined; however, we observed an OR of 9.0, a result suggestive of a greater than multiplicative effect. The current study had several limitations. We frequency matched case patients with control patients by 5-year age groups, and despite random selection within strata, the case patients tended to be slightly younger than the control patients within each stratum. Therefore, matching on age at breast carcinoma diagnosis was not achieved perfectly; hence, for the multivariate analysis, we adjusted for age. Another limitation of studies that obtain data from medical records is that they may be prone to information bias and misclassification. Even a thorough histologic review, such as the one we performed, does not fully exclude misclassification of primary and metastatic lung carcinoma. Smoking histories were more complete for the case patients than for the control patients, possibly because physicians treating patients with breast carcinoma did not routinely obtain a complete smoking history. The current study also lacked information on smoking during the time between breast carcinoma diagnosis and lung carcinoma diagnosis. Case patients may have stopped smoking after their breast carcinoma diagnosis, may have stopped and restarted smoking, or even may have started smoking in the interval between diagnoses. Assuming that women stopped smoking immediately after their breast carcinoma diagnosis, the observed risk of smoking would decrease; whereas, assuming that women who continued to smoke indicates that the observed risk is underestimated. Likewise, information on XRT was limited to yes or no, without quantification. Although no effect was associated with the year of XRT (1970 or later vs. before 1970; OR, 0.80; 95% CI, ), dosimetry information would have strengthened the study. Moreover, because breast-conserving surgery with XRT has been used widely only since 1985, our cohort is only now attaining 10 years after treatment. Following this cohort and examining their risk for subsequent primary lung carcinoma as they age would refine our analysis further. The low absolute risk of lung carcinoma after breast carcinoma is weak motivation to modify breast carcinoma treatment but not to advocate smoking cessation among breast carcinoma survivors. The current study indicates caution when considering radiation treatment that includes the lung fields of smok-

7 Lung Carcinoma in Breast Carcinoma Survivors/Ford et al ers. It substantiates and quantifies the significant effect of smoking and XRT on lung carcinoma after the successful treatment of breast carcinoma. Large numbers of women are at risk: in 1985, the midpoint of the study period, 28% of women in the United States were current smokers and 18% were former smokers. 30 Moreover, the Surgeon General s Report on Women and Smoking, 2001 states that women now account for 39% of all smoking-related deaths each year in the United States, a proportion that has more than doubled since These large percentages indicate that many breast carcinoma survivors are potentially at risk from the enhanced effect of XRT and smoking. In addition, the age of women developing breast carcinoma coincides with the cohort of women with the highest prevalence of ever smoking. Studies of the current type reinforce the need for greater cancer prevention efforts on the part of treating oncologists to strongly urge patients with breast carcinoma who smoke to quit smoking or to encourage attending smoking-cessation programs. An increasing number of women are surviving breast carcinoma; therefore, researchers and clinicians should understand and circulate information concerning the possible heightened risk of another primary malignancy after a malignancy is cured. Also, more women are undergoing lumpectomy and irradiation, especially patients who are diagnosed with ductal carcinoma in situ and who are considered to be in a good-risk disease group (i.e., patients with small, localized lesions with low-grade pathology classifications) Lumpectomy and radiation may then put more women, especially cigarette smokers, at risk of developing lung carcinoma. Current treatment increasingly employs breastconservation surgery and XRT. Given the relation between smoking and XRT and the risk of lung carcinoma after breast carcinoma, quitting smoking or attending smoking cessation programs should be recommended strongly. Future research should include multicenter, prospective studies that integrate complete data on smoking behavior during the induction period between breast carcinoma and lung carcinoma, radiation dosimetry, the major genes for breast carcinoma (BRCA1, BRCA2, ATM), and genetic markers of susceptibility (radiosensitivity, DNA repair). Because of the small numbers of minority women represented in this and similar studies, it would be especially beneficial to evaluate ethnically diverse populations. REFERENCES 1. Hanlon, FR. Multiple primary carcinomas. Am J Cancer. 1931;15: Curtis RE, Boice JD Jr., Kleinerman RA, Flannery JT, Fraumeni JF Jr. Summary: multiple primary cancers in Connecticut, Natl Cancer Inst Monogr. 1985;68: Li FP. Second cancers. In: DeVita VT, Hellman S, Rosenberg SA, editors. Cancer: principles and practice of oncology. Philadelphia: Lippincott-Raven, 1985: Harvey EB, Brinton LA. Second cancer following cancer of the breast in Connecticut, Natl Cancer Inst Monogr. 1985;68: Neugut AI, Robinson E, Lee WC, Murray T, Karwoski K, Kutcher GJ. Lung cancer after radiation therapy for breast cancer. Cancer. 1993;71: Inskip PD, Stovall M, Flannery JT. Lung cancer risk and radiation dose among women treated for breast cancer. J Natl Cancer Inst. 1994;86: Neugut AI, Murray T, Santos J, et al. Increased risk of lung cancer after breast cancer radiation therapy in cigarette smokers. Cancer. 1994;73: Prochazka M, Granath F, Ekbom A, Shields PG, Hall P. Lung cancer risks in women with previous breast cancer. Eur J Cancer. 2002;38: Committee on the Biological Effects of Ionizing Radiation. Health effects of exposure to low levels of ionizing radiation (BEIR V). Washington, DC: National Academy Press, 1990: United Nations Scientific Committee on the Effects of Atomic Radiation. UNSCEAR report to the General Assembly. Sources and effects of ionizing radiation. New York: United Nations, Boice JD Jr., Land CE, Preston DL. Ionizing radiation. In: Schottenfeld D, Fraumeni JF, editors. Cancer epidemiology and prevention. New York: Oxford University Press, 1996: Angele S, Hall J. The ATM gene and breast cancer: is it really a risk factor? Mutat Res. 2000;462: Bishay K, Ory K, Olivier MF, Lebeau J, Levalois C, Chevillard S. DNA damage-related RNA expression to assess individual sensitivity to ionizing radiation. Carcinogenesis. 2001;22: Hendry JH. Genomic instability: potential contributions to tumor and normal tissue response, and second tumours, after radiotherapy. Radiother Oncol. 2001;59: Jongmans W, Hall J. Cellular responses to radiation and risk of breast cancer. Eur J Cancer. 1999;35: Loeffler JS, Harris JR, Dahlberg WK, Little JB. In vitro radiosensitivity of human diploid fibroblasts derived from women with unusually sensitive clinical responses to definitive radiation therapy for breast cancer. Radiat Res. 1990; 121: World Health Organization. International classification of diseases, injuries, and causes of death. Ninth revision (ICD- 9). Geneva: World Health Organization, Balch C, Singletary E, Bland KI. Clinical decision making in early breast cancer. Ann Surg. 1993;217: Ginsberg RJ, Vokes EE, Raben A. Non-small cell lung cancer. In: DeVita VT, Hellman S, Rosenberg SA, editors. Cancer: principles and practice of oncology. Philadelphia: Lippincott-Raven, 1997: Advisory Committee to the Surgeon General. Smoking and health-report of the Advisory Committee to the Surgeon General of the Public Health Service. U.S. Department of Health, Education and Welfare, Public Health Service Publication No Washington, DC: U.S. Government Printing Office, 1964:24 43.

8 1464 CANCER October 1, 2003 / Volume 98 / Number Doll R, Hill AB. Mortality in relation to smoking: ten years observations in British doctors. Br Med J. 1964;1: Doll R, Gray R, Hafner B. Mortality in relation to smoking: 22 years observations on female British doctors. Br Med J. 1980;280: Centers for Disease Control and Prevention (CDC). Cigarette smoking among adults -United States, Morbid Mortal Wkly Rep. 1997;46: van Leeuwen FE, Somers R, Taal BG, et al. Increased risk of lung cancer, non-hodgkin s lymphoma and leukemia following Hodgkin s disease. J Clin Oncol. 1989;7: van Leeuwen FE, Klokman WJ, Stovall M, et al. Roles of radiotherapy and smoking in lung cancer following Hodgkin s disease. J Natl Cancer Inst. 1995;87: Boivin JF. Smoking, treatment for Hodgkin s disease, and subsequent lung cancer risk. J Natl Cancer Inst. 1995;87: Travis LB, Gospodarowicz M, Curtis RE, et al. Lung cancer following chemotherapy and radiotherapy for Hodgkin s disease. J Natl Cancer Inst. 2002;94: Inskip PD, Boice JD Jr. Radiotherapy-induced lung cancer among women who smoke [published erratum appears in Cancer. 1994;73:2456]. Cancer. 1994;73: Obedian E, Fischer DB, Haffty BG. Second malignancies after treatment of early-stage breast cancer: lumpectomy and radiation therapy versus mastectomy. J Clin Oncol. 2000;18: American Cancer Society. Cancer facts and figures, Atlanta: American Cancer Society, Office of the Surgeon General, Public Health Service, Centers for Disease Control and Prevention (CDC). Women and smoking: a report of the Surgeon General National Center for Chronic Disease Prevention and Health Promotion, United States Office on Smoking and Health. Rockville, MD: U.S. Dept. of Health and Human Services, Public Health Service, Office of the Surgeon General, Available at URL: db/local.sgen.sgrep.womsmk/screen/tocdisplay/da/1/s/ 61927/action/Toc 32. Fisher B, Redmond C, Poisson R, et al. Eight-year results of a randomized clinical trial comparing total mastectomy and lumpectomy with or without irradiation in the treatment of breast cancer [published erratum appears in N Engl J Med. 1994;330:1467]. N Engl J Med. 1989;320: Fisher B, Dignam J, Wolmark N, et al. Lumpectomy and radiation therapy for the treatment of intraductal breast cancer: findings from the National Surgical Adjuvant Breast and Bowel Project B-17. J Clin Oncol. 1998;16: