Malignant Breast Tumors After Radiotherapy for a First Cancer During Childhood
http://www.100md.com
《临床肿瘤学》
the National Institute of Health and Medical Research, Cancer Epidemiology Research Unit (U605 INSERM), Physics Department, Department of Paediatrics, and Department of Pathology, Institut Gustave Roussy, Villejuif
Department of Paediatrics, Institut Curie, Paris
Department of Radiotherapy, Institut Jean Godinot, Reims
Department of Radiotherapy, Centre Lacassagne, Nice
Department of Radiotherapy, Centre Claudius-Regaud, Toulouse, France
Childhood Cancer Research Group, Radcliffe Infirmary, Oxford, United Kingdom
ABSTRACT
Web Table 1. Chemotherapy: Drugs Administered to 1,167 of 1,814 Female Children Treated for a First Cancer in Childhood
Web Table 3. Breast Cancers According to the Mean Radiation Dose to the Breast in a Cohort of 1,814 Patients Treated for a First Cancer During Childhood
PATIENTS AND METHODS: In a cohort of 1,814 3-year female survivors treated between 1946 and 1986 in eight French and English centers, data on chemotherapy and radiotherapy were collected. Individual estimation of radiation dose to each breast was performed for the 1,258 patients treated by external radiotherapy; mean dose to breast was 5.06 Gy (range, 0.0 to 88.0 Gy) delivered in 20 fractions (mean).
RESULTS: Mean follow-up was 16 years; 16 patients developed a clinical BC, 13 after radiotherapy. The cumulative incidence of BC was 2.8% (95% CI, 1.0% to 4.5%) 30 years after the FC and 5.1% (95% CI, 2.1% to 8.2%) at the age of 40 years. The annual excess incidence increased as age increased, whereas the standardized incidence ratio decreased. On average, each Gray unit received by any breast increased the excess relative risk of BC by 0.13 (< 0.0 to 0.75). After stratification on castration and attained age, and adjusting for radiation dose, FC type, and chemotherapy, a higher risk of a subsequent BC was associated with Hodgkin’s disease (relative risk, 7.0; 95% CI, 1.4 to 30.9).
CONCLUSION: The reported high risk of BC after childhood Hodgkin’s disease treatment seems to be due not only to a higher radiation dose to the breasts, but also to a specific susceptibility.
INTRODUCTION
We report here the results of a subcohort study designed to evaluate the long-term risk of a subsequent BC after fractionated high doses of external-beam radiotherapy. Radiation doses to each breast were separately estimated for every child with specialized software. An attempt was made to assess the role of the FC type in determining the risk of subsequent BC, controlling for treatment and particularly for the radiation dose to the breasts.
PATIENTS AND METHODS
these 4,400 children, only the female patients were selected because among males, no BC was observed. One hundred fifty-one patients were excluded because they had received brachytherapy or information on radiation doses was not available. Among the 1,814 remaining women, 1,258 had received external-beam radiotherapy. The end of the period at risk was the date of the last contact, death, or the last time the patient was known to be alive; for those with a BC, this was the date of the diagnosis. The follow-up data were obtained using registries for the English patients or, for the French patients, from hospital clinical files actualized by physicians who performed an active follow-up by writing to patients and linking their database with the national death certificate data. Twelve percent of the cohort (n = 210) was lost to follow-up.
Information on treatments was abstracted from the clinical notes and radiotherapy files in the participating centers and the diagnoses of the first and second malignant neoplasms were histologically confirmed. The International Classification of the Diseases for Oncology was used to classify the tumors.22
Radiation Dosimetry
The radiation doses were estimated in the middle of the right and left nipples for each of the 1,258 patients treated with ionizing radiation. The dose to 149 other anatomic sites including the thymus, spleen, gonads, and 91 skeletal sites was also estimated. A computer program called Dos_EG, described elsewhere,23,24 was developed for these calculations.
In the estimation, all treatments delivered on the same day were considered as one fraction. This definition is considered acceptable, given that there was no hyperfractionated treatment in the cohort.
Among the 1,258 patients who received external radiotherapy, the mean number of fractions was 20 and the average radiation dose for the 1,258 patients was 5.06 Gy for a breast (Table 2). Figure 1 shows the distribution of the radiation dose to breast in Grays (median, 0.85 Gy). Because of variations in the radiotherapy instruments and in the treatment modalities of the FC, the mean dose to the breast varied according to the calendar period: 3.5 Gy before 1960, 4.4 Gy during the 1960s, 6.3 Gy during the 1970s, and 3.6 Gy during 1980 to 1985.
Measurement of Chemotherapy
Chemotherapy doses were not available for four of the 1,167 patients who received chemotherapy: 360 had chemotherapy alone with (n = 326) or without (n = 34) surgery. Drugs were grouped into five categories according to their known mechanism of action in the cell: electrophilic agents, spindle inhibitors, nucleotide synthesis inhibitors (NSI), topoisomerase II inhibitors, and other compounds. To quantify the total amount of each drug category administered, the dose of each cytotoxic agent was converted into moles per square meter, rather than into milligrams per square meter. A detailed description of the chemotherapy administered to cohort patients has been published.18 The median doses for each drug category and main drugs administered to the children were calculated (Web Table 1).
Statistical Methods
For analysis, the follow-up ended on January 1, 1993, for patients treated in the French centers and January 1, 1991, for those treated in United Kingdom. We used the competing method to estimate the cumulative incidence of breast tumors.25 The expected numbers of BC (number of tumors) per 5-year calendar periods and per 5-year age classes were calculated from 1975 to 1995 Francim data (based on the French cancer registries) for the French patients26 and on the England and Wales cancer registry data for the United Kingdom patients.27 The SIR was defined as the ratio between the observed and expected number of breast tumors. The relative risk (RR) was defined in a similar manner: the expected number of breast tumors was estimated from a given reference category within the cohort; an individual breast was used as the statistical unit.
Both the SIR and RR of BC were modeled assuming that the number of BC (number of tumors) followed Poisson distribution. Statistical tests were carried out using deviance of nested models.28
The basic model for calculating the excess relative risk (ERR) of BC as a function of the radiation dose was that generally used in radiation epidemiology29: ERR = Cst ({alpha}dose + dose2) [exp(–{gamma} dose)], where the constant (Cst) summarizes the role of dichotomous variables such as having had a given type of FC (yes or no) or chemotherapy; the negative exponential term models the cell sterilization for high doses.
We used an empirical approach to investigate the role of dose fractionation and introduced a threshold number of fractions to test the similarity of the coefficients for the doses delivered with a high (dose1) and low (dose2) number of fractions in a linear model, using the following equation: ERR = Cst ({alpha}2dose2 + {alpha}1dose1).
The simple linear model of the risk as a function of the total dose is nested in this model: ERR = Cst [{alpha}2(dose – dose1) + {alpha}1dose1] = Cst ({alpha}2dose – {alpha}2dose1 + {alpha}1dose1) = Cst [{alpha}2dose + ({alpha}1 – {alpha}2) dose1].
The effect of the dose rate was considered significant with this model if the decrease in deviance between model 2 fractionated dose and model 1 total dose limited to the linear term was greater than 3.84 (ie, the value of {chi}21 for P = .05). The standard deviation of the coefficient for doses delivered with fewer than 20 fractions (ie, {alpha}1 – {alpha}2) was estimated as follows: standard deviation .
These analyses were done using AMFIT Software.29 The 95% CIs were estimated for parameters using the maximum likelihood method.30
RESULTS
During the follow-up, 16 patients had developed a clinical BC (0.95 expected), excluding carcinoma in situ: 11 cancers arose in the left breast and three in the right; in two patients, this information was not available. The histological type was invasive ductal carcinoma in 12 patients (nine of them had a grade II or III according to the Scarff, Bloom, and Richardson scale, a histoprognostic grading system of invasive breast tumors based on the degree of differentiation of the tumor, mitotic index, and anisonucleosis, grade I to III, where I is a good prognosis), malignant phyllodes tumor in three patients, and tumor of an unknown type in one patient. Among the carcinomas, six patients presented with axillary involvement.
Thirteen patients developed a BC after radiotherapy, either with or without chemotherapy, with a follow-up between 13 and 37 years. One patient developed a BC (at 9 years of follow-up) for an FC treated by surgery alone; two patients developed a BC after chemotherapy (follow-up, 18 and 19 years, respectively). All but one cancer occurred in patients with a dose less than 20 Gy to the breast containing the tumor. The treatment characteristics of the children who developed a secondary breast tumor are listed in Table 3.
The cumulative incidence of BC after a 30-year follow-up was 2.8% (95% CI, 1.0% to 4.5%) and 10.7% (95% CI, 1.4% to 19.9%) after 40 years. Figure 2 depicts the cumulative incidence by attained age for the entire cohort. The annual incidence of BC increased dramatically to reach 537 x 10–5 person years (95% CI, 269 to 942) when women were 30 to 39 years old. The annual excess incidence increased considerably to 515 x 10–5 person years (95% CI, 246 to 919) for those aged 30 to 39 years, but the SIR declines from 185 (95% CI, 31 to 573) in the 10- to 19-year age group to 23.2 (95% CI, 12 to 41) for the 30- to 39-year age group (Table 4).
A total of 222 patients underwent a surgical or radiologic castration. None of them developed a secondary BC later (P = .1).
Radiotherapy, measured as ever versus never, was not associated with an increased risk of BC (RR, 1.3; 95% CI, 0.4 to 5.9). The ERR of occurrence of a BC was found to be nonsignificantly linearly linked to the total dose of radiation (ERR/Gy, 0.13; 95% CI, < 0.0 to 0.75; P = .06) received by the breast, after adjustment for castration, chemotherapy, and FC type. When model 2 with a threshold at 20 fractions was used, a significant effect of the dose fractionation was observed (P = .05): the radiation dose delivered in less than 20 fractions led to a higher risk (ERR/Gy = 0.12; 95% CI, –0.14 to 0.38) than when it was administered in 20 fractions or more (ERR/Gy = –0.02; 95% CI, –0.06 to 0.03).
Overall, chemotherapy was associated with a nearly significant RR of 2.7 (95% CI, 0.9 to 7.6; P = .07) when stratified on castration and attained age, and adjusted for radiation dose and FC type compared with no chemotherapy (Web Table 2). When we examined the role of the different classes of chemotherapeutic agents, we observed that electrophilic agents, spindle inhibitors, and topoisomerase II inhibitors had no significant effect on the risk of developing a BC in this cohort (Web Table 2). When stratification on castration and attained age plus adjustment on other drug categories, FC type, age at FC, and radiation dose to the breast were made, the RR for NSI (ie, for methotrexate) was found to be 4.6 (95% CI, 0.7 to 19.9; Table 5). Only 62 patients received mechlorethamine, vincristine, procarbazine, and prednisone, of whom two developed a BC. There was no evidence for a role of the number of moles per square meter for any of the drug categories or of the total number of moles per square meter administered.
Four of the 16 BCs occurred among the 123 patients treated for Hodgkin’s disease (SIR, 71; 95% CI, 22 to 164), three occurred among the 276 treated for a neuroblastoma (SIR, 47; 95% CI, 12 to 122), three occurred among the 373 treated for a Wilms’ tumor (SIR, 28; 95% CI, 7.0 to 73), two occurred among the 209 patients treated for soft tissue sarcoma (SIR, 16; 95% CI, 2.7 to 51), and one in each of the following FC types: brain, gonad, and non-Hodgkin’s lymphoma (Table 3). A heterogeneity was evidenced in SIR between FC (P = .02).
After Hodgkin’s disease, none of the four BCs occurred before 16 years after first diagnosis, and breasts of these patients received a mean dose of 11.84 Gy, whereas for all other patients, the mean dose was 4.39 Gy. The latent periods after neuroblastoma and Wilms’ tumors were, respectively, 13 to 37 years and 18 to 36 years, and the mean doses received by the breasts were 5.85 Gy (v 4.95 Gy when excluding neuroblastoma patients) and 7.04 Gy (v 4.51 Gy for non-Wilms’ tumor patients). When patients treated for these three types of malignancies were excluded, the mean dose administered to the breasts decreased to 3.06 Gy. When age at FC, attained age, castration, radiation dose to the breast (in Grays), and chemotherapy as yes versus no were taken into account, the risk for developing a BC was higher in any breast of patients previously treated for Hodgkin’s disease (RR, 7.01; 95% CI, 1.4 to 30.9) or, nonsignificantly, for neuroblastoma (RR, 2.9; 95% CI, 0.6 to 11.6) than in those treated for any other FC combined (Table 5).
Of the 264 patients who underwent a surgical or radiologic splenectomy (> 20 Gy to the spleen), three developed a BC. No evidence for a spontaneous higher risk of BC (RR, 1.8; 95% CI, 0.6 to 5.5) was found. Similarly, a modification of the risk of BC was not evidenced among the 242 patients (of whom three had a secondary BC) who had received a mean dose to active bone marrow exceeding 10 Gy (RR, 1.4; 95% CI, 0.3 to 5.9). Age at FC does not affect the risk of BC, whereas risk of BC increased with increasing doses to the breasts up to 20 Gy (P = .06; Table 5).
DISCUSSION
The women lost to follow-up (n = 210) were not different from the others in terms of the date of the diagnosis (1974 on average). They did, however, receive fewer moles of drugs per square meter (13 v 30 moles/m2; P = .04), but the average radiation dose to the breasts was the same (3.4 v 3.4 Gy). Given that the cumulative dose of chemotherapy can be linked to relapses, the difference in the amount of chemotherapy administered appears to indicate that we were more likely to lose to follow-up patients who did not experience relapse (ie, patients with less aggressive therapy), and this could lead to an overestimation of the long-term incidence of BC. However, the likelihood of overestimation is probably limited given the low rate of patients lost to follow-up in our cohort (12%).
The first limitation is the relatively small size of our cohort, which included 1,814 patients and only 16 BCs. This allows us to satisfactorily only risk factors associated with substantially increased risks. A possible bias was the heightened medical surveillance of women who survived a childhood cancer, leading to a greater likelihood of detecting other diseases, including second malignancies. Because there is no French national cancer registry, the follow-up of cohorts such as ours must be active, which means that this bias could never be eliminated. Furthermore, screening for BC is likely to increase the observed incidence of breast tumors.31 In our opinion, this potential bias is low because no routine BC screening program existed during our study follow-up period, and the size and extent of the 16 BCs diagnosed in our cohort were more aggressive (Scarff, Bloom, and Richardson scale grade II or III) than those usually found during screening for prevention.
In our cohort, the cumulative incidence of BC at 40 years of age was 5.1% (95% CI, 2.1 to 8.2) corresponding to 16.9-fold the rate expected (95% CI, 9.9 to 26.6) in the general population. Although no patients with leukemia are included in our cohort, this global SIR is within the range observed for BC in hospital-based studies,4,10,32 but higher than that observed in registry-based cohorts.6,8 However, an overall comparison between BC SIR observed in cohorts is very difficult to interpret because it is highly dependent on the duration of follow-up, the average attained age, and the expected BC incidence, which is unstable before 45 years of age because it is based on small numbers, and because of differences in the FC types between cohorts.
For the SIR with attained age, we observed a decrease from 185 (95% CI, 31 to 573) in the 10- to 19-year age group to 23.2 (95% CI, 12 to 41) in the 30- to 39-year age group. One BC (0.35 expected) was observed in our cohort after 40 years of age, but only 60 women had reached this age during the follow-up period. Variation in BC SIR with attained age or time since FC was examined in detail in only one study, which obtained similar results.8 Because of the strong increase in expected BC incidence with attained age, this reduction in SIR nevertheless corresponds to a substantial increase in excess incidence with attained age, at least during the first four decades.8
We confirm previously reported8,11-13,16,17 higher BC incidence after Hodgkin’s disease (RR, 7.01 compared with BC risk after other FC) and demonstrate, for the first time, that this excess is not due to higher radiation doses and/or chemotherapy.
In addition, we found that survivors of childhood neuroblastoma could be at a greater risk (RR, 2.9) of developing a BC. To date, this result has not been reported in any other study. Although Wilms’ tumor was the most frequent FC in our cohort (n = 373), it was not associated with an increased BC risk.10,33 However, one BC was observed among Hodgkin’s disease and neuroblastoma patients who had no radiotherapy or a low dose to breasts (Web Table 3). Thus, the observed increased risks of BC after such FC suggest either a specific susceptibility for developing a BC or a particular susceptibility to radiation and/or chemotherapy, or both.
Although the number of patients treated by radiotherapy in our study is relatively limited (n = 1,258), the range of the doses available for the investigation of a dose-response relationship was wide (from 0.0015 to 88.0 Gy received by the breast): 514 breasts received less than 0.2 Gy and 197 received more than 20 Gy. Radiotherapy was not found to be a risk factor for BC occurrence when considered as a dichotomous variable. Nevertheless, this risk was found to be nonsignificantly linearly linked to the radiation dose received by any breast with an ERR/Gy of 0.13 (95% CI, < 0.0 to 0.75). This result confirms established effects of radiation exposure1,34 and has been described in many studies. To our knowledge, no other study using individual radiation dose estimation to breasts has provided an estimate of a coefficient for BC occurrence after radiotherapy treatment for a malignancy. Our result is consistent with that observed after radiotherapy for benign diseases in childhood: hemangioma,35-37 tinea capitis,38 and multiple diagnosis x-rays.39,40 On the contrary, it is lower than that described for the Hiroshima-Nagasaki survivors.41,42
The higher risk observed for the same total dose to a breast delivered in less than 20 fractions (ERR/Gy = 0.12) compared with delivery in more fractions (ERR/Gy = –0.02) appears to be in contradiction with the linear dose-response we estimated with a negative exponential term. Nevertheless, it is noteworthy that model 1 (total dose) is not adequate for fractionated irradiation.43 Women subjected to repeated fluoroscopy examinations during the follow-up of tuberculosis had a BC risk that was similar to that of female atomic bomb survivors who had received a similar total dose.40,44
We were unable to identify any substantial decrease in the doses delivered to the breasts after changes of treatment instruments and energy (low-energy x-rays, high-energy x-rays, electrons, cobalt). Seven BCs occurred after low-energy x-ray therapy. A follow-up of these patients (mean, 26 years; range, 3 to 46 years) longer than that attained in patients treated by the other instruments (mean, 14 years; range, 3 to 36 years) can explain this high number.
In this cohort, after stratification on castration and attained age and adjustment on radiation dose and FC type, the RR associated with chemotherapy was weak (RR, 2.7; 95% CI, 0.9 to 7.6) and chemotherapeutic drug classes were not found to increase the risk of BC except for the NSI class (only methotrexate), which seemed more likely to be associated with an increased risk of BC. However, only two BCs occurred after the administration of this cytotoxic agent. This increased risk due to chemotherapy is in agreement with one study45 in which alkylating agents and vinca alkaloids, but not NSI, were associated with a more frequent occurrence of BC, but not with another study,10 which failed to determine such an association. In these two studies, adjustment for radiotherapy was only addressed as yes or no.
The distribution of laterality of the 16 BCs observed in this cohort is different from that observed in the French breast cancer registries. Thus, 11 of the 14 BCs (approximately 80%) for which information was available were located in the left breast (the percentage is usually approximately 50%).26 Regarding the histologic type, three among the 14 BCs with this information are infrequent malignant phyllodes tumors; the percentage in the French BC registries is 0.18%.26
We also investigated the potential effect of radiation doses to organs influencing the immunologic status (spleen or active bone marrow) on the occurrence of BC. We did not find any role like that found by others after splenectomy, following a childhood cancer9,10 as after trauma.46,47
In conclusion, our study confirms that the previously reported excess BC observed after Hodgkin’s disease in childhood is due not only to chemotherapy and higher radiation dose to the breast, but also to a susceptibility that appears to be specific. Children treated for a neuroblastoma could also be at higher risk of developing BC. These results have to be confirmed by studies including more occurrences of BC.
Authors’ Disclosures of Potential Conflicts of Interest
Web Table 2. Risks for a Second Breast Tumor After a Childhood Cancer According to Chemotherapy
Acknowledgment
Lorna Saint-Ange edited the manuscript. We thank Armelle Leszczynski-Kramar, Gisèle Da Silva, and Veronique Mosserie for their help in data management, and all of the physicians and physicists who participated in the elaboration of the study or data collection at Gustave Roussy Institute, Thames Cancer Registry, Institut Godinot, Institut Curie, Centre Regaud, and Centre Lacassagne.
NOTES
Supported by grants from Europe Against Cancer (Contract 91CV01090-0) and Radiation Protection (Contracts F13P-CT92-0064 and F14P-CT95-0009) programs of the EEC, Institut Gustave Roussy, INSERM, CNAMTS, EDF, SNECMA, FRAMATOME, ARC, and the committees Champagne Ardennes, Essonne, and Val d’Oise of the Ligue Nationale Contre le Cancer.
Presented at the 40th Annual Meeting of the American Society of Clinical Oncology, New Orleans, LA, June 5-8, 2004.
Web Tables 1 to 3 are included in the full-text version of this article, available online only at www.jco.org. They are not included in the PDF (via Adobe Acrobat Reader) version.
Authors’ disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
1. United Nations: United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) Ionising Radiation: Sources and Biological Effects. 1994. Report to the General Assembly, with annexes. New York, NY, United Nations Sales Publications, 1994
2. Habrand JL, Bondiau PY, Dupuis O, et al: Effets tardifs de la radiotherapie chez l’enfant. Cancer/Radiother 1:810-816, 1997
3. Cleary SF, Link MP, Donaldson SS: Hodgkin’s disease in the very young. Int J Radiat Oncol Biol Phys 28:77-83, 1994
4. Tucker MA, Meadows AT, Boice JD, et al: Cancer risk following treatment of childhood cancer, in Boice JD, Fraumeni JF (eds): Radiation Carcinogenesis Epidemiology. New York, NY, Raven Press, 1984, pp 211-224
5. Meadows AT, Baum E, Fossati-Bellani F, et al: Second malignant neoplasms in children: An update from the Late Effects Study Group. J Clin Oncol 3:532-538, 1985
6. Hawkins MM, Draper GJ, Kingston JE: Incidence of second primary tumours among childhood cancer survivors. Br J Cancer 56:339-347, 1987
7. de Vathaire F, Schweisguth O, Rodary C, et al: Long-term risk of second malignant neoplasm after a cancer in childhood. Br J Cancer 59:448-452, 1989
8. Olsen JH, Garwicz S, Hertz H, et al: Second malignant neoplasms after cancer in childhood or adolescence. BMJ 307:1030-1036, 1993
9. Green DM, Zevon MA, Reese PA, et al: Second malignant tumours following treatment during childhood and adolescence for cancer. Med Pediatr Oncol 22:1-10, 1994
10. Neglia JP, Friedman DL, Yasui Y, et al: Second malignant neoplasms in five-years survivors of childhood cancer: Childhood Cancer Survivor Study. J Natl Cancer Inst 93:618-629, 2001
11. Hancock SL, Tucker MA, Hoppe RT: Breast cancer after treatment of Hodgkin’s disease. J Natl Cancer Inst 85:25-31, 1993
12. Bhatia S, Robison LL, Oberlin O, et al: Breast cancer and other second neoplasms after childhood Hodgkin’s disease. N Engl J Med 334:745-751, 1996
13. Sankila R, Garwicz S, Olsen JH, et al: Risk of subsequent malignant neoplasms among 1,641 Hodgkin’s disease patients diagnosed in childhood and adolescence: A population-based cohort study in the five Nordic countries. J Clin Oncol 14:1442-1446, 1996
14. Green DM, Hyland A, Barcos MP, et al: Second malignant neoplasms after treatment for Hodgkin’s disease in childhood or adolescence. J Clin Oncol 18:1492-1499, 2000
15. Aisenberg AC, Finkelstein DM, Doppke KP, et al: High risk of breast carcinoma after irradiation of young women with Hodgkin’s disease. Cancer 79:1203-1210, 1997
16. Swerdlow AJ, Barber JA, Hudson GV, et al: Risk of second malignancy after Hodgkin’s disease in a collaborative british cohort: The relation to age at treatment. J Clin Oncol 18:498-509, 2000
17. Bhatia S, Yasui Y, Robison LL, et al: High risk of subsequent neoplasms continues with extended follow-up of childhood Hodgkin’s disease: Report from the Late Effects Study. J Clin Oncol 21:4386-4394, 2003
18. de Vathaire F, Hawkins M, Campbell S, et al: Second malignant neoplasm after a first cancer in childhood: Temporal pattern of risk according to type of treatment. Br J Cancer 79:1884-1893, 1999
19. Le Vu B, de Vathaire F, Shamsaldin A, et al: Radiation dose, chemotherapy and risk of osteosarcoma after solid tumors during childhood. Int J Cancer 77:370-377, 1998
20. Little MP, de Vathaire F, Shamsaldin A, et al: Risks of brain cancer following treatment for cancer in childhood. Int J Cancer 78:269-275, 1998
21. de Vathaire F, Hardiman C, Shamsaldin A, et al: Thyroid carcinomas after irradiation for a first cancer during childhood. Arch Intern Med 13:2713-2719, 1999
22. Fritz A, Percy C, Jack A, et al (eds): International Classification of the Diseases for Oncology (ed 2). Geneva, Switzerland, WHO, 1991
23. Diallo I, Lamon A, Shamsaldin A, et al: Estimation of the radiation dose delivered at any point in the body for individual patients in external beam radiotherapy. Radiother Oncol 38:269-271, 1996
24. Shamsaldim A, Grimaud E, Hardiman C, et al: Dose distribution throughout the body for radiotherapy for Hodgkin’s disease in childhood. Radiother Oncol 49:85-90, 1998
25. Gooley TA, Leisenring W, Crowley J, et al: Estimation of failure probabilities in the presence of competing risks: New representations of old estimators. Stat Med 18:698-706, 1999
26. Menegoz F, Cherie-Challine L: Le cancer en France: Incidence et mortalite: Situation en 1995—Evolution entre 1975 et 1995. La Documentation Franaise, Paris, France, Reseau FRANCIM 2000, 1998
27. Office for National Statistics: Cancer 1971-1997 CD-ROM. London, United Kingdom, ONS, 1999
28. Breslow NE, Day NE: Fitting models to grouped data. Statistical methods in cancer research. The design and analysis of cohort studies. Lyon, France, International Agency for Research on Cancer, IARC Scientific Publication 82, 1987
29. Preston DL, Lubin JH, Pierce DA: EPICURE User’s guide. Seattle, WA, HiroSoft International Corp, 1993
30. Venzon DJ, Moolgavkar SH: A method for computing profile-likelihood-based confidence intervals. Appl Stat 37:87-94, 1988
31. Coleman MP: Trends in breast cancer incidence, survival, and mortality. Lancet 356:590-591, 2000
32. Metayer C, Lynch CF, Clark EA, et al: Second cancers among long-term survivor’s of Hodgkin’s disease diagnosed in childhood and adolescence. J Clin Oncol 18:2435-2443, 2000
33. Hawkins MM: Second primary tumors following radiotherapy for childhood cancer. Int J Radiat Oncol Biol Phys 19:1297-1301, 1990
34. UNSCEAR Sources and Effects of Ionizing Radiation; Report to the General Assembly, with Scientific Annexes. New York, NY, UNSCEAR, 2000
35. Lundell M, Mattsson A, Hakulinen T, et al: Breast cancer after radiotherapy for skin hemangioma in infancy. Radiat Res 145:225-230, 1996
36. Lundell MN, Mattson A, Karlsson P, et al: Breast cancer risk after radiotherapy in infancy: A pooled analysis of two Swedish cohorts of 17,202 infants. Radiat Res 151:626-632, 1999
37. Lindberg S, Karlsson P, Arvidsson B, et al: Cancer incidence after radiotherapy for skin haemangioma during infancy. Acta Oncol 34:735-740, 1995
38. Modan B, Chetrit A, Alfandary E, et al: Increased risk of breast cancer after low-dose irradiation. Lancet 1:629-631, 1989
39. Hoffman DA, Lonstein JE, Morin MM: Breast cancer in women with scoliosis exposed to multiple diagnostic X rays. J Natl Cancer Inst 81:1307-1312, 1989
40. Boice JD Jr, Preston D, Davis FG, et al: Frequent chest X-ray fluoroscopy and breast cancer incidence among tuberculosis patients in Massachusetts. Radiat Res 125:214-222, 1991
41. Tokunaga M, Land CE, Tokuoka S, et al: Incidence of female breast cancer among atomic bomb survivors, 1950-1985. Radiat Res 138:209-223, 1994
42. Land CE: Studies of cancer and radiation dose among atomic bomb survivors: The example of breast cancer. JAMA 274:402-407, 1995
43. Little MP: Comparison of the risks of cancer incidence and mortality following radiation therapy for benign and malignant disease with the cancer risks observed in the Japanese A-bomb survivors. Int J Radiat Biol 77:431-464, 2001
44. Little MP, Boice JD Jr: Comparison of breast cancer incidence in the Massachusetts tuberculosis fluoroscopy cohort and in the Japanese atomic bomb survivors. Radiat Res 151:218-224, 1999
45. Garwicz S, Anderson H, Olsen HJ, et al: Second malignant neoplasms after cancer in childhood and adolescence: A population-based case-control study in the 5 Nordic countries. Int J Cancer 88:672-678, 2000
46. Mellemkjaer L, Olsen JH, Linet MS, et al: Cancer risk after splenectomy. Cancer 75:577-583, 1995
47. Linet MS, Nyren O, Gridley G, et al: Risk of cancer following splenectomy. Int J Cancer 66:611-616, 1996(Catherine Guibout, Elisab)
Department of Paediatrics, Institut Curie, Paris
Department of Radiotherapy, Institut Jean Godinot, Reims
Department of Radiotherapy, Centre Lacassagne, Nice
Department of Radiotherapy, Centre Claudius-Regaud, Toulouse, France
Childhood Cancer Research Group, Radcliffe Infirmary, Oxford, United Kingdom
ABSTRACT
Web Table 1. Chemotherapy: Drugs Administered to 1,167 of 1,814 Female Children Treated for a First Cancer in Childhood
Web Table 3. Breast Cancers According to the Mean Radiation Dose to the Breast in a Cohort of 1,814 Patients Treated for a First Cancer During Childhood
PATIENTS AND METHODS: In a cohort of 1,814 3-year female survivors treated between 1946 and 1986 in eight French and English centers, data on chemotherapy and radiotherapy were collected. Individual estimation of radiation dose to each breast was performed for the 1,258 patients treated by external radiotherapy; mean dose to breast was 5.06 Gy (range, 0.0 to 88.0 Gy) delivered in 20 fractions (mean).
RESULTS: Mean follow-up was 16 years; 16 patients developed a clinical BC, 13 after radiotherapy. The cumulative incidence of BC was 2.8% (95% CI, 1.0% to 4.5%) 30 years after the FC and 5.1% (95% CI, 2.1% to 8.2%) at the age of 40 years. The annual excess incidence increased as age increased, whereas the standardized incidence ratio decreased. On average, each Gray unit received by any breast increased the excess relative risk of BC by 0.13 (< 0.0 to 0.75). After stratification on castration and attained age, and adjusting for radiation dose, FC type, and chemotherapy, a higher risk of a subsequent BC was associated with Hodgkin’s disease (relative risk, 7.0; 95% CI, 1.4 to 30.9).
CONCLUSION: The reported high risk of BC after childhood Hodgkin’s disease treatment seems to be due not only to a higher radiation dose to the breasts, but also to a specific susceptibility.
INTRODUCTION
We report here the results of a subcohort study designed to evaluate the long-term risk of a subsequent BC after fractionated high doses of external-beam radiotherapy. Radiation doses to each breast were separately estimated for every child with specialized software. An attempt was made to assess the role of the FC type in determining the risk of subsequent BC, controlling for treatment and particularly for the radiation dose to the breasts.
PATIENTS AND METHODS
these 4,400 children, only the female patients were selected because among males, no BC was observed. One hundred fifty-one patients were excluded because they had received brachytherapy or information on radiation doses was not available. Among the 1,814 remaining women, 1,258 had received external-beam radiotherapy. The end of the period at risk was the date of the last contact, death, or the last time the patient was known to be alive; for those with a BC, this was the date of the diagnosis. The follow-up data were obtained using registries for the English patients or, for the French patients, from hospital clinical files actualized by physicians who performed an active follow-up by writing to patients and linking their database with the national death certificate data. Twelve percent of the cohort (n = 210) was lost to follow-up.
Information on treatments was abstracted from the clinical notes and radiotherapy files in the participating centers and the diagnoses of the first and second malignant neoplasms were histologically confirmed. The International Classification of the Diseases for Oncology was used to classify the tumors.22
Radiation Dosimetry
The radiation doses were estimated in the middle of the right and left nipples for each of the 1,258 patients treated with ionizing radiation. The dose to 149 other anatomic sites including the thymus, spleen, gonads, and 91 skeletal sites was also estimated. A computer program called Dos_EG, described elsewhere,23,24 was developed for these calculations.
In the estimation, all treatments delivered on the same day were considered as one fraction. This definition is considered acceptable, given that there was no hyperfractionated treatment in the cohort.
Among the 1,258 patients who received external radiotherapy, the mean number of fractions was 20 and the average radiation dose for the 1,258 patients was 5.06 Gy for a breast (Table 2). Figure 1 shows the distribution of the radiation dose to breast in Grays (median, 0.85 Gy). Because of variations in the radiotherapy instruments and in the treatment modalities of the FC, the mean dose to the breast varied according to the calendar period: 3.5 Gy before 1960, 4.4 Gy during the 1960s, 6.3 Gy during the 1970s, and 3.6 Gy during 1980 to 1985.
Measurement of Chemotherapy
Chemotherapy doses were not available for four of the 1,167 patients who received chemotherapy: 360 had chemotherapy alone with (n = 326) or without (n = 34) surgery. Drugs were grouped into five categories according to their known mechanism of action in the cell: electrophilic agents, spindle inhibitors, nucleotide synthesis inhibitors (NSI), topoisomerase II inhibitors, and other compounds. To quantify the total amount of each drug category administered, the dose of each cytotoxic agent was converted into moles per square meter, rather than into milligrams per square meter. A detailed description of the chemotherapy administered to cohort patients has been published.18 The median doses for each drug category and main drugs administered to the children were calculated (Web Table 1).
Statistical Methods
For analysis, the follow-up ended on January 1, 1993, for patients treated in the French centers and January 1, 1991, for those treated in United Kingdom. We used the competing method to estimate the cumulative incidence of breast tumors.25 The expected numbers of BC (number of tumors) per 5-year calendar periods and per 5-year age classes were calculated from 1975 to 1995 Francim data (based on the French cancer registries) for the French patients26 and on the England and Wales cancer registry data for the United Kingdom patients.27 The SIR was defined as the ratio between the observed and expected number of breast tumors. The relative risk (RR) was defined in a similar manner: the expected number of breast tumors was estimated from a given reference category within the cohort; an individual breast was used as the statistical unit.
Both the SIR and RR of BC were modeled assuming that the number of BC (number of tumors) followed Poisson distribution. Statistical tests were carried out using deviance of nested models.28
The basic model for calculating the excess relative risk (ERR) of BC as a function of the radiation dose was that generally used in radiation epidemiology29: ERR = Cst ({alpha}dose + dose2) [exp(–{gamma} dose)], where the constant (Cst) summarizes the role of dichotomous variables such as having had a given type of FC (yes or no) or chemotherapy; the negative exponential term models the cell sterilization for high doses.
We used an empirical approach to investigate the role of dose fractionation and introduced a threshold number of fractions to test the similarity of the coefficients for the doses delivered with a high (dose1) and low (dose2) number of fractions in a linear model, using the following equation: ERR = Cst ({alpha}2dose2 + {alpha}1dose1).
The simple linear model of the risk as a function of the total dose is nested in this model: ERR = Cst [{alpha}2(dose – dose1) + {alpha}1dose1] = Cst ({alpha}2dose – {alpha}2dose1 + {alpha}1dose1) = Cst [{alpha}2dose + ({alpha}1 – {alpha}2) dose1].
The effect of the dose rate was considered significant with this model if the decrease in deviance between model 2 fractionated dose and model 1 total dose limited to the linear term was greater than 3.84 (ie, the value of {chi}21 for P = .05). The standard deviation of the coefficient for doses delivered with fewer than 20 fractions (ie, {alpha}1 – {alpha}2) was estimated as follows: standard deviation .
These analyses were done using AMFIT Software.29 The 95% CIs were estimated for parameters using the maximum likelihood method.30
RESULTS
During the follow-up, 16 patients had developed a clinical BC (0.95 expected), excluding carcinoma in situ: 11 cancers arose in the left breast and three in the right; in two patients, this information was not available. The histological type was invasive ductal carcinoma in 12 patients (nine of them had a grade II or III according to the Scarff, Bloom, and Richardson scale, a histoprognostic grading system of invasive breast tumors based on the degree of differentiation of the tumor, mitotic index, and anisonucleosis, grade I to III, where I is a good prognosis), malignant phyllodes tumor in three patients, and tumor of an unknown type in one patient. Among the carcinomas, six patients presented with axillary involvement.
Thirteen patients developed a BC after radiotherapy, either with or without chemotherapy, with a follow-up between 13 and 37 years. One patient developed a BC (at 9 years of follow-up) for an FC treated by surgery alone; two patients developed a BC after chemotherapy (follow-up, 18 and 19 years, respectively). All but one cancer occurred in patients with a dose less than 20 Gy to the breast containing the tumor. The treatment characteristics of the children who developed a secondary breast tumor are listed in Table 3.
The cumulative incidence of BC after a 30-year follow-up was 2.8% (95% CI, 1.0% to 4.5%) and 10.7% (95% CI, 1.4% to 19.9%) after 40 years. Figure 2 depicts the cumulative incidence by attained age for the entire cohort. The annual incidence of BC increased dramatically to reach 537 x 10–5 person years (95% CI, 269 to 942) when women were 30 to 39 years old. The annual excess incidence increased considerably to 515 x 10–5 person years (95% CI, 246 to 919) for those aged 30 to 39 years, but the SIR declines from 185 (95% CI, 31 to 573) in the 10- to 19-year age group to 23.2 (95% CI, 12 to 41) for the 30- to 39-year age group (Table 4).
A total of 222 patients underwent a surgical or radiologic castration. None of them developed a secondary BC later (P = .1).
Radiotherapy, measured as ever versus never, was not associated with an increased risk of BC (RR, 1.3; 95% CI, 0.4 to 5.9). The ERR of occurrence of a BC was found to be nonsignificantly linearly linked to the total dose of radiation (ERR/Gy, 0.13; 95% CI, < 0.0 to 0.75; P = .06) received by the breast, after adjustment for castration, chemotherapy, and FC type. When model 2 with a threshold at 20 fractions was used, a significant effect of the dose fractionation was observed (P = .05): the radiation dose delivered in less than 20 fractions led to a higher risk (ERR/Gy = 0.12; 95% CI, –0.14 to 0.38) than when it was administered in 20 fractions or more (ERR/Gy = –0.02; 95% CI, –0.06 to 0.03).
Overall, chemotherapy was associated with a nearly significant RR of 2.7 (95% CI, 0.9 to 7.6; P = .07) when stratified on castration and attained age, and adjusted for radiation dose and FC type compared with no chemotherapy (Web Table 2). When we examined the role of the different classes of chemotherapeutic agents, we observed that electrophilic agents, spindle inhibitors, and topoisomerase II inhibitors had no significant effect on the risk of developing a BC in this cohort (Web Table 2). When stratification on castration and attained age plus adjustment on other drug categories, FC type, age at FC, and radiation dose to the breast were made, the RR for NSI (ie, for methotrexate) was found to be 4.6 (95% CI, 0.7 to 19.9; Table 5). Only 62 patients received mechlorethamine, vincristine, procarbazine, and prednisone, of whom two developed a BC. There was no evidence for a role of the number of moles per square meter for any of the drug categories or of the total number of moles per square meter administered.
Four of the 16 BCs occurred among the 123 patients treated for Hodgkin’s disease (SIR, 71; 95% CI, 22 to 164), three occurred among the 276 treated for a neuroblastoma (SIR, 47; 95% CI, 12 to 122), three occurred among the 373 treated for a Wilms’ tumor (SIR, 28; 95% CI, 7.0 to 73), two occurred among the 209 patients treated for soft tissue sarcoma (SIR, 16; 95% CI, 2.7 to 51), and one in each of the following FC types: brain, gonad, and non-Hodgkin’s lymphoma (Table 3). A heterogeneity was evidenced in SIR between FC (P = .02).
After Hodgkin’s disease, none of the four BCs occurred before 16 years after first diagnosis, and breasts of these patients received a mean dose of 11.84 Gy, whereas for all other patients, the mean dose was 4.39 Gy. The latent periods after neuroblastoma and Wilms’ tumors were, respectively, 13 to 37 years and 18 to 36 years, and the mean doses received by the breasts were 5.85 Gy (v 4.95 Gy when excluding neuroblastoma patients) and 7.04 Gy (v 4.51 Gy for non-Wilms’ tumor patients). When patients treated for these three types of malignancies were excluded, the mean dose administered to the breasts decreased to 3.06 Gy. When age at FC, attained age, castration, radiation dose to the breast (in Grays), and chemotherapy as yes versus no were taken into account, the risk for developing a BC was higher in any breast of patients previously treated for Hodgkin’s disease (RR, 7.01; 95% CI, 1.4 to 30.9) or, nonsignificantly, for neuroblastoma (RR, 2.9; 95% CI, 0.6 to 11.6) than in those treated for any other FC combined (Table 5).
Of the 264 patients who underwent a surgical or radiologic splenectomy (> 20 Gy to the spleen), three developed a BC. No evidence for a spontaneous higher risk of BC (RR, 1.8; 95% CI, 0.6 to 5.5) was found. Similarly, a modification of the risk of BC was not evidenced among the 242 patients (of whom three had a secondary BC) who had received a mean dose to active bone marrow exceeding 10 Gy (RR, 1.4; 95% CI, 0.3 to 5.9). Age at FC does not affect the risk of BC, whereas risk of BC increased with increasing doses to the breasts up to 20 Gy (P = .06; Table 5).
DISCUSSION
The women lost to follow-up (n = 210) were not different from the others in terms of the date of the diagnosis (1974 on average). They did, however, receive fewer moles of drugs per square meter (13 v 30 moles/m2; P = .04), but the average radiation dose to the breasts was the same (3.4 v 3.4 Gy). Given that the cumulative dose of chemotherapy can be linked to relapses, the difference in the amount of chemotherapy administered appears to indicate that we were more likely to lose to follow-up patients who did not experience relapse (ie, patients with less aggressive therapy), and this could lead to an overestimation of the long-term incidence of BC. However, the likelihood of overestimation is probably limited given the low rate of patients lost to follow-up in our cohort (12%).
The first limitation is the relatively small size of our cohort, which included 1,814 patients and only 16 BCs. This allows us to satisfactorily only risk factors associated with substantially increased risks. A possible bias was the heightened medical surveillance of women who survived a childhood cancer, leading to a greater likelihood of detecting other diseases, including second malignancies. Because there is no French national cancer registry, the follow-up of cohorts such as ours must be active, which means that this bias could never be eliminated. Furthermore, screening for BC is likely to increase the observed incidence of breast tumors.31 In our opinion, this potential bias is low because no routine BC screening program existed during our study follow-up period, and the size and extent of the 16 BCs diagnosed in our cohort were more aggressive (Scarff, Bloom, and Richardson scale grade II or III) than those usually found during screening for prevention.
In our cohort, the cumulative incidence of BC at 40 years of age was 5.1% (95% CI, 2.1 to 8.2) corresponding to 16.9-fold the rate expected (95% CI, 9.9 to 26.6) in the general population. Although no patients with leukemia are included in our cohort, this global SIR is within the range observed for BC in hospital-based studies,4,10,32 but higher than that observed in registry-based cohorts.6,8 However, an overall comparison between BC SIR observed in cohorts is very difficult to interpret because it is highly dependent on the duration of follow-up, the average attained age, and the expected BC incidence, which is unstable before 45 years of age because it is based on small numbers, and because of differences in the FC types between cohorts.
For the SIR with attained age, we observed a decrease from 185 (95% CI, 31 to 573) in the 10- to 19-year age group to 23.2 (95% CI, 12 to 41) in the 30- to 39-year age group. One BC (0.35 expected) was observed in our cohort after 40 years of age, but only 60 women had reached this age during the follow-up period. Variation in BC SIR with attained age or time since FC was examined in detail in only one study, which obtained similar results.8 Because of the strong increase in expected BC incidence with attained age, this reduction in SIR nevertheless corresponds to a substantial increase in excess incidence with attained age, at least during the first four decades.8
We confirm previously reported8,11-13,16,17 higher BC incidence after Hodgkin’s disease (RR, 7.01 compared with BC risk after other FC) and demonstrate, for the first time, that this excess is not due to higher radiation doses and/or chemotherapy.
In addition, we found that survivors of childhood neuroblastoma could be at a greater risk (RR, 2.9) of developing a BC. To date, this result has not been reported in any other study. Although Wilms’ tumor was the most frequent FC in our cohort (n = 373), it was not associated with an increased BC risk.10,33 However, one BC was observed among Hodgkin’s disease and neuroblastoma patients who had no radiotherapy or a low dose to breasts (Web Table 3). Thus, the observed increased risks of BC after such FC suggest either a specific susceptibility for developing a BC or a particular susceptibility to radiation and/or chemotherapy, or both.
Although the number of patients treated by radiotherapy in our study is relatively limited (n = 1,258), the range of the doses available for the investigation of a dose-response relationship was wide (from 0.0015 to 88.0 Gy received by the breast): 514 breasts received less than 0.2 Gy and 197 received more than 20 Gy. Radiotherapy was not found to be a risk factor for BC occurrence when considered as a dichotomous variable. Nevertheless, this risk was found to be nonsignificantly linearly linked to the radiation dose received by any breast with an ERR/Gy of 0.13 (95% CI, < 0.0 to 0.75). This result confirms established effects of radiation exposure1,34 and has been described in many studies. To our knowledge, no other study using individual radiation dose estimation to breasts has provided an estimate of a coefficient for BC occurrence after radiotherapy treatment for a malignancy. Our result is consistent with that observed after radiotherapy for benign diseases in childhood: hemangioma,35-37 tinea capitis,38 and multiple diagnosis x-rays.39,40 On the contrary, it is lower than that described for the Hiroshima-Nagasaki survivors.41,42
The higher risk observed for the same total dose to a breast delivered in less than 20 fractions (ERR/Gy = 0.12) compared with delivery in more fractions (ERR/Gy = –0.02) appears to be in contradiction with the linear dose-response we estimated with a negative exponential term. Nevertheless, it is noteworthy that model 1 (total dose) is not adequate for fractionated irradiation.43 Women subjected to repeated fluoroscopy examinations during the follow-up of tuberculosis had a BC risk that was similar to that of female atomic bomb survivors who had received a similar total dose.40,44
We were unable to identify any substantial decrease in the doses delivered to the breasts after changes of treatment instruments and energy (low-energy x-rays, high-energy x-rays, electrons, cobalt). Seven BCs occurred after low-energy x-ray therapy. A follow-up of these patients (mean, 26 years; range, 3 to 46 years) longer than that attained in patients treated by the other instruments (mean, 14 years; range, 3 to 36 years) can explain this high number.
In this cohort, after stratification on castration and attained age and adjustment on radiation dose and FC type, the RR associated with chemotherapy was weak (RR, 2.7; 95% CI, 0.9 to 7.6) and chemotherapeutic drug classes were not found to increase the risk of BC except for the NSI class (only methotrexate), which seemed more likely to be associated with an increased risk of BC. However, only two BCs occurred after the administration of this cytotoxic agent. This increased risk due to chemotherapy is in agreement with one study45 in which alkylating agents and vinca alkaloids, but not NSI, were associated with a more frequent occurrence of BC, but not with another study,10 which failed to determine such an association. In these two studies, adjustment for radiotherapy was only addressed as yes or no.
The distribution of laterality of the 16 BCs observed in this cohort is different from that observed in the French breast cancer registries. Thus, 11 of the 14 BCs (approximately 80%) for which information was available were located in the left breast (the percentage is usually approximately 50%).26 Regarding the histologic type, three among the 14 BCs with this information are infrequent malignant phyllodes tumors; the percentage in the French BC registries is 0.18%.26
We also investigated the potential effect of radiation doses to organs influencing the immunologic status (spleen or active bone marrow) on the occurrence of BC. We did not find any role like that found by others after splenectomy, following a childhood cancer9,10 as after trauma.46,47
In conclusion, our study confirms that the previously reported excess BC observed after Hodgkin’s disease in childhood is due not only to chemotherapy and higher radiation dose to the breast, but also to a susceptibility that appears to be specific. Children treated for a neuroblastoma could also be at higher risk of developing BC. These results have to be confirmed by studies including more occurrences of BC.
Authors’ Disclosures of Potential Conflicts of Interest
Web Table 2. Risks for a Second Breast Tumor After a Childhood Cancer According to Chemotherapy
Acknowledgment
Lorna Saint-Ange edited the manuscript. We thank Armelle Leszczynski-Kramar, Gisèle Da Silva, and Veronique Mosserie for their help in data management, and all of the physicians and physicists who participated in the elaboration of the study or data collection at Gustave Roussy Institute, Thames Cancer Registry, Institut Godinot, Institut Curie, Centre Regaud, and Centre Lacassagne.
NOTES
Supported by grants from Europe Against Cancer (Contract 91CV01090-0) and Radiation Protection (Contracts F13P-CT92-0064 and F14P-CT95-0009) programs of the EEC, Institut Gustave Roussy, INSERM, CNAMTS, EDF, SNECMA, FRAMATOME, ARC, and the committees Champagne Ardennes, Essonne, and Val d’Oise of the Ligue Nationale Contre le Cancer.
Presented at the 40th Annual Meeting of the American Society of Clinical Oncology, New Orleans, LA, June 5-8, 2004.
Web Tables 1 to 3 are included in the full-text version of this article, available online only at www.jco.org. They are not included in the PDF (via Adobe Acrobat Reader) version.
Authors’ disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
1. United Nations: United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) Ionising Radiation: Sources and Biological Effects. 1994. Report to the General Assembly, with annexes. New York, NY, United Nations Sales Publications, 1994
2. Habrand JL, Bondiau PY, Dupuis O, et al: Effets tardifs de la radiotherapie chez l’enfant. Cancer/Radiother 1:810-816, 1997
3. Cleary SF, Link MP, Donaldson SS: Hodgkin’s disease in the very young. Int J Radiat Oncol Biol Phys 28:77-83, 1994
4. Tucker MA, Meadows AT, Boice JD, et al: Cancer risk following treatment of childhood cancer, in Boice JD, Fraumeni JF (eds): Radiation Carcinogenesis Epidemiology. New York, NY, Raven Press, 1984, pp 211-224
5. Meadows AT, Baum E, Fossati-Bellani F, et al: Second malignant neoplasms in children: An update from the Late Effects Study Group. J Clin Oncol 3:532-538, 1985
6. Hawkins MM, Draper GJ, Kingston JE: Incidence of second primary tumours among childhood cancer survivors. Br J Cancer 56:339-347, 1987
7. de Vathaire F, Schweisguth O, Rodary C, et al: Long-term risk of second malignant neoplasm after a cancer in childhood. Br J Cancer 59:448-452, 1989
8. Olsen JH, Garwicz S, Hertz H, et al: Second malignant neoplasms after cancer in childhood or adolescence. BMJ 307:1030-1036, 1993
9. Green DM, Zevon MA, Reese PA, et al: Second malignant tumours following treatment during childhood and adolescence for cancer. Med Pediatr Oncol 22:1-10, 1994
10. Neglia JP, Friedman DL, Yasui Y, et al: Second malignant neoplasms in five-years survivors of childhood cancer: Childhood Cancer Survivor Study. J Natl Cancer Inst 93:618-629, 2001
11. Hancock SL, Tucker MA, Hoppe RT: Breast cancer after treatment of Hodgkin’s disease. J Natl Cancer Inst 85:25-31, 1993
12. Bhatia S, Robison LL, Oberlin O, et al: Breast cancer and other second neoplasms after childhood Hodgkin’s disease. N Engl J Med 334:745-751, 1996
13. Sankila R, Garwicz S, Olsen JH, et al: Risk of subsequent malignant neoplasms among 1,641 Hodgkin’s disease patients diagnosed in childhood and adolescence: A population-based cohort study in the five Nordic countries. J Clin Oncol 14:1442-1446, 1996
14. Green DM, Hyland A, Barcos MP, et al: Second malignant neoplasms after treatment for Hodgkin’s disease in childhood or adolescence. J Clin Oncol 18:1492-1499, 2000
15. Aisenberg AC, Finkelstein DM, Doppke KP, et al: High risk of breast carcinoma after irradiation of young women with Hodgkin’s disease. Cancer 79:1203-1210, 1997
16. Swerdlow AJ, Barber JA, Hudson GV, et al: Risk of second malignancy after Hodgkin’s disease in a collaborative british cohort: The relation to age at treatment. J Clin Oncol 18:498-509, 2000
17. Bhatia S, Yasui Y, Robison LL, et al: High risk of subsequent neoplasms continues with extended follow-up of childhood Hodgkin’s disease: Report from the Late Effects Study. J Clin Oncol 21:4386-4394, 2003
18. de Vathaire F, Hawkins M, Campbell S, et al: Second malignant neoplasm after a first cancer in childhood: Temporal pattern of risk according to type of treatment. Br J Cancer 79:1884-1893, 1999
19. Le Vu B, de Vathaire F, Shamsaldin A, et al: Radiation dose, chemotherapy and risk of osteosarcoma after solid tumors during childhood. Int J Cancer 77:370-377, 1998
20. Little MP, de Vathaire F, Shamsaldin A, et al: Risks of brain cancer following treatment for cancer in childhood. Int J Cancer 78:269-275, 1998
21. de Vathaire F, Hardiman C, Shamsaldin A, et al: Thyroid carcinomas after irradiation for a first cancer during childhood. Arch Intern Med 13:2713-2719, 1999
22. Fritz A, Percy C, Jack A, et al (eds): International Classification of the Diseases for Oncology (ed 2). Geneva, Switzerland, WHO, 1991
23. Diallo I, Lamon A, Shamsaldin A, et al: Estimation of the radiation dose delivered at any point in the body for individual patients in external beam radiotherapy. Radiother Oncol 38:269-271, 1996
24. Shamsaldim A, Grimaud E, Hardiman C, et al: Dose distribution throughout the body for radiotherapy for Hodgkin’s disease in childhood. Radiother Oncol 49:85-90, 1998
25. Gooley TA, Leisenring W, Crowley J, et al: Estimation of failure probabilities in the presence of competing risks: New representations of old estimators. Stat Med 18:698-706, 1999
26. Menegoz F, Cherie-Challine L: Le cancer en France: Incidence et mortalite: Situation en 1995—Evolution entre 1975 et 1995. La Documentation Franaise, Paris, France, Reseau FRANCIM 2000, 1998
27. Office for National Statistics: Cancer 1971-1997 CD-ROM. London, United Kingdom, ONS, 1999
28. Breslow NE, Day NE: Fitting models to grouped data. Statistical methods in cancer research. The design and analysis of cohort studies. Lyon, France, International Agency for Research on Cancer, IARC Scientific Publication 82, 1987
29. Preston DL, Lubin JH, Pierce DA: EPICURE User’s guide. Seattle, WA, HiroSoft International Corp, 1993
30. Venzon DJ, Moolgavkar SH: A method for computing profile-likelihood-based confidence intervals. Appl Stat 37:87-94, 1988
31. Coleman MP: Trends in breast cancer incidence, survival, and mortality. Lancet 356:590-591, 2000
32. Metayer C, Lynch CF, Clark EA, et al: Second cancers among long-term survivor’s of Hodgkin’s disease diagnosed in childhood and adolescence. J Clin Oncol 18:2435-2443, 2000
33. Hawkins MM: Second primary tumors following radiotherapy for childhood cancer. Int J Radiat Oncol Biol Phys 19:1297-1301, 1990
34. UNSCEAR Sources and Effects of Ionizing Radiation; Report to the General Assembly, with Scientific Annexes. New York, NY, UNSCEAR, 2000
35. Lundell M, Mattsson A, Hakulinen T, et al: Breast cancer after radiotherapy for skin hemangioma in infancy. Radiat Res 145:225-230, 1996
36. Lundell MN, Mattson A, Karlsson P, et al: Breast cancer risk after radiotherapy in infancy: A pooled analysis of two Swedish cohorts of 17,202 infants. Radiat Res 151:626-632, 1999
37. Lindberg S, Karlsson P, Arvidsson B, et al: Cancer incidence after radiotherapy for skin haemangioma during infancy. Acta Oncol 34:735-740, 1995
38. Modan B, Chetrit A, Alfandary E, et al: Increased risk of breast cancer after low-dose irradiation. Lancet 1:629-631, 1989
39. Hoffman DA, Lonstein JE, Morin MM: Breast cancer in women with scoliosis exposed to multiple diagnostic X rays. J Natl Cancer Inst 81:1307-1312, 1989
40. Boice JD Jr, Preston D, Davis FG, et al: Frequent chest X-ray fluoroscopy and breast cancer incidence among tuberculosis patients in Massachusetts. Radiat Res 125:214-222, 1991
41. Tokunaga M, Land CE, Tokuoka S, et al: Incidence of female breast cancer among atomic bomb survivors, 1950-1985. Radiat Res 138:209-223, 1994
42. Land CE: Studies of cancer and radiation dose among atomic bomb survivors: The example of breast cancer. JAMA 274:402-407, 1995
43. Little MP: Comparison of the risks of cancer incidence and mortality following radiation therapy for benign and malignant disease with the cancer risks observed in the Japanese A-bomb survivors. Int J Radiat Biol 77:431-464, 2001
44. Little MP, Boice JD Jr: Comparison of breast cancer incidence in the Massachusetts tuberculosis fluoroscopy cohort and in the Japanese atomic bomb survivors. Radiat Res 151:218-224, 1999
45. Garwicz S, Anderson H, Olsen HJ, et al: Second malignant neoplasms after cancer in childhood and adolescence: A population-based case-control study in the 5 Nordic countries. Int J Cancer 88:672-678, 2000
46. Mellemkjaer L, Olsen JH, Linet MS, et al: Cancer risk after splenectomy. Cancer 75:577-583, 1995
47. Linet MS, Nyren O, Gridley G, et al: Risk of cancer following splenectomy. Int J Cancer 66:611-616, 1996(Catherine Guibout, Elisab)