- Email: [email protected]
Thyroid Adenomas After Solid Cancer in Childhood
International Journal of
Radiation Oncology biology
physics
www.redjournal.org
Clinical Investigation: Pediatrics
Thyroid Adenomas After Solid Cancer in Childhood Nadia Haddy, PhD,*,y,z Chiraz El-Fayech, MD,*,y,z Catherine Guibout, PhD,*,y,z Elisabeth Adjadj, PhD,*,y,z Ce´cile Thomas-Teinturier, MD,*,x Odile Oberlin, MD,*,y Cristina Veres, MSc,*,y,z He´le`ne Pacquement, MD,jj Angela Jackson, MSc,*,y,z Martine Munzer, MD,{ Tan Dat N’Guyen, MD,{ Pierre-Yves Bondiau, MD,** Delphine Berchery, MD,yy Anne Laprie, MD,yy Andre´ Bridier, PhD,y Dimitri Lefkopoulos, PhD,y Martin Schlumberger, MD,y,z Carole Rubino, MD, PhD,*,y,z Ibrahima Diallo, PhD,*,y,z and Florent de Vathaire, PhD*,y,z *Radiation Epidemiology Group, INSERM, Villejuif, France; yInstitut Gustave Roussy, Villejuif, France; zUniv. Paris-Sud, Villejuif, France; xHoˆpital Biceˆtre, Biceˆtre, France; jjInstitut Curie, Paris, France; {Institut Jean Godinot, Reims, France; **Centre Antoine Lacassagne, Nice, France; and yyCentre Claudius Re´gaud, Toulouse, France Received Oct 13, 2011, and in revised form Feb 24, 2012. Accepted for publication Mar 19, 2012
Summary Based on a cohort of 3254 2-year survivors of childhood cancer, the risk of thyroid adenoma increases with the radiation dose received by the thyroid during childhood cancer treatment, and reaches a plateau for doses higher than approximately 10 Gy. Chemotherapy administration was found to modify the relationship between the radiation dose to the thyroid
Purpose: Very few childhood cancer survivor studies have been devoted to thyroid adenomas. We assessed the role of chemotherapy and the radiation dose to the thyroid in the risk of thyroid adenoma after childhood cancer. Methods and Materials: A cohort of 3254 2-year survivors of a solid childhood cancer treated in 5 French centers before 1986 was established. The dose received by the isthmus and the 2 lobes of the thyroid gland during each course of radiation therapy was estimated after reconstruction of the actual radiation therapy conditions in which each child was treated as well as the dose received at other anatomical sites of interest. Results: After a median follow-up of 25 years, 71 patients had developed a thyroid adenoma. The risk strongly increased with the radiation dose to the thyroid up to a few Gray, plateaued, and declined for high doses. Chemotherapy slightly increased the risk when administered alone but also lowered the slope of the dose-response curve for the radiation dose to the thyroid. Overall, for doses up to a few Gray, the excess relative risk of thyroid adenoma per Gray was 2.8 (90% CI: 1.2-6.9), but it was 5.5 (90% CI: 1.9-25.9) in patients who had not received chemotherapy or who had received only 1 drug, and 1.1 (90% CI: 0.4-3.4) in the children who had received more than 1 drug (PZ.06, for the difference). The excess relative risk per Gray was
Reprint requests to: Florent de Vathaire, PhD, Radiation Epidemiology Group, Unit 1018, INSERM, Institut Gustave Roussy, Rue Camille Desmoulins, 94805 Villejuif, France. Tel: þ33-1-42-11-54-57; Fax: þ33-1-4211-53-15; E-mail: [email protected] Supported by the Ligue Nationale Contre le Cancer, the Institut de Recherche en Sante´ Publique, the Programme Hospitalier de Recherche Clinique, the Agence Franc¸aise de Se´curite´ Sanitaire et Produit de Sante´, Electricite´ de France, and the Fondation Wyeth for childhood and adolescent health. Conflict of interest: none. Int J Radiation Oncol Biol Phys, Vol. 84, No. 2, pp. e209ee215, 2012 0360-3016/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.ijrobp.2012.03.044
Supplementary material for this article can be found at www.redjournal.org. AcknowledgmentsdThe authors are grateful to M. Labbe´ and C. Paoletti for their help in data management; to all the physicians and physicists who participated in the elaboration of the study at the Gustave Roussy Institute (Villejuif), Institut Godinot (Reims), Institut Curie (Paris), Centre Regaud (Toulouse), and Centre Lacassagne (Nice); and to L. Saint Ange for editing.
International Journal of Radiation Oncology Biology Physics
e210 Haddy et al. gland and the risk of thyroid adenoma by decreasing its initial slope.
also higher for younger children at the time of radiation therapy than for their older counterparts and was higher before attaining 40 years of age than subsequently. Conclusions: The overall pattern of thyroid adenoma after radiation therapy for a childhood cancer appears to be similar to that observed for thyroid carcinoma. Ó 2012 Elsevier Inc.
Introduction
Radiation dosimetry
Thyroid tissue is one of the most radiosensitive human tissues (1). The relationship between the radiation dose received by the thyroid gland during radiation therapy for a childhood cancer and the subsequent emergence of a differentiated thyroid carcinoma has been investigated in several studies (2-6). Thyroid adenomas after radiation therapy require investigation because of the medical and surgical impact of this benign pathology. Additionally, in animal models, thyroid adenomas could be precancerous lesions. Thyroid cancers could therefore develop from benign thyroid tumors and not necessarily be de novo. To date, few epidemiological studies have been published on thyroid adenomas after radiation therapy for a childhood cancer. We report the incidence and risk factors for thyroid adenomas in a cohort of 3254 childhood cancer survivors treated before 1986 and followed for 25 years on average.
External beam radiation therapy always involves unwanted irradiation of healthy organs remote from the target volume. This is mainly because of secondary radiations resulting from beam scattering from 1) the tissue volume within the geometric limits of the primary beam, 2) the beam limiting devices, and 3) any object within the primary beam. Radiation leakage through the protective shielding also contributes to the doses delivered to organs remote from the target volume. The radiation dose was estimated in the middle of the 2 lobes and in thyroid isthmus for patients who had received external radiation therapy. Our in-house software “Dos_EG,” which takes into account the previously mentioned dose components used for this calculation (8). The doses at 185 other anatomical sites including the thymus, spleen, gonads, and 91 skeletal sites were also estimated. To establish the effects of fractionation, we considered all treatments given on the same day as 1 fraction. This definition was acceptable because no hyperfractionated treatment had been delivered to the patients in the cohort. The radiation dose to the thyroid was defined as the average of the radiation dose received in the center of the isthmus and of the 2 lobes. The mean dose of radiation to the thyroid in the 2288 patients who had received external radiation therapy was 5.8 Gy, but the median value was only 0.8 Gy (range <0.001-72).
Methods and Materials Patients A retrospective cohort of 4568 children treated in 8 centers in France and the United Kingdom was constituted comprising patients who were alive 2 (French centers) or 3 (UK centers) years after a first solid cancer diagnosed before age 16 (before 15 years of age in the UK centers) and before 1986 (5). An updated analysis of the incidence of differentiated thyroid carcinoma in this cohort was recently published (6). The present analysis is restricted to the 3382 patients treated in France. Among these patients, 26 treated for a thyroid cancer and 86 treated with brachytherapy were excluded. The 3254 remaining patients were included, of whom 2288 had been treated with external beam radiation therapy. Information on treatments was abstracted from the clinical chart and radiation therapy files of the participating centers. The follow-up of the 3254 patients was initially assessed using medical records from treatment centers, and later via a selfcompleted questionnaire sent on Sept. 1, 2005. This comprehensive questionnaire provided information on the socioeconomic status, quality of life, and health outcomes and it was partially based on that of the British Childhood Cancer Survivor Study (7). A total of 2449 patients were still alive and were therefore considered eligible to receive the questionnaire. It was mailed to the 2095 patients for whom the most recent address was obtained from the National Health Insurance System and who had returned a signed consent agreement. This agreement included an authorization to contact their medical practitioner and medical facilities. A total of 1823 (74% of 2449) patients returned the completed questionnaire by Dec. 31, 2010. The diagnoses of the first cancer and thyroid adenoma were histologically confirmed.
Chemotherapy quantification Drugs were grouped into categories according to their known mechanisms of action in the cell: vinca alkaloids, antimetabolites, alkylating agents, anthracyclines, cytotoxic antibiotics, epipodophyllotoxins, and other drugs. Chemotherapy information could not be found for 18 patients. To quantify the total amount of drug administered in each class, the dose of each cytotoxic agent was converted into moles per square meter.
Statistical methods French patients were followed from 2 years of survival after diagnosis of the first cancer until the occurrence of thyroid adenoma, thyroid surgery for any other reason, response to the self-questionnaire, or the date of last medical contact. The risk factors for thyroid adenoma were identified using Cox’s proportional hazard model and Poisson regression (9). Multivariate analysis was performed to evaluate the radiation doses effect on thyroid adenomas. This analysis was adjusted on age at first cancer in years, attained age, gender, chemotherapy (none or 1 drug vs 2 drugs or more), and first cancer type (Hodgkin disease, brain tumor, vs others). The 95% confidence intervals were estimated for parameters using the methods of maximum of likelihood (10). Because exposure to radiation is not supposed to reduce the risk of thyroid
Volume 84 Number 2 2012
Thyroid adenomas after radiation therapy e211
adenoma, 90% confidence intervals were used for all risks or incidence figures presenting results on effects of radiation dose. Standard radiobiological models were used to model the variation in the relative risk (RR) and the absolute excess risk (AER) according to the radiation doses.
Results The median duration of follow-up was 25 years after the first cancer, and 904 people were followed for 30 years or more, among whom 720 had received radiation therapy. From 4 to 41 years after the first cancer, 71 patients had developed a thyroid adenoma and 67 of them had received radiation therapy. Of the 71 adenomas, 46 had arisen in patients who had returned the self-questionnaire. The histological size (main diameter of the largest nodule) of the adenomas ranged from 8 to140 mm (median Z 20 mm) and the clinical size ranged from 2 to 50 mm (median Z 19 mm). Five were classified as polymorphous adenoma, 3 as oxyphilic adenoma, and the others as follicular adenoma. The cumulative incidence of thyroid adenoma 40 years after the diagnosis of the primary cancer was 4.3% (95% CI: 3.3-5.6); this rate was 5.5% (95% CI: 4.2-7.2) after radiation therapy. The cumulative incidence at 40 years of age was 3.5% (95% CI: 2.8-4.6) (Supplementary Table), and 4.6% (95% CI: 3.4-6.3) at 50 years of age, when attained age was taken into account. Forty-five adenomas had developed among the 1027 women who had received radiation therapy, and 22 among the 1261 men. At the age of 40, the cumulative incidence of thyroid adenoma was 5.5 (95%: 4.1-7.4) in women and 1.9 (95% CI: 1.2-3.1) in men (Table 1) and it was, respectively, 6.6 (95% CI: 4.8-9.0) and 2.6 (95% CI: 1.7-4.2) among those who had received radiation Table 1 Incidence of thyroid adenoma according to demographical characteristics of 3254 2-y childhood cancer survivors Characteristics Gender Men Women Age at first cancer 0-1 2-4 5-9 10þ Year of diagnosis 1940-1969 1970-1974 1975-1979 1980-1985 Treatment No RT, no CT RT, no CT No RT, CT CT and RT RT and 1 drug RT and 2 drugs or more
therapy. After adjustment on age at diagnosis, type of chemotherapy administered for the first cancer and the radiation dose to the thyroid, the incidence of thyroid adenomas was 2.9-fold higher (95% CI: 1.8-4.8) in women than in men. The incidence of thyroid adenoma varied considerably according to the type of first cancer. The cumulative incidence at age 40 ranged from 0 in bone sarcoma and retinoblastoma survivors to 10.7% (95% CI: 6.6-17.0) in those who had been treated for Hodgkin disease. However, when adjusted for treatment variables and using nephroblastoma survivors as the reference group, the differences in the RR of its onset were substantially reduced and were only of borderline significance (Supplementary Table). When both the time since the childhood cancer and attained age were taken into account, no clear time trend was observed in the RR of developing thyroid adenoma (Table 2). Compared with that observed in patients who had not received radiation therapy, the RR of thyroid adenoma was 18.4 (90% CI: 7.4-45.4) in patients who had received from 5 to 19 Gy to the thyroid and 11.5 (90% CI: 2.7-41.9) in those who had received 40 Gy or more. These findings remained similar when the analysis was restricted to patients who had returned their questionnaire (Table 3). Figure 1 shows the cumulative incidence according to the thyroid radiation dose category. A model with a linear coefficient of the dose plus an exponential term for high radiation doses to the thyroid fitted the data more adequately than a purely linear model (P<.0001). Adding a quadratic term did not improve the fit of the model (PZ.9). The excess relative risk per Gy (ERR/Gy) of developing thyroid adenoma was found to be 2.8 (90% CI: 1.2-6.9) for the entire cohort (Fig. 2) and 3.3 (90% CI: 1.3-9.4) when restricted to patients who had returned the self-questionnaire. The linear parameter of the dose response was significantly higher (PZ.02) when radiation therapy had been delivered before age 5 (ERR/Gy Z 3.4, 90% CI: 1.5-8.3) than when it had been delivered at an older age (ERR/Gy Z 1.0, 90% CI: 0.3-3.1). It was also significantly higher (PZ.04) before attaining age 40 (ERR/Gy Z 3.0,
Thyroid Cumulative incidence adenomas/patients at 40 y old (95% CI) 22/1831 49/1423
1.9% (1.2-3.1) 5.5% (4.1-7.4)
17/905 23/757 16/797 15/795
3.5% 6.0% 3.2% 2.2%
(2.1-5.7) (3.8-9.5) (1.9-5.5) (1.2-4.0)
35/628 14/626 12/842 10/1158
6.3% 2.8% 2.7% 2.3%
(4.5-8.8) (1.5-5.2) (1.5-4.7) (1.1-4.9)
0/208 21/647 4/758 46/1641 18/256 28/1321
4.6% 1.2% 4.4% 10.2% 3.2%
0 (2.9-7.2) (0.4-3.3) (3.2-6.0) (6.4-16.0) (2.1-4.7)
Abbreviations: CI Z confidence interval; CT Z computed tomography; RT Z radiation therapy.
Table 2 Incidence of thyroid adenoma, according to the time since first cancer treatment and to attained age in a cohort of 3254 patients treated for a cancer during childhood Differentiated thyroid adenomas Time scale
Patients still Annual incidence Adjusted relative followed up n 1000 risk (95% CI)*
Years since first cancer 2-9 3254 1 10-19 2696 36 20-29 2181 23 30-39 904 10 40 227 1 Attained age in years 2-19 3254 16 20-29 2524 31 30-39 1649 19 40-49 605 4 50 121 1
0.04 1.5 1.4 2.0 0.9
(0.01-0.2) (1.0-2.0) (0.9-2.1) (1.0-3.4) (0.05-3.9)
0.03 1 1 1.4 0.7
(0.002-0.1) (ref) (0.6-1.7) (0.6-2.7) (0.04-3.1)
0.5 1.4 1.7 1.3 1.8
(0.3-0.7) (1.1-2.0) (1.1-2.6) (0.4-3.0) (0.01-8.0)
0.3 1 1.2 1.0 1.5
(0.2-0.6) (ref) (0.7-2.1) (0.3-2.5) (0.08-7.0)
Abbreviation: CI Z confidence interval. * Relative risk, adjusted on gender, age at diagnosis, radiation dose to thyroid, chemotherapy (none or 1 drug vs 2 drugs or more), and first cancer type (Hodgkin disease, brain tumor, vs others).
International Journal of Radiation Oncology Biology Physics
e212 Haddy et al.
Table 3 Thyroid adenomas according to the radiation dose to the thyroid in a cohort of 3254 patients treated for a first cancer during childhood Dose to the thyroid (Gy) No RT
>0-0.9
Thyroid dose: mean, Gy 0 0.3 Whole cohort (nZ3254) Adenomas/patients 4/966 14/1235 RR (90% CI)* 1y 2.1 (0.8-5.4) Patients who returned the self-questionnaire (nZ1827) Adenomas/patients 3/622 9/703 RR (90% CI)* 1y 2.1 (0.7-6.4)
1-4.9
5-19.9
20-39.9
40
2.3
11.8
27.3
50.0
12/458 7.8 (3.0-20.3)
25/351 18.4 (7.4-45.4)
14/202 18.9 (6.9-51.5)
2/42 11.5 (2.7-41.9)
9/207 9.5 (3.1-28.6)
17/182 20.8 (7.3-59.5)
7/90 17.5 (5.2-59.2)
1/19 8.6 (1.3-59.4)
Abbreviations: CI Z confidence interval; RR Z relative risk. * Relative risk, adjusted on age at first cancer in years, attained age, gender, chemotherapy (none or 1 drug vs 2 drugs or more), and first cancer type (Hodgkin disease, brain tumor, vs others). y Reference category.
90% CI: 1.3-7.3) than after (ERR/Gy Z 0.8, 90% CI: 0.1-3.2). A comparable dose-response modification was not observed for follow-up, whatever the threshold tested. Because of a lack of convergence, it was not possible to investigate the effect of gender as a dose-response modifier. Among the 2288 patients who received radiation therapy, 1079 were treated in 20 or less fractions, of whom 35 later developed a thyroid adenoma and 1209 were treated in more than 20 fractions, of whom 32 developed a thyroid adenoma. We were not able to evidence a role of the fractionation in the reduction of risk (PZ.4): the ERR/Gy was 3.0 (90%:1.3-7.7) for the radiation dose delivered in 20 fractions or less and 2.5 (90% CI: 1.0-6.5) for radiation dose delivered in more than 20 fractions. Nevertheless it has to be noted that the slope was nonsignificantly reduced (ERR/ Gy Z 1.5, 90% CI: 0.4-4.9) if thyroid radiation dose received in more than 40 fractions (5 thyroid adenomas in 216 patients).
Of the 491 patients who had received more than 20 Gy to the spleen or had undergone a splenectomy, 22 had developed a thyroid adenoma. When adjusted for all other factors, including the radiation dose to the thyroid, neither a splenectomy nor radiation therapy to the spleen was found to be an individual risk factor for thyroid adenoma (PZ.3), nor a (thyroid) radiation doseresponse modifier (P>.5). Similar results were observed for the age-weighted radiation dose to the active bone marrow, the integral radiation dose, the radiation dose to the breasts and to the ovaries. Neither a hemilateral ovariectomy nor radiation delivered to the ovary (>20 Gy), nor a total ovariectomy, modified the risk of thyroid adenoma among women (P>.5) or the dose-response for the radiation dose to the thyroid (P>.5). Overall, chemotherapy (yes/no) did not increase the risk of thyroid adenoma (RR Z 0.7, 95% CI: 0.4-1.2), but patients who had received more than 1 drug had a borderline significantly
Fig. 1. Cumulative incidence of thyroid adenoma by attained age, in 4 categories of average radiation dose to the thyroid. Vertical bars represent the 90% CI.
Volume 84 Number 2 2012
Thyroid adenomas after radiation therapy e213
Fig. 2. Thyroid adenoma RR according to radiation dose to the thyroid. The circle are of observed RR for 6 dose categories and the curve is the linear-exponential dose-response model for RR calculated as RR Z 1 þ 2.8 dose exp (0.052 dose). Vertical lines Z 90% CIs for RR. Relative risk, adjusted on age at first cancer in years, attained age, gender, chemotherapy (none or 1 drug vs 2 drugs or more), and first cancer type (Hodgkin disease, brain tumor, vs others). RR Z relative risk. (PZ.1) decreased risk of thyroid adenoma (RR Z 0.7, 95% CI: 0.4-1.2), when adjusted for the radiation dose to thyroid and the type of childhood cancer (Table 4). Having received more than 1 chemotherapy drug also modified (almost significantly, PZ.06) the relationship between the radiation dose to the thyroid and the risk of thyroid adenoma, by decreasing the linear term; the ERR/Gy was 5.5 (90% CI: 1.9-25.9) among the 1237 patients who had received 0 or 1 drug, compared with an ERR/Gy Z 1.1 (90% CI: 0.4-3.4) among the 2017 patients who had received more than 1 drug. Topoisomerase II inhibitors were the only drugs that exerted an independent effect or by acting as a radiation dose-response modifier (Table 4).
When each drug was taken into account individually in the multivariate analysis, melphalan was associated with a reduced risk of thyroid adenoma, whereas rifomycin and vincristine were associated with an increased risk. However, given the potential role of chemotherapy, acting both as a main risk factor and as a dose-response modifier and the huge number of potential drug combinations in our cohort, it was not possible to correctly investigate the role of each drug and this result has to be considered with caution. The size of the thyroid adenoma did not vary significantly according to age at diagnosis (PZ.2), chemotherapy (PZ.7), follow-up (PZ.2), or the radiation dose to the thyroid (PZ.7), but
Table 4 Thyroid adenoma incidence, according to chemotherapy administration in a cohort of 3254 patients treated for a cancer during childhood Model without interaction Chemotherapy Chemotherapy
RR (95% CI)*
Any chemotherapy No 1 Yes 0.7 (0.4-1.2) Nb drugs 0 or 1 1 2 drugs or þ 0.7 (0.4-1.2) Topoisomerase II inhibitor No 1 Yes 0.6 (0.4-1.0)
Model with interaction
Thyroid radiation dose
Chemotherapy
Thyroid radiation dose
ERR/Gy (90% CI)
RR (95% CI)*
ERR/Gy (90% CI)
2.7 (1.2-6.8)
1 1.3 (0.3-21.6)
5.0 (2.7-12.7) 2.4 (1.0-6.2)
.5
2.8 (1.2-7.0)
1 2.6 (0.6-17.4)
5.5 (1.9-25.9) 1.1 (0.4-3.4)
.06
2.9 (1.3-7.3)
1 4.1 (0.7-33.6)
13.7 (1.8-38.5) 1.2 (0.5-3.4)
.02
P value for interaction
Abbreviations: CI Z confidence interval; ERR Z excess RR; Nb Z Number; RR Z relative risk. * Relative risk, adjusted on gender, age at diagnosis, attained age, radiation dose to thyroid, and first cancer type (Hodgkin disease, brain tumor, vs others).
e214 Haddy et al.
International Journal of Radiation Oncology Biology Physics
was on average larger in men, 36 mm (SEM Z 4.1 mm), than in women, 21 mm (SEM Z 2.1 mm). This difference remained significant (PZ.001) after adjustment on other factors.
That chemotherapy could independently increase the risk of thyroid adenoma, but could also reduce the effect of radiation doses (a dose-response modifier effect) is probably one of most important results of our study. A similar result was evidenced in the last publication by the CCSS, in which the ERR/Gy was 5.0 in patients who had not received radiation therapy and 1.1 in patients who had (P value for the difference Z .07) (4), but not in the other publications on thyroid cancer after childhood cancer (2, 3, 6). If confirmed, this result should not be interpreted as a protective effect of the association of these 2 treatment modalities. It is probably rather due to the fact that the percentage of patients who develop a thyroid tumor after chemotherapy is the same as the proportion who develop a thyroid tumor after thyroid radiation exposure and that the addition of the carcinogens does not increase the risk, which would be the case if the 2 carcinogens were considered independently. The fact that, in our cohort, the effect of radiation therapy seems to be focused on patients who had received topoisomerase II inhibitors should be cautiously interpreted and needs to be confirmed. Unlike that evidenced in the same cohort for thyroid carcinoma (6), but consistent with that observed in the CCSS (2-4) for thyroid carcinoma, we failed to find evidence either of an independent effect or a dose-response modifier effect of a splenectomy or radiation delivered to the spleen (>20 Gy) or of high radiation doses to the pituitary gland on the risk of thyroid adenoma. Similar to what we observed for thyroid carcinoma (6), the incidence of thyroid adenoma seemed to be significantly lower among survivors of central nervous system tumors, and higher among survivors of Hodgkin disease, than in the other childhood cancer survivors. Hodgkin disease is the type of childhood cancer exposed to the highest risk of thyroid cancer in all the childhood cancer survivor studies (2-4, 6, 11-14), but whether a specific susceptibility to thyroid tumors exists in these patients is a matter of debate. Indeed, until now, no other study has published the RR of developing a thyroid tumor per type of childhood cancer adjusted on the radiation dose to the thyroid and other risk factors. We failed to find evidence of a role played by radiation dose fractionation. In fact, childhood cancer survivors constitute a population that is probably far too homogenous to allow the investigation of this issue. Our finding of a similarity between the risk of thyroid adenomas and thyroid carcinomas after radiation therapy for a childhood cancer is in agreement with that found in other populationsdin particular, in patients who had received radiation therapy for benign reasons (15-18). In conclusion, our study shows that the risk of developing a thyroid adenoma after radiation therapy for childhood cancer has a pattern that is comparable to that observed for thyroid carcinoma: the risk increased with a few Gray, then it decreased, and younger children were more sensitive to this risk than older children. Chemotherapy was found to play a role both independently and as a dose-response modifier, but this will have to be confirmed.
Discussion Based on a cohort of 3254 2-year childhood cancer survivors treated before 1986 and followed for 25 years on average, we showed that the risk of thyroid adenoma increased with the radiation dose received by the thyroid during childhood cancer treatment, and plateaued at doses exceeding approximately 10 Gy. The risk of thyroid adenoma per unit of radiation dose to the thyroid was higher if radiation therapy had been delivered before age 5, and before the attained age of 40 years. Chemotherapy administration was found to slightly increase the risk of thyroid adenoma in subjects who had not received external radiation therapy and to modify the relationship between the radiation dose to the thyroid gland and the risk of thyroid adenoma by decreasing its initial slope. In our cohort, a surveillance bias probably exists because physicians know that long-term childhood cancer survivors treated with radiation therapy are at high risk of developing a thyroid adenoma. This surveillance bias is likely to increase the observed incidence of thyroid adenoma among these survivors. We think that this bias does not substantially affect the dose-response relationship because we did not observe any correlation between the level of radiation doses to the thyroid gland and the size of the thyroid adenomas. Another issue is that the follow-up of our cohort is partially retrospective and was done using a questionnaire for 1823 patients and hospital medical records for 1431 patients (deceased patients included). Even if only thyroid adenomas validated by a copy of the pathological record were included in this analysis, this difference could introduce a degree of bias. A model that includes a linear term plus a negative exponential term for cell killing is the best fit for our data on thyroid adenomas. This type of dose-response was also observed for thyroid carcinomas both in this cohort (6) and in the Childhood Cancer Survivor Study (CCSS) (2, 3). In the last publication of the CCSS, more complex models, including a quadratic term rather than a linear one and an exponential quadratic term rather than an exponential term fitted the data better than a linear exponential model. This did not modify the overall shape of the dose-response for thyroid carcinomas, which increased up to a few Gray, then plateaued, and decreased for higher doses (4). The initial slope of the relationship between the radiation dose to the thyroid and the risk of thyroid adenoma estimated at ERR/ Gy Z 2.8 (90% CI: 1.2-6.9) is consistent with the risk estimated in the same cohort for thyroid carcinomas, ERR/Gy Z 2.0, 95% CI: 0.6-6.4 (6), and in the CCSS for thyroid cancer. In our study, a younger age at radiation therapy increased the risk of radiation-induced thyroid adenoma. This is similar to that observed for differentiated thyroid carcinoma in all childhood cancer survivor studies (2-4, 6, 11). We also observed a significant decrease in the ERR/Gy with increasing attained age. Because attained age and follow-up since radiation therapy are closely linked, that we did not observe a dose-response modifier effect of follow-up is probably due to a lack of power. Our results are in agreement with what we observed for carcinoma in the publication on this cohort (6), but not with that observed in the CCSS (2-4).
References 1. United Nations. Ionising radiation: Sources and biological effects. Scientific Committee on the Effects of Atomic Radiation 1994. Report to the General Assembly, with annexes. New York: United Nations Sales Publications; 1994. 2. Sigurdson AJ, Ronckers CM, Mertens AC, et al. Primary thyroid cancer after a first tumour in childhood (the Childhood Cancer
Volume 84 Number 2 2012
3.
4.
5.
6.
7.
8.
9.
Survivor Study): a nested case-control study. Lancet 2005;365: 2014-2023. Ronckers CM, Sigurdson AJ, Stovall M, et al. Thyroid cancer in childhood cancer survivors: a detailed evaluation of radiation dose response and its modifiers. Radiat Res 2006;166:618-628. Bhatti P, Veiga LH, Ronckers CM, et al. Risk of second primary thyroid cancer after radiotherapy for a childhood cancer in a large cohort study: an update from the childhood cancer survivor study. Radiat Res 2010;174:741-752. De Vathaire F, Hardiman C, Shamsaldin A, et al. Thyroid carcinoma following irradiation for a first cancer during childhood. Arch Intern Med 1999;159:2713-2719. de Vathaire F, Haddy N, Hawkins M, et al. Thyroid carcinoma following childhood cancer: role of radiation therapy, splenectomy and chemotherapy. J Clin Endocrinol Metab. In press. Hawkins MM, Lancashire ER, Winter DL, et al. The British Childhood Cancer Survivor Study: objectives, methods, population structure, response rates and initial descriptive information. Pediatr Blood Cancer 2008;50:1018-1025. Diallo I, Lamon A, Shamsaldin A, et al. Estimation of the radiation dose delivered at any point in the body for individual patient in external beam radiotherapy. Radiother Oncol 1996;38: 269-297. Breslow NE, Day NE, and Statistical methods in cancer research. The design and analysis of cohort studies. IARC Scientific Publication No. 82. Lyon, France: International Agency for Research on Cancer; 1987.
Thyroid adenomas after radiation therapy e215 10. Moolgavkar SH, Venzon DJ. A method for computing profile likelihood based confidence bounds. Ann Stat 1987;15:346-359. 11. Taylor AJ, Croft AP, Palace AM, et al. Risk of thyroid cancer in survivors of childhood cancer: results from the British Childhood Cancer Survivor Study. Int J Cancer 2009;125:2400-2405. 12. O’Brien MM, Donaldson SS, Balise RR, et al. Second malignant neoplasms in survivors of pediatric Hodgkin’s lymphoma treated with low-dose radiation and chemotherapy. J Clin Oncol 2010;28:1232-1239. 13. Olsen JH, Mo¨ller T, Anderson H, et al. cancer incidence in 47,697 patients treated for childhood cancer in the Nordic countries. J Natl Cancer Inst 2009;101:806-813. 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 2000;18:1492-1499. 15. Schneider AB, Ron E, Lubin J, et al. Dose response relationships for radiation-induced thyroid cancer and thyroid nodule: evidence for the prolonged effects of radiation on the thyroid. J Clin Endocrinol Metab 1993;77:362-369. 16. Shore RE, Hildreth N, Dvoretsky P, et al. Benign thyroid adenomas among persons X-irradiated in infancy for enlarged thymus glands. Radiat Res 1993;134:217-223. 17. Wong FL, Ron E, Gierlowski T, et al. Benign thyroid tumors: general risk factors and their effects on radiation risk estimation. Am J Epidemiol 1996;144:728-733. 18. Ron E, Modan B, Preston D, et al. Thyroid neoplasms following lowdose radiation in childhood. Radiat Res 1989;120:516-531.
Radiation Oncology biology
physics
www.redjournal.org
Clinical Investigation: Pediatrics
Thyroid Adenomas After Solid Cancer in Childhood Nadia Haddy, PhD,*,y,z Chiraz El-Fayech, MD,*,y,z Catherine Guibout, PhD,*,y,z Elisabeth Adjadj, PhD,*,y,z Ce´cile Thomas-Teinturier, MD,*,x Odile Oberlin, MD,*,y Cristina Veres, MSc,*,y,z He´le`ne Pacquement, MD,jj Angela Jackson, MSc,*,y,z Martine Munzer, MD,{ Tan Dat N’Guyen, MD,{ Pierre-Yves Bondiau, MD,** Delphine Berchery, MD,yy Anne Laprie, MD,yy Andre´ Bridier, PhD,y Dimitri Lefkopoulos, PhD,y Martin Schlumberger, MD,y,z Carole Rubino, MD, PhD,*,y,z Ibrahima Diallo, PhD,*,y,z and Florent de Vathaire, PhD*,y,z *Radiation Epidemiology Group, INSERM, Villejuif, France; yInstitut Gustave Roussy, Villejuif, France; zUniv. Paris-Sud, Villejuif, France; xHoˆpital Biceˆtre, Biceˆtre, France; jjInstitut Curie, Paris, France; {Institut Jean Godinot, Reims, France; **Centre Antoine Lacassagne, Nice, France; and yyCentre Claudius Re´gaud, Toulouse, France Received Oct 13, 2011, and in revised form Feb 24, 2012. Accepted for publication Mar 19, 2012
Summary Based on a cohort of 3254 2-year survivors of childhood cancer, the risk of thyroid adenoma increases with the radiation dose received by the thyroid during childhood cancer treatment, and reaches a plateau for doses higher than approximately 10 Gy. Chemotherapy administration was found to modify the relationship between the radiation dose to the thyroid
Purpose: Very few childhood cancer survivor studies have been devoted to thyroid adenomas. We assessed the role of chemotherapy and the radiation dose to the thyroid in the risk of thyroid adenoma after childhood cancer. Methods and Materials: A cohort of 3254 2-year survivors of a solid childhood cancer treated in 5 French centers before 1986 was established. The dose received by the isthmus and the 2 lobes of the thyroid gland during each course of radiation therapy was estimated after reconstruction of the actual radiation therapy conditions in which each child was treated as well as the dose received at other anatomical sites of interest. Results: After a median follow-up of 25 years, 71 patients had developed a thyroid adenoma. The risk strongly increased with the radiation dose to the thyroid up to a few Gray, plateaued, and declined for high doses. Chemotherapy slightly increased the risk when administered alone but also lowered the slope of the dose-response curve for the radiation dose to the thyroid. Overall, for doses up to a few Gray, the excess relative risk of thyroid adenoma per Gray was 2.8 (90% CI: 1.2-6.9), but it was 5.5 (90% CI: 1.9-25.9) in patients who had not received chemotherapy or who had received only 1 drug, and 1.1 (90% CI: 0.4-3.4) in the children who had received more than 1 drug (PZ.06, for the difference). The excess relative risk per Gray was
Reprint requests to: Florent de Vathaire, PhD, Radiation Epidemiology Group, Unit 1018, INSERM, Institut Gustave Roussy, Rue Camille Desmoulins, 94805 Villejuif, France. Tel: þ33-1-42-11-54-57; Fax: þ33-1-4211-53-15; E-mail: [email protected] Supported by the Ligue Nationale Contre le Cancer, the Institut de Recherche en Sante´ Publique, the Programme Hospitalier de Recherche Clinique, the Agence Franc¸aise de Se´curite´ Sanitaire et Produit de Sante´, Electricite´ de France, and the Fondation Wyeth for childhood and adolescent health. Conflict of interest: none. Int J Radiation Oncol Biol Phys, Vol. 84, No. 2, pp. e209ee215, 2012 0360-3016/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.ijrobp.2012.03.044
Supplementary material for this article can be found at www.redjournal.org. AcknowledgmentsdThe authors are grateful to M. Labbe´ and C. Paoletti for their help in data management; to all the physicians and physicists who participated in the elaboration of the study at the Gustave Roussy Institute (Villejuif), Institut Godinot (Reims), Institut Curie (Paris), Centre Regaud (Toulouse), and Centre Lacassagne (Nice); and to L. Saint Ange for editing.
International Journal of Radiation Oncology Biology Physics
e210 Haddy et al. gland and the risk of thyroid adenoma by decreasing its initial slope.
also higher for younger children at the time of radiation therapy than for their older counterparts and was higher before attaining 40 years of age than subsequently. Conclusions: The overall pattern of thyroid adenoma after radiation therapy for a childhood cancer appears to be similar to that observed for thyroid carcinoma. Ó 2012 Elsevier Inc.
Introduction
Radiation dosimetry
Thyroid tissue is one of the most radiosensitive human tissues (1). The relationship between the radiation dose received by the thyroid gland during radiation therapy for a childhood cancer and the subsequent emergence of a differentiated thyroid carcinoma has been investigated in several studies (2-6). Thyroid adenomas after radiation therapy require investigation because of the medical and surgical impact of this benign pathology. Additionally, in animal models, thyroid adenomas could be precancerous lesions. Thyroid cancers could therefore develop from benign thyroid tumors and not necessarily be de novo. To date, few epidemiological studies have been published on thyroid adenomas after radiation therapy for a childhood cancer. We report the incidence and risk factors for thyroid adenomas in a cohort of 3254 childhood cancer survivors treated before 1986 and followed for 25 years on average.
External beam radiation therapy always involves unwanted irradiation of healthy organs remote from the target volume. This is mainly because of secondary radiations resulting from beam scattering from 1) the tissue volume within the geometric limits of the primary beam, 2) the beam limiting devices, and 3) any object within the primary beam. Radiation leakage through the protective shielding also contributes to the doses delivered to organs remote from the target volume. The radiation dose was estimated in the middle of the 2 lobes and in thyroid isthmus for patients who had received external radiation therapy. Our in-house software “Dos_EG,” which takes into account the previously mentioned dose components used for this calculation (8). The doses at 185 other anatomical sites including the thymus, spleen, gonads, and 91 skeletal sites were also estimated. To establish the effects of fractionation, we considered all treatments given on the same day as 1 fraction. This definition was acceptable because no hyperfractionated treatment had been delivered to the patients in the cohort. The radiation dose to the thyroid was defined as the average of the radiation dose received in the center of the isthmus and of the 2 lobes. The mean dose of radiation to the thyroid in the 2288 patients who had received external radiation therapy was 5.8 Gy, but the median value was only 0.8 Gy (range <0.001-72).
Methods and Materials Patients A retrospective cohort of 4568 children treated in 8 centers in France and the United Kingdom was constituted comprising patients who were alive 2 (French centers) or 3 (UK centers) years after a first solid cancer diagnosed before age 16 (before 15 years of age in the UK centers) and before 1986 (5). An updated analysis of the incidence of differentiated thyroid carcinoma in this cohort was recently published (6). The present analysis is restricted to the 3382 patients treated in France. Among these patients, 26 treated for a thyroid cancer and 86 treated with brachytherapy were excluded. The 3254 remaining patients were included, of whom 2288 had been treated with external beam radiation therapy. Information on treatments was abstracted from the clinical chart and radiation therapy files of the participating centers. The follow-up of the 3254 patients was initially assessed using medical records from treatment centers, and later via a selfcompleted questionnaire sent on Sept. 1, 2005. This comprehensive questionnaire provided information on the socioeconomic status, quality of life, and health outcomes and it was partially based on that of the British Childhood Cancer Survivor Study (7). A total of 2449 patients were still alive and were therefore considered eligible to receive the questionnaire. It was mailed to the 2095 patients for whom the most recent address was obtained from the National Health Insurance System and who had returned a signed consent agreement. This agreement included an authorization to contact their medical practitioner and medical facilities. A total of 1823 (74% of 2449) patients returned the completed questionnaire by Dec. 31, 2010. The diagnoses of the first cancer and thyroid adenoma were histologically confirmed.
Chemotherapy quantification Drugs were grouped into categories according to their known mechanisms of action in the cell: vinca alkaloids, antimetabolites, alkylating agents, anthracyclines, cytotoxic antibiotics, epipodophyllotoxins, and other drugs. Chemotherapy information could not be found for 18 patients. To quantify the total amount of drug administered in each class, the dose of each cytotoxic agent was converted into moles per square meter.
Statistical methods French patients were followed from 2 years of survival after diagnosis of the first cancer until the occurrence of thyroid adenoma, thyroid surgery for any other reason, response to the self-questionnaire, or the date of last medical contact. The risk factors for thyroid adenoma were identified using Cox’s proportional hazard model and Poisson regression (9). Multivariate analysis was performed to evaluate the radiation doses effect on thyroid adenomas. This analysis was adjusted on age at first cancer in years, attained age, gender, chemotherapy (none or 1 drug vs 2 drugs or more), and first cancer type (Hodgkin disease, brain tumor, vs others). The 95% confidence intervals were estimated for parameters using the methods of maximum of likelihood (10). Because exposure to radiation is not supposed to reduce the risk of thyroid
Volume 84 Number 2 2012
Thyroid adenomas after radiation therapy e211
adenoma, 90% confidence intervals were used for all risks or incidence figures presenting results on effects of radiation dose. Standard radiobiological models were used to model the variation in the relative risk (RR) and the absolute excess risk (AER) according to the radiation doses.
Results The median duration of follow-up was 25 years after the first cancer, and 904 people were followed for 30 years or more, among whom 720 had received radiation therapy. From 4 to 41 years after the first cancer, 71 patients had developed a thyroid adenoma and 67 of them had received radiation therapy. Of the 71 adenomas, 46 had arisen in patients who had returned the self-questionnaire. The histological size (main diameter of the largest nodule) of the adenomas ranged from 8 to140 mm (median Z 20 mm) and the clinical size ranged from 2 to 50 mm (median Z 19 mm). Five were classified as polymorphous adenoma, 3 as oxyphilic adenoma, and the others as follicular adenoma. The cumulative incidence of thyroid adenoma 40 years after the diagnosis of the primary cancer was 4.3% (95% CI: 3.3-5.6); this rate was 5.5% (95% CI: 4.2-7.2) after radiation therapy. The cumulative incidence at 40 years of age was 3.5% (95% CI: 2.8-4.6) (Supplementary Table), and 4.6% (95% CI: 3.4-6.3) at 50 years of age, when attained age was taken into account. Forty-five adenomas had developed among the 1027 women who had received radiation therapy, and 22 among the 1261 men. At the age of 40, the cumulative incidence of thyroid adenoma was 5.5 (95%: 4.1-7.4) in women and 1.9 (95% CI: 1.2-3.1) in men (Table 1) and it was, respectively, 6.6 (95% CI: 4.8-9.0) and 2.6 (95% CI: 1.7-4.2) among those who had received radiation Table 1 Incidence of thyroid adenoma according to demographical characteristics of 3254 2-y childhood cancer survivors Characteristics Gender Men Women Age at first cancer 0-1 2-4 5-9 10þ Year of diagnosis 1940-1969 1970-1974 1975-1979 1980-1985 Treatment No RT, no CT RT, no CT No RT, CT CT and RT RT and 1 drug RT and 2 drugs or more
therapy. After adjustment on age at diagnosis, type of chemotherapy administered for the first cancer and the radiation dose to the thyroid, the incidence of thyroid adenomas was 2.9-fold higher (95% CI: 1.8-4.8) in women than in men. The incidence of thyroid adenoma varied considerably according to the type of first cancer. The cumulative incidence at age 40 ranged from 0 in bone sarcoma and retinoblastoma survivors to 10.7% (95% CI: 6.6-17.0) in those who had been treated for Hodgkin disease. However, when adjusted for treatment variables and using nephroblastoma survivors as the reference group, the differences in the RR of its onset were substantially reduced and were only of borderline significance (Supplementary Table). When both the time since the childhood cancer and attained age were taken into account, no clear time trend was observed in the RR of developing thyroid adenoma (Table 2). Compared with that observed in patients who had not received radiation therapy, the RR of thyroid adenoma was 18.4 (90% CI: 7.4-45.4) in patients who had received from 5 to 19 Gy to the thyroid and 11.5 (90% CI: 2.7-41.9) in those who had received 40 Gy or more. These findings remained similar when the analysis was restricted to patients who had returned their questionnaire (Table 3). Figure 1 shows the cumulative incidence according to the thyroid radiation dose category. A model with a linear coefficient of the dose plus an exponential term for high radiation doses to the thyroid fitted the data more adequately than a purely linear model (P<.0001). Adding a quadratic term did not improve the fit of the model (PZ.9). The excess relative risk per Gy (ERR/Gy) of developing thyroid adenoma was found to be 2.8 (90% CI: 1.2-6.9) for the entire cohort (Fig. 2) and 3.3 (90% CI: 1.3-9.4) when restricted to patients who had returned the self-questionnaire. The linear parameter of the dose response was significantly higher (PZ.02) when radiation therapy had been delivered before age 5 (ERR/Gy Z 3.4, 90% CI: 1.5-8.3) than when it had been delivered at an older age (ERR/Gy Z 1.0, 90% CI: 0.3-3.1). It was also significantly higher (PZ.04) before attaining age 40 (ERR/Gy Z 3.0,
Thyroid Cumulative incidence adenomas/patients at 40 y old (95% CI) 22/1831 49/1423
1.9% (1.2-3.1) 5.5% (4.1-7.4)
17/905 23/757 16/797 15/795
3.5% 6.0% 3.2% 2.2%
(2.1-5.7) (3.8-9.5) (1.9-5.5) (1.2-4.0)
35/628 14/626 12/842 10/1158
6.3% 2.8% 2.7% 2.3%
(4.5-8.8) (1.5-5.2) (1.5-4.7) (1.1-4.9)
0/208 21/647 4/758 46/1641 18/256 28/1321
4.6% 1.2% 4.4% 10.2% 3.2%
0 (2.9-7.2) (0.4-3.3) (3.2-6.0) (6.4-16.0) (2.1-4.7)
Abbreviations: CI Z confidence interval; CT Z computed tomography; RT Z radiation therapy.
Table 2 Incidence of thyroid adenoma, according to the time since first cancer treatment and to attained age in a cohort of 3254 patients treated for a cancer during childhood Differentiated thyroid adenomas Time scale
Patients still Annual incidence Adjusted relative followed up n 1000 risk (95% CI)*
Years since first cancer 2-9 3254 1 10-19 2696 36 20-29 2181 23 30-39 904 10 40 227 1 Attained age in years 2-19 3254 16 20-29 2524 31 30-39 1649 19 40-49 605 4 50 121 1
0.04 1.5 1.4 2.0 0.9
(0.01-0.2) (1.0-2.0) (0.9-2.1) (1.0-3.4) (0.05-3.9)
0.03 1 1 1.4 0.7
(0.002-0.1) (ref) (0.6-1.7) (0.6-2.7) (0.04-3.1)
0.5 1.4 1.7 1.3 1.8
(0.3-0.7) (1.1-2.0) (1.1-2.6) (0.4-3.0) (0.01-8.0)
0.3 1 1.2 1.0 1.5
(0.2-0.6) (ref) (0.7-2.1) (0.3-2.5) (0.08-7.0)
Abbreviation: CI Z confidence interval. * Relative risk, adjusted on gender, age at diagnosis, radiation dose to thyroid, chemotherapy (none or 1 drug vs 2 drugs or more), and first cancer type (Hodgkin disease, brain tumor, vs others).
International Journal of Radiation Oncology Biology Physics
e212 Haddy et al.
Table 3 Thyroid adenomas according to the radiation dose to the thyroid in a cohort of 3254 patients treated for a first cancer during childhood Dose to the thyroid (Gy) No RT
>0-0.9
Thyroid dose: mean, Gy 0 0.3 Whole cohort (nZ3254) Adenomas/patients 4/966 14/1235 RR (90% CI)* 1y 2.1 (0.8-5.4) Patients who returned the self-questionnaire (nZ1827) Adenomas/patients 3/622 9/703 RR (90% CI)* 1y 2.1 (0.7-6.4)
1-4.9
5-19.9
20-39.9
40
2.3
11.8
27.3
50.0
12/458 7.8 (3.0-20.3)
25/351 18.4 (7.4-45.4)
14/202 18.9 (6.9-51.5)
2/42 11.5 (2.7-41.9)
9/207 9.5 (3.1-28.6)
17/182 20.8 (7.3-59.5)
7/90 17.5 (5.2-59.2)
1/19 8.6 (1.3-59.4)
Abbreviations: CI Z confidence interval; RR Z relative risk. * Relative risk, adjusted on age at first cancer in years, attained age, gender, chemotherapy (none or 1 drug vs 2 drugs or more), and first cancer type (Hodgkin disease, brain tumor, vs others). y Reference category.
90% CI: 1.3-7.3) than after (ERR/Gy Z 0.8, 90% CI: 0.1-3.2). A comparable dose-response modification was not observed for follow-up, whatever the threshold tested. Because of a lack of convergence, it was not possible to investigate the effect of gender as a dose-response modifier. Among the 2288 patients who received radiation therapy, 1079 were treated in 20 or less fractions, of whom 35 later developed a thyroid adenoma and 1209 were treated in more than 20 fractions, of whom 32 developed a thyroid adenoma. We were not able to evidence a role of the fractionation in the reduction of risk (PZ.4): the ERR/Gy was 3.0 (90%:1.3-7.7) for the radiation dose delivered in 20 fractions or less and 2.5 (90% CI: 1.0-6.5) for radiation dose delivered in more than 20 fractions. Nevertheless it has to be noted that the slope was nonsignificantly reduced (ERR/ Gy Z 1.5, 90% CI: 0.4-4.9) if thyroid radiation dose received in more than 40 fractions (5 thyroid adenomas in 216 patients).
Of the 491 patients who had received more than 20 Gy to the spleen or had undergone a splenectomy, 22 had developed a thyroid adenoma. When adjusted for all other factors, including the radiation dose to the thyroid, neither a splenectomy nor radiation therapy to the spleen was found to be an individual risk factor for thyroid adenoma (PZ.3), nor a (thyroid) radiation doseresponse modifier (P>.5). Similar results were observed for the age-weighted radiation dose to the active bone marrow, the integral radiation dose, the radiation dose to the breasts and to the ovaries. Neither a hemilateral ovariectomy nor radiation delivered to the ovary (>20 Gy), nor a total ovariectomy, modified the risk of thyroid adenoma among women (P>.5) or the dose-response for the radiation dose to the thyroid (P>.5). Overall, chemotherapy (yes/no) did not increase the risk of thyroid adenoma (RR Z 0.7, 95% CI: 0.4-1.2), but patients who had received more than 1 drug had a borderline significantly
Fig. 1. Cumulative incidence of thyroid adenoma by attained age, in 4 categories of average radiation dose to the thyroid. Vertical bars represent the 90% CI.
Volume 84 Number 2 2012
Thyroid adenomas after radiation therapy e213
Fig. 2. Thyroid adenoma RR according to radiation dose to the thyroid. The circle are of observed RR for 6 dose categories and the curve is the linear-exponential dose-response model for RR calculated as RR Z 1 þ 2.8 dose exp (0.052 dose). Vertical lines Z 90% CIs for RR. Relative risk, adjusted on age at first cancer in years, attained age, gender, chemotherapy (none or 1 drug vs 2 drugs or more), and first cancer type (Hodgkin disease, brain tumor, vs others). RR Z relative risk. (PZ.1) decreased risk of thyroid adenoma (RR Z 0.7, 95% CI: 0.4-1.2), when adjusted for the radiation dose to thyroid and the type of childhood cancer (Table 4). Having received more than 1 chemotherapy drug also modified (almost significantly, PZ.06) the relationship between the radiation dose to the thyroid and the risk of thyroid adenoma, by decreasing the linear term; the ERR/Gy was 5.5 (90% CI: 1.9-25.9) among the 1237 patients who had received 0 or 1 drug, compared with an ERR/Gy Z 1.1 (90% CI: 0.4-3.4) among the 2017 patients who had received more than 1 drug. Topoisomerase II inhibitors were the only drugs that exerted an independent effect or by acting as a radiation dose-response modifier (Table 4).
When each drug was taken into account individually in the multivariate analysis, melphalan was associated with a reduced risk of thyroid adenoma, whereas rifomycin and vincristine were associated with an increased risk. However, given the potential role of chemotherapy, acting both as a main risk factor and as a dose-response modifier and the huge number of potential drug combinations in our cohort, it was not possible to correctly investigate the role of each drug and this result has to be considered with caution. The size of the thyroid adenoma did not vary significantly according to age at diagnosis (PZ.2), chemotherapy (PZ.7), follow-up (PZ.2), or the radiation dose to the thyroid (PZ.7), but
Table 4 Thyroid adenoma incidence, according to chemotherapy administration in a cohort of 3254 patients treated for a cancer during childhood Model without interaction Chemotherapy Chemotherapy
RR (95% CI)*
Any chemotherapy No 1 Yes 0.7 (0.4-1.2) Nb drugs 0 or 1 1 2 drugs or þ 0.7 (0.4-1.2) Topoisomerase II inhibitor No 1 Yes 0.6 (0.4-1.0)
Model with interaction
Thyroid radiation dose
Chemotherapy
Thyroid radiation dose
ERR/Gy (90% CI)
RR (95% CI)*
ERR/Gy (90% CI)
2.7 (1.2-6.8)
1 1.3 (0.3-21.6)
5.0 (2.7-12.7) 2.4 (1.0-6.2)
.5
2.8 (1.2-7.0)
1 2.6 (0.6-17.4)
5.5 (1.9-25.9) 1.1 (0.4-3.4)
.06
2.9 (1.3-7.3)
1 4.1 (0.7-33.6)
13.7 (1.8-38.5) 1.2 (0.5-3.4)
.02
P value for interaction
Abbreviations: CI Z confidence interval; ERR Z excess RR; Nb Z Number; RR Z relative risk. * Relative risk, adjusted on gender, age at diagnosis, attained age, radiation dose to thyroid, and first cancer type (Hodgkin disease, brain tumor, vs others).
e214 Haddy et al.
International Journal of Radiation Oncology Biology Physics
was on average larger in men, 36 mm (SEM Z 4.1 mm), than in women, 21 mm (SEM Z 2.1 mm). This difference remained significant (PZ.001) after adjustment on other factors.
That chemotherapy could independently increase the risk of thyroid adenoma, but could also reduce the effect of radiation doses (a dose-response modifier effect) is probably one of most important results of our study. A similar result was evidenced in the last publication by the CCSS, in which the ERR/Gy was 5.0 in patients who had not received radiation therapy and 1.1 in patients who had (P value for the difference Z .07) (4), but not in the other publications on thyroid cancer after childhood cancer (2, 3, 6). If confirmed, this result should not be interpreted as a protective effect of the association of these 2 treatment modalities. It is probably rather due to the fact that the percentage of patients who develop a thyroid tumor after chemotherapy is the same as the proportion who develop a thyroid tumor after thyroid radiation exposure and that the addition of the carcinogens does not increase the risk, which would be the case if the 2 carcinogens were considered independently. The fact that, in our cohort, the effect of radiation therapy seems to be focused on patients who had received topoisomerase II inhibitors should be cautiously interpreted and needs to be confirmed. Unlike that evidenced in the same cohort for thyroid carcinoma (6), but consistent with that observed in the CCSS (2-4) for thyroid carcinoma, we failed to find evidence either of an independent effect or a dose-response modifier effect of a splenectomy or radiation delivered to the spleen (>20 Gy) or of high radiation doses to the pituitary gland on the risk of thyroid adenoma. Similar to what we observed for thyroid carcinoma (6), the incidence of thyroid adenoma seemed to be significantly lower among survivors of central nervous system tumors, and higher among survivors of Hodgkin disease, than in the other childhood cancer survivors. Hodgkin disease is the type of childhood cancer exposed to the highest risk of thyroid cancer in all the childhood cancer survivor studies (2-4, 6, 11-14), but whether a specific susceptibility to thyroid tumors exists in these patients is a matter of debate. Indeed, until now, no other study has published the RR of developing a thyroid tumor per type of childhood cancer adjusted on the radiation dose to the thyroid and other risk factors. We failed to find evidence of a role played by radiation dose fractionation. In fact, childhood cancer survivors constitute a population that is probably far too homogenous to allow the investigation of this issue. Our finding of a similarity between the risk of thyroid adenomas and thyroid carcinomas after radiation therapy for a childhood cancer is in agreement with that found in other populationsdin particular, in patients who had received radiation therapy for benign reasons (15-18). In conclusion, our study shows that the risk of developing a thyroid adenoma after radiation therapy for childhood cancer has a pattern that is comparable to that observed for thyroid carcinoma: the risk increased with a few Gray, then it decreased, and younger children were more sensitive to this risk than older children. Chemotherapy was found to play a role both independently and as a dose-response modifier, but this will have to be confirmed.
Discussion Based on a cohort of 3254 2-year childhood cancer survivors treated before 1986 and followed for 25 years on average, we showed that the risk of thyroid adenoma increased with the radiation dose received by the thyroid during childhood cancer treatment, and plateaued at doses exceeding approximately 10 Gy. The risk of thyroid adenoma per unit of radiation dose to the thyroid was higher if radiation therapy had been delivered before age 5, and before the attained age of 40 years. Chemotherapy administration was found to slightly increase the risk of thyroid adenoma in subjects who had not received external radiation therapy and to modify the relationship between the radiation dose to the thyroid gland and the risk of thyroid adenoma by decreasing its initial slope. In our cohort, a surveillance bias probably exists because physicians know that long-term childhood cancer survivors treated with radiation therapy are at high risk of developing a thyroid adenoma. This surveillance bias is likely to increase the observed incidence of thyroid adenoma among these survivors. We think that this bias does not substantially affect the dose-response relationship because we did not observe any correlation between the level of radiation doses to the thyroid gland and the size of the thyroid adenomas. Another issue is that the follow-up of our cohort is partially retrospective and was done using a questionnaire for 1823 patients and hospital medical records for 1431 patients (deceased patients included). Even if only thyroid adenomas validated by a copy of the pathological record were included in this analysis, this difference could introduce a degree of bias. A model that includes a linear term plus a negative exponential term for cell killing is the best fit for our data on thyroid adenomas. This type of dose-response was also observed for thyroid carcinomas both in this cohort (6) and in the Childhood Cancer Survivor Study (CCSS) (2, 3). In the last publication of the CCSS, more complex models, including a quadratic term rather than a linear one and an exponential quadratic term rather than an exponential term fitted the data better than a linear exponential model. This did not modify the overall shape of the dose-response for thyroid carcinomas, which increased up to a few Gray, then plateaued, and decreased for higher doses (4). The initial slope of the relationship between the radiation dose to the thyroid and the risk of thyroid adenoma estimated at ERR/ Gy Z 2.8 (90% CI: 1.2-6.9) is consistent with the risk estimated in the same cohort for thyroid carcinomas, ERR/Gy Z 2.0, 95% CI: 0.6-6.4 (6), and in the CCSS for thyroid cancer. In our study, a younger age at radiation therapy increased the risk of radiation-induced thyroid adenoma. This is similar to that observed for differentiated thyroid carcinoma in all childhood cancer survivor studies (2-4, 6, 11). We also observed a significant decrease in the ERR/Gy with increasing attained age. Because attained age and follow-up since radiation therapy are closely linked, that we did not observe a dose-response modifier effect of follow-up is probably due to a lack of power. Our results are in agreement with what we observed for carcinoma in the publication on this cohort (6), but not with that observed in the CCSS (2-4).
References 1. United Nations. Ionising radiation: Sources and biological effects. Scientific Committee on the Effects of Atomic Radiation 1994. Report to the General Assembly, with annexes. New York: United Nations Sales Publications; 1994. 2. Sigurdson AJ, Ronckers CM, Mertens AC, et al. Primary thyroid cancer after a first tumour in childhood (the Childhood Cancer
Volume 84 Number 2 2012
3.
4.
5.
6.
7.
8.
9.
Survivor Study): a nested case-control study. Lancet 2005;365: 2014-2023. Ronckers CM, Sigurdson AJ, Stovall M, et al. Thyroid cancer in childhood cancer survivors: a detailed evaluation of radiation dose response and its modifiers. Radiat Res 2006;166:618-628. Bhatti P, Veiga LH, Ronckers CM, et al. Risk of second primary thyroid cancer after radiotherapy for a childhood cancer in a large cohort study: an update from the childhood cancer survivor study. Radiat Res 2010;174:741-752. De Vathaire F, Hardiman C, Shamsaldin A, et al. Thyroid carcinoma following irradiation for a first cancer during childhood. Arch Intern Med 1999;159:2713-2719. de Vathaire F, Haddy N, Hawkins M, et al. Thyroid carcinoma following childhood cancer: role of radiation therapy, splenectomy and chemotherapy. J Clin Endocrinol Metab. In press. Hawkins MM, Lancashire ER, Winter DL, et al. The British Childhood Cancer Survivor Study: objectives, methods, population structure, response rates and initial descriptive information. Pediatr Blood Cancer 2008;50:1018-1025. Diallo I, Lamon A, Shamsaldin A, et al. Estimation of the radiation dose delivered at any point in the body for individual patient in external beam radiotherapy. Radiother Oncol 1996;38: 269-297. Breslow NE, Day NE, and Statistical methods in cancer research. The design and analysis of cohort studies. IARC Scientific Publication No. 82. Lyon, France: International Agency for Research on Cancer; 1987.
Thyroid adenomas after radiation therapy e215 10. Moolgavkar SH, Venzon DJ. A method for computing profile likelihood based confidence bounds. Ann Stat 1987;15:346-359. 11. Taylor AJ, Croft AP, Palace AM, et al. Risk of thyroid cancer in survivors of childhood cancer: results from the British Childhood Cancer Survivor Study. Int J Cancer 2009;125:2400-2405. 12. O’Brien MM, Donaldson SS, Balise RR, et al. Second malignant neoplasms in survivors of pediatric Hodgkin’s lymphoma treated with low-dose radiation and chemotherapy. J Clin Oncol 2010;28:1232-1239. 13. Olsen JH, Mo¨ller T, Anderson H, et al. cancer incidence in 47,697 patients treated for childhood cancer in the Nordic countries. J Natl Cancer Inst 2009;101:806-813. 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 2000;18:1492-1499. 15. Schneider AB, Ron E, Lubin J, et al. Dose response relationships for radiation-induced thyroid cancer and thyroid nodule: evidence for the prolonged effects of radiation on the thyroid. J Clin Endocrinol Metab 1993;77:362-369. 16. Shore RE, Hildreth N, Dvoretsky P, et al. Benign thyroid adenomas among persons X-irradiated in infancy for enlarged thymus glands. Radiat Res 1993;134:217-223. 17. Wong FL, Ron E, Gierlowski T, et al. Benign thyroid tumors: general risk factors and their effects on radiation risk estimation. Am J Epidemiol 1996;144:728-733. 18. Ron E, Modan B, Preston D, et al. Thyroid neoplasms following lowdose radiation in childhood. Radiat Res 1989;120:516-531.