Originally published as JCO Early Release 10.1200/JCO.2006.09.0936 on March 19 2007

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hodgson, D. C. Articles by Travis, L. B. Search for Related Content PubMed PubMed Citation Articles by Hodgson, D. C. Articles by Travis, L. B. Social Bookmarking                            What's this?

Long-Term Solid Cancer Risk Among 5-Year Survivors of Hodgkin's Lymphoma

David C. Hodgson, Ethel S. Gilbert, Graça M. Dores, Sara J. Schonfeld, Charles F. Lynch, Hans Storm, Per Hall, Froydis Langmark, Eero Pukkala, Michael Andersson, Magnus Kaijser, Heikki Joensuu, Sophie D. Fosså, Lois B. Travis

From the Princess Margaret Hospital, University Health Network, and the Department of Radiation Oncology, University of Toronto, Canada; Division of Cancer Epidemiology and Genetics and the Division of Cancer Prevention, National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Bethesda, MD; The University of Iowa, Iowa City, IA; Danish Cancer Society, Copenhagen, Denmark; Karolinska Institute, Stockholm, Sweden; The Norwegian Cancer Registry; Norwegian Radium Hospital, Oslo, Norway; Finnish Cancer Registry, Institute for Statistical and Epidemiological Cancer Research; and the Helsinki University Central Hospital, Helsinki, Finland

Address reprint requests to David Hodgson MD, MPH, FRCPC, Department of Radiation Oncology, Princess Margaret Hospital, 610 University Ave, Toronto, Ontario, Canada M5G 2M9; e-mail: David.Hodgson{at}rmp.uhn.on.ca

	ABSTRACT

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Purpose Hodgkin's lymphoma (HL) survivors are known to be at substantially increased risk of solid cancers (SC). However, no investigation has used multivariate modeling to estimate the relative risk (RR), excess absolute risk (EAR), and cumulative incidence for specific attained ages and ages at HL diagnosis.

Patients and Methods We identified 18,862 5-year HL survivors from 13 population-based cancer registries in North America and Europe. Poisson regression was used to evaluate the effects of age at diagnosis, attained age, latency, sex, treatment, and year of diagnosis on the RR and EAR of SC.

Results Among 1,490 identified SC, 850 were estimated to be in excess. For most cancer sites, both RR and EAR decreased with age at HL diagnosis and showed strong dependencies on attained age. For a patient diagnosed at age 30 years and survived to 40 years, modeled risks were significantly elevated for cancers of the breast (RR = 6.1), other supradiaphragmatic sites (RR = 6.0), and infradiaphragmatic sites (RR = 3.7); the largest RR (20-fold) was observed for malignant mesothelioma. Thirty-year cumulative risks of SC for men and women diagnosed at 30 years were 18% and 26%, respectively, compared with 7% and 9%, respectively, in the general population. For young HL patients, risks of breast and colorectal cancers were elevated 10 to 25 years before the age when routine screening would be recommended in the general population.

Conclusion Multivariable modeling demonstrates for the first time temporal changes in SC risk not evident in unadjusted analyses, and can facilitate the development of individualized risk assessment and the creation of screening strategies for early detection.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Second primary cancers are a major cause of morbidity and mortality among long-term survivors of Hodgkin's lymphoma (HL).^1,2 There are, however, few estimates of the cumulative incidence of second cancers 25 years after HL diagnosis, and most estimates do not account for competing causes of death.^1-6 Further, risk estimates often group together survivors of considerably different ages and do not reveal the significant age-related differences in the excess incidence of solid cancers (SC). Evaluation of long-term site-specific risks and changes in risk beyond 20 years of follow-up are sparse, due to the constraints of sample sizes in most series.^4,6,7

A better understanding of the age- and sex-specific absolute risks of second cancers would facilitate the development of individualized risk assessment, meaningful communication of risks to HL survivors, and the development of screening strategies to facilitate early detection. No study has used multivariable methods to model changes in the absolute excess risk (EAR) of SC with prolonged follow-up, taking into account the effect of age at HL diagnosis, attained age, time since HL diagnosis, calendar year of HL diagnosis, sex, and initial treatment. Further, age-specific estimates of the cumulative incidence of second cancer among adult HL survivors, adjusted for competing risks and contrasted with the expected incidence in the general population, do not exist.

In this study, we applied for the first time multivariable modeling to a cohort of 18,862 5-year survivors of HL to evaluate the relative risk (RR) and EAR of specific SC change over time for HL survivors diagnosed at specified ages, and to estimate the age-specific cumulative incidence of SC in survivors with prolonged follow-up.

	PATIENTS AND METHODS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Patients diagnosed with a first primary HL (years of diagnosis, 1970-2001) and who survived at least 1 year after diagnosis (N = 28,421) were identified from nationwide population-based cancer registries in Norway, Sweden, Denmark, Finland, and nine areas of the United States using the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) Program. The SEER Program registries cover approximately 13% of the US population, and include the states of Connecticut (1973+), Hawaii (1973+), Iowa (1973+), New Mexico (1973+), and Utah (1973+), and cover the metropolitan areas of San Francisco-Oakland (1973+), Detroit (1973+), Seattle-Puget Sound (1974+), and Atlanta (1975+).^8,9 Most analyses were restricted to 5-year HL survivors, including a subgroup of patients described in an earlier report,⁵ with extended follow-up. The previous report did not included multivariate modeling.⁵

Information on patient vital status, date of death, cause of death, and the occurrence of second invasive cancers was obtained from each cancer registry. Data regarding initial course of therapy (chemotherapy, radiation therapy [RT], and combined-modality therapy [CMT]) were provided by all registries except Sweden; data were not available, however, for specific cytotoxic drugs, RT fields, or the use of salvage therapy.

The follow-up period began 1 year after the date of HL diagnosis and ended on the date of death, date of diagnosis of second cancer, or on the study end date (December 31, 2002), whichever occurred first. Person-years (PY) and second cancers were categorized by sex, calendar year of HL diagnosis (1970 to 1984 or 1985 to 1996), initial treatment modality, registry, and 5-year intervals of attained age, attained calendar year, time since HL diagnosis, and age at HL diagnosis. Cancer incidence rates specific for each registry, sex, and 5-year age and calendar year intervals were multiplied by the corresponding PY at risk to estimate the number of cancer cases expected in each stratum.

In general, O and E are used to denote observed and expected numbers of SC. O_i, E_i, and PY_i denote observed cases, expected cases, and PY, respectively, in a specified cell of the PY categorization defined in the previous paragraph. Analyses focused on the RR, defined as the ratio of risks in HL patients and the general population, and EAR, defined as the difference in risks in HL patients and the general population and expressed as the number of excess cases per 10,000 PY. Poisson regression methods¹⁰ were used to estimate the RR and EAR as functions of sex, age at HL diagnosis, attained age, and other variables. The statistical expectation of O_i was assumed to be E_i RR (x_i) or E_i + PY_i EAR (x_i), where x is a vector of variables on which the RR or EAR depends. Methods are similar to those used in a study of testicular cancer survivors.¹¹

For modeling, broad categories of SC were defined based on anatomic site, estimated radiation dose within traditional supradiaphragmatic (mantle) and infradiaphragmatic RT fields (Appendix 1, online only), and whether differences existed in the nature of the dependencies of risk on age at HL diagnosis and attained age. The selected categories were female breast cancer, thyroid cancer, other supradiaphragmatic sites (head and neck, esophageal, and respiratory cancers), and infradiaphragmatic sites (nonesophageal gastrointestinal cancers and urinary cancers). The designation "supradiaphragmatic sites" excluded cancers of the thyroid and female breast. Most analyses were restricted to 5-year survivors, with RRs and EARs pertaining to the period 10 years after HL diagnosis. Two-sided P values are used throughout.

Cumulative risks of a SC were calculated using a competing risk analysis similar to that used in a study of SC after testicular cancer diagnosis,¹¹ accounting for age at HL diagnosis, attained age, and competing risks from HL mortality, noncancer mortality, and intervening diagnosis of leukemia or non-Hodgkin's lymphoma. Further detail on statistical methods is provided in the Appendix (online only).

	RESULTS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
A total of 18,862 5-year survivors of HL were identified (Table 1). Median age at HL diagnosis was 30 years, and median duration of follow-up was 12.2 years (range, 5.0 to 32.9 years). SC were diagnosed in 1,490 patients compared with 625.4 expected (O/E = 2.38; 95% CI, 2.26 to 2.51).

View larger version (16K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 1. Relative risk (RR) and excess absolute risk (EAR) according to age at Hodgkin's lymphoma (HL) diagnosis and attained age. (A) RR of female breast cancer; (B) RR of supra- and infradiaphragmatic solid cancers (both sexes); (C) EAR of female breast cancer; (D) EAR of supra- and infradiaphragmatic solid cancers (average for men and women).

Both the RR and EAR for thyroid cancer (data not shown) showed exceptionally strong decreases with age at HL diagnosis (P < .001). However, in contrast with sites shown in Figure 1, there was no evidence that either measure depended on attained age (P > .5).

These results were based on analyses that simultaneously evaluated both age at HL diagnosis and attained age, which is important because these variables are highly correlated. For example, without adjustment for age at HL diagnosis, the EAR for female breast cancer decreased instead of increased with attained age, and without adjustment for attained age, the EAR for supra- and infradiaphragmatic sites increased instead of decreased with age at HL diagnosis.

Modeled RRs for patients diagnosed with HL at age 30 years and, for most sites, attained ages 40 to 60 years are listed in Table 3, Table 4, and Table 5. Risks were significantly elevated for all cancer sites evaluated except bladder and prostate. Relative risks greater than tenfold were found for cancers of pleura, soft tissue, and bone, and for cancers of unspecified primary site. The estimated total number of excess SC was 850; 417 (49%) of these occurred in women. The leading sites of excess cancer were the lungs (26% of excess cases), breasts (42% of excess cases in women), and colorectal sites (7% of excess cases).

We also examined the association between the RR of SC and several additional variables. The RR for female breast cancer decreased with increased time beyond 10 years from HL diagnosis (P < .001; Table 4), parallel to the decline with attained age as illustrated in Figure 1A. For both the supra- and infradiaphragmatic sites, RR was elevated for all time intervals, with little indication of a trend toward increasing or decreasing RR with latency beyond 10 years of follow-up (P > .5 and P = .12 for supra- and infradiaphragmatic sites, respectively). Again, without adjustment for age at HL diagnosis, results would have differed; the RR for supra- and infradiaphragmatic sites would have shown a significant increase with increasing latency beyond 10 years (P = .01), whereas the RR for breast cancer would have shown no evidence of a decline (P = .33).

Significantly increased RRs of cancers of female breast, supra-, and infradiaphragmatic sites occurred in patients treated initially with RT alone, chemotherapy alone, and CMT (Table 5). Initial treatment that included RT was associated with significantly increased RR of breast (P = .012) and supradiaphragmatic SC (P < .001) compared with treatment with chemotherapy alone. For infradiaphragmatic sites, there was no significant association between initial treatment and RR. When comparing HL patients diagnosed from 1970 to 1984 with those patients diagnosed from 1985 to 1996, there was no evidence of a significant change in SC risk in the later time period (Table 5); this finding was unchanged when the period analysis was limited to patients whose initial treatment included radiation therapy, although the relative risk of breast cancer was slightly higher in the later time period.

For men diagnosed with HL at ages 20, 30, and 40 years, the 30-year cumulative incidence of SC was 10.5%, 18.3%, and 27.1%, respectively (Fig 2), compared with 2.4%, 6.9%, and 17.4%, in the general population. For women diagnosed at the same ages, cumulative incidence of SC was 24.3%, 26.1%, and 26.7%, respectively, compared with 4.5%, 8.9% and 15.4%, in the general population. If trends in risks with attained age (Fig 1) continue, the cumulative incidence of second cancer by age 80 years is predicted to be 46%, 40%, and 35% for men diagnosed with HL at ages 20, 30, and 40 years, respectively; corresponding risks for women diagnosed would be 66%, 51%, and 35%.

View larger version (12K): [in this window] [in a new window] [PowerPoint Slide for Teaching]   Fig 2. (A) Cumulative incidence of all solid cancers among 10,619 male 5-year survivors of Hodgkin's lymphoma (HL) compared with men of the same age in the general population (GP). (B) Cumulative incidence for 8,243 female 5-year survivors compared with women of the same age in the GP.

	DISCUSSION

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

It is noteworthy that in this large cohort study, unadjusted analyses suggested substantially different temporal patterns of SC risk than adjusted analyses. This may explain why we found little evidence of increasing RR for SC of supra- and infradiaphragmatic sites, and a decrease in RR for female breast cancer with latency beyond 10 years, whereas prior unadjusted analyses found increasing RR of SC with latency 10 years.^3-5,7,12 Similarly, in contrast with unadjusted analyses that have shown that EARs of non-breast SC increase or remain stable with increasing age at HL diagnosis,^4-7,12 we found that after adjusting for attained age the EAR decreased with older age at diagnosis for non-breast SC as well as for female breast cancer.

Age-related variations in baseline and excess risks are reflected in the progressive divergence between the expected and observed cumulative incidence of SC among survivors of different age groups. The excess cumulative incidence of SC was strongly dependent on sex, age at HL diagnosis, and attained age, and was most pronounced for women diagnosed at young ages. Women diagnosed with HL at age 20 were estimated to have a 30-year cumulative incidence of SC that was 20% higher than expected, and they had the potential to experience a 50-year cumulative incidence of 66%. In contrast, young men had a lower incidence of SC, 8% higher than expected at 30 years.

Understanding temporal and age-related trends in SC risk could inform screening recommendations for HL survivors. In North America, colorectal cancer screening is recommended for average-risk adults starting at age 50 years.^13,14 Among young HL survivors, the absolute risk of colorectal cancer was comparable to an average 50- to 54-year-old by age 35 to 40 years, long before screening would generally be recommended. Our results suggest for the first time that among HL survivors, colorectal cancer screening should be considered before the age of routine screening in the general population, although a better understanding of the accuracy of existing screening tests in HL survivors is needed. Similarly, the absolute risk of breast cancer among young female survivors increased to a level comparable to an average 50-year-old woman within 5 to 10 years following HL diagnosis, a finding that supports recommendations to initiate breast cancer screening within this interval.¹⁵ The risk of lung cancer is also elevated among HL survivors, and a recent study suggests that CT screening may increase their quality-adjusted survival, particularly among those survivors who smoke.¹⁶ We did not compare lung cancer risk among HL survivors to that seen among other high risk groups, since we did not have data on patients’ smoking habits, and also because appropriate high risk comparison groups have not been clearly established.

The strong role of radiation in the excess risk for SC is demonstrated by the increased risks of breast cancer and supradiaphragmatic cancers in patients whose initial treatment included RT. Similar observations in prior studies^3-6,17-19 have influenced HL management, with greater use of chemotherapy as initial treatment, and more selective use of smaller involved-field RT.^20-22 It is notable, however, that the risk for cancer at all major evaluated sites was significantly elevated after initial treatment with chemotherapy alone (about 3.3-fold). These findings may be due in part to subsequent RT use not documented in our database, and to the use of alkylating agents, which have been associated with increased risks of lung cancer^17,23 and other SC.^6,24 We also found no evidence that patients treated after 1984 had a lower risk of all SC taken together than those patients who were diagnosed from 1970 to 1984. However, since this comparison was limited to 5 to 15 years after HL diagnosis, it does not provide a complete assessment of long-term risk. In particular, the transition from extended-mantle RT to smaller-volume involved-field RT occurred well after 1984,^20,21,25-27 and one would not expect the anticipated reduction in SC risk due to smaller RT fields to be apparent in the latter period of this study. These findings highlight the importance of quantifying the risk of chemotherapy-induced cancers,²⁸ especially in view of carcinogenicity in laboratory animals,^29,30 and highlight a better understanding of the extent to which the SC risk in HL patients is biologically mediated in ways unrelated to treatment.^19,31

To our knowledge, this is the first study to identify a significantly increased risk of cancer of the pleura (RR = 19.5) following treatment for HL. Previously, only case reports have documented mesothelioma diagnoses, typically more than 10 years after RT for HL.^32-35 Similarly, few studies have identified a significant increase in the RR of brain cancers among HL survivors.^5,6 RT fields that encompass lymph nodes in the upper neck may include part of the posterior fossa, delivering an average radiation dose of 0.39 Gy to brain (Appendix Table A1, online only). Our findings are compatible with evidence that meningiomas and other CNS cancers may develop after low-dose RT.^36,37

This study has limitations that warrant consideration. We lack detailed comprehensive treatment data that would help refine our SC risk estimates. Although the models in this article describe the dependencies of the RR and EAR on age at HL diagnosis and attained age, they should not be interpreted as defining these patterns precisely. Caution is needed in making predictions of second cancer risk beyond 30 years of follow-up for patients who are young when diagnosed with HL, since the estimated changes in risk seen with advanced attained age were based on those patients diagnosed with HL at older ages.

Nevertheless, our results indicate that while there is a decrease in the risk of non-breast SC among elderly HL survivors, many survivors remain at increased risks for SC for most of their remaining lives. Ongoing research should help to identify HL treatment modifications to reduce SC risk while maintaining efficacy. For current survivors, there is a need to investigate interventions to reduce the morbidity and mortality caused by second cancers and to explore possible genetic modifiers of treatment-related effects.^15,38

	AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
The authors indicated no potential conflicts of interest.

	AUTHOR CONTRIBUTIONS

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Conception and design: David C. Hodgson, Ethel S. Gilbert, Lois B. Travis

Financial support: Lois B. Travis

Administrative support: Lois B. Travis

Provision of study materials or patients: Per Hall, Froydis Langmark, Eero Pukkala, Michael Andersson, Magnus Kaijser, Heikki Joensuu, Sophie D. Fosså, Lois B. Travis

Collection and assembly of data: Charles F. Lynch, Hans Storm, Per Hall, Froydis Langmark, Eero Pukkala, Michael Andersson, Magnus Kaijser, Heikki Joensuu, Sophie D. Fosså, Lois B. Travis

Data analysis and interpretation: David C. Hodgson, Ethel S. Gilbert, Graça M. Dores, Sara J. Schonfeld, Hans Storm, Sophie D. Fosså, Lois B. Travis

Manuscript writing: David C. Hodgson, Ethel S. Gilbert, Graça M. Dores, Sara J. Schonfeld, Hans Storm, Per Hall, Sophie D. Fosså, Lois B. Travis

Final approval of manuscript: David C. Hodgson, Ethel S. Gilbert, Graça M. Dores, Sara J. Schonfeld, Charles F. Lynch, Hans Storm, Eero Pukkala, Michael Andersson, Magnus Kaijser, Heikki Joensuu, Sophie D. Fosså, Lois B. Travis

	Appendix

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
Statistical methods. The follow-up period began 1 year after the date of Hodgkin's lymphoma (HL) diagnosis and ended on the date of death, date of diagnosis of second cancer, or the study end date (December 31, 2002), whichever occurred first. Thus, third or higher order cancers were not considered in this study. Person-years (PY) and second cancers were categorized by sex, calendar year of HL diagnosis (1970 to 1984 or 1985 to 1996), initial treatment modality, registry, and by 5-year intervals of attained age, attained calendar year, time since HL diagnosis, and age at HL diagnosis. Cancer incidence rates specific for each registration area, sex, and 5-year age and calendar year intervals were multiplied by the corresponding accumulated PY at risk to estimate the number of cancer cases expected in each stratum.

In general, O and E are used to denote observed and expected numbers of incident second cancers. O_ax,a,k, E_ax,a,k, and PY_ax,a,k are used to denote observed cases, expected cases, and PY, respectively, in a specified category identified by age at HL diagnosis (ax), attained age (a), and other variables of interest (k). Analyses were based on Poisson regression methods, in which it is assumed that the number of incident solid cancers follows a Poisson distribution with the mean given by the product of the PY and the cancer-specific incidence rate for each cell of a multi-way person-year table (Breslow NE, Day NE: Statistical methods in cancer research, volume II: The design and analysis of cohort studies. IARC Sci Publ 82, 1987). Analyses that included attained age, time since diagnosis, and age at diagnosis as continuous variables were based on midpoints of 5-year intervals. For example, the attained age group of 60 to 64 years was assigned a value of 62.5.

Parameter estimates were computed with maximum likelihood methods. Hypothesis tests and CIs were based on likelihood ratio tests and direct evaluation of the profile likelihood. Two-sided P values are used throughout. Analyses were implemented with the AMFIT module of the software package EPICURE (Preston DL, Lubin JH, Pierce DA: EPICURE User's Guide., Wash: HiroSoft International Corporation, 1991; Seattle, WA).

Both the excess relative risk (ERR = relative risk [RR] –1) and the excess absolute risk (EAR) were evaluated. The statistical expectation of O_ax,a,k was assumed to be:

(A1)

(A2)

For detailed modeling, broad categories of solid cancers were defined based on anatomic site, estimated radiation dose within traditional supradiaphragmatic (mantle) and infradiaphragmatic RT fields, and whether there were differences in the nature of the dependencies of risk on age at HL diagnosis and attained age. The selected categories were female breast cancer, thyroid cancer, other supradiaphragmatic sites (head and neck, esophageal, and respiratory cancers), and infradiaphragmatic sites (nonesophageal gastrointestinal cancers, and urinary cancers). The phrase "supradiaphragmatic sites" as used in this article always excludes female breast and thyroid cancer. Less detailed analyses were conducted for cancers not included in the listed categories.

Solid cancer risks in the first 10 years following HL diagnosis were generally lower than risks in later latency periods, with particularly low risks in the first 5 years. Thus, most analyses were restricted to 5-year survivors of HL, and a parameter was included in the model that allowed the ERR or EAR in the 5- to 9-year period to differ from that in the later intervals.

Both the ERR and EAR were expressed as functions of age at HL diagnosis in years (ax), attained age in years (a), and time since diagnosis (t = 1 for 5-9 years; t = 0 for 10 years), as indicated in Equation 3. Greek letters denote parameters to be estimated.

The model used for initial evaluation and for statistical testing of dependencies of the ERR and EAR on ax and a is as follows, and is referred to as the "a priori model":

(A3)

This model has been used in analyses of solid cancer risks following testicular cancer (Travis LB, Fossa SD, Schonfeld SJ, et al: Second cancers among 40,576 testicular cancer patients: Focus on long-term survivors. J Natl Cancer Inst 97:1354-1365, 2005), and in other radiation-exposed cohorts (Preston DL, Mattsson A, Holmberg E, et al: Radiation effects on breast cancer risk: A pooled analysis of eight cohorts. Radiat Res 158:220-235, 2002; Preston DL, Shimizu Y, Pierce DA, et al: Studies of mortality of atomic bomb survivors, Report 13: Solid cancer and noncancer disease mortality—1950 to 1997. Radiat Res 160:381-407, 2003).

In subsequent analyses, we evaluated models with several alternative functions of ax and a, including functions in which the dependencies applied only to various specified age ranges, and also including models that substituted time since HL diagnosis for attained age. The fits to the data of various models were judged by comparing deviance statistics, and were also checked by comparing deviances to models that estimated the ERR or EAR for categories defined by age at HL diagnosis and attained age.

For female breast and thyroid cancer, the a priori model provided an adequate fit to the data although for the breast cancer EAR model; it was necessary to include both a linear and quadratic term for age at HL diagnosis (ax) in the exponential term of equation 3. For the supra- and infradiaphragmataic sites, the following alternative models fitted the data much better than the a priori model (P = .002 for ERR; P < .001 for EAR). These models allow for modification of the ERR or EAR only for ages at HL diagnosis younger than 40 years, and, for the ERR model, only for attained ages older than 60 years.

(A4)

(A5)

where ax40 = ax –40 for ax less than 40, otherwise ax40 = 0; a60+ = a/60 for a 60, otherwise a60 = 1. The same dependencies of the ERR and EAR on age at HL diagnosis and attained age were found to be appropriate for both supra- and infradiaphragmatic sites, although the overall risk () was different.

Heterogeneity among the five population-based cancer registries was investigated by comparing fits of models with country-specific estimates of the main effect parameter to that of models with a single value for all countries. Heterogeneity in parameters expressing the dependence of risk on age at HL diagnosis or attained age was also investigated. In a previous article on solid cancer after testicular cancer (Travis LB, Fossa SD, Schonfeld SJ, et al: Second cancers among 40,576 testicular cancer patients: Focus on long-term survivors. J Natl Cancer Inst 97:1354-1365, 2005), we found that rates for the North American patients (United States and Canada) were lower than that for other countries, so we introduced a correction factor for this. In this study, risks for the United States were not consistently lower than those for the Nordic registries, although there was evidence of heterogeneity among the remaining Nordic registries. We did not apply such a correction in the current study and results must be regarded as an average for all population-based cancer registries.

For analyses of site-specific cancers other than female breast and thyroid (Table 3), we used the estimated values of β₁ and β₂ from the ERR model for supra- and infradiaphragmatic sites (equation 4), after testing whether data were compatible with these values. With this approach, the ratios of ERRs for different cancer sites are the same for all ages at HL diagnosis and all attained ages. Results in Tables 3, 4, and 5 are presented for age 30 years at HL diagnosis. Results in Tables 3 and 5 are for attained age 40 years; however, because the ERR for the supra- and infradiaphragmatic sites is constant for attained ages younger than 60 years, the ERR values for these sites apply to attained ages 40 to 60 years. The number of excess cancers was estimated as the sum over categories defined by ax, a, and t of the terms

(A6)

Analyses according to sex, treatment, and year of HL diagnosis (Table 5) were adjusted for age at HL diagnosis and attained age. The analysis of time since HL diagnosis (Table 4) was adjusted for age at HL diagnosis. Analyses comparing solid cancer risks for HL patients diagnosed before and after 1985 were restricted to the 5 to 14 year period after HL diagnosis because few patients diagnosed after 1985 received follow-up for more than 15 years.

We investigated how the increased risk of solid cancer might have an impact on recommendations for cancer screening by estimating the age at which HL survivors had absolute risks of screen-detectable cancers (breast and colorectal) comparable to those risks seen in the general population for whom screening is recommended. Age-specific rates of breast and colorectal cancer in the general population (baseline risks) were obtained as the average rates (weighted by PY) for all registries for the 1985+ period. The total absolute risk was obtained as the sum of the baseline rate and the EAR. For breast cancer, the EAR was obtained from the model described above (equation 2 and the information shortly thereafter). For colorectal cancers, the EAR was obtained by multiplying ERR (Table 3) by age-specific baseline rates. Combining the two sexes gave similar results to sex-specific analyses.

Cumulative probabilities of 5-year survivors of HL subsequently developing solid cancers were calculated using an approach similar to that used by Travis et al (Travis LB, Fossa SD, Schonfeld SJ, et al: Second cancers among 40,576 testicular cancer patients: Focus on long-term survivors. J Natl Cancer Inst 97:1354-1365, 2005). The approach takes into account the dependency of absolute risks on both age at HL diagnosis and attained age and the dependency on competing risks from HL mortality, noncancer mortality, and any intervening diagnosis of leukemia or other nonsolid cancer. All calculations were sex-specific and required both baseline rates and excess risk models for various disease categories. Baseline rates for cancer incidence and noncancer mortality were obtained as the average age- and sex-specific rates (weighted by PY) for all registries from 1985 through 2001. Excess solid cancer was estimated using the EAR models indicated above and in Table 2, with an additional model developed to address the residual cancers not included in the categories noted above (the information that follows equations 1 and 2).

Separate estimates of the cumulative risks for female breast cancer, supradiaphragmatic sites, infradiaphragmatic sites, thyroid cancer, and the residual category of all other solid cancers were obtained and then summed to estimate the cumulative risk for all solid cancers. Models were developed to address competing risks from HL mortality, infectious disease mortality, mortality from remaining nonmalignant diseases, and non-Hodgkin's lymphoma incidence. The EAR model developed by Schonfeld et al (Schonfeld SJ, Gilbert ES, Dores GM, et al: Acute myeloid leukemia following Hodgkin's lymphoma: A population-based study of 35,511 patients. J Natl Cancer Inst 98:215-218, 2006) was used for leukemia incidence. It was assumed that there was no excess risk of mortality from external causes.

	ACKNOWLEDGMENTS

 
We thank Jeremy Miller, Information Management Services (Rockville, MD), for expert computer support and data management, Cathy Kasper, M.D. Anderson Cancer Center (Houston, TX), for estimation of radiation doses, and Denise Duong, National Cancer Institute (Bethesda, MD), for typing support.

	NOTES

 
published online ahead of print at www.jco.org on March 19, 2007.

Supported by the Intramural Research Program of the National Institutes of Health, the National Cancer Institute, Division of Cancer Epidemiology and Genetics, and by a career development award from the Ministry of Health and Long Term Care of Ontario, Canada (D.C.H.).

Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.

	REFERENCES

TOP ABSTRACT INTRODUCTION PATIENTS AND METHODS RESULTS DISCUSSION AUTHORS' DISCLOSURES OF... AUTHOR CONTRIBUTIONS Appendix REFERENCES

 
1. Aleman BM, van den Belt-Dusebout AW, Klokman WJ, et al: Long-term cause-specific mortality of patients treated for Hodgkin's disease. J Clin Oncol 21:3431-3439, 2003[Abstract/Free Full Text]

2. Hoppe RT: Hodgkin's disease: Complications of therapy and excess mortality. Ann Oncol 8:115-118, 1997[CrossRef][Medline]

3. van Leeuwen FE, Klokman WJ, Hagenbeek A, et al: Second cancer risk following Hodgkin's disease: A 20-year follow-up study. J Clin Oncol 12:312-325, 1994[Abstract]

4. van Leeuwen FE, Klokman WJ, Veer MB, et al: Long-term risk of second malignancy in survivors of Hodgkin's disease treated during adolescence or young adulthood. J Clin Oncol 18:487-497, 2000[Abstract/Free Full Text]

5. Dores GM, Metayer C, Curtis RE, et al: Second malignant neoplasms among long-term survivors of Hodgkin's disease: A population-based evaluation over 25 years. J Clin Oncol 20:3484-3494, 2002[Abstract/Free Full Text]

6. 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[Abstract/Free Full Text]

7. Ng AK, Bernardo MV, Weller E, et al: Second malignancy after Hodgkin disease treated with radiation therapy with or without chemotherapy: Long-term risks and risk factors. Blood 100:1989-1996, 2002[Abstract/Free Full Text]

8. Parkin DM, Muir CS: Comparability and quality of data, in Parkin DM, Muir CS, Whelan SL, et al (eds): Cancer Incidence in Five Continents. Lyon, France, IARC Sci Publ 120:45-173, 1992

9. Parkin DM, Plummer M: Comparability and quality of data, in Parkin DM, Muir CS, Whelan SL, et al (eds): Cancer Incidence in Five Continents. Lyon, France, International Agency for Research on Cancer 155:57-73, 2002

10. Breslow NE, Day NE: Statistical methods in cancer research, volume II: The design and analysis of cohort studies. Lyon, France, IARC Sci Publ 1-406, 1987

11. Travis LB, Fossa SD, Schonfeld SJ, et al: Second cancers among 40,576 testicular cancer patients: Focus on long-term survivors. J Natl Cancer Inst 97:1354-1365, 2005[Abstract/Free Full Text]

12. Foss Abrahamsen A, Andersen A, Nome O, et al: Long-term risk of second malignancy after treatment of Hodgkin's disease: The influence of treatment, age and follow-up time. Ann Oncol 13:1786-1791, 2002[Abstract/Free Full Text]

13. Pignone M, Rich M, Teutsch SM, et al: Screening for colorectal cancer in adults at average risk: A summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 137:132-141, 2002[Abstract/Free Full Text]

14. Canadian Task Force on Preventive Health Care: Colorectal cancer screening: Recommendation statement from the Canadian Task Force on Preventive Health Care. CMAJ 165:206-208, 2001[Free Full Text]

15. Mauch P, Ng A, Aleman B, et al: Report from the Rockefellar Foundation Sponsored International Workshop on reducing mortality and improving quality of life in long-term survivors of Hodgkin's disease: July 9-16, 2003, Bellagio, Italy. Eur J Haematol Suppl 68-76, 2005

16. Das P, Ng AK, Earle CC, et al: Computed tomography screening for lung cancer in Hodgkin's lymphoma survivors: Decision analysis and cost-effectiveness analysis. Ann Oncol 17:785-793, 2006[Abstract/Free Full Text]

17. Travis LB, Gospodarowicz M, Curtis RE, et al: Lung cancer following chemotherapy and radiotherapy for Hodgkin's disease. J Natl Cancer Inst 94:182-192, 2002[Abstract/Free Full Text]

18. Travis LB, Hill DA, Dores GM, et al: Breast cancer following radiotherapy and chemotherapy among young women with Hodgkin disease. JAMA 290:465-475, 2003[Abstract/Free Full Text]

19. Travis LB, Hill D, Dores GM, et al: Cumulative absolute breast cancer risk for young women treated for Hodgkin lymphoma. J Natl Cancer Inst 97:1428-1437, 2005[Abstract/Free Full Text]

20. Anselmo AP, Cavalieri E, Osti FM, et al: Intermediate stage Hodgkin's disease: Preliminary results on 210 patients treated with four ABVD chemotherapy cycles plus extended versus involved field radiotherapy. Anticancer Res 24:4045-4050, 2004[Abstract/Free Full Text]

21. Bonadonna G, Bonfante V, Viviani S, et al: ABVD plus subtotal nodal versus involved-field radiotherapy in early-stage Hodgkin's disease: Long-term results. J Clin Oncol 22:2835-2841, 2004[Abstract/Free Full Text]

22. Engert A, Schiller P, Josting A, et al: Involved-field radiotherapy is equally effective and less toxic compared with extended-field radiotherapy after four cycles of chemotherapy in patients with early-stage unfavorable Hodgkin's lymphoma: Results of the HD8 trial of the German Hodgkin's Lymphoma Study Group. J Clin Oncol 21:3601-3608, 2003[Abstract/Free Full Text]

23. Swerdlow AJ, Schoemaker MJ, Allerton R, et al: Lung cancer after Hodgkin's disease: A nested case-control study of the relation to treatment. J Clin Oncol 19:1610-1618, 2001[Abstract/Free Full Text]

24. Birdwell SH, Hancock SL, Varghese A, et al: Gastrointestinal cancer after treatment of Hodgkin's disease. Int J Radiat Oncol Biol Phys 37:67-73, 1997[CrossRef][Medline]

25. Bates NP, Williams MV, Bessell EM, et al: Efficacy and toxicity of vinblastine, bleomycin, and methotrexate with involved-field radiotherapy in clinical stage IA and IIA Hodgkin's disease: A British National Lymphoma Investigation pilot study. J Clin Oncol 12:288-296, 1994[Abstract]

26. Hudson MM, Greenwald C, Thompson E, et al: Efficacy and toxicity of multiagent chemotherapy and low-dose involved-field radiotherapy in children and adolescents with Hodgkin's disease. J Clin Oncol 11:100-108, 1993[Abstract]

27. Vose JM, Bierman PJ, Anderson JR, et al: CHLVPP chemotherapy with involved-field irradiation for Hodgkin's disease: Favorable results with acceptable toxicity. J Clin Oncol 9:1421-1425, 1991[Abstract]

28. Travis LB: Therapy-associated solid tumors. Acta Oncol 41:323-333, 2002[CrossRef][Medline]

29. International Agency for Research on Cancer: IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Overall Evaluations of Carcinogenicity: An Updating of IARC Monographs Volumes 1 to 42. Lyon, France, World Health Organisation International Agency for Research on Cancer, 1987 (suppl 7)

30. International Agency for Research on Cancer: IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Pharmaceutical Drugs, Volume 50. Lyon, France, World Health Organisation International Agency for Research on Cancer, 1990

31. Yahalom J: Breast cancer after Hodgkin disease: Hope for a safer cure. JAMA 290:529-531, 2003[Free Full Text]

32. Antman KH, Corson JM, Li FP, et al: Malignant mesothelioma following radiation exposure. J Clin Oncol 1:695-700, 1983[Abstract]

33. Weissmann LB, Corson JM, Neugut AI, et al: Malignant mesothelioma following treatment for Hodgkin's disease. J Clin Oncol 14:2098-2100, 1996[Abstract/Free Full Text]

34. Henley JD, Loehrer PJ Sr, Ulbright TM: Deciduoid mesothelioma of the pleura after radiation therapy for Hodgkin's disease presenting as a mediastinal mass. Am J Surg Pathol 25:547-548, 2001[CrossRef][Medline]

35. Kramer G, Gans S, Rijnders A, et al: Long term survival of a patient with malignant pleural mesothelioma as a late complication of radiotherapy for Hodgkin's disease treated with 90-yttrium-silicate. Lung Cancer 27:205-208, 2000[CrossRef][Medline]

36. Ron E, Modan B, Boice JD Jr, et al: Tumors of the brain and nervous system after radiotherapy in childhood. N Engl J Med 319:1033-1039, 1988[Abstract]

37. Sadetzki S, Chetrit A, Freedman L, et al: Long-term follow-up for brain tumor development after childhood exposure to ionizing radiation for tinea capitis. Radiat Res 163:424-432, 2005[CrossRef][Medline]

38. Travis LB, Rabkin CS, Brown LM, et al: Cancer survivorship: Genetic susceptibility and second primary cancers—Research strategies and recommendations. J Natl Cancer Inst 98:15-25, 2006[Abstract/Free Full Text]

Submitted September 7, 2006; accepted January 23, 2007.

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This article has been cited by other articles:

				M. L. De Bruin, L. D. A. Dorresteijn, M. B. van't Veer, A. D. G. Krol, H. J. van der Pal, A. C. Kappelle, W. Boogerd, B. M. P. Aleman, and F. E. van Leeuwen Increased Risk of Stroke and Transient Ischemic Attack in 5-Year Survivors of Hodgkin Lymphoma J Natl Cancer Inst, July 1, 2009; 101(13): 928 - 937. [Abstract] [Full Text] [PDF]

				M. L. De Bruin, J. A. Burgers, P. Baas, M. B. van 't Veer, E. M. Noordijk, M. W. J. Louwman, J. M. Zijlstra, H. van den Berg, B. M. P. Aleman, and F. E. van Leeuwen Malignant mesothelioma after radiation treatment for Hodgkin lymphoma Blood, April 16, 2009; 113(16): 3679 - 3681. [Abstract] [Full Text] [PDF]

				J. D. Rizzo, R. E. Curtis, G. Socie, K. A. Sobocinski, E. Gilbert, O. Landgren, L. B. Travis, W. D. Travis, M. E. D. Flowers, D. L. Friedman, et al. Solid cancers after allogeneic hematopoietic cell transplantation Blood, January 29, 2009; 113(5): 1175 - 1183. [Abstract] [Full Text] [PDF]

				M. Lustberg and C. L. Shapiro Physical Challenges of Adult Cancer Survivors ASCO Educational Book, January 1, 2009; 2009(1): 564 - 569. [Abstract] [Full Text] [PDF]

				K. A. Goodman, E. Riedel, V. Serrano, S. Gulati, C. H. Moskowitz, and J. Yahalom Long-Term Effects of High-Dose Chemotherapy and Radiation for Relapsed and Refractory Hodgkin's Lymphoma J. Clin. Oncol., November 10, 2008; 26(32): 5240 - 5247. [Abstract] [Full Text] [PDF]

				K. Hemminki, P. Lenner, J. Sundquist, and J. Lorenzo Bermejo Risk of Subsequent Solid Tumors After Non-Hodgkin's Lymphoma: Effect of Diagnostic Age and Time Since Diagnosis J. Clin. Oncol., April 10, 2008; 26(11): 1850 - 1857. [Abstract] [Full Text] [PDF]

				H. Brenner, A. Gondos, and D. Pulte Ongoing improvement in long-term survival of patients with Hodgkin disease at all ages and recent catch-up of older patients Blood, March 15, 2008; 111(6): 2977 - 2983. [Abstract] [Full Text] [PDF]

				L J Worrillow, A G Smith, K Scott, M Andersson, A J Ashcroft, G M Dores, B Glimelius, E Holowaty, G H Jackson, G L Jones, et al. Polymorphic MLH1 and risk of cancer after methylating chemotherapy for Hodgkin lymphoma J. Med. Genet., March 1, 2008; 45(3): 142 - 146. [Abstract] [Full Text] [PDF]

				L. Lee, M. Pintilie, D. C. Hodgson, P. E. Goss, and M. Crump Screening mammography for young women treated with supradiaphragmatic radiation for Hodgkin's lymphoma Ann. Onc., January 1, 2008; 19(1): 62 - 67. [Abstract] [Full Text] [PDF]

				A. K. Ng Second Cancers after Hodgkin's Lymphoma: Long-term Risks ASCO Educational Book, January 1, 2008; 2008(1): 418 - 422. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hodgson, D. C. Articles by Travis, L. B. Search for Related Content PubMed PubMed Citation Articles by Hodgson, D. C. Articles by Travis, L. B. Social Bookmarking                            What's this?