Journal of the National Cancer Institute Advance Access originally published online on August 8, 2007
JNCI Journal of the National Cancer Institute 2007 99(16):1248-1256; doi:10.1093/jnci/djm081
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Published by Oxford University Press 2007.
ARTICLES |
Comparative Mortality for 621 Second Cancers in 29356 Testicular Cancer Survivors and 12420 Matched First Cancers
Affiliations of authors: Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Bethesda, MD (CS, MH, BEC, LMB, RH, MG, LBT); Department of Clinical Research, Rikshospitalet-Radiumhospitalet Medical Centre, Oslo, Norway (SDF)
Correspondence to: Catherine Schairer, PhD, National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Executive Plaza South, Rm 8020, Bethesda, MD 20892 (e-mail: schairec{at}exchange.nih.gov).
 |
ABSTRACT |
|---|
 Top Abstract Context and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
Background: Testicular cancer survivors, many of whom have undergone radiotherapy, are at substantial risk of second cancers. Treatment for testicular cancer may limit treatment options for second cancers, thereby adversely affecting survival after the second cancer. However, no data on outcomes of testicular cancer survivors with second cancers compared to patients with comparable first cancers exist.
Methods: Among 29356 white testicular cancer patients reported to the Surveillance, Epidemiology, and End Results (SEER) program (1973–2002), 621 developed a second cancer with known stage and were matched to a random sample of 12420 white male first cancer patients in the SEER program by cancer site, stage, diagnosis year, and age at diagnosis. Mortality was ascertained through 2002. Cancer-specific and all-cause mortality following second cancers were compared with those of matched first cancers, and rate ratios (RRs) were estimated using proportional hazards analysis. Survival functions were calculated using product-limit estimates.
Results: During the study period, 284 testicular cancer survivors with second cancers died, 191 from their second cancer; 5443 matched first cancer patients died, 3929 from their first cancer. Rate ratios for cancer-specific and all-cause mortality for second cancers compared with matched first cancers were 1.05 (95% confidence interval [CI] = 0.90 to 1.23) and 1.09 (95% CI = 0.96 to 1.23), respectively. However, among testicular cancer patients who were diagnosed during 1973–1979, an era in which radiation therapy was given at high doses and to the chest area, all-cause mortality following second cancers at sites below the diaphragm (79 deaths) and second lung cancers (29 deaths) was statistically significantly higher than that from matched first cancers (RR = 1.44, 95% CI = 1.13 to 1.83, and RR = 1.65, 95% CI = 1.12 to 2.42, respectively).
Conclusions: Mortality from second cancers following testicular cancer was similar to matched first cancers, except for selected tumors in the radiotherapy field among testicular cancer patients who were diagnosed during 1973–1979, a time when radiotherapy doses for treatment of testicular cancer were high and chest irradiation was an option in standard practice.
Prior knowledge Survivors of testicular cancer are at high risk of developing second cancers. Study design Comparison of cancer and all-cause mortality among testicular cancer survivors who developed second cancers and patients with matched first cancers using data from the Surveillance, Epidemiology, and End Results program from 1973 to 2002. Separate analyses were performed for testicular cancer survivors who were diagnosed with testicular cancer during 1973–1979, when high-dose radiation therapy was common and irradiation to the chest area was a treatment option. Contribution Cancer-specific and all-cause mortality were similar between the groups overall. However, for diagnoses during 1973–1979, all-cause mortality among testicular cancer survivors with second cancers of the lung and at sites below the diaphragm was higher than that among patients with first cancers at these anatomic sites. Implications The mortality of testicular cancer survivors with second cancers was similar to that of patients with matched first cancers, but it was higher among testicular cancer survivors who were diagnosed with testicular cancer during 1973–1979 who developed second cancers at sites that were likely to have been in the radiotherapy treatment field during treatment for testicular cancer. Limitations The number of second cancers was small, and details of the treatments patients received and lifestyle risk factors were not available.
|
Treatment for testicular cancer is one of the major medical successes of the 20th century, due largely to the introduction of platinum-based chemotherapy in the early 1970s (1). Most men with testicular cancer are now cured, with a 10-year relative survival of 95% (2,3). Because testicular cancer is generally diagnosed during young adulthood (4), most patients remain at risk of second cancers for many years (4,5). For example, it is estimated that approximately one-third of men who are diagnosed with a testicular seminoma at age 35 years will have developed a second cancer 40 years later (4).
Despite the substantial risk of developing a second tumor following testicular cancer, it is not known whether survival associated with a second cancer differs from that associated with a first cancer of the same anatomic site. Prior treatment for testicular cancer may limit management options for second cancers and thus may adversely affect survival (6,7). In particular, long-term effects associated with prior tissue irradiation may influence the radiotherapy dose that can be delivered (6), and surgical approaches may be limited by radiation-induced fibrosis (8) and impaired wound healing (9). In addition, mechanisms of therapy-related carcinogenesis, such as genomic instability caused by mutations in DNA repair genes (10), may negatively affect the response of second cancers to treatment (11).
To determine whether mortality following a second cancer among testicular cancer survivors differs from mortality following comparable first cancers in the general population, we compared cancer-specific and all-cause mortality among 621 white testicular cancer survivors with second cancers to mortality among a random sample of 12420 age-matched white male patients with first cancers of the same anatomic site, stage, and calendar year of diagnosis. To evaluate any influence of different management strategies on patient mortality, we did not match patients by cancer treatment. All study subjects were identified from population-based cancer registries within the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program.
|
Subjects and Methods |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
The underlying cohort from which second cancer patients were identified consisted of 29356 white men who were diagnosed with a first primary invasive testicular germ cell tumor [International Classification of Diseases for Oncology, Third Edition (ICD-O-3) codes C62.0–C62.9; histology codes 9060–9062 (seminoma) and 9070–9073, 9080–9085, and 9100–9102 (nonseminoma) (12)] between January 1, 1973, and December 31, 2002, that was reported to 15 population-based SEER program registries and who survived for at least 2 months (13). Testicular cancer with histology code 9063 (spermatocytic seminoma) was excluded. The SEER program registries include approximately 26% of the US population (3). The cohort was limited to white men due to the rarity of this tumor among other racial and ethnic groups. Excluding second testicular cancers and Kaposi sarcoma, 621 testicular cancer patients subsequently developed a second primary invasive nonlymphohematopoietic cancer; stage was recorded for 590 (95%) of these patients. An additional 31 testicular cancer patients developed leukemia other than chronic lymphocytic leukemia. Second cancers were diagnosed from March 1, 1973, to December 31, 2002, in nine SEER registries (states of Connecticut, Iowa, New Mexico, Utah, and Hawaii and the metropolitan areas of Detroit, Michigan, Atlanta, Georgia, San Francisco-Oakland, California, and Seattle [Puget Sound], Washington); from March 1, 1978, to December 31, 2002, in rural Georgia; from March 1, 1979, to December 31, 2002, in the state of New Jersey; from March 1, 1988, to December 31, 2002, in San Jose-Monterey and Los Angeles, California; and from March 1, 1995, to December 31, 2002, in the states of Kentucky and Louisiana. A total of 621 second cancers were included in the analysis (590 nonlymphohematopoietic cancers with recorded stage and 31 non–chronic lymphocytic leukemias).
We compared the age, calendar year, and stage distribution of the 621 second cancers in testicular cancer patients with the distribution for first nonlymphohematopoietic invasive cancers with recorded stage and non–chronic lymphocytic leukemias identified from the same population-based registries. Second cancers were diagnosed on average at an earlier age (55.4 years versus 65.9 years), a later calendar year (1995 versus 1992), and an earlier stage (46% versus 40% localized for nonlymphohematopoietic tumors other than prostate cancer) than comparable first cancers. Therefore, for each of the 621 second cancers, we selected 20 first cancers at random without replacement with the same values for the following variables: 1) SEER program site recode (based on ICD-O-3 anatomic site and histology) (13), 2) SEER program historic stage A (local, regional, or distant) (14), 3) age at diagnosis (matched within 5-year intervals), and 4) calendar year of diagnosis (matched within 5-year intervals). For some rare sites (i.e., rectum, brain, anus, retroperitoneum, other male genital organs, ureter, other urinary, acute monocytic leukemia, other leukemia, and aleukemic, subleukemic, and leukemia not otherwise specified), criteria for age and calendar year were relaxed to identify a sufficient number of matches. A total of 12420 matched first cancers were included in the analysis; 95% of both the second cancers and the matched first cancers were histologically confirmed.
We attributed cause of death for the 621 second cancers and 12420 first cancers to the specific incident cancer if one of the following criteria was met: 1) the SEER program mortality site recode for cause of death (based on International Classification of Diseases, Eighth, Ninth, and Tenth Revisions) (15–17) and the site recode for the cancer were identical; 2) the site recodes for mortality and cancer were not identical, but one was a general version of the other (e.g., "other oral cavity and pharynx" compared with tongue) or the two sites are difficult to distinguish (e.g., sigmoid colon versus rectum); 3) the mortality site recode represented the same organ system (as specified by ICD-O-3) as the cancer site recode and there were no higher order cancers (third or higher cancers for testicular cancer survivors; second or higher cancers for first cancer patients) with the same site recode as the cause of death; 4) the mortality site recode was not in the same organ system as the cancer site recode, but the sites were in close proximity (e.g., "other oral cavity and pharynx" and esophagus; pleura and lung), and there were no subsequent cancers with the same site recode as the cause of death; or 5) the mortality site recode was "miscellaneous cancer" or "unknown neoplasm," and there were no subsequent cancers identified in the SEER program database. In 91% of the instances among which the cause of death was determined to be due to the cancer of interest, criteria specified in categories 1 or 2 were met for both the second and first cancers. Third-order (n = 44), fourth-order (n = 1), and fifth-order (n = 1) cancers were identified among 46 (7%) testicular cancer patients with a second cancer, and second-order (n = 781), third-order (n = 54), fourth-order (n = 5), and fifth-order (n = 1) cancers were identified among 841 (7%) patients with a first cancer.
We used a similar algorithm to identify death due to higher order cancers. Some deaths due to cancer could not reasonably be attributed to any of the incident cancers identified in the SEER program database and may either reflect metastatic sites that were erroneously coded as new cancers or subsequent cancers that were not reported to SEER program registries. Deaths from higher order and unidentified cancers were included only in analyses of all-cause mortality and were not considered as deaths from the second or matched first cancers.
For some analyses, we divided second cancers following testicular cancer into four groups: 1) those that were located in standard subdiaphragmatic radiotherapy fields for testicular cancer during the study period (N = 367): stomach (n = 11), small intestine (n = 1), colon (n = 49), rectum (n = 33), anus (n = 3), liver (n = 4), other biliary (n = 2), pancreas (n = 27), retroperitoneum (n = 3), other digestive organs (n = 1), prostate (n = 130), other male genital organs (n = 1), bladder (n = 64), kidney (n = 35), ureter (n = 2), and other urinary organs (n = 1); 2) those that were located in the thoracic radiotherapy field for testicular cancer (17) (N = 102): esophagus (n = 8), lung (n = 89), and pleura (n = 5); 3) other nonhematologic cancers (N = 121): oral cavity and pharynx (n = 28), larynx (n = 10), soft tissue including heart (n = 10), malignant melanoma (n = 57), brain and other nervous system (n = 1), thyroid (n = 14), and miscellaneous malignant cancer (n = 1); 4) leukemias (N = 31): acute lymphoblastic leukemia (n = 1), other lymphocytic leukemia (n = 3), acute myeloid leukemia (n = 18), chronic myeloid leukemia (n = 4), acute monocytic leukemia (n = 2), other acute leukemia (n = 1), aleukemic, subleukemic, and not otherwise specified (n = 2). First cancers were similarly divided into these groups.
The SEER program routinely collects data on the initial course of cancer treatment in terms of broad categories, such as radiotherapy, chemotherapy, or surgery. Data on radiotherapy fields or specific cytotoxic drugs are not recorded, and information is not collected with regard to subsequent courses of treatment. Substantial changes in the treatment of testicular cancer have occurred since the late 1970s, with the elimination of thoracic radiotherapy and reductions in the size and administered doses of irradiation to subdiaphragmatic fields (18). Therefore, in the absence of detailed data on specific treatment regimens, we accounted for gradual changes in testicular cancer treatment over time by grouping second cancers according to the calendar year of antecedent testicular cancer diagnosis, i.e., 1973–1979, 1980–1989, and 1990–2002. For example, in a prior analytic study of testicular cancer patients undertaken in population-based cancer registries (5), it was estimated that approximately 16% of patients received chest radiotherapy in the 1970s but that few, if any, received thoracic irradiation after 1979. We therefore also conducted selected analyses of second cancers in the radiotherapy field for testicular cancer among those whose first course of treatment for testicular cancer was known to include radiotherapy.
Statistical Analysis
Mortality was assessed from the date of second cancer diagnosis and the date of first cancer diagnosis for the matched sample to whichever date occurred first: death, loss to follow-up, or end of study (December 31, 2002).
Hazard rate ratios (RRs) of death from second cancers (henceforth referred to as cancer-specific mortality) and from all causes (henceforth referred to as all-cause mortality) were compared with those in the sample of first cancers using Cox proportional hazards regression models (19,20). All models were stratified on the matching factors, namely site of cancer (when analyses included more than one site), stage at diagnosis (localized, regional, distant for nonlymphohematopoietic malignancies other than prostate cancer; localized, regional/distant for prostate cancer, and unstaged for non–chronic lymphocytic leukemias), age at diagnosis (in 5-year intervals), and calendar year of diagnosis (in 5-year intervals). Analyses were also conducted with additional stratification on initial treatment for second and first cancers (surgery alone, radiation with no chemotherapy, chemotherapy with no radiation, radiation and chemotherapy, other). The number of months since diagnosis of the second and first cancers was used as the time metric for the proportional hazards analyses. The proportional hazards assumption was assessed by plotting the log-minus-log of the survival distribution functions; no serious violations were noted. The Wald test was used to test for heterogeneity in rate ratios according to categories of covariates (20).
Survival functions were calculated according to months since diagnosis of the second cancer and the sample of first cancers using product-limit estimates without stratification on the matching variables. However, the two groups were exactly balanced on the matching variables. The log-rank test was used to test for the statistical significance of differences in survival functions (20).
Pearson's chi-square tests were used to compare the treatment distributions of second cancers and the sample of first cancers (20). All statistical tests were two-sided, with a .05 level for statistical significance.
|
Results |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
The 621 second cancers were diagnosed an average of 10.6 years after the testicular cancer diagnosis (range = 2 months to 28.9 years). The majority (56.8%) of second cancers occurred among men with localized testicular cancer (Table 1). For 389 patients (62.6%), initial management of testicular cancer included radiotherapy; 84 (13.5%) were administered chemotherapy alone.
|
The average duration of follow-up after second cancer diagnosis was 3.9 years (range = <1 month to 22.8 years) and that after first cancer diagnosis was 3.8 years (range = <1 month to 29.2 years). Overall, 284 second cancer patients died during the follow-up period, with 191 (67%) of the deaths due to the second cancer. A total of 5443 first cancer patients died during the follow-up period; 3929 (72%) of the deaths were attributed to the first cancer.
By design, distributions of the matching variables (stage, age at diagnosis, and calendar year of diagnosis) were similar for second cancers and the matched first cancers (Table 2). Approximately 46% of all tumors (excluding prostate cancer and leukemia) were localized at diagnosis, 35.4% were regional, and 18.5% were distant. Approximately 78% of cancers were diagnosed from ages 40–69 years, and 65% were diagnosed since 1994. Initial course of treatment, which was not a matching factor, differed statistically significantly between the second and first cancers (P = .04); a higher percentage of second cancers were treated by surgery only (49.6% versus 44.4%) and a smaller percentage with radiotherapy and chemotherapy (8.5% versus 11.7%).
|
We compared cancer-specific and all-cause mortality and median survival among second cancer and matched first cancer patients (Table 3). Rate ratios are adjusted for stage, age, calendar period of diagnosis, and cancer site (when analyses included more than one site). The rate ratios for overall cancer-specific and all-cause mortality were 1.05 (95% confidence interval [CI] = 0.90 to 1.23) and 1.09 (95% CI = 0.96 to 1.23), respectively. Small and non–statistically significant elevations in mortality from second cancers of the colon/rectum/anus, pancreas, kidney, lung, and acute myeloid leukemia were evident; conversely, there were non–statistically significant reductions in mortality from second cancers of the stomach and bladder but a non–statistically significant increase for all-cause mortality after a second bladder cancer. Adjustment for differences in initial course of treatment for first and second cancers did not substantially change the rate ratio estimates (data not shown). Although median survival differed by cancer site, there were no statistically significant differences in survival between first and second cancers of the same anatomic site.
|
We further examined findings for lung cancer according to tumor histology (small-cell versus non–small-cell), given the poorer prognosis of the former group (21). Nine (7%) second lung cancers and 280 (16%) first lung cancers were small cell in type (P = .15). The rate ratio of death due to a second small-cell lung cancer was nearly twice that observed for a first small-cell lung cancer (RR = 1.95, 95% CI = 0.92 to 4.13); that corresponding to non–small-cell lung cancer was 1.12 (95% CI = 0.96 to 1.47) (Pheterogeneity = .17). Results for all-cause mortality showed a similar pattern of greater rate ratio for second compared to first small-cell lung cancer than for second compared to first non–small-cell lung cancer.
No statistically significant differences were observed for cancer-specific and all-cause mortality for all sites combined according to prior testicular cancer histology (seminoma, nonseminoma), stage, age at diagnosis, initial course of treatment, years since testicular cancer diagnosis, age, year of diagnosis, or initial course of treatment of second and matching first cancers. However, rate ratios of cancer-specific and all-cause mortality in second cancer patients declined with calendar period of testicular cancer diagnosis. For example, rate ratios for second cancer–specific mortality were 1.29 (95% CI = 1.02 to 1.64), 1.06 (95% CI = 0.83 to 1.35), and 0.70 (95% CI = 0.49 to 1.02) among testicular cancer patients who were diagnosed during 1973–1979, 1980–1989, and 1990–2002, respectively (Pheterogeneity = .02). Similar patterns were evident for all-cause mortality. Rate ratios for cancer-specific and all-cause mortality were generally elevated regardless of time since testicular cancer diagnosis among those diagnosed during 1973–1979.
Elevations in second cancer–specific and all-cause mortality among men who were diagnosed with testicular cancer in 1973–1979 were derived largely from sites that were in the radiotherapy treatment fields for testicular cancer during that era (i.e., subdiaphragmatic and thoracic cancers) (Table 4). Rate ratios for all-cause mortality were 1.44 (95% CI = 1.13 to 1.83) and 1.46 (95% CI = 1.02 to 2.10) for these groups, respectively. Seventy-five percent of subdiaphragmatic cancers and 81% of thoracic cancers were seen in survivors of testicular seminoma, which was more frequently treated with radiotherapy than nonseminoma (85% versus 29%). Among those who were known to have received radiotherapy as part of the first course of treatment for testicular cancer, the rate ratio for all-cause mortality following second subdiaphragmatic and thoracic cancers combined compared with similar first cancers was higher following nonseminoma (RR = 2.47; 95% CI = 1.07 to 5.69) than seminoma (RR = 1.35; 95% CI = 1.06 to 1.73), although differences according to testicular cancer histology were not statistically significant (P = .17).
|
With regard to specific second cancer sites in the radiotherapy fields for testicular cancer diagnosed during 1973–1979, non–statistically significant increases in mortality were observed following second cancers of the colon/rectum/anus or the pancreas (Table 4). A statistically significant increase in cancer-specific mortality was noted for second kidney cancer, although it was based on only two deaths (RR = 5.8, 95% CI = 1.10 to 31.1). The rate ratios for mortality following a second lung cancer were also statistically significantly elevated compared with comparable first lung cancers (e.g., for all-cause mortality, RR = 1.65, 95% CI = 1.12 to 2.42); the difference in survival distributions for first and second lung cancers was of borderline statistical significance (P = .08). No elevations in mortality were observed following second cancers of the stomach, bladder, or prostate (data not shown).
Additional adjustment for initial treatment of second and matched first cancers did not substantially change the rate ratio for all-cause mortality for subdiaphragmatic sites combined or for specific subdiaphragmatic sites. However, adjustment for initial treatment resulted in a slight reduction in the rate ratio for all-cause mortality following lung cancer, from 1.65 (95% CI = 1.12 to 2.42) to 1.53 (95% CI = 1.01 to 2.33). Second lung cancers among testicular cancer survivors diagnosed during 1973–1979 were somewhat less frequently treated with radiotherapy than were first lung cancers (40% versus 57%) (P = .07).
|
Discussion |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
The current investigation represents the first large population-based study of testicular cancer survivors to examine mortality/survival after a second cancer diagnosis compared with first cancers of the same type. For all sites combined, cancer-specific and all-cause mortality following a second cancer were not statistically significantly greater than mortality for comparable first cancers. This was also true for most specific cancer sites and subgroups. However, we noted a pattern of increased cancer-specific and all-cause mortality for second cancers in the radiotherapy field for testicular cancer diagnosed during 1973–1979, specifically cancers of the lung, colon/rectum/anus, pancreas, and kidney. Cancer-specific and all-cause mortality for second cancers in the radiotherapy field following testicular cancer diagnosed in 1980 or later were comparable to those of first cancers.
Changes in the treatment of testicular cancer have occurred during the past 25 years (18). With the advent of platinum-based chemotherapy in the early 1970s, increasing numbers of nonseminoma patients were managed with retroperitoneal lymph node dissection and chemotherapy rather than with radiotherapy. Radiotherapy doses were generally higher for treatment of nonseminoma than seminoma (4). Whereas subdiaphragmatic radiotherapy has persisted as the standard treatment for seminoma patients, doses and field sizes have been reduced (18,22–25), with a commensurate lowering of doses received by various organs (4). Prophylactic irradiation of the mediastinum in seminoma patients has declined over time (22). These changes in treatment of testicular cancer may explain the pattern of increased cancer-specific and all-cause mortality for selected sites in the radiotherapy field following testicular cancer diagnosed during 1973–1979 but not later. In support of this hypothesis, among those who were known to have received radiation treatment for testicular cancer, we found some evidence of higher mortality among second cancers in the radiotherapy field following nonseminoma than seminoma testicular cancer, consistent with the higher doses of radiation received by patients with nonseminomas.
An analysis of cause-specific or all-cause mortality that is not adjusted for type of initial treatment for the second cancer (or matched first cancer) more fully reflects the impact of the initial testicular cancer diagnosis than an analysis adjusted for initial treatment of the second and first cancer because the management of the initial testicular cancer may have limited the range of possible treatment options for the second cancer. Nevertheless, we also conducted analyses adjusted for the type of treatment of the second cancer (and matched first cancers) to determine whether increased mortality could be explained by differences in initial treatment for the second cancers. We found little evidence that concern over complications from radiation fibrosis (26) following testicular cancer treatment discouraged surgical treatment of second cancers, but there was some indication that radiotherapy treatment for second thoracic cancers was less frequent than for matched first cancers among those whose testicular cancer was diagnosed during 1973–1979. This situation may reflect the reluctance of radiation oncologists to reirradiate tissue that had previously received high doses of radiation (6). These differences in initial course of treatment accounted for only a small part of the elevation in mortality following second lung cancers in this group. Differences in subsequent treatment, the use of lower than optimal radiation doses due to concern regarding normal tissue tolerance to reirradiation, or second tumors that are either biologically more aggressive or less responsive to treatment could also contribute to the observed excess mortality following second cancers in the radiotherapy field for testicular cancer diagnosed during this early time period. For example, widespread genomic instability has been reported in lung tumors after radiotherapy for Hodgkin lymphoma (27).
Previous studies in testicular cancer patients addressed mortality due to second cancers compared with expected rates in the general population, but not mortality from comparable first cancers (22,28–30). Other studies of survival after a second cancer diagnosis compared with first cancers were limited to a specific second cancer following a wide variety of first cancers or following a specific first cancer other than testicular cancer (31–34). A previous study concluded that acute myeloid leukemia that occurs after chemotherapy or radiotherapy for a wide variety of first cancers is either biologically more aggressive or less responsive to treatment than first primary acute myeloid leukemia (31). Although we did not identify a statistically significant increase in mortality from acute myeloid leukemia following testicular cancer compared with first cancers, the numbers were small, precluding definitive conclusions. Similar to the overall lung cancer results, a previous study found that operable second non–small-cell lung cancers had a prognosis similar to comparable first lung cancers (32).
The study had several potential limitations, including the small number of second cancers for each specific cancer site. Thus, for most specific cancer sites, we were unable to rule out adverse survival effects for second cancers relative to comparable first cancers. The lack of complete and precise treatment information limited our ability to determine whether differences in mortality for selected second cancers compared with first cancers were due to differences in dose of radiation, type and dose of chemotherapy, extent of surgery, adjuvant treatment modalities, or treatment beyond the initial course of therapy. The SEER program does not collect information about lifestyle risk factors that may be associated with cancer incidence and prognosis, e.g., tobacco, alcohol, diet, and obesity. Thus, although we assumed that these factors and more subtle differences in tumor characteristics than are evident in the broad staging classification or morphologic classification used by SEER were distributed similarly between the second and first cancer cohorts, we cannot be sure that this assumption is correct. It is possible that within the broad staging scheme that we used for matching, the second cancers had on average more favorable prognostic profiles than the first cancers due to increased surveillance in testicular cancer survivors, resulting in some underestimation of the rate ratios. In addition, subsequent cancers are not recorded in the SEER database for those who move from their original SEER reporting area, a factor most likely to be important for metropolitan areas, for the longest follow-up observation periods, and for young adults (35). However, the results are not likely to be materially affected by this last point because we found no major differences in mortality for second cancers compared with first cancers according to age at or time since testicular cancer diagnosis. It is also unlikely that the development of more (or less) aggressive second malignancies (and thus, survival after such a diagnosis) would be related to the propensity of a testicular cancer survivor to migrate from SEER areas. Lack of precision in cause-of-death coding may have impaired our ability to link cause of death to a certain cancer, but this happened for only a small portion of patients and was not differential between second and first cancers. Finally, there were very few statistically significant findings, even though our reported P values were not adjusted for multiple comparisons. Except in cases for which we had prior hypotheses, such as that radiation treatment for testicular cancer could adversely affect survival from second cancers, we regarded the P values less as formal hypothesis tests than as descriptive statistics that could point to interesting findings.
Our study also had a number of strengths. All patients were identified from population-based registries, which use standardized methods to collect data for first and second cancers. We were also able to identify a relatively large number of second cancers after testicular cancer and match them to comparable first cancers.
In summary, cancer-specific and all-cause mortality following a second cancer diagnosis in testicular cancer patients were not statistically significantly greater than mortality in comparable first cancers for all cancer sites combined and for most specific cancer sites. We did note a pattern of increased cancer-specific and all-cause mortality for selected sites in the radiotherapy field for testicular cancer diagnosed during 1973–1979, suggesting that treatment of testicular cancer by radiotherapy during that era may have negatively impacted survival from subsequent cancers. Survival for second cancers following testicular cancer that was diagnosed during 1980–2002 was similar to that of comparable first cancers, suggesting that treatment regimens for testicular cancer used since 1980 have not adversely affected survival from subsequently diagnosed cancers.
|
Funding |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
Intramural Research Program of the National Cancer Institute, Division of Cancer Epidemiology and Genetics.
|
NOTES |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
The authors are solely responsible for the study design, data collection and analysis, interpretation of the data, and the preparation of the manuscript.
We are indebted to Jeremy Miller at Information Management Services, Rockville, MD, for expert computer support and data management.
|
REFERENCES |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
(1) Einhorn LH. Testicular cancer as a model for a curable neoplasm: The Richard and Hinda Rosenthal Foundation Award Lecture. Cancer Res (1981) 4:3274–80.
(2) Coleman MP, Gatta G, Verdecchia A, Esteve J, Sant M, Storm H, et al. EUROCARE-3 summary: cancer survival in Europe at the end of the 20th century. Ann Oncol (2003) 14(Suppl 5):v128–49.[CrossRef][Medline]
(3) Ries LAG, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L, et al. SEER cancer statistics review. 1975–2000. (2003) Bethesda (MD): National Cancer Institute. Available at: http://seer.cancer.gov. [Last accessed: October 23, 2006.].
(4) Travis LB, Fosså SD, Schonfeld SJ, McMaster ML, Lynch CF, Storm H, et al. Second cancers among 40,576 testicular cancer patients: focus on long-term survivors. J Natl Cancer Inst (2005) 97:1354–65.
(5) Travis LB, Andersson M, Gospodarowicz M, van Leeuwen FE, Bergfeldt K, Lynch CF, et al. Treatment-associated leukemia following testicular cancer. J Natl Cancer Inst (2000) 92:1165–71.
(6) Nieder C, Milas L, Ang KK. Tissue tolerance to reirradiation. Semin Radiat Oncol (2000) 10:200–9.[CrossRef][Web of Science][Medline]
(7) Robinson E, Neugut AI. Clinical aspects of multiple primary neoplasms. Cancer Detect Prev (1989) 13:287–92.[Web of Science][Medline]
(8) Fajardo LF. The pathology of ionizing radiation as defined by morphologic patterns. Acta Oncol (2005) 44:13–22.[Web of Science][Medline]
(9) Tibbs MK. Wound healing following radiation therapy: a review. Radiother Oncol (1997) 42:99–106.[CrossRef][Web of Science][Medline]
(10) Allan JM, Travis LB. Mechanisms of therapy-related carcinogenesis. Nat Rev Cancer (2005) 4:943–55.[CrossRef]
(11) Claij N, te Riele H. Microsatellite instability in human cancer: a prognostic marker for chemotherapy? Exp Cell Res (1999) 246:1–10.[CrossRef][Web of Science][Medline]
(12) World Health Organization. International Classification of Diseases for Oncology. (2000) 3rd ed. Geneva (Switzerland): World Health Organization.
(13) Surveillance, Epidemiology, and End Results (SEER) Program. SEER*Stat Database: Incidence—SEER 18 Regs excluding AK/AZ, Nov 2004 Sub for White/Black/Other (1973–2002 varying)—Linked to County Attributes—Total U.S. 1969–2002 Counties. In: National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released December 2004, based on the November 2004 submission. Available at: http://seer.cancer.gov. [Last accessed: October 23, 2006.].
(14) Seiffert JE, ed. SEER program: comparative staging guide for cancer (1993) Bethesda (MD): National Institutes of Health (NIH publication no. 93-3640).
(15) World Health Organization. Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death. Vol 1. 8th Rev (1967) Geneva (Switzerland): World Health Organization.
(16) World Health Organization. Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death. Vol 1. 9th Rev (1977) Geneva (Switzerland): World Health Organization.
(17) World Health Organization. International Statistical Classification of Diseases and Related Health Problems. Vol 1. 10th rev (1992) Geneva (Switzerland): World Health Organization.
(18) Thomas G. Over 20 years of progress in radiation oncology: seminoma. Semin Radiat Oncol (1997) 7:135–45.[CrossRef][Web of Science][Medline]
(19) Cox DR, Oakes D. Analysis of survival data. (1984) London (U.K.): Chapman and Hall.
(20) SAS Institute Inc. SAS Online Doc, version 8. (1999) Cary (NC): SAS Institute Inc.
(21) Govindan R, Page N, Morgensztern D, Read W, Tierney R, Vlahiotis A, et al. Changing epidemiology of small-cell lung cancer in the United States over the last 30 years: analysis of the Surveillance, Epidemiologic, and End Results database. J Clin Oncol (2006) 24:4539–44.
(22) Hanks GE, Peters T, Owen J. Seminoma of the testis: long-term beneficial and deleterious results of radiation. Int J Radiat Oncol Biol Phys (1992) 24:913–9.[Web of Science][Medline]
(23) Fosså SD, Horwich A, Russell JM, Roberts JT, Cullen MH, Hodson NJ, et al. Optimal planning target volume for stage I testicular seminoma: a Medical Research Council randomized trial. J Clin Oncol (1999) 17:1146–54.
(24) Jones W, Fosså S, Mead GM, Roberts TJ, Sokal M, Horwich A, et al. A randomized trial of two radiotherapy schedules in the adjuvant treatment of stage I seminoma (MRC TE18): preliminary report. In: Germ cell Tumours V—Harnden P, Joffe JK, Jones WG, eds. (2002) London (U.K.): Springer. 235–6.
(25) Jones WG, Fosså SD, Mead GM, Roberts TJ, Sokal M, Horwich A, et al. A randomized trial of 30 Gy vs 20 Gy in the adjuvant treatment of stage I testicular seminoma (MRC TE18, EORTC 30942, ISRCTN18525328). J Clin Oncol (2005) 23:1200–8.
(26) Dormand EL, Banwell PE, Goodacre TE. Radiotherapy and wound healing. Int Wound J (2005) 2:112–27.[CrossRef][Medline]
(27) Behrens C, Travis LB, Wistuba II, Davis S, Maitra A, Clarke EA, et al. Molecular changes in second primary lung and breast cancers after therapy for Hodgkin's disease. Cancer Epidemiol Biomarkers Prev (2000) 9:1027–35.
(28) Horwich A, Bell J. Mortality and cancer incidence following radiotherapy for seminoma of the testis. Radiother Oncol (1994) 30:193–8.[CrossRef][Web of Science][Medline]
(29) Fosså SD, Aass N, Harvei S, Tretli S. Increased mortality rates in young and middle-aged patients with malignant germ cell tumours. Br J Cancer (2004) 90:607–12.[CrossRef][Web of Science][Medline]
(30) Zagars GK, Ballo MT, Lee AK, Strom SS. Mortality after cure of testicular seminoma. J Clin Oncol (2004) 22:640–7.
(31) Neugut AI, Robinson E, Nieves J, Murray T, Tsai W-Y. Poor survival of treatment-related acute nonlymphocytic leukemia. JAMA (1990) 264:1006–8.
(32) Koppe MJ, Zoetmulder FAN, van Zandwijk N, Hart AAM, Baas P, EJTh Rutgers. The prognostic significance of a previous malignancy in operable non-small cell lung cancer. Lung Cancer (2001) 32:47–53.[CrossRef][Web of Science][Medline]
(33) Bhattacharyya N, Nayak VK. Survival outcomes for second primary head and neck cancer: a matched analysis. Otolaryngol Head Neck Surg (2005) 132:63–8.[CrossRef][Web of Science][Medline]
(34) Robinson E, Neugut AI, Murray T, Rennert G. A comparison of the clinical characteristics of first and second primary head and neck cancers. Cancer (1991) 68:189–92.[CrossRef][Web of Science][Medline]
(35) Curtis RE, Freedman DM, Ron E, Ries LAG, Hacker DG, Edwards BK, et al, eds. New malignancies among cancer survivors: SEER cancer registries, 1973–2000 (NIH publication no. 05-5302). (2006) Bethesda (MD): National Cancer Institute.
Manuscript received November 16, 2006; revised May 25, 2007; accepted June 26, 2007.
Related Article in JNCI
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2007 99: 1213.
This article has been cited by other articles:
|
|
 |
|
  X. Garcia-del-Muro, P. Maroto, J. Guma, J. Sastre, M. Lopez Brea, J. A. Arranz, N. Lainez, D. S. de Prado, J. Aparicio, J. M. Piulats, et al. Chemotherapy As an Alternative to Radiotherapy in the Treatment of Stage IIA and IIB Testicular Seminoma: A Spanish Germ Cell Cancer Group Study J. Clin. Oncol., November 20, 2008; 26(33): 5416 - 5421. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. R. Feldman, G. J. Bosl, J. Sheinfeld, and R. J. Motzer Medical Treatment of Advanced Testicular Cancer JAMA, February 13, 2008; 299(6): 672 - 684. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. D. Fossa Second Cancers among Survivors of Testicular Cancer: The Emerging Picture ASCO Educational Book, January 1, 2008; 2008(1): 423 - 426. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Outcomes of Second Cancers in Men with Histories of Testicular Cancer Journal Watch Oncology and Hematology, September 4, 2007; 2007(904): 1 - 1. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||