Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
RADIOGENIC CANCER AT SPECIFIC SITES 242 5 Radiogenic Cancer at Specific Sites LEUKEMIA The induction of leukemia by ionizing radiation has been well documented in humans and laboratory animals. The types of leukemia induced and their rates of induction vary markedly, depending on the species, strain, age at irradiation, sex, and physiological state of the exposed individuals. They also depend on the dose, dose rate, anatomical distribution, and LET of the radiation, among other variables. The early literature has been summarized elsewhere (NRC80, UN77, UN82, UN86, UN88). Human Data The most extensive human data on the dose-incidence relationship come from studies of the Japanese atomic-bomb survivors and patients treated with x rays for ankylosing spondylitis. In the atomic-bomb survivors of the Life Span Study Cohort, a total of 202 deaths from leukemia were recorded for the period from 1950 to 1985, during which there were an estimated 2,185,335 person- years of follow-up. Analyzed in terms of absorbed dose to the bone marrow as estimated with the new DS86 dosimetry, the dose response for Nagasaki rises less steeply than for Hiroshima, especially in the dose range below 0.5 Gy, but the difference between the two cities is smaller with the DS86 dosimetry than with the T65D dosimetry and is no longer significant (Sh87). For the combined data, the rate of mortality is significantly elevated at 0.4 Gy and above but not at lesser doses. At bone marrow doses of 3-4 Gy, the estimated dose-response curve peaks and
RADIOGENIC CANCER AT SPECIFIC SITES 243 turns downward (Figure 5-1). As noted below, this pattern is characteristic of the leukemia response in other irradiated populations. The saturation of the leukemia response at high doses has been attributed to the reduced survival of potentially transformed myeloblasts in the range above 3-4 Gy (Un86). FIGURE 5-1 Cumulative leukemia mortality in Hiroshima and Nagasaki as a function of the estimated dose equivalent to the bone marrow under DS86. By 1985, there were 51 cases in the 0 Sv category and 31 cases in the 0.01-0.1 Sv stratum. Based on a simple linear dose-response model, which in the opinion of RERF analysts fit the LSS data for leukemia mortality as well as a linear- quadratic model and better than a simple quadratic model, the excess relative risk per Sievert was estimated to range from 4.24 to 5.21, and the number of excess deaths per 104 person-year-Sv was estimated to range from 2.40 for a neutron RBE of 20 to 2.95 for an RBE of 1 (Sh87). The excess mortality from leukemia reached a peak within 10 years after irradiation and has persisted at a diminished level (Figure 5-2). No excess cases of chronic lymphocytic leukemia have been observed (Pr87a). Among 14,106 patients who were followed for up to 48 years after a single course of x-ray therapy for ankylosing spondylitis, 39 deaths from leukemia were recorded versus a total of 12.29 expected cases (ratio of observed to expected deaths, 3.17) (Da87). The excess deaths became detectable within two years after irradiation, reached a peak within the first 5 years, and declined thereafter; however, the excess death rate remained significantly elevated (relative risk, 1.87) for more than 15 years, after which it appeared to persist with little change (Da87). The relative risk did not vary significantly with age at the time of treatment, but it was higher in males (3.43) than in females (1.79). The relative risk also varied with the hematologic type of the disease, being higher for those with acute myeloid
RADIOGENIC CANCER AT SPECIFIC SITES 244 FIGURE 5-2 Relative risk of mortality from leukemia and all cancers other than leukemia in A-bomb survivors, 1950-1982, in relation to time after irradiation. The number of deaths in each interval of follow-up and 99% confidence intervals are indicated (Pr87).
RADIOGENIC CANCER AT SPECIFIC SITES 245 leukemia than for those with other types of leukemia. It was not elevated for those with chronic lymphatic leukemia (Table 5-1). TABLE 5-1 Observed, as Compared with Expected, Numbers of Deaths from Leukemia in Persons Treated with Spinal Irradiation for Ankylosing Spondylitisa Type of Leukemia Number of Deathsb Expected Ratio of Observed/ Observed Expected Myeloid leukemia Acute 17 4.34 3.92 Chronic 3 2.05 1.46 Unspecified 4 0.71 5.63 All types 24 7.10 3.38 Lymphatic leukemia Acute 2 0.93 2.15 Chronic 2 2.38 0.84 Unspecified 3 0.38 7.89 All types 7 3.69 1.89 Unspecified leukemia 3 0.28 10.71 All types 36 11.29 3.19 a From Darby et al. (Da87). b Observed and expected deaths from leukemia occurring more than one year after first treatment at ages less than 85 years by age at first treatment and by type of leukemia as recorded on the death certificate. Retreated patients were included for 12 months following treatment. Analyzed in relation to the average dose to the bone marrow, which was estimated to be 3.21 Gy, the excess relative risk amounted to 0.98/Gy, or 0.45 additional cases of leukemia per 104 PYGy (Sm82). The smaller magnitude of the risk per Gy in patients with ankylosing spondylitis, compared with that in atomic-bomb survivors, may be ascribable to the younger average age of atomic-bomb survivors at the time of exposure and to the fact that they received instantaneous whole-body irradiation, whereas in the patients with ankylosing spondylitis only a portion of the active marrow was irradiated and the dose was received in fractionated exposures that usually totaled more than 5 Gy within a given treatment field (Le88). Muirhead and Darby have proposed different models of leukemia risk for the spondylitics and the A-bomb survivors. They proposed a relative risk model for the spondylitics and an absolute risk model for the atomic-bomb survivors (Mu87). In an international case-control study of 30,000 women treated with fractionated doses of radiation for carcinoma of the uterine cervix, the risk was estimated to be increased by about 70%/Gy, corresponding to an excess of 0.48 cases of leukemia/104 PYGy (Bo87, Bo88). As in the
RADIOGENIC CANCER AT SPECIFIC SITES 246 A-bomb survivors mentioned previously, the excess cases were confined to leukemias other than those of the chronic lymphatic type. The relative risk was maximal within the first 5 years after irradiation, was larger in women who were irradiated when they were under age 45 than in those who were irradiated when they were over age 45, and reached a peak at an average bone marrow dose of 2.5-5.0 Gy, above which it decreased (Figure 5-3). The data conformed to a linear-exponential model in which the total risk equaled the sum of incremental risks to individually irradiated masses of marrow. The latter risks, in turn, were taken to increase linearly with the mass exposed and inversely with the total mass of marrow in the body; they were also taken to increase curvilinearly in a manner consistent with the dose-dependent killing of marrow cells (Bo87). In view of the decreased risk per Gy at high doses, it is not surprising that the average risk per Gy in the women of this series was appreciably lower than that which has been observed in women treated with smaller doses of x rays for benign gynecologic disorders (Bo86). FIGURE 5-3 Relative risk of acute leukemia and chronic myeloid leukemia in women treated with radiation for carcinoma of the uterine cervix, as influenced by the average dose to the bone marrow. A better fit was obtained with a linear exponential model (Î£W) which considered the weighted dose to each marrow component as opposed to the average dose over all compartments (d) (Bo87). The incidence of leukemia has been observed to be elevated similarly in patients treated with radiation for cancers of other sites (Bo84, Cu84, Wa84). An association between previous diagnostic irradiation and adult
RADIOGENIC CANCER AT SPECIFIC SITES 247 onset myeloid or monocytic leukemia has been suggested by three case-control studies (St62, Gu64, Gi72); however, the data in the first and largest of the three studies (St62) have since been reinterpreted to argue against a causal relationship on the grounds that ''the 'extra' examinations all happened within 5 years of the onset" of symptoms of leukemia (St73). No association between previous diagnostic irradiation and adult-onset myeloid or monocytic leukemia was observed in a fourth case-control study (Li80). On the basis of extrapolation from the leukemogenic effects of irradiation in atomic-bomb survivors and other relatively heavily irradiated groups, it has been estimated that about 1% of all leukemia cases in the general population may be attributable to diagnostic radiography (Ev86). The risk has not been confined to acutely irradiated populations, such as those mentioned above. Early cohorts of radiologists in the United States (Le63, Ma84), the United Kingdom (Co58), and the People's Republic of China (Wa88), who were exposed to x rays occupationally in the days preceding modern safety standards, also have shown an increased incidence of acute leukemia and chronic granulocytic leukemia. These diseases have, likewise, been observed to occur with increased frequency in patients previously injected with radium-224 or Thorotrast (NRC80). Because of uncertainty about the doses to the bone marrow in the occupationally and internally irradiated populations, it is not clear how their risks per unit dose compare with those in the more acutely irradiated populations described above. An excess number of cases of leukemia have been observed in children who were exposed to diagnostic x-irradiation in utero; the excess is larger per unit dose than that in children who were irradiated during postnatal life. The magnitude of the excess and the extent to which it may signify an unusually high susceptibility of the embryo and fetus are discussed in Chapter 6 of this report. Reports of an increased incidence of leukemia in children residing in the vicinity of nuclear installations in the United Kingdom are reviewed in Chapter 7. Committee Analysis For purposes of risk estimation, the Committee's analysis was restricted to the total mortality from leukemias of all hematologic types combined, excluding chronic lymphocytic leukemia. Modeling in terms of the various types of leukemia was not possible because of limitations in the available data. The different types vary markedly in the age distributions of their occurrence in the general population and in their relative frequencies with time after irradiation, depending on age at the time of exposure. To this extent, the Committee's risk model for leukemia is a gross simplification. For both the Life Span Study (LSS) and the Ankylosing Spondylitis
RADIOGENIC CANCER AT SPECIFIC SITES 248 (ASS) data, essentially comparable fits could be obtained using either additive or relative risk models, although somewhat different modifying effects were required in the two models and the relative risk model was consistently more parsimonious. It must be remembered that follow-up of the LSS cohort did not begin until five years after exposure, by which time the peak in the excess rate had already occurred in the ASS data. Despite this and other differences between the two studies, the modifying effects are reasonably consistent. The preferred model from the ASS data is a relative risk model with a decreasing effect in time after exposure. However, the addition of an effect of age at exposure significantly improves the fit of the LSS data. The magnitude of this effect and also the effect of time after exposure depends on whether exposure occurred before or after age 20. The ASS cohort did not include individuals younger than 20 years of age at exposure, so the age factor could not be tested in that data set. Dose-response in the LSS data was significantly improved by the addition of a quadratic term in dose. (Here, the linear term includes both the gamma and neutron components, the latter weighted by the assumed RBE of 20; the quadratic component includes only the gamma component.) The "cross-over dose" (the dose at which the linear and quadratic contributions are equal) was estimated to be about 0.9 Gy. However, ratios of log likelihood estimates are biased and for these data the uncertainty is very large (see Annex 4F). Similarly, the "dose rate effectiveness factor" (DREF, the ratio of the fitted slopes of the pure linear and the linear-quadratic models) is estimated as 2 but again with a very large uncertainty. The final preferred model for leukemia mortality used in the risk projections is given by equation 4-3 reproduced below. This model is plotted as a function of attained age in Figure 5-4 and excess risk as a function of time after exposure for males is shown in Figure 5-5. The abrupt changes in risk with age at the time of irradiation that are specified in the model reflect simplifying compromises in model fitting and are not based on hypotheses concerning the biological mechanism of age-dependent changes in susceptibility. Insofar as different types of leukemia vary in age distribution in the general population, their causative mechanisms and temporal distributions in irradiated populations might be expected to vary as well. This leukemia model is based on LSS data, which do not include information prior to five years post exposure. A number of fitted models
RADIOGENIC CANCER AT SPECIFIC SITES 249 were tested but these produced rather varied and unreliable risk estimates in extrapolations to this early, first 5-year period. Sources of data, other than that from A-bomb survivors, provide some guidance on this point. The cervical cancer study by Boice et al. (Bo87) indicates that excess leukemia cases were observed only within the first five years post exposure. On the other hand, the spondylitic cohort shows a mixture of excess cases before and after five years post exposure (Da87). In that study, 14 cases with 1.6 expected were observed in the first five years, and 25 cases with 10.7 expected after five years post exposure. One could then reasonably argue that nearly one-half of the excess leukemias would be observed within the first five years after exposure. The Committee chose to model the 2- to 5-year post-exposure period by extrapolating to two years the excess relative risk observed for the 5- to 10-year post-exposure period. This method resulted in an approximately 15% increase in the lifetime risks. The Committee's extrapolation procedure for the 2- to 5- year post-exposure period may lead to an underestimate of the actual risk, and this should be kept in mind when interpreting the Committee's risk estimates for leukemia. FIGURE 5-4 The relative risk of leukemia due to low LET radiation for both sexes by attained age from age 7 to age 75 for exposure at various ages.
RADIOGENIC CANCER AT SPECIFIC SITES 250 FIGURE 5-5 Excess leukemia deaths by time after exposure to low-LET radiation for U.S. males at various ages of exposure. Leukemia Studies in Animals In mice, rats, dogs, swine, and other laboratory animals, a variety of lymphoid and myeloid leukemias have been induced by irradiation (UN77, UN86, NRC80). In such animals, the dose-incidence relationship has been observed to vary from one type of leukemia to another, but in no instance does it conform to a simple, linear nonthreshold function. The most extensively studied of the experimental leukemias are T-cell neoplasms that arise in the mouse thymus. The induction of these growths is inhibited drastically by shielding a portion of the hemopoietic marrow (UN77) and may involve the activation of a latent leukemia virus (Rad LV) (Yo86). The dose-incidence curve for the disease is of the threshold type in mice of certain strains (UN86). In the range of 0.5-1.0 Gy, the RBE of fast neutrons for induction of these neoplasms has been observed to range from a value of 1.0-2.0 with single or fractionated exposures to a value exceeding 10 with continuous, duration-of-life irradiation (UN77, UN86 Fe87).
RADIOGENIC CANCER AT SPECIFIC SITES 251 Less thoroughly investigated are experimentally induced myeloid leukemias, which have been observed in mice (Up70, Ma78, Hu87), dogs (Fr73), and swine (Ho70) that were subjected to various regimens of external or internal irradiation. The dose-incidence curve for myeloid leukemia in mice rises with increasing dose of acute whole-body x or gamma radiation, passes through a maximum at 2-3 Gy of x or gamma rays (lower dose of neutrons), and decreases at higher doses (Figure 5-6); in the dose range below 1 Gy, the shape of the curve appears to vary among strains (UN86, U187). The downturn in the dose-incidence curve at doses above 2-3 Gy is consistent with the reduction in numbers of potentially transformed myelopoietic cells surviving such doses (Gr65, Ba78, Ro78, Ma78, UN86). In the low to intermediate dose range, the curve rises more steeply with fast neutrons than with x rays or gamma rays (Up70, Mo82, U187, Pr87a), and on fractionation or protraction, the incidence per Gy decreases markedly with x or gamma irradiation but decreases less markedly, if at all, with fast neutron irradiation (Figure 5-6). As a result, the neutron RBE increases with decreasing dose rate, from a value of 2-3 at dose rates exceeding 0.1 Gy/minute to a value as high as 16 at dose rates of less than 0.01 Gy/minute (Up70). Various models have been fitted to the observed dose- incidence data, all of which have included cell-killing terms to account for the diminution of the response at intermediate to high dose levels (UN86). Although the data do not exclude a linear dose term in the low to intermediate dose range, all models also include higher power dose terms to account for the fact that the incidence per Gy of low-LET radiation increases with increasing dose at high dose rates in the intermediate dose range but is substantially reduced at low dose rates (UN86). The induction of myeloid leukemia, in contrast to induction of thymic lymphoma, is not inhibited disproportionately by shielding part of the hemopoietic system (Up64). The incidence of myeloid leukemia per Gy has been observed to be increased in mice in which granulocyte turnover is accelerated by injection of turpentine and decreased in mice in which granulocyte turnover is reduced by the elimination of microflora, implying that induction of the disease is promoted by proliferation of granulocyte precursors (Up64). Susceptibility to the induction of lymphoid and myeloid leukemias also varies among mice of different strains and in relation to age at the time of irradiation (UN77). There is no evidence, however, that susceptibility in mice is unusually high during prenatal life; on the contrary, the data imply that it may be substantially reduced at that time of life (Up66, Si81, UN86). Whereas the incidence of lymphoid and myeloid leukemias is typically increased by whole-body irradiation in most strains of mice, depending on the conditions of irradiation, the incidence of reticulum cell
RADIOGENIC CANCER AT SPECIFIC SITES 252 neoplasms in such animals has usually been observed to decrease with increasing dose (UN77, UN86). FIGURE 5-6 Lifetime incidence of myeloid leukemia (in excess of control incidence) in male mice of different strains, in relation to dose and dose rate of whole body neutron-, x-, or g-irradiation. RFM mice (U187): acute neutron irradiation (curve 1); acute Î³-irradiation (curve 2); CBA mice (Mo82, Mo83a, Mo83b). acute neutron irradiation (curve 3); acute x-irradiation (curve 4); protracted Î³- irradiation (curve 5). RF/Up mice (Up70): acute neutron irradiation (curve 6); protracted neutron irradiation (curve 7); acute x-irradiation (curve 8); protracted Î³-irradiation (curve 9). Summary The risks of acute leukemia and of chronic myeloid leukemia are increased by irradiation of hemopoietic cells, the magnitude of the increase depending on the dose of radiation, its distribution in time and space, and the age and sex of the exposed individuals, among other variables. The mean latent period preceding the clinical onset of the leukemia also varies, depending on the hematologic type of the disease as well as age at the time of irradiation. The data do not suffice to define the dose-incidence relationship precisely, but the dose-response curve for the total excess cases of leukemia appears to increase in slope with increasing mean dose to the marrow, to pass through a maximum in the dose range of 3-4 Gy, and to decrease with a further increase in the dose.
RADIOGENIC CANCER AT SPECIFIC SITES 253 Age at exposure is an important modifier of risk. From the LSS data it is clear that risks are initially higher for those exposed at under 20 years of age but decrease somewhat more rapidly with time after exposure than for those exposed at older ages. There was no clear indication that the risks for those under 10 years of age were significantly greater than for persons 10-20 years old at the time of exposure. When data become available that will allow the analysis of human leukemia in terms of specific hematologic types, it may be possible to develop more precise risk models that capture the age and time modifying factors in more detail. BREAST Introduction The sensitivity of the mammary gland to the carcinogenic effects of ionizing radiation was first demonstrated in x-irradiated mice in 1936 (Fu36a, Fu36b). and has since been described in other species of laboratory animal, including guinea pigs, dogs, and rats (Sh86a). An increase in the incidence of breast cancer in irradiated humans was first recognized in 1965 in women who had received repeated fluoroscopic examinations (Ma65), and subsequently in Japanese atomic-bomb survivors in 1968 (Wa68). During recent decades, mammary cancer has been studied extensively in irradiated animals and in several large series of irradiated women. Although a number of questions about radiation-induced breast cancer still remain, the data are consistent with the following generalizations: 1. The development of overt cancer from the radiogenically damaged mammary target cells is critically dependent upon the hormonal status of the cells over time. 2. Radiation-related breast cancers are similar in age distribution and histopathological types to breast cancers resulting from other or unknown causes. 3. Women who are irradiated at less than 20 years of age are at a higher relative risk for breast cancer than those who are irradiated later in life. 4. The epidemiological data reveal little or no decrease in the yield of tumors when the total radiation dose is received in multiple exposures rather than in a single, brief exposure. Parallel Analyses of Breast Cancer Incidence and Mortality The Committee had available for analysis the original data from two mortality series and three incidence series. The mortality series were the
RADIOGENIC CANCER AT SPECIFIC SITES 254 Canadian Tuberculosis Fluoroscopy (CAN-TB) Study (Mi89) (473 deaths) and the subcohort of the Radiation Effects Research Foundation (RERF) Life Span Study (LSS) of atomic bomb survivors for which DS86 doses were available with follow-up through 1985 (151 deaths) (Sh87, Sh88). The incidence series included data on women in the LSS for whom DS86 doses were available with follow-up through 1980 (367 cases), data on women in the New York Acute Postpartum Mastitis Study (NY-APM) (118 cases) (Sh86b), and data on women in the Massachusetts Tuberculosis Fluoroscopy (MASS-TB) cohort (65 cases) (Hr89). In the Committee's analyses of breast cancer, the data from the first 5 years of follow-up have been omitted. As there were no cases of breast cancer in women less than 25 years of age, expression of risk in the 0-24 age group was excluded in the analysis. This made virtually no difference in the risk modeling. In the LSS data, breast dose equivalents were computed by using an assumed relative biological effectiveness (RBE) of 20 for neutrons. As discussed in Annex 4E, women who received doses in excess of 4 Gray (Gy) were excluded from both the incidence and mortality analyses of the LSS data. For the NY-APM and MASS-TB cohorts, women with doses in excess of 6.5 and 4 Gy, respectively, were excluded from the analyses. Second breast primaries were not included in the analyses. Since for the NY-APM series, the dose received by each breast could differ, the follow-up time was computed in terms of breast-years using the procedures described in Shore et al. (Sh86b). All results are presented in terms of person-years. Breast- years in the NY-APM series were converted to person-years by dividing by 2. For the LSS incidence data, the person-years were adjusted for the effects of migration by using the factors given by Tokunaga et al. (To87). The AMFIT computer program, described in Annex 4C, was used to fit various models of the radiation effects for each of the individual series, and separately for the combined mortality and combined incidence data sets. The patterns seen in the combined analyses were generally present in the individual series, and results are presented only for the combined analyses. Those studies which depart significantly from the results of the combined analyses are described in Annex 4E. The patterns of breast cancer mortality or incidence, in the absence of radiation exposure, were first modeled for each of the populations from which the cohorts were drawn. These background rates were then either multiplied by a function of dose, age at exposure, and time since exposure (relative risk model) or added to an appropriate function of these covariates (additive excess risk model). Details of the procedures used in modeling the background rate for the various cohorts are described in Annex 4E.
RADIOGENIC CANCER AT SPECIFIC SITES 255 The Committee's Preferred Model The Committee has investigated a number of models for lifetime excess risk of breast cancer incidence and mortality, and its preferred models are described here in general terms. More detailed information on the parameter estimates for the preferred models and on issues which arose as these models were developed is presented in Annex 4E. The Committee's preferred models for both incidence and mortality are relative risk models in which the excess relative risk is linear in dose and varies with both age-at-exposure and time-since exposure. A relative risk model was chosen for the incidence data because it was found that the A-bomb survivors and the U.S. relative risks for breast cancer did not differ significantly (p =0.3) while the additive excess risks among the A-bomb survivors were significantly lower than those in the NY-APM and MASS-TB cohorts (p = 0.0001). The choice between relative and absolute risk models for the mortality data was less clear-cut. Within the CAN-TB cohort the estimated risk per Gy for women treated in Nova Scotia was about six times that for women treated in other provinces. This difference is highly significant (p <0.001). Women treated in Nova Scotia faced the x-ray beam and thus received higher doses than other women in the CAN-TB cohort. However, the analyses described in Annex 4E indicate the higher risk observed among Nova Scotia women is not attributable to nonlinearities in the dose response. Since there is currently no explanation for the differences within the CAN-TB cohort and since the Committee was generally interested in low dose effects it was decided to use the data on the CAN-TB cohort without the Nova Scotia women as the basis for risk estimates in the parallel analysis. Although the relative risk estimate for mortality in the LSS was about three times the estimate in the non-Nova Scotia CAN-TB cohort the difference was not statistically significant (p = 0.1). The estimated absolute excess risks were about equal in the two cohorts. Since the relative risks for the two cohorts were not significantly different and because of the evidence against equal excess absolute risks in the incidence data, the Committee's preferred model for breast cancer mortality is a relative risk model in which the level of risk was determined from the combined LSS-non-Nova Scotia CAN-TB data. It should be noted that women in the LSS and NY-APM study received acute exposures whereas the women in both TB series received highly fractionated exposures, usually over several years. Despite this difference
RADIOGENIC CANCER AT SPECIFIC SITES 256 in the pattern of exposure, on a relative risk scale there are no significant differences among cohorts in risk of breast cancer incidence or mortality. For the incidence data, the variation in relative risk with age-at-exposure has been modeled as a step function with separate values for different age-at- exposure groups under 20, 20-40, and over 40 years of age. The relative risk for the under 20 age group was estimated as four times that in the 20-40 age group while the risk for those over 40 was only about 40% of that in the 20-40 age group. As discussed in Annex 4E, 15- to 19-year-old women in the NY-APM cohort had a significantly lower risk of radiation-induced breast cancer than young women in the other two cohorts (p = 0.01). Therefore the model for those less than 20 years of age included a separate parameter for the New York cohort. In the mortality analyses, the excess relative risk for the 10-14 year olds was found to be significantly greater than that for older women. For those under 10 years of age, the relative risk was somewhat smaller but not significantly so (p = 0.2) and the groups were combined. For the older women, the relative risks appeared to decrease with increasing age-at-exposure (p = 0.05). The Committee's final model allows for the increased risks seen among the young women, for the decreasing trend in risk with age-at-exposure in older women, and for a discontinuous drop in relative risk estimate at 15 years of age. The use of additional steps for other age groups did not significantly improve the fit of the model for mortality due to breast cancer. In the incidence data there is evidence that the excess relative risk varies with time (p =0.01). In particular it was found that, for women over the age of 25, it increases to a maximum value at about 15 to 20 years after exposure and decreases slowly thereafter. In the Committee's preferred model the change in the log relative risk with time-since-exposure was modeled as a linear spline in log time with a single knot at 15 years after exposure. There is some evidence (p = 0.12) of a similar pattern in the mortality data with the maximum relative risk occurring between 20 and 25 years after exposure. In the Committee's preferred model for the mortality data the log relative risk is modeled as a quadratic in log time. This model fit somewhat better than a linear spline. The Committee found no evidence in either the mortality or incidence data that the temporal pattern was affected by dose or age-at-exposure (though it should be borne in mind that tests for such effects lack power). For both incidence and mortality the models assume that there is no excess risk during the first five years after exposure and no excess risk occurs among women under the age of 25. In fact there is no evidence of a significant excess risk for at least 10 years post exposure. The Committee's preferred risk model for breast cancer mortality is given in equation 4-5, reproduced below.
RADIOGENIC CANCER AT SPECIFIC SITES 257 To illustrate the variation in excess risk with age-at-exposure and time in the preferred models, Figures 5-7 through 5-10 present risk estimates for specific ages-at-exposure by attained age and by time-since-exposure for breast cancer mortality. Figures 5-11 and 5-12 illustrate the Committee's model for breast cancer incidence by attained age in terms of relative risk and the estimated number of excess cases per 10,000 person year Gy. The excess absolute risks for incidence are based upon fitted background rates derived from the Connecticut Tumor registry, while the excess absolute mortality estimates are based upon fitted Canadian mortality rates. It was assumed that the exposure took place in 1980 and that temporal trends in the age-specific baseline rates do not occur after 1985. With the fitted models, excess absolute incidence rates increase until the women reach the age of 50, after which they decrease. The youngest women generally have the highest absolute risks. The general pattern of risk predicted by the mortality models is similar, with absolute risks increasing with time for women under age 50 and decreasing after age 50. In the Committee's final model, the relative risks at a given time after exposure are the same for all women under 20 years old at exposure. In fact, the data indicate that women who were 10-14 years old at exposure have higher relative risks than women who were older. However, risks of women who were less than 10 years old at exposure are poorly estimated. The data from this age group of atomic bomb survivors are as yet not adequate for precise characterization of risk. The contrast between the incidence and mortality predictions with regard to the excess risks for women under 10 years of age at exposure is not surprising in view of the limited data currently available for this group. Among the cohorts available, only the RERF cohort had an appreciable amount of information on risks in women who were very young when exposed. The youngest women in the RERF cohort are just now reaching the age at which one would expect an appreciable incidence of breast cancer. Projections based upon the Committee's models for the youngest age group should be interpreted with caution. Additional follow-up is clearly important in order to clarify our understanding of excess breast cancer risks in women exposed under the age of 10.
RADIOGENIC CANCER AT SPECIFIC SITES 258 FIGURE 5-7 The relative risk of female breast cancer mortality due to low-LET radiation by attained age for exposure at various ages. FIGURE 5-8 Excess breast cancer deaths due to low-LET radiation by attained age for U.S. females for exposure at various ages.
RADIOGENIC CANCER AT SPECIFIC SITES 259 FIGURE 5-9 The relative risk due to low-LET radiation of female breast mortality as a function of time after exposure for exposure at various ages. Risk is not projected for an attained age greater than 75 years. FIGURE 5-10 Excess breast cancer deaths as a function of time after exposure to low-LET radiation for U.S. females of various ages. No data are presented for an attained age greater than 75 years.
RADIOGENIC CANCER AT SPECIFIC SITES 260 FIGURE 5-11 Excess relative risk of breast cancer incidence due to low-LET radiation by attained age for U.S. females exposed at various ages. FIGURE 5-12 Excess cases of breast cancer by attained age for U.S. females exposed to low- LET radiation at various ages.
RADIOGENIC CANCER AT SPECIFIC SITES 261 Experimental Data on Cancer Latency and Dose Response Latency The prolonged persistence of radiogenically initiated cells, suggested by the long latency of cancer in women who were exposed when they were children, has been confirmed experimentally. When rats are subjected to small doses of x rays or fission-spectrum neutrons and gamma rays, which alone produce few or no mammary neoplasms, and are then grafted with pituitary tumors that secrete high levels of prolactin, mammary neoplasms shortly appear in a high incidence (Yo77, Yo78). The time between irradiation and elevation of prolactin levels can be extended from a few days to as long as 12 months with little change in lag period from increased prolactin to appearance of tumors or in final tumor incidence. Life-span studies of irradiated rats have illustrated that there is a marked inverse relationship between radiation dose and the latency of mammary neoplasms (Sh80, Sh82, Sh86a). This relationship may be related to the number of radiogenically initiated cells. The development of quantitative normal cell transplantation techniques has allowed the identification of a subpopulation of rat mammary cells that, when stimulated with appropriate hormones, can give rise to clonal multicellular glandular units (Cl85a). Following acute exposure, the acute post-irradiation survival and repair capacities of these mammary clonogens have been defined by transplantation assays (Cl85a). When grafts containing about 120 clonogens which had survived 7 Gy of 137Cs gamma irradiation were transplanted to rats with marked prolactin and glucocorticoid deficiencies, mammary cancers arose in approximately 50% of the graft sites, indicating that there was one neoplastically initiated cell per 240-300 grafted clonogens (Cl86a). Consideration of these and other experimental data in rats suggests that there is a shortening of cancer latency as well as an increase in cancer incidence as irradiated mammary clonogen numbers are increased. The data also are consistent with the conclusion that a considerable period of hormonal promotion/progression is necessary for the development of overt cancer from radiation-initiated mammary target cells. In the women in the studies analyzed by the Committee, the normally functioning endocrine system supplied sufficient hormone, and a radiation dose-related shortening of latency was not observed. In the relatively small groups of experimental rats, more intense normal stimulation was often necessary to reveal radiogenic initiation, and an abbreviation of latency with increased dose was found. Dose Response A number of investigators have suggested that the low-LET radiation dose- response relationship for mammary tumors in rats is linear and that
RADIOGENIC CANCER AT SPECIFIC SITES 262 there is little effect of dose fractionation or protraction (Sh66, Sh86a). This appears to be true in Sprague-Dawley rats over the dose range of 0.28 to 4.0 Gy when the experiments are terminated at 10-12 months (Bo60, Sh57, Sh86a). In life-span studies of Sprague-Dawley rats (Sh80) and rats of the ACI strain (Sh82), inspection of the final numbers of animals with neoplasms has suggested a linear dose response over the ranges of 0.28-0.85 Gy and 0.37-3.0 Gy of x rays, respectively. Because of the effect of dose on the time of appearance of mammary neoplasms, more complex analyses that included both tumor incidence and latency have been employed (Sh80, Sh82). The results obtained are somewhat difficult to relate to human data. The designs of rat experiments have differed from laboratory to laboratory, hormonal manipulations were often used, experimental groups were often small, and benign fibroadenomas were often grouped with adenocarcinomas. The promotional effects of hormones on the induction of fibroadenomas differ from those on the induction of carcinomas (Cl78, Sh66, Sh82). Ullrich's study of mammary carcinogenesis in otherwise untreated BALB/c mice exposed to 137Cs gamma rays (low LET) at different dose rates and different dose fractions revealed a linear-quadratic dose-response relationship over the dose range 0-0.25 Gy administered at 0.35 Gy/minute (Ul87b). The ratio of the linear to the quadratic dose term is very low, 2.3; that is, the dose at which the effects governed by the linear dose component equals those governed by the quadratic function is 0.023 Gy (Ul87b). When the exposures were delivered at a low dose rate of 0.083 Gy/day, the dose response followed the linear term. When a total dose of 0.25 Gy was delivered in daily fractions of 0.01 Gy at a high dose rate, the carcinoma response fell on the linear curve, but when the same total dose was delivered in 0.05-Gy fractions at a high dose rate, the carcinoma incidence fell near the linear-quadratic curve (Ul87b). Although these experimental results with mice are of considerable theoretical interest, their quantitative application to humans is problematic. As noted above, of the recent analyses of breast cancer mortality among irradiated women, only the Canadian fluoroscopy series (Mi89), with the Nova Scotia series included, presented any evidence of a positive quadratic component in the dose-response relationship. In that series, the ratio of the linear to the quadratic coefficient derived from all of the combined data is 205; and for those exposed to less than 6 Gy this ratio is 613 (Mi89). These ratios are 89and 266-times larger respectively than the ratios from Ullrich's mouse data. Furthermore, the role of the mouse mammary tumor virus or its genomic equivalent in radiogenic mammary neoplasia is not clear (Sh86a). Thus, the nature of the dose response of mammary cancer to low-LET radiation deserves continuing investigation with respect to its
RADIOGENIC CANCER AT SPECIFIC SITES 263 underlying mechanisms, particularly at low doses, low dose rates, and small fractions. Neutrons and Mammary Cancer The results of the reevaluation of the atomic-bomb and radiation doses received by individuals in Hiroshima and Nagasaki (Ro87) have precluded the likelihood that the LSS data will yield useful information on the relative carcinogenic effect of neutrons (Sh87, Sh88). Experimental studies have, however, shown that per unit dose, neutrons have a significantly higher mammary neoplasm-inducing potential than low-LET radiations, and that this greater relative neoplastic potential is increased at small doses (Sh86a). An RBE of 20-60 was calculated for fission-spectrum neutrons from life-time mammary neoplasm incidences in Sprague-Dawley rats irradiated with 0.05-2.5 Gy (Vo72). In this study, 0.05 Gy of fission-spectrum neutrons (average energy about 1 million electron volts [MeV]) was as effective in terms of the final yield of mammary tumors as were higher doses; that is, the tumor yield appeared to reach a plateau at 0.05 Gy. In contrast, in an experiment with the same rat line exposed to 14-MeV neutrons, mammary tumor incidence and the number of tumors per rat at 11 months after exposure was a near linear function of dose over the range 0.025-0.4 Gy (Mo77). The 14-MeV neutrons were about half as effective as reported for 0.43-MeV neutrons (Mo77). The RBE for mammary tumor induction by neutrons with different energies rank as follows: 2.0 MeV (fission spectrum) > 14 MeV > 0.025 MeV (thermal) (Ka85). Compared with x rays, the RBE of fission neutrons was Ë18. In two of the most complete experimental life-span studies involving Sprague-Dawley and ACI rats, the latter with and without estrogen (diethylstilbestrol) supplementation (Sh80, Sh82), the analyses involved effects on both latency and incidence. In the estrogen-supplemented ACI rats, a significant increase in early tumor incidence was seen after they received 0.01 Gy of 0.43-MeV neutrons (Sh82). At low doses, the RBE increased in inverse proportion to the square root of the neutron dose and exceeded 100 at 0.01 Gy. Unfortunately, the proper interpretation of many of the rat experiments is difficult because benign fibroadenomas were combined with carcinomas. These neoplasms differ fundamentally (Br85). For example, irradiated and unirradiated, but otherwise untreated ACI rats develop both mammary fibroadenomas and carcinomas. When either irradiated or unirradiated rats are given estrogen, the mammary tumors are virtually all adenocarcinomas (Sh82). The effect of total dose and dose rate of fission-spectrum neutrons on mammary cancer incidence has been investigated in BALB/c mice (Ul84).
RADIOGENIC CANCER AT SPECIFIC SITES 264 Lifetime mammary cancer incidence after exposure at a rate of 0.05-0.25 Gy/minute increased linearly with dose to a total of 0.1 Gy. At greater doses, tumor incidence increased less markedly over the range of 0-0.5 Gy, the dose- cancer response relationship was best fit by a model in which effect increased as the square root of dose (Ul84). When the neutron dose rate was reduced to 0.1 Gy/day or less, the mammary cancer incidence was approximately twice that at higher dose rates at total doses of up to 0.05 Gy. At higher total doses, cancer incidence plateaued, being very similar at 0.4 and 0.1 Gy (Ul84). When mammary cells from neutron-irradiated mouse glands were transplanted into gland-free fat pads, they gave rise to ductal dysplasias, that is, precancerous lesions (Ul86). Most such lesions derived from mice exposed to 0.025- or 0.2- Gy neutrons at 0.01 Gy/minute regressed by 16 weeks after grafting. In contrast, most such lesions derived from mice exposed to the same total doses at 0.01 Gy/day persisted, suggesting the promotion of initiated cells during the long exposure (Ul86). Finally, the decreasing effectiveness per unit dose of higher neutron doses given either at a high or a low dose rate may be related to the high sensitivity of the mouse ovaries to radiation damage. In summary, the experimental data show that the neoplastic effect of neutrons on mammary tissue is higher than previously considered, particularly at low radiation doses (â¤ 0.01 Gy). Furthermore, in contrast to both human and experimental data with low-LET radiations, exposure to neutrons at low dose rates may be more damaging than exposures at high dose rates. The fine structure of the dose response for the production of mammary tumors at low dose and dose rate requires further mechanistic analysis. Hormones and Breast Cancer The growth, development, and function of the normal mammary gland is dependent upon hormonal regulation (Cl79, He88, Ro79, Ru82, Sh86a). The spectrum of hormones involved includes the steroids estrogen, progesterone, and glucocorticoid; the peptides prolactin, growth hormone, and placental lactogen; and perhaps other hypophyseal factors (Cl78, Kl87, Ru82). The actions of these hormones at a given time depend on the past hormonal exposure of the mammary tissue as well as the concurrent titers of other hormones. Hence, a given hormone may potentiate breast neoplasia under one set of circumstances and suppress it under others. Many breast carcinomas in women retain responsiveness to hormonal therapy (Cl77, Sh71). Although hormones are important to the promotion and progression of initiated mammary cells, experimental studies have unequivocally shown that radiogenic initiation is a scopal effect directly on mammary target cells (Cl86a, Sh71).
RADIOGENIC CANCER AT SPECIFIC SITES 265 Gonadal estrogens play a dual role in mammary growth. They act directly as mitogenic agents on cells in the mammary gland and indirectly to induce the secretion of prolactin by the anterior pituitary gland. In women, estrogens are the primary mitogenic hormones; prolactin may facilitate the mitogenic action of estrogens and promotes differentiation and function (He88). In rats, prolactin is the primary mammary mitogen and agent for the promotion and progression of cancers (Cl79). Estrogens also induce progesterone receptors in mammary cells; progesterone, in turn, acts in synergy with estrogen and prolactin to control mammary growth and differentiation (Ro79, Ru82). Finally, glucocorticoids are essential for milk secretion. The most efficient hormonal combination for promotion/progression of radiation-initiated rat mammary cells is a combination of elevated prolactin coupled with a glucocorticoid deficiency (Cl85c). Given the profound role that is played by hormones in mammary cells, it is likely that most of the conditions that have been shown to enhance or suppress human breast cancer risk are mediated through effects on hormone levels and, in turn, on the number and condition of the mammary target cells. Most significant among these conditions are the ages at menarche and menopause. In addition, at first full-term pregnancy, the lactational history, and body weight are significant. Circulating estrogens, progestins, and prolactin increase in women at menarche and decrease in the perimenopausal period (Pi83). Both an early age at menarche and a late age at menopause predispose an individual to breast cancer (Ho83, Ma73). Both of these conditions lengthen the total period of time during which the breast is subjected to the mitogenic stimuli of gonadal steroids. For example, the relative risk of breast cancer increased nearly linearly (p < 0.004) to 2.2 in women in Shanghai who entered menarche at ages â¤12 years compared with women who reached menarche at ages â¥18 years (Yu88). Surgically induced menopause (ovariectomy) before age 35 is strongly protective, but breast cancer risk is detectably reduced by ovariectomy at as late as 45-50 years of age (Ma73). In irradiated rats, even in the presence of high levels of prolactin plus glucocorticoid deficiency, ovariectomy reduces mammary cancer risk (Cl85c). Breast cancer is most markedly reduced by the occurrence of a full-term pregnancy at an early age. Women of 15-16 years of age who carry a child to full term have 35-40% the risk of breast cancer of nulliparous women and Ë30% the risk of women who first give birth when they are over 30 years of age (Ma73). When corrected for age of first full-term pregnancy and lactation duration, multiparity was found to decrease risks further; for example, the relative breast cancer risk of women in Shanghai with five children was 0.39 compared with those with just one child (Yu88).
RADIOGENIC CANCER AT SPECIFIC SITES 266 Previous lactation decreases the breast cancer risk, particularly premenopausal disease (By85). In a study of Caucasian women in the United States, the risk of premenopausal breast cancer among those who had ever nursed a child compared with that among those who had not nursed a child was 0.49 (Mc86a). Among women in Shanghai of average age (Ë50.5 years), those who had lactated for a total duration of â¥9 years had a breast cancer risk of 0.37 compared with those who had lactated for â¤3 years (Yu88). The effects of diet, and hence of body weight and body fat, are presumably likely hormonally mediated and appear to account to a large extent for differences in the geographic distribution of breast cancer rates. The incidence of breast cancer is five-to sixfold greater in North America and northern Europe than it is in Asia and Africa (Ma73). This effect is most likely related to life- style, and especially to the effect of diet on body weight and body fat. Recent experimental evidence suggests that total caloric intake is a more important risk factor for breast cancer than is the fat concentrations of the diet (He88). Among women in Shanghai, the breast cancer risk of those who weighed â¥60 kg was 2.4 times that of women who weighed â¤45 kg (Yu88). Second-and third- generation American offspring of Oriental immigrants have increased breast cancer risks (Ho83); Hawaiians of Chinese ancestry have a risk pattern similar to that of Hawaiian Caucasians (Ma73). Age at menarche in Japanese women is inversely related to body weight, and age at menopause is directly related to body weight (Ho83). In Japan, age at menarche decreased 6 months per decade during the period 1900 to 1945; age at menopause increased 1 year during the same period. There was a marked upswing in the age at menarche among Japanese women born between 1930 and 1940; this birth cohort reached the age of puberty during the severe food shortage during and after World War II (Ho83). Thus, the incidence of radiogenic breast cancer in individuals in the various age cohorts of the Hiroshima-Nagasaki Life Span Study are being measured against a shifting background of breast cancer risk from other causes. Fatty tissues contain aromatizing enzymes that convert adrenal androgens into estrogens. This leads to continued hormonal stimulation of the mammary glands after menopause, and likely accounts in part for the greater risk among the generally heavier postmenopausal women of North America and northern Europe than in the lighter Oriental population (Ho83, Ma73). These findings are consistent with the conclusion that radiogenic initiation and the expression of such radiogenic damage in the formation of overt breast cancer is dependent on the number of mammary target cells and their degree of differentiation at the time of exposure and on subsequent promotion and progression of the initiated cells under hormonal
RADIOGENIC CANCER AT SPECIFIC SITES 267 control. Those conditions that induce functional mammary differentiation, and hence, that reduce target cell numbers (Ru82, Cl86a), for example, early and multiple pregnancies and lactation, reduce the risk of breast cancer. Those conditions that reduce or block full functional differentiation and increase mitogenesis, for example, nulliparity and glucocorticoid deficiency, increase the risk. Summary 1. Animal experiments and human studies indicate that the induction of breast cancer is hormonally mediated and that hormonal status plays a critical role in radiation carcinogenesis of the breast. 2. A strong linear component is seen in the dose-response relationship for radiation-induced cancer. 3. There is little evidence of reduction in risk associated with dose fractionation in the human cohorts considered, even though these cohorts included both fractionated and acute exposures. 4. There is no evidence in humans of the occurrence of radiation- induced cancers until after the age of 25, which is about the youngest age at which breast cancer is seen in the general population. 5. Age at exposure strongly influences susceptibility, with risk being highest among women under 20 years of age at the time of exposure, suggesting that the time of puberty corresponds to a period of elevated risk. The level of risk among those under age 10 at the time of exposure is uncertain. There is little evidence of any excess risk in women over age 40 at the time of exposure. 6. The low relative risk for women in the NY-APM study exposed between the ages of 15 and 17 suggests that the occurrence of a full- term pregnancy before age 20 may reduce susceptibility to radiation-induced breast cancer. 7. There is no evidence that radiogenic breast cancers appear during the first 10 years following exposure, but after this time the number of such cancers appears to increase rapidly. On the relative-risk scale, the data suggest that the incidence peaks at 15 to 20 years after exposure and the mortality about 5 years later. Observations to date indicate that the absolute risk continues to increase until 50 but may decrease at older ages. 8. Although animal data suggest that there is a relationship between dose and latency, the human data show no such relationship. LUNG There are three main sources of epidemiologic data on the induction of human lung cancer by radiation; (1) the survivors of atomic-bomb
RADIOGENIC CANCER AT SPECIFIC SITES 268 explosions in Hiroshima and Nagasaki, (2) patients who were treated with x rays for ankylosing spondylitis, and (3) uranium miners and other underground miners exposed chronically to high-LET (alpha) radiation from inhaled 222Rn and its progeny. Each of these populations has received detailed long-term follow-up to ascertain the health risks associated with these diverse types of exposure. Japanese Atomic-Bomb Survivors Since the publication of the BEIR III report (NRC80), there have been three follow-up reports on the Life Span Study (LSS) of survivors of the atomic bomb explosions in Hiroshima and Nagasaki (Ka82, Pr87a, Sh88) in whom lung cancer has been a prominent late effect. These reports provide analyses of the excess cases of lung cancer in the period 1950-1985, or a maximum of 40 years after the detonations. In the 1950-1978 LSS report (Ka82), the absolute risk of lung cancer showed a significant increase during the period from 1950-1974, and this increase continued during the interval from 1974 to 1978. Overall, the absolute risk of lung cancer occurring between 1950 and 1978, based on T65D dosimetry, was 0.61 lung cancer deaths/104 person-year Gy (PYGy) (90% confidence interval, 0.37-0.86). When the relative risk and 90% confidence interval for lung cancer were compared with similar computations for other radiation-related cancers (viz., breast, stomach, and colon), the relative risks for these four sites fell within the same confidence intervals, suggesting that relative risks may not differ by target organ. In the 1950-1982 LSS report (Pr87a), which used the T65 dosimetry, mortality from cancers of the trachea, bronchus and lung was significantly associated with the radiation dose (p < 0.001). The estimated relative risk at 100 rad (1 Gy) was 1.33 (90% confidence interval, 1.19-1.50) and the average excess risk was 0.82 cancers per 104 PYGy (90% confidence interval, 0.48-1.19). When the relative risk at 100 rad (1 Gy) for cancer of the trachea, bronchus, and lung was examined in 4-year intervals to 1982, the values remained approximately constant from 1955 to 1982. Darby et al. (Da85) conducted a parallel analysis of the cancer mortality seen in patients with ankylosing spondylitis and Japanese atomic-bomb survivors who received a T65 dose of at least 100 rad (1 Gy); statistically significant excess deaths from lung cancer were observed when the results of these two studies were compared. In the ankylosing spondylitis series, using lung cancer deaths that occurred more than 3 years after treatment, the relative risk was 1.41 (90% confidence interval, 1.20-1.65); for the Japanese atomic- bomb survivors who received more than 100 rad (1 Gy) compared with the group who received <9 rad (0.09 Gy), the relative risk was 1.94 (90% confidence interval, 1.53-2.45).
RADIOGENIC CANCER AT SPECIFIC SITES 269 In the most recent LSS report (Sh88), cancer mortality was analyzed for the period 1950-1985 as a function of the revised DS86 doses in a DS86 subcohort of approximately 76,000 individuals. In general, the number of excess deaths from all cancers other than leukemia was observed to increase proportional to the background cancer rate for attained age, with the result that the relative risks tended to remain constant for specific age cohorts, except for those who were 0-9 years old at the time of bombing. Lung cancer deaths deviated somewhat from this pattern because of a slightly decreasing trend with time. Using the DS86 organ dose values, the relative risk of lung cancer mortality at 100 rad (1 Gy) was 1.63 (90% confidence interval, 1.35-1.97); the absolute risk was 1.68 excess lung cancer deaths/106 PYR (90% confidence interval, 0.97-2.49). Patients with Ankylosing Spondylitis in England and Wales Results of earlier studies on over 14,000 patients treated with x rays for ankylosing spondylitis in England and Wales from 1935 to 1955 were summarized in the BEIR III report (NRC80). The BEIR III Committee estimated the average dose to the bronchus in such patients to be approximately 197 rad (1.97 Gy) on the assumption that 80% of the bronchial epithelium was irradiated. Because no dose values for individuals were available, no dose- response models were tested at that time. Smith and Doll (Sm82) extended the follow-up to January 1, 1970. To avoid uncertainties caused by multiple radiation exposures, their analysis was restricted to patients receiving a single course of treatment. There were 133,874 person-years at risk available for analysis, 83% of which were from male subjects. Cancers arising in the heavily irradiated sites became prominent beginning about 9 years after the first treatment and continued at an elevated level up to 20 years or more after treatment. Lung cancer was the most frequent type of cancer observed in the heavily irradiated sites, accounting for 37 of the 92 excess cancers (ratio of observed to expected cancer, 1.42; p < 0.001). The follow-up was subsequently extended to January 1, 1983, by Darby and colleagues (Da87) who observed 224 lung cancer deaths occurring at times >5 years after the first treatment, while 184.5 lung cancer deaths were expected (ratio of observed to expected deaths, 1.21; p < 0.01) during the interval from 5 years to 24.9 years after the first exposure. The ratio of observed to expected deaths was 0.97 for lung cancers occurring after 25 years. No indication was given of the contribution of smoking to the risk of lung cancer in this population. Dosimetric estimates have since been published by Lewis et al. (Le88), who used Monte Carlo techniques to calculate the average doses received
RADIOGENIC CANCER AT SPECIFIC SITES 270 by various organs and tissues, but individual dose estimates are still unavailable. For the lungs, the average dose was reported to be 1.79 Gy, while for the main bronchi, the average dose was 6.77 Gy. Because dose estimates for individuals are not available for this cohort, it was not possible to analyze the follow-up results in terms of various dose-response models. Cervical Cancer Patients In a multicenter international study of 182,040 women treated for cancer of the uterine cervix, the relative risk of lung cancer was observed to be increased after radiation therapy (Bo85). On the basis of preliminary dosimetry, the lung was estimated to have received an average dose of 35 rad (range, 10 to 60 rad) and the observed relative risk was 3.7 (0.001 < p < 0.01). When the relative risk was examined as a function of time after irradiation, the pattern suggested an influence of misclassified metastases and confounding by cigarette smoking. Underground Miners Detailed epidemiologic studies on radiation-induced lung cancer are being conducted on underground miners of uranium and other minerals who are exposed chronically to alpha radiation from inhaled 222Rn, 220Rn, and their radioactive progeny. The exposure of the miner cohorts differs from the exposures of the Japanese atomic-bomb survivors and patients with ankylosing spondylitis in being chronic, low-dose-rate irradiation from internally deposited alpha emitters, as opposed to single, acute, or short-term fractionated, high-dose- rate, low-LET external irradiation. Because of the complexity of the dosimetry of radon in the respiratory tract, it has been customary to measure and record miner exposures in terms of working levels (WL), and cumulative exposures in working level months (WLM). It is difficult, therefore, to compare the results of the studies directly unless an appropriate value for the absorbed dose, per WLM is estimated. This involves a number of dosimetric uncertainties, and reported values of dose (rad) to the bronchial epithelium per WLM have differed by a factor of 10 or more (NCRP84). A number of analyses of cohorts of underground miners concerning radiation and lung cancer have been conducted to examine dose-response relationships and various dose-effect modifying factors (UN77, ICRP87, NCRP84, NRC88). The BEIR IV Committee (NRC88) examined lung cancer risks associated with four principal underground mining populations: two Canadian uranium miner cohorts at Eldorado, Beaverlodge, Saskatchewan, and Ontario; Swedish iron miners at Malmberget; and Colorado Plateau uranium miners. The data base analyzed by the BEIR
RADIOGENIC CANCER AT SPECIFIC SITES 271 IV Committee contained a total of 360 lung cancer deaths and 425, 614 person- years at risk. TABLE 5-2 Comparison of Estimates of Lifetime Risk of Lung Cancer Mortality due to a Lifetime Exposure to Radon Progenya Study Excess Lifetime Lung Cancer Mortality (deaths/106 person WLM) BEIR IV 1988 350 ICRP 1987 170-230b 360c NCRP 1984 130 BEIR III 1980 730 UNSCEAR 1977 200-450 a Adapted from NAS88 and ICRP87. b Relative risk with ICRP population. c Relative risk with 1980 U.S. population as in BEIR IV. The analyses were based on a descriptive analytical model of the pooled data in which the excess relative risk eventually decreases with time after exposure and also depends on age at risk. This is in contrast to the analyses of the Japanese atomic-bomb survivors, in which a definite decline in relative risk has not yet been observed. When the BEIR IV model was used to project lifetime risks of lung cancer from lifetime exposures to radon progeny, an overall value of 350 excess lung cancer deaths/106 person-WLM was obtained. Annex 4G provides an analytical description of the risk model developed by the BEIR IV Committee and summary tables of their risk estimates for lifetime exposure to radon progeny. Additional information regarding the data and methods of these analyses are described in detail in the BEIR IV report (NRC88). Table 5-2 compares the risk values derived by the BEIR IV Committee with values derived by United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) (UN77), the BEIR III Committee (NRC80), the National Council on Radiation Protection and Measurements (NCRP) (NCRP84), and the International Commission on Radiological Protection (ICRP87). Considering that the risk estimates in Table 5-2 are based, for the most part, on essentially the same epidemiological studies of underground miners, the range of estimates is fairly broad. This is largely due to the difference in the models used to project lifetime risks. For example, the BEIR IV Committee used the time and age dependent relative risk model outlined in Annex 4G, while the ICRP used a simple relative risk projection. In contrast, the NCRP used an absolute risk model with exponentially
RADIOGENIC CANCER AT SPECIFIC SITES 272 decreasing risks with time after exposure. The BEIR III Committee used an absolute risk model that is constant over time. One reason the BEIR IV Committee's model is more elaborate then the others is that original data from miners studies were made available to this Committee while the other studies had to rely on published summaries which contain little information on how risks change with increasing follow-up time as the population at risk ages. Effect of Smoking Several attempts have been made to study the influence of cigarette smoking on the carcinogenic effects of irradiation in the Japanese atomic-bomb survivors. Prentice et al. (Pr83) assembled several subsets of known smokers into a study population of 40,498 subjects. The T65DR doses were used along with questionnaire results on smoking habits to analyze 281 lung cancer deaths in this cohort. Using a Cox proportional hazards model and stratifying on city, sex, age at the time of the bombing, and survey date, the relative risk of lung cancer in nonsmokers rose from 1.0 to 1.1 to 2.3 for exposure doses of <10, 10-100, and > 100 rad (<0.1, 0.1-1.0, and > 1.0 Gy), respectively, whereas the corresponding relative risks for cigarette smokers were 2.4, 2.4, and 3.6. Neither a multiplicative nor an additive interaction for lung cancer mortality could be distinguished clearly. Kopecky et al. (Ko86) examined the combined effects of irradiation and cigarette smoking in a cohort of 29,332 Japanese atomic-bomb survivors among whom there were 351 lung cancer deaths. An additive excess-risk model fit the data without either superadditivity or subadditivity; no corresponding test of a multiplicative model was presented. The BEIR IV Committee (NRC88) also examined the question of the possible combined effects of cigarette smoking and exposure to radiation, reviewing the available information for underground miners as well as the A- bomb survivors and analyzing three populations in detail. The data sets used were from case-control studies of New Mexico uranium miners, a cohort study of Colorado uranium miners with follow-up through 1982, and Japanese atomic- bomb survivors. In discussing the results of these analyses, the BEIR IV Committee noted that analyses of this type normally included only some measure of radiation exposure and duration or intensity of cigarette use. Other factors that need to be considered in later analyses include age at first exposure, dose rate, sex, diet, and genetic predisposition. The BEIR IV Committee noted that although a multiplicative model for the interaction between exposure to radon progeny and cigarette smoking appears to have received the greater support in the literature, a submultiplicative model may provide a more accurate description of the underlying
RADIOGENIC CANCER AT SPECIFIC SITES 273 relationship. The Committee's analysis of the Japanese atomic-bomb survivor data indicated that neither an additive nor a multiplicative model could be rejected on statistical grounds (NRC88), a finding consistent with the earlier observations of Prentice et al. (Pr83) and Blot et al. (Bl84). The present Committee's analysis of lung cancer mortality relies heavily on death certificate information from the Life Span Study (LSS) of Japanese A- bomb survivors whose deaths were classified as due to cancers of the respiratory system, International Classification of Disease (ICD) codes 160-163. Use of this broad classification minimizes the loss of case information due to any lack of precision in the death certificates. Deaths due to lung cancer occur relatively late in life and thus in a rather narrow age range. The LSS data are limited in this regard in that the reliability of Japanese death certificates apparently diminishes rapidly for persons over 75 years of age. Therefore, deaths occurring after that age were not used in the Committee's analyses. Moreover, those exposed as children are still too young to provide reliable information on lung cancer mortality. The Committee's preferred risk model for respiratory cancer mortality is given in Equation 4-4 which is reproduced below. The Committee's analysis of respiratory cancer in A-bomb survivors showed little effect of age at exposure but did show a decrease with time after exposure (Figure 5-13). Therefore the relative risk also decreases with attained age as shown in Figure 5-14. The exponential coefficient for decreasing risk with time after exposure in the Committee's model is relatively large, â1.44, but has a rather broad standard error, Â±0.91. The change in deviance when time after exposure is included in the preferred model is modest, 1.75. Even so, the similarity of diminishing risk with time after exposure observed in the ankylosing spondylitic study reinforces the Committee's view that the modifying effect of time should be included in the preferred model. Because respiratory cancer is mainly a disease of old age, most of the excess mortality projected by the Committee's relative risk model occurs among those exposed rather late in life (Figure 5-15). The Committee also modeled additive excess risks for respiratory cancer. Under this model, mortality increased more strongly with attained age, age at exposure, and time after exposure for lung cancer than for other cancers, but was virtually identical for males and females. The additive model involved risks that increased as the square of the time after exposure (starting 10 years after) and as the 2.7 power of age at exposure. The additive and relative-risk models gave virtually indistinguishable fits to
RADIOGENIC CANCER AT SPECIFIC SITES 274 FIGURE 5-13 Relative risk of lung cancer mortality in males as a function of time after exposure to low-LET radiation. FIGURE 5-14 Relative risk of lung cancer in U.S. males by attained age at various ages due to low-LET radiation exposure.
RADIOGENIC CANCER AT SPECIFIC SITES 275 the observed mortality data, but the relative-risk model was preferred on the grounds of its simplicity and consistency with the Committee's treatment of cancer risks at other sites. FIGURE 5-15 Excess lung cancer mortality due to low-LET radiation for U.S. males by attained age for various ages of exposure. Studies in Laboratory Animals The studies of populations described above are invaluable in providing information on the response of the lungs to certain kinds of exposure to ionizing radiation. However, except for radon progeny, there are no human data available on lung cancer due to internally deposited radionuclides. Therefore a number of life-span studies have been, and are being, conducted in various species of laboratory animals to supplement and extend the human data. Recent major summaries of research on the carcinogenic response of the respiratory tracts of laboratory animals to ionizing radiation include those by Kennedy and Little (Ke78), ICRP (ICRP80), Bair (Ba86), Thompson and Mahaffey (Th86), and the BEIR IV Committee (NRC88). These summaries provide a wealth of information on the influence of various factors on the temporal and spatial characteristics of the dose received
RADIOGENIC CANCER AT SPECIFIC SITES 276 by tissues of the respiratory tract from inhaled radionuclides and the resulting biological effects, particularly at long times after inhalation exposure. A few of these results are described below. Various factors that influence the lifetime risks of inhaled beta-emitting radionuclides are being examined in a series of life-span studies in beagle dogs exposed once, briefly, by inhalation, to different beta-emitters in relatively soluble or insoluble forms (Mc86). The effect of dose protraction is being studied by using radionuclides with radioactive half-lives ranging from about 3 days to 29 years, encapsulated in fused aluminosilicate particles. Recent results from these studies have demonstrated that when the delivery of a dose of beta radiation to the lung is protracted from days to years, the pulmonary carcinogenic response is reduced by a factor of about 3 (Ha83a, Gr87). The induction of lung tumors in mice from external x-irradiation has been compared with that from neutron irradiation by Ullrich and Storer (Ul79). The mice were sacrificed 9 months after they received 100-900 rad of x-irradiation or 5 to 150 rad of neutron irradiation localized to the thorax. The relationship between the number of lung tumors (adenomas) per mouse and the x-ray dose could be described adequately by a linear-quadratic model with a shallow initial slope or by a threshold model with a dose-squared response above the threshold. In contrast, the tumorigenic response of the lung to neutron irradiation could be described by a linear or threshold model, with the linear response being above the threshold. The relative biological effectiveness of the neutron irradiation increased with decreasing neutron dose: from 25 at 25 rad to 40 at 10 rad. Lafuma et al. (1989) reported on the effectiveness of various radiation exposure modalities (radon-daughter inhalation, fission-neutron irradiation, or gamma irradiation) for inducing lung carcinomas in Sprague-Dawley rats. The observed equivalence ratio for radon daughter to neutrons was approximately 15 WLM to 10 mGy neutrons, and the ratio of neutron effectiveness to gamma rays from 60Co was 50 or more at a gamma dose of 1 Gy (La89). Coggle et al. (Co85) examined the tumorigenic effectiveness of uniform versus nonuniform external x-irradiation of the mouse thorax. The nonuniform irradiation was produced by 72 1-mm microbeams that irradiated about 20% of the total lung volume. Although a smaller study by these investigators had previously suggested that nonuniform x-irradiation might be more tumorigenic than uniform x-irradiation, the larger study demonstrated a nearly equal tumorigenic response of the lung to uniform and nonuniform x-irradiation. A study of the lifetime relative biological effectiveness of chronic beta irradiation of the lung versus chronic alpha irradiation of the lung for
RADIOGENIC CANCER AT SPECIFIC SITES 277 the production of pulmonary carcinomas, primarily bronchioloalveolar carcinomas, was reported recently by Boecker et al. (Bo88b). Beagle dogs were exposed once, briefly, by inhalation, to the beta emitter 91Y or to three different aerodynamic particle sizes of the alpha-emitter 239PuO2. A proportional hazards model was used to estimate the relative risk coefficients for these various radiation exposure modalities. Comparison of the linear risk coefficients among the four studies indicated that all three exposure regimens with 239PuO2 were more effective in producing lung cancer than was 91Y. The ratios of the relative risk coefficients for 239Pu/91Y ranged from 10 to 18 for different sizes of 239PuO , with the more uniform irradiation being more effective in producing 2 cancer. Other studies on the lifetime biological effects of inhaled radon progeny, plutonium, and other actinide radionuclides were discussed in detail in the BEIR IV report (NRC88). Studies on the effects of inhaled radon progeny inhalation in laboratory animals, primarily rats and dogs, are being conducted in the United States and France. Although earlier studies focused primarily on acute effects, more recent studies have been directed to lung cancers resulting from chronic inhalation exposures. The use of laboratory animals has made it possible to study the impact of various modifying factors on the resulting lung cancer incidence, thereby broadening the knowledge obtained from human epidemiology studies. Factors discussed in this regard in the BEIR IV report include the effects of cumulative exposure, exposure rate, unattached fraction of radon progeny, disequilibrium of radon progeny, and concomitant exposure to cigarette smoke. Major studies on the long-term effects of inhaled 238PuO2 or 239PuO 2 in beagle dogs are continuing at two laboratories, the Battelle Pacific Northwest Laboratories, Richland, WA (Pa86), and the Lovelace Inhalation Toxicology Research Institute, Albuquerque, NM (Mc86). The lack of any human dose- response data for internally deposited plutonium radionuclides makes these studies particularly valuable for estimating the potential human health risks associated with such internal depositions. For instance, these studies provide firm evidence for the risk to the lung, skeleton, and liver following inhalation of 238PuO and allows comparison with the risk to the lung following inhalation of 2 239PuO . Differences in risk are caused by differences in tissue distribution and 2 retention of these two plutonium radionuclides. Studies of the pulmonary tumorigenic responses of rats to inhaled low levels of 239PuO2, by Sanders et al. (Sa88) indicate that the observed dose- response function is best fitted by a quadratic function, with a ''practical" threshold of about 100 rad. These studies and those in progress by Lundgren et al. (Lu87a, Lu87b) will provide an important basis for comparing the relative effectiveness of low doses of brief external x-irradiation
RADIOGENIC CANCER AT SPECIFIC SITES 278 with those of chronic beta or chronic alpha irradiation. The Bayesian analysis performed in the BEIR IV report (NRC88) to estimate the human bone cancer risk from internally deposited 239Pu, comparing available data for alpha- emitting radionuclides in humans and in laboratory animals, illustrates one way in which such data are useful. Summary Absolute radiogenic risks of radiation-induced lung cancer are similar for both sexes although baseline lung cancer risks are much higher for males than they are for females. Consequently, the excess relative risk for irradiated females is higher than for males. Because the relative risk attributable to radiation appears to be essentially constant with respect to age at the time of exposure the relative risk model is preferred for risk projection purposes. However, there is no clear evidence as to whether relative risks or additive risks are more consistent between populations. The constancy of additive risks between the sexes would support the use of an additive model for the comparing of populations. Data suggesting greater than additive interactions between radiation and smoking are equivocal and are largely restricted to the effects of chronic exposure to radon progeny, which are of uncertain relevance to the effects of brief exposure to low-LET irradiation. Studies in laboratory animals have provided much of what is known about the influence of various modifying factors on the long-term biological effects of chronic beta or chronic alpha irradiation on the respiratory tract. They have shown that prolongation of beta irradiation of the lung from a period of days to years reduces its tumorigenic effectiveness by a factor of about 3, and that chronic alpha irradiation of the lung from inhaled 239PuO2 is 10 to 20 times more carcinogenic than is chronic beta irradiation. STOMACH The best-known and, perhaps, strongest evidence of a relationship between ionizing radiation and cancer of the human stomach comes from the follow-up of studies of the Japanese atomic-bomb survivors. The most recent such analysis, using the new dosimetry on the combined cohort from Hiroshima and Nagasaki, shows that there is a highly significant radiation-related relative risk of mortality from stomach cancer; e.g., the average excess relative risk at 1 Gy is 0.23 in terms of the kerma at a survivor's location. Females have higher absolute and relative risks than males, although neither difference is statistically significant (Sh87). Overall, stomach cancer is one of the most common types of cancer seen in this cohort, and its average excess risk of 2.09 deaths/104 PYGy is, with the
RADIOGENIC CANCER AT SPECIFIC SITES 279 single exception of leukemia, the largest excess observed among specific cancer sites. As noted above, however, the relative risk is not large, since this is a commonly occurring tumor in the general population. A similar association between radiation exposure and stomach cancer was observed in a recent follow-up study of patients irradiated for peptic ulcer at the University of Chicago (Gr84). The original study involved 1,457 patients treated with radiotherapy over a 1- to 2-week period between 1937 and 1955 and 763 nonirradiated patients with ulcers who served as controls (Cl74). In the majority of the radiation-treated patients, the stomach received an estimated organ dose of 16-17 Gy. Patient follow-up was continued through 1962, but no significant increase in tumors was noted. Twenty-two years later Griem and colleagues (Gr84) expanded the study population to 2,049 cases by including all patients with peptic ulcers treated with radiation through the mid-1960s. They then used the hospital identification numbers of the study population members to search the University of Chicago Tumor Registry and update their tumor histories. Their preliminary evaluation of the registry data found a radiation- related relative risk for stomach cancer of 3.7, with a corresponding estimated excess risk of 5.5 Ã 10â4 stomach cancers per person Gy, based on a life-table analysis. The findings from the Japanese atomic-bomb survivors and the University of Chicago ulcer patients appear, at first, to be in marked contrast to the results of the investigation of second primary tumors among more than 82,000 women who underwent treatment for cervical cancer (Bo85). In the radiation-treated patient series of this study, an excess of 60 stomach cancers was predicted on the basis of risk estimates for radiation-induced stomach cancers prevalent in the literature at that time. However, only three excess stomach cancers were actually seen. Furthermore, the relative risk did not seem to vary to any extent either with age or time elapsed since the initiation of radiotherapy. The authors speculated that the failure to duplicate the effects observed among Japanese atomic-bomb survivors may have been due to differences in the age or sex composition of the study populations, even though the estimated average organ dose of 2 Gy for the patients with cervical cancer was somewhat higher than doses which produced significant effects in the Japanese atomic-bomb survivors. However, a further follow-up of this study population (Bo88b) suggests that an association between radiation exposure and stomach cancer is present in the patients with cervical cancer, i.e., when a subset of the original study population consisting of women with second primary cancers and their matched controls were selected for further study, it was found that an exposure of several Gy led to a significant (p < 0.05) relative risk of 2.1 (Bo88b). The issue becomes complicated when one takes into account the data
RADIOGENIC CANCER AT SPECIFIC SITES 280 on patients with ankylosing spondylitis given a single treatment course with x rays. Previous analyses of the mortality data from the Court-Brown and Doll (Co65) cohort of over 14,000 patients with ankylosing spondylitis who had been treated with x rays indicated the presence of an elevated risk of stomach cancer among those patients for whom a sufficient period of time had elapsed since treatment. For example, Land indicated that 31 deaths from stomach cancer were observed, while only 20.1 were expected in those patients 9 or more years after their first exposure (La86). However, a more recent analysis of this data base (Da87) shows that, overall, 64 deaths from stomach cancer have been observed, compared with 63.2 expected. The authors reported relative risks of 1.01 (9 observed versus 8.88 expected deaths), 1.2 (44 observed versus 36.5 expected deaths), and 0.62 (11 observed versus 17.8 expected deaths) for the respective time intervals of less than 5 years, 5-24.9 years, and 25 or more years since their first treatment. Doses to the stomach were quite variable, ranging somewhat uniformly between 0 and 5 Gy (Le88). These values, which are based on different time intervals and a longer period of follow-up than many of the previous analyses, and some of the other results reported above, suggest that the relationship between radiation and stomach cancer may be more complicated than was formerly recognized. A number of investigators have established that radiation is capable of causing adenocarcinomas in the mouse stomach (Wa86), although the stomach appears to be less susceptible to the carcinogenic effects of ionizing radiation than are many other similarly exposed organs in the laboratory mouse. Doses required to produce an effect are often quite large. For example, Hirose (Hi69) considered a single x-ray exposure of 20 Gy to be appropriate. Summary In human populations, as well as in laboratory animals, the risks of stomach cancer have been observed to be increased by irradiation. The existing data, although not sufficient to define the dose-incidence relationship precisely, are consistent with observations on atomic-bomb survivors, in whom the relative risk of mortality from stomach cancer is estimated to approximate 1.19 per Gy. The Committee's risk model for cancer of the digestive system is based on the mortality of A-bomb survivors in the Life Span Study (LSS), ICD codes 150-159. Stomach cancer was the predominant cause of death in this group. The Committee's preferred risk model contains terms for sex and age at exposure as shown in Equation 4-7 which is reproduced below.
RADIOGENIC CANCER AT SPECIFIC SITES 281 where I(S) equals 1 for females and 0 for males and The data indicate that females are at higher relative risk for cancer of the digestive system than males and there is a comparatively higher risk for those younger than 30 years when exposed. The Committee notes that this is not the usual pattern in which risk is highest for those exposed as children, diminishing for adolescents and young adults. Instead, the data indicate an abrupt decrease in risk for those over 30 years of age when exposed. In the Committee's judgment, this change is not simply a reflection of artifacts in the data, although a biological basis for it is unknown. Although the relative risk diminishes for ages greater than 30, the baseline risk for digestive cancers increases rapidly with age, so that most of the excess risk occurs after middle age (Figure 5-16). THYROID Introduction The incidence of thyroid carcinoma has been observed to be increased in a number of irradiated populations. Carcinoma of the thyroid was the first of the solid tumors noted to occur at increased frequency in the Japanese atomic bomb survivors (Ho63, So63). Even earlier, however, an increased incidence of thyroid cancer had been reported among persons exposed to therapeutic x rays as infants (Du50, Si55) and among individuals on the Marshall Islands who were exposed to radioactive fallout (Co80, Co84). Continuing studies of the atomic bomb survivors (Pr82, Wa83) and other irradiated populations (Co76, Ro77, Du80, NCRP85, Ro84a, Sc85, Sh84b) suggest the following generalizations: 1. Susceptibility to radiation-induced thyroid cancer is greater early in childhood than at any time later in life. In those exposed before puberty, however, the tumors usually do not become apparent until after sexual maturation. 2. Females are two to three times more susceptible than males to radiogenic as well as spontaneous thyroid cancer.
RADIOGENIC CANCER AT SPECIFIC SITES 282 FIGURE 5-16 Excess deaths due to cancer of the digestive system in U.S. females by attained age due to exposure to low-LET radiation at various ages. Persons exposed at less than age 25 are expected to follow the same mortality pattern as those exposed at age 25. 3. Radiogenic cancer of the thyroid is frequently preceded or accompanied by benign thyroid nodules; the frequency of hypothyroidism and simple goiter also is increased in those exposed to large doses of radiation when young. 4. Radiogenic carcinomas of the human thyroid are generally papillary growths; relatively few are of follicular or mixed histopathology. 5. The development of overt cancer from thyroid epithelial cells is dependent on hormonal stimulation, with the result that any condition leading to sustained elevation of thyroid-stimulating- hormone levels increases the risk of thyroid neoplasia. The following aspects of thyroid carcinoma are discussed below: (1) the Committee's analyses and risk estimates from two sets of subjects exposed to external radiation, (2) thyroid physiology as related to radiogenic thyroid cancer, (3) successive phases in the development of thyroid neoplasia, and (4) the special problem of radioiodide. This discussion is limited
RADIOGENIC CANCER AT SPECIFIC SITES 283 to the epithelium of the thyroid follicles; there is no evidence that radiogenic neoplasms develop from the parafollicular C cells. For reviews of radiogenic thyroid neoplasia, see references Cl86c, Co80, Du80, NCRP85. Risk of Human Thyroid Cancer from External Radiation The incidence of thyroid cancer was evaluated in two groups of patients who were exposed to low-LET radiations for benign conditions: children in the Israel Tinea Capitis Study (Ro84a) (10,834 irradiated subjects of ages 0-15 years when exposed and 16,226 nonirradiated subjects for comparison) and the Rochester Thymus Study (He75) (Sh85) (Sh84b) (2,652 irradiated subjects who were less than 1 year old when exposed and 4,823 nonirradiated subjects who were siblings of the irradiated subjects). The Israeli series included 39 cases of thyroid cancer among the irradiated group and 16 cases among the comparison group. In the Rochester cohort there were 37 cancer cases among the irradiated subjects and 1 in the nonirradiated group. It should be noted that the Committee's analyses are based on new individual dose estimates for individuals in the tinea study that take account of possible movement by the patients during irradiation. The mean doses remain the same as in an earlier analysis of these data which also included this feature (Ro84a). Doses in the Rochester thymus study are rough estimates of individual doses. It is likely that these dose estimates will be refined in the future. The Committee's analyses were carried out with the use of the AMFIT program (see Annex 4C) and were restricted to cases occurring 5 years or more after exposure. Since all subjects in the cohorts described above were irradiated during childhood, and because concerns about bias in ascertainment of this often nonfatal cancer among the LSS subjects mitigated against their use in risk estimation, the Committee's risk estimates for adults are based solely on extrapolations from the childhood exposures. Background Rates In the Israeli study, the 16 thyroid cancer cases that occurred in the nonexposed group were used to estimate background rates. However, since the Rochester study had only one cancer case among the nonexposed, background rates were modeled on the basis of a standardized incidence ratio relative to sex-, age-, and calendar period-specific rates in the Connecticut Cancer Registry. In the final models, the Rochester background rates were estimated to be about 97% of the Connecticut rates and about one half of the rates in the Israeli cohort.
RADIOGENIC CANCER AT SPECIFIC SITES 284 TABLE 5-3 Estimated Risk of Thyroid Cancer Incidence Under Various Modeling Assumptions Ethnic Agea (yr) Sex Latency (yr) Additive Riskb Relative Riskc Origin Israel Tinea Study Israeli 0â4 f 20 15.1 23.6 Israeli 0â4 f 30 25.5 23.6 Israeli 0â4 m 20 4.7 23.6 Israeli 0â4 m 30 7.9 23.6 Israeli 5â15 f 20 7.2 8.3 Israeli 5â15 f 30 10.8 8.3 Israeli 5â15 m 20 2.3 8.3 Israeli 5â15 m 30 3.4 8.3 Non-Isr 0â4 f 20 45.1 68.7 Non-Isr 0â4 f 30 75.7 68.7 Non-Isr 0â4 m 20 14.1 68.7 Non-Isr 0â4 m 30 23.7 68.7 Non-Isr 5â15 f 20 21.4 23.1 Non-Isr 5â15 f 30 32.3 23.1 Non-Isr 5â15 m 20 6.7 23.1 Non-Isr 5â15 m 30 10.1 23.1 Rochester Thymus Study â 0.5 f 20 1.8 19.2 â 0.5 f 30 1.8 6.7 â 0.5 m 20 0.6 19.2 â 0.5 m 30 0.6 6.7 a Age at exposure. b Excess is cases per 10,000 PYGy at age 40. c Relative risk at 1 Gy at age 40. NOTE: In the preferred model, constant relative risk is for Israelis exposed after age 5. Dose-Response Relationships and Their Modification A highly significantly elevated risk of thyroid cancer following radiation exposure during childhood was demonstrated in both the Israeli and Rochester cohorts (Table 5-3). Moreover, there was no evidence of significant nonlinearity in the shape of the dose-response function in either cohort. The level and pattern of the excess thyroid cancer incidence appears to depend on a number of the factors which are discussed briefly below. Cohort effects: The additive cancer excess among children over 5 years old at the time of exposure in the Israeli cohort was 9 times the excess in the Rochester cohort (p < 0.01). A cohort indicator was thus included in the final model. The relative risk in Israeli children over the age of five at the time of exposure was about 2.7 times as large as the relative risk in the
RADIOGENIC CANCER AT SPECIFIC SITES 285 Rochester cohort. Due in part to the small number of cases, this difference was not statistically significant. Additive versus relative risk models: For the Israeli cohort, a constant relative risk model with age at exposure and ethnic origin effects fit the data as well as a time dependent excess risk model with sex, age at exposure, and ethnic origin effects. For the Rochester data, after allowing for a 10-year minimal latent period, the excess risk appeared to be relatively constant with time since exposure. In terms of deviance this model fit the data better than constant or simple time dependent relative risk models. Sex: The excess absolute risk of thyroid cancer in irradiated females was approximately three times greater than that in irradiated males (p = 0.002). There was no significant sex difference (p > 0.9) in the excess relative risk. The thyroid cancer background rates among unirradiated females in both cohorts were about 3 times those among unirradiated males. Age at exposure: In the Israeli study, age at irradiation ranged from less than 1 year to 15 years, which allowed a limited assessment of the effect of age at exposure on cancer incidence. The effect was pronounced. When the data were subdivided into three age subgroups (0-4, 5-9, and 10-15 years), the estimated number of excess cancer cases was 31/10,000 person-year-Gy (PYGy) at age 40 for children exposed under the age of 5 as compared with 10/10,000 PYGy among those exposed over the age of 5 years. It should be noted that individual doses were estimated by using a model that assumed that children exposed at younger ages received higher doses; however, an age at exposure effect was also demonstrated when the number of radiation treatments was used as an indicator of dose. A comparable analysis of the Rochester cohort was not possible, as all exposed subjects were irradiated when they were less than 1 year of age. Time since exposure: Because it has been shown that radiation-induced thyroid cancer usually does not occur until five or more years after exposure, the results from the first five years after exposure were eliminated from the analysis. The Israeli data suggested that there is a continual increase in excess risk over the entire study period. The increase in risk was more than linear, that is, was proportional to time since exposure to the 1.4 power, but the trend was of borderline statistical significance (p = 0.08). In contrast, the Rochester data are well described by a constant-excess-risk model. There was no evidence of a significant time trend in the Israeli relative-risk data, but there was a significant decrease in relative risk over time in the Rochester cohort (p = <0.001). This decrease in relative risk was substantially greater than linear, that is, was proportional to time since exposure to the -2.8 power. The effect of this decrease in the relative risk
RADIOGENIC CANCER AT SPECIFIC SITES 286 was to make the excess risk roughly constant. However, despite allowance for this trend the constant-excess-risk model provides a better fit to the Rochester data. Ethnic origin: Within the Israeli cohort, thyroid cancer risks differed among different Jewish ethnic subgroups. Persons born in Israel had one-third the risk (excess risk, 3.4) of those born in Asia (mostly in the Middle East) or North Africa (excess risk, 10.2). Since the cohort of those born in Israel was restricted to those whose fathers were born in Asia or northern Africa, this difference in risk appears to originate from environmental rather than genetic bases. The westernization of the Israeli life-style, and hence of those born in Israel, may be the reason why their excess cancer risk was closer to that of the Rochester cohort. Recommended model: The choice of a model to be used in projecting thyroid cancer risks is difficult, because both available studies are limited to childhood exposures, and the levels of risk, as well as the latency distributions, in the two studies are inconsistent. Model choice is further complicated by the fact that thyroid cancer incidence rates are highly dependent on both the method of ascertainment and the criteria for surgery and hence confirmation of diagnoses used in a given study area. These differences profoundly affect excess-risk estimates. Thus, it is unlikely that projections based on an excess- absolute-risk model can provide a reliable indication of lifetime risk when applied to different populations. In view of these considerations, projections of lifetime thyroid cancer risks are based on the relative-risk model for Israeli-born children who were over 5 years old at the time of irradiation. This is a constant relative-risk model in which the relative risk at 1 Gy for all ages is 8.3 (Table 5-3). The distribution of this estimate is so highly skewed that the usual asymptotic standard error provides little useful information and a likelihood based confidence interval (Co74a) provides a more accurate summary of its variability. The likelihood based 95% confidence interval for this relative risk coefficient is 2 to 31 per Gy. Since there was no sex differential with the relative risk model, an equal risk is predicted for both sexes. Israeli-born children over the age of 5 were the preferred reference population because they had an intermediate risk within the Israeli cohort and their risk was relatively similar to the overall Rochester risk estimate. These lifetime risk estimates are less than those estimated from data from non-Israeli-born children exposed when they were less than age 5, but are greater than estimates based on the Rochester data. On the basis of unpublished data on A-bomb survivors made available to the Committee by RERF, we conclude that the risk of radiation-induced thyroid cancer in adults is, at most, one-half that in children.
RADIOGENIC CANCER AT SPECIFIC SITES 287 Thyroid Neoplasia from Internally Deposited Radionuclides Interpretation of the effects of internally deposited 131I and other iodine radionuclides has been complicated by dosimetric problems (Du80). From 73 to 96% of the energy absorption from these nuclides is from beta particles (NCRP85), and the fraction of energy absorbed within the gland depends critically on gland size and geometry and on the beta particle energy. For example, it is estimated that about 70% of the total beta energy from resident 131I is absorbed in the thyroid gland of a mouse or rat, 90% in a human infant gland, and >95% is absorbed in an adult human gland. Because of geometric considerations, follicular cells near the gland surface will absorb only 35-40% of the 131I beta dose delivered to cells in the center of a mouse thyroid (Du80). The issue is further complicated by biological factors. About 90% of the iodine in the thyroid is contained in the colloid in the follicular lumena. The efficiency of iodide uptake differs among follicles, as does the efficiency of hormone secretion. Thus, relative dose distribution varies from follicle to follicle and over time. Dose is also a function of electron energy at the cellular as well as the glandular level. The low-energy electrons released by 125I decay in colloidal TG deposit most of their energy at the lumenal end of the follicular cell, with little reaching the nucleus located at the basal pole. Finally, total dose depends on the resident time of the nuclide within the gland. This in turn is a function of dietary iodine content and hormonal status. These factors, coupled with limited numbers of measurements of nuclide concentrations over time, small and variable numbers of animals, use of various goitrogens or diets as promoters, and the lack of standardization of rat age, sex, strain, and husbandry, render quantitative interpretations and comparisons of the earlier experimental data difficult (NCRP85). Estimates of the relative effectiveness per unit of radiation dose from 131I varied from 1/15th to 1/2 of that from external x-irradiation in the older experimental literature (Du80, NCRP85). The effects of internally deposited radioiodides have been investigated in three categories of human subjects: patients who received relatively large therapeutic doses of 131I (Do74, Ho84a, Ho84b, Ho80a), patients who received much smaller doses of 131I for diagnostic purposes (Ho80b, Ho80c, Ho88), and those on the Marshall Islands who were exposed to iodine radionuclides in fallout from a bomb test (Co80,Co84). The majority of subjects who received therapeutic 131I were suffering from thyrotoxicosis. In this disease, the rate of thyroid hormone synthesis and release is accelerated, and thyroid stimulating hormone (TSH) or long-acting thyroid-stimulating protein (LATS) is generally elevated. Hence, the maximum 131I uptake by the thyroid is increased and the residence time
RADIOGENIC CANCER AT SPECIFIC SITES 288 in the gland is reduced. The total radiation doses were often very large (approximately 50-100 Gy); that is, they were designed to cause extensive cell death; indeed, the probability of development of hypothyroidism within 2 years after treatment increased nearly linearly with dose over the range 0.9-8.3 megabecquerels 131I/gram of thyroid (Cl86a). Furthermore, the incidence of thyroid cancer among patients with thyrotoxicosis may be as much as 10 times that among the general population (NCRP85). Hence, the choice of an appropriate unirradiated control group is difficult (Ho84a). Finally, given the long latency of radiogenic cancers, the periods of follow-up have been brief (means of 8-15 years). When the observed cancer incidence in a prospective study of 1,005 131I- treated patients with thyrotoxicosis (mean follow-up of 15 years) was compared to that of 2,141 surgically treated patients, the risk was 9.1-fold greater (p < 0.05) in the irradiated group (Ho84a). When compared with the Connecticut Cancer Registry, however, the relative risk from 131I (3.8) was not statistically significant. After a mean follow-up of 8 years, the thyroid cancer incidence ratio of 21,714 131I-treated patients to 11,732 surgically treated patients was 2.6, and was not significant (D074). A relative risk (insignificant) of 1.01 was found when the cancer incidence in 4,557 131I-treated patients followed for an average of 9.5 years was compared with data from the Swedish Cancer Registry (Ho84b). An initial follow-up study of 10,133 subjects who received diagnostic doses of 131I at the Radiumhemmet, Stockholm, Sweden, yielded no evidence of an increase in thyroid cancer risk (Ho80b, Ho80c). A more recent study included 35,074 Swedish subjects (28,180 women and 6,884 men) from six institutions who had survived 5 years or more after a diagnostic dose of 131I (Ho88). The mean dose was 1.92 MBq of 131I (range: 0.04-35.52 MBq), the mean radiation dose was approximately 0.5 Gy, the mean age at exposure was about 44 years (range: 1-74 years), the mean follow-up was 20 years during the period 1951-1984, and the data were compared with the Swedish Cause-of- Death Registry. A total of 50 thyroid cancers were found in the 131I group compared with an expected number of 39.37 cases, yielding an overall standardized incidence ratio of 1.27 observed to 1.0 expected cancers (95% confidence interval, 0.94-1.67). Of the 50 observed cancers, 10 were either medullary or poorly differentiated and 1 was a sarcoma. Medullary carcinomas have not been seen to be associated with radiation exposure. Six thyroid cancers occurred among men who were 50-74 years old at the time of exposure; this subgroup yielded the only significantly increased standardized incidence ratio 3.14 (95% confidence ratio, 1.15-6.84). Sixty-eight percent of the cancers occurred among 31% of the subjects who had received a diagnostic dose of 131I because of suspected thyroid cancer. Of these 34 cases, 15 cancers (44%) became clinically apparent 5-9 years after exposure, suggesting that they were occult
RADIOGENIC CANCER AT SPECIFIC SITES 289 at the time of the 131I diagnostic procedure. In summary, the results of these studies do not support the conclusion that diagnostic doses of 131I significantly increase the risk of thyroid cancer (Ho88). People on the Marshall Islands were exposed to fallout from the thermonuclear BRAVO bomb test on Bikini atoll on March 1, 1954 (Co80). The atoll Bikini is approximately 95, 100, and 300 miles from Alingnae, Rongelap, and Utirik atolls, respectively. The radiation dose to the thyroid glands of the residents of the Utirik atoll was in part from external gamma rays from fallout dust (1.75, 0.69, and 0.14 Gy for those on Rongelap, Alingnae, and Utirik atolls respectively) and in part from inhaled and ingested radioiodides. Doses of the ingested radioiodides were calculated from the 131I content of pooled urine samples collected 15 days after the first exposure (Co80, Co84); the dose contributions from the short lived radionuclides 132I, 133I and 135I were assumed to be equal to 2-3 times the 131I dose. Two-thirds of those on Rongelap atoll and 5% of those on Alingnae atoll suffered nausea within 48 hours. Half of the Rongelap atoll natives developed partial epilation beginning 2 weeks after exposure, indicating significant total-body and body surface doses. By 8 years after exposure, two boys who were 1 year of age when they were irradiated were diagnosed with myxedema (Co80). Nine years after exposure, the first thyroid nodule was noted in a 12-year-old girl. The seriousness of the situation was apparent by 1965, and prophylactic thyroid hormone treatment was then initiated in residents of Rongelap atoll; prophylaxis was initiated 4 years later in residents of Alingnae atoll. How this treatment has influenced the development of neoplasia is unknown. However, insofar as increased thyroid hormone levels would be expected to reduce TSH release, the rate of malignant progression of the progeny of radiation-initiated thyroid cells would be expected to have been reduced (Cl86b, Du80). The thyroid status of the Marshall Islanders 27 years after exposure is summarized in Table 5-4. Although the dose estimation is open to question, the prevalence of hypothyroidism, thyroid nodules and proven thyroid cancer all appears to increase with dose (Co84). These studies have recently been extended to residents of more distant atolls who were not previously considered to have been at risk (Ha87). The 14 atolls were chosen to include all northern atolls that could possibly have been in the fallout path and as many southern atolls as feasible; the latter were chosen as sources of unexposed controls. Alingnae atoll was excluded, as it was uninhabited at the time of the study (1983-1985). Study subjects included 2,273 persons who were alive and residing on 1 of the 14 atolls at the time of the BRAVO test. For purposes of this study, a thyroid nodule or neoplasm was defined as a solitary discreet nodule of at least 1 cm in diameter; nodules of less than 1 cm were not considered as positive, nor were multinodular goiters, diffuse hyperplasias, or cases of Graves' disease.
RADIOGENIC CANCER AT SPECIFIC SITES 290 TABLE 5-4 Prevalence of Thyroid Abnormalities Among Marshall Islanders 27 Years (1981) After Exposure to Fallout (Co84) Percent with Condition Group and Number of Dose (Gy) Hypothyroid Nodules Cancera Age, 1954 Subjects Rongelap 1 yr 6 â¥15 83.3 66.7 0 2-9 16 8-15 25.0 81.2 6.2 â¥10 45 3.4-8 8.9 13.3 6.7 Alingnae <10 7 2.8-4.5 0 28.6 0 â¥10 12 1.4-1.9 8.3 33.3 0 Utirik <10 64 0.6-1.0 0 7.8 1.6 â¥10 100 0.3-0.6 1.0 12.0 2.0 Controls <10 229 â 0.4 2.6 0.9 â¥10 371 â 0.3 7.8 0.8 aValues are conservative estimates; unoperated nodules were considered benign, and occult carcinomas were excluded. Included as positive were a number of Rongelap and Utirik atoll residents who had previously been thyroidectomized for a thyroid nodule. Among the residents of the 12 atolls who had previously been considered to be unexposed, the thyroid nodule prevalence varied from 0.9 to 10.6%. The null hypothesis of no differences among these populations was rejected (p < 0.025). The age- adjusted prevalence of thyroid nodules bore a highly significant inverse relationship to the distance of the residence from Bikini atoll (p < 0.002). As suggested by the control data in Table 4-9, previous analyses had assumed a prevalence of non-radiation-related thyroid nodules of 6-7% (Co84). In the current analysis, the prevalence of spontaneous thyroid nodules was taken to be 2.45%, which was the mean prevalence of residents of the two southernmost atolls, which were far from the path of fallout (Ha87). Logistic regression analysis showed a significant dependence of nodule prevalence on distance from Bikini atoll, age at exposure, sex, and the angle of deviation in latitude from Bikini atoll. Women were 3.7-fold more susceptible to nodule formation than men. The effect of distance in angle of deviation in latitude from Bikini atoll was attributed to shifting wind patterns that at first carried the fallout approximately due east from Bikini atoll but that later carried it southwest from Utirik atoll. Using the age-adjusted spontaneous thyroid nodule prevalence of 2.45%, a new estimate of risk was calculated based on the nodule prevalence in persons residing on Rongelap and Utirik atolls. The absolute
RADIOGENIC CANCER AT SPECIFIC SITES 291 risk coefficient so calculated was 11 excess nodule cases/104 PYGy (Ha87). This estimate is approximately 33% greater than previous estimates, largely because of the difference in assumed spontaneous prevalence. These calculations are subject to the same limitations imposed by post hoc dosimetry as all previous estimates. Furthermore, there is a wide variation among reports of the spontaneous thyroid cancer incidence among different Polynesian populations. For example, reported incidences among the female Polynesians of Cook Island, New Zealand, and Hawaii were 18.6, 2.5, and 9.3/105 person-year Gy, respectively; however, the total numbers of cancer cases were only 4, 6, and 19 respectively (He85). The SEER report also gives a high estimate for Hawaiian Polynesian women: 19.2 cases of cancer/105 person-year Gy (Se81). The above analysis of the data on the Marshall Islanders should thus be interpreted with caution. A survey of the thyroid status of school children in Lincoln Co., Nevada, and Washington Co., Utah, was performed annually from 1965 to 1971 (Ra74, Ra75). A total of 1,378 children were identified who had lived as infants in these counties during the period 1952-1955 when there was estimated to be fallout from atmospheric atomic bomb tests. Most of the dose from this fallout was assumed to be from the ingestion of 131I-contaminated milk, in which the nuclide was metabolically concentrated. Cumulative radiation doses to the thyroids of children residing in southwestern Utah were estimated to average as high as 1 Gy. A cohort of 1,313 children in the same schools who had moved to the counties after the cessation of atmospheric atomic-bomb testing and another cohort of 2,140 children from a county in Arizona that was remote from the fallout path were chosen as unexposed controls. In the original reports of this study, no significance was attached to the relatively modest differences in thyroid abnormalities noted among the exposed and unexposed groups (Ra74, Ra75). The data from the first report (Ra74) have recently been reanalyzed and are shown in Table 5-5 (Ro84b). Although there was no increase in thyroid cancer incidence in the presumably exposed populations, there was a suggestive 20-30% greater prevalence of all thyroid abnormalities in exposed versus unexposed children. It is also important to note, however, that the lower 90% confidence limits of the prevalence ratios are individually and collectively less than or equal to 1.0. It is also of importance to note that the follow-up period was approximately 14 years from the time of exposure. In a second report of this study, the prevalence of nodular goiter among these three groups was compared (Ra75). The prevalences were 8.7, 4.6, and 4.7 per 1,000 children in the exposed group, the unexposed Utah-Nevada group, and the Arizona group, respectively (Ra75, Ro84b). The exposed/unexposed prevalence ratio was 1.9 (90% confidence limits,
TABLE 5-5 Prevalence of Thyroid Abnormalities in Fallout-Exposed and Unexposed Children (Ro84) Children, Group and Number Prevalence of Thyroid Abnormalities per 1,000 Subjects Cancer Benign Neoplasm Adolescent Goiter Thyroiditis Hyperthyroid. Misc. Total 1,378 exposed Utah-Nev. 0 4.4 16.0 13.1 3.6 37.0 1,313 unexposed, Utah-Nev. 0.8 3.0 9.1 12.9 2.3 28.2 2,140 unexposed, Arizona 0.5 2.8 15.4 8.4 1.9 29.0 Ratio, exposed:unexposed, and 90% conf. lim. 0 1.5 1.2 1.3 1.8 1.3 0, 5.4 0.6, 3.5 0.8, 1.9 0.8, 2.1 0.6, 4.8 1.0, 1.7 Ratio, excluding Arizona, and 90% conf. lim. 0 1.4 1.7 1.0 1.6 1.2 0, 8.4 0.5, 4.5 1.0, 3.2 0.6, 1.8 0.5, 6.0 0.9, 1.9 RADIOGENIC CANCER AT SPECIFIC SITES 292
RADIOGENIC CANCER AT SPECIFIC SITES 293 1.0-3.5) (Ro84b). Again, the lower confidence limit was 1.0. The analyst concludes that although the data showed a weak but positive radiation effect, in the absence of better dosimetric information they revealed little about the effects of such exposure (Ro84b). In contrast to human studies, a large scale animal experiment showed little difference between the effects of x rays and 131I. In this study, oncogenic effects of 131I on the thyroid were compared with those of x rays using 3,000 female rats of the Long Evans strain (Le82). The carcinogenesis experiment was preceded by detailed dosimetric studies (Le79). The thyroids of 6 week-old rats were exposed to 0, 0.94, 4.10, or 10.60 Gy of highly localized x rays or were injected intraperitoneally with 0, 0.48, 1.9, or 5.4 ÂµCi of 131I. The 131I doses were chosen to yield radiation doses to the gland of 0.80, 3.30, and 8.50 Gy, respectively. Two additional groups of rats received 4.10 Gy of x rays to the pituitary alone or to the pituitary and the thyroid to yield 10 equal groups in all. A total of 2,762 animals that died or were killed from 6 months until the termination of the experiment at 24 months after exposure were included in the analysis. The incidence of thyroid cancer increased as a positive exponential function of dose following administration of either x rays or 131I; in both cases, however, the coefficient of dose in the exponent was significantly less than 1.0 (Le82). The ratios of x-ray-induced cancer to 131I-induced cancer at 0.80, 3.30, and 8.50 Gy were 1.3, 1.0 and 0.9, respectively, suggesting nearly equal effectiveness per unit dose. The 24 month thyroid cancer risk per 0.01 Gy for both 131I and x rays in the 0.80-0.90 Gy range was 1.9 Ã 10â4 which is similar to the estimated human life time risk from x rays (Le82). In contrast, the slopes of the dose-response relationship for adenoma and total tumor had upward curvature, and appeared to rise more rapidly in the x-ray-treated groups. The parameters of the dose-response relationships following administration of 131I or x rays did not differ significantly, however. Irradiation of the pituitary gland did not alter the results. The National Council on Radiation Protection and Measurements (NCRP) reviewed the data and analyses on radiation induced thyroid cancers that were available through 1985 and recommend the use of a Specific Risk Estimate (SRE) according to the following formula (NCRP85): SRE = RFSAY, where R is the absolute risk in excess thyroid cancer cases per 10 4 person- year-Gy; F is the dose effectiveness factor, which is assumed to be 1 for external radiation, 132I, 133I, and 135I and 1/3 for 125I and 131I; S is the sex factor taken to be 4/3 for women and 2/3 for men; A is the age factor which is equal to 1 for those <18 years of age and 1/2 for those >18 years of
RADIOGENIC CANCER AT SPECIFIC SITES 294 age at time of exposure; and Y is the anticipated mean number of years at risk. The absolute risk, R, chosen for this calculation is assumed to be that for an ethnically homogeneous population of children (<18 years of age) of equal numbers of each sex who were exposed to external radiation and corrected for a minimum 5 year latency. The SRE calculated in this way is the risk of development of thyroid cancer during the rest of an individual's life. If a thyroid cancer mortality risk is desired, the SRE is multiplied by L, the lethality factor, assumed to be 1/10. The NCRP uses an absolute risk factor, R, of 2.5 thyroid cancers/104 PYGy for doses in the range of 0.06-15.0 Gy (NCRP85). Parallel and combined analyses of six cohorts of children and two cohorts of adults exposed to external radiation and one cohort of children and one cohort of adults exposed to 131I have been reported recently (La87). Data from the study of the large Long Evans strain of rat were also included. A constant relative potency for neoplasia induction by 131I as compared with that by external x rays was assumed across ethnic and sex cohorts and ages across species lines. The risk ratio estimate so derived for 131I compared to x rays was 0.66 (95% confidence limits, 0.14-3.15) and did not differ significantly from 1.0 (La87). Physiology of Radiogenic Thyroid Cancer Thyroid neoplasia has been an attractive model in experimental carcinogenesis because thyroid epithelial cell proliferation and function affect susceptibility to thyroid neoplasia and can be readily manipulated, the thyroid and pituitary hormones that regulate and/or reflect thyroid cell proliferation and function are easily measured, thyroid tissue or cells are readily transplantable, and thyroid neoplasia is a significant human risk (Cl86b, Du80). The rate of proliferation of thyroid cells is regulated by the concentration of thyroid-stimulating hormone (TSH) in blood. Synthesis and release of TSH from the anterior pituitary gland is stimulated by TSH-releasing hormone (TRH), which is synthesized in the hypothalamus and reaches the TSH- secreting cells via the hypothalamic-hypophyseal portal system. TSH levels reaching the thyroid via the general circulation cause the synthesis and release of thyroid hormone and the proliferation of thyroid follicular cells. Serum thyroid hormone reaching the hypothalamus inhibits TRH release, thus modulating hypophyseal TSH release and, in turn, maintaining thyroid hormone titers within normal levels. This long-loop feedback regulation is supplemented by neural input via the hypothalamus and by additional short-loop feedback systems which operate under special circumstances (Cl86b). The prime functions of the thyroid follicular cells are the synthesis,
RADIOGENIC CANCER AT SPECIFIC SITES 295 storage, and release of the thyroid hormones thyroxine (T4) and 3-5-3' triodothyronine (T3). T4 synthesis and T3 synthesis occur in three phases: (1) uptake and concentration of inorganic iodide, (2) the preceding or concurrent synthesis of thyroglobulin (TG), and (3) iodine organification and iodothyronine formation in the TG molecule. The iodinated TG is then either hydrolyzed to release T3 and T4 for secretion or is stored in the thyroid follicular lumina as colloid (De65). Feedback regulation of the thyroid is vulnerable to disruption by natural, therapeutic, or experimental means at virtually every step (Cl86b, Du80). The goitrogenic effect of iodide deficiency has been recognized since antiquity, and experimental hypothyroidism is readily induced by diets low in iodine. Pharmacological disruption of the iodide concentration by perchlorate and of iodide oxidation and iodothyronine synthesis by thiocarbamides and other goitrogens are common experimental techniques used to block T4 and T3 synthesis. Partial or total destruction of the thyroid epithelial cells can be induced by administration of radioiodide. A TSH-mimicking molecule, long- acting thyroid-stimulating protein (LATS), results in hyperthyroidism in some humans. These observations have been experimentally exploited in studies of thyroid radiobiology and carcinogenesis. Phases of Thyroid Carcinogenesis The events in thyroid carcinogenesis can be divided into three phases: (1) an acute phase, including early radiation injury, neoplastic initiation, and intracellular repair; (2) a latent phase, from the acute phase until overt tumor formation; and (3) the phase of tumor growth (Cl86b). The Acute Phase The first step in radiogenic thyroid cancer induction is initiation, that is, the formation of one or more heritable precancerous changes in one to many thyroid cells (Cl86b, Du80). About 1-2% of young rat thyroid epithelial cells are clonogenic, that is, they are capable of forming new clonal thyroid follicles under TSH stimulation (Cl85a); these proliferation-competent cells are presumed to be the cells of origin of thyroid neoplasms. Clonal follicular units have been used as an endpoint in a quantitative transplantation assay of the relative numbers, acute response to radiation, postirradiation repair capacity, and frequency of neoplastic initiation in the thyroid clonogens (Cl85a, Mu84). The evidence indicates that 98-99% of the thyroid epithelial cells are nonclonogenic, that is, they are capable of but a few rounds of mitosis in response to TSH (Du80). Radiogenic cellular damage after doses in the carcinogenic range of less than 20 Gy is usually expressed during mitosis. There is little biochemical
RADIOGENIC CANCER AT SPECIFIC SITES 296 evidence of acute impairment of secretory function after radiation exposure; that is, the irradiated animals remain euthyroid for a time. Hence, TSH levels are within the normal range, mitotic activity remains low, and cellular damage is not immediately expressed (Du80). The Latent Phase Whether radiogenic damage is expressed in frank tumor formationâand if so, whenâdepends on the interaction of internal environmental factors with the initiated thyroid cells. Under normal circumstances, the rate of thyroid epithelial cell division is low, but it is not nil (Do67, Do77). In euthyroid rats, TSH in concert with other factors is normally present at concentrations that are sufficient to stimulate a small portion of grafted clonogens to follicle formation (Mu80b). In the early latent phase, the cells in the irradiated thyroid are thus subjected, a few at a time, to mitosis-triggering stimuli. TSH-triggered normal cells respond with normal mitosis. Altered but reproductively viable cells, including initiated cells, also proliferate in response to triggering stimuli. Triggered and terminally damaged cells may pass successfully through one or a few mitoses before death, or they may survive without division but with persistence of the secretory function for several months or longer. Ultimately, however, the proportion of terminally damaged cells decreases through cell death; as a result, compensatory triggering stimuli increase to bring about the replacement of lost cells. Depending on radiation dose, and hence, the fraction of the population made up of terminally damaged cells, this process of triggering and cell death continues very slowly, perhaps undetectably, over many months, or it may accelerate in a cascade fashion after a slow beginning (Cl86b). In rats there is a threshold between 2.5 and 5.0 Gy for histologically detectable radiogenic damage (Do67). After a dose of 10 Gy, however, when only a small fraction (approximately 7%) of clonogens would be expected to retain reproductive capacity (Mu80a), although the animals may remain euthyroid, partial glandular atrophy coupled with epithelial cell hypertrophy and interstitial fibrosis occurs with time (Do67). Similar changes are observed after injection of 30 micro-Ci of 131I (Do77). Higher doses of x rays (greater than 15-20 Gy) result in widespread evidence of epithelial cell damage (Do67) with cell degeneration, follicle disruption, and interstitial and vascular fibrosis. These changes are qualitatively similar following external radiation or internal radiation by 131I (Do77, Ga63, Li63). They occur soon after exposure to single doses in excess of 20 Gy, but are delayed for weeks to months in smaller animals and for years in large species after a dose of about 20 Gy. If such damage is extensive, hypothyroidism develops, but neoplasia is a less common result in extensively damaged
RADIOGENIC CANCER AT SPECIFIC SITES 297 glands than in glands in which 5 to 50% of the epithelial clonogens survive (i.e., at doses of about 6-11 Gy) (Mu80a) (NCRP85). From the standpoint of carcinogenesis, the important processes during the latent phase include amplification of the radiation-initiated clonogen population under repeated mitosis-triggering stimuli. During this process, insofar as repeated rounds of DNA synthesis and mitosis play a role in neoplastic changes in initiated cells, promotion and progression, as well as clonal expansion, occur in the initiated cells. In endocrine-responsive cell populations, progression is frequently associated with quantitative or qualitative changes in hormone responsiveness (Cl75, Fu75). The Phase of Tumor Growth Radiation-induced thyroid tumors first appear as localized hyperplastic nodules. They are often multifocal, suggesting that they originated from randomly distributed initiated cells. Adenomas are most common, occurring 10-16 months after exposure in rats and in increasing frequency with time thereafter (Do63). Carcinomas appear after 18-30 months in rats and are frequently found within or associated with adenomas. The development of the thyroid cell transplantation technique has permitted studies of carcinogenesis in vivo in terms of surviving clonogenic cells, that is, thyroid carcinoma and total tumor incidence per 0 or 5 Gy of x- irradiated grafted clonogen have been investigated in thyroidectomized rats maintained on an iodine-deficient diet (Mu84, Wa88). On a cellular basis, the radiogenic initiation frequency was high. For example, cancers developed in 34% of the transplant sites, each of which were grafted with 11 surviving clonogens irradiated with a dose of 5 Gy. This corresponds to one initiated clonogen for about every 32 clonogens grafted (Wa88). In summary, although both benign and malignant thyroid neoplasms arise from the relatively small radiation-initiated cell subpopulation, this occurs gradually over time as the result of neoplastic promotion, progression, and clonal amplification under the mitogenic stimulation of TSH. The intensity of the TSH stimulation depends in turn on the functional capacity of the entire thyroid follicular cell population, a significant fraction of which, although it retains secretory capacity, may die during mitosis from radiation injury. Summary Thyroid cancer is well established as a late consequence of exposure to ionizing radiation from both external and internal sources in humans and experimental animals. The histopathology of radiogenic thyroid cancer indicates that it appears to arise exclusively from the follicular epithelium. It is relatively indolent and causes death infrequently (mortality/incidence
RADIOGENIC CANCER AT SPECIFIC SITES 298 ratio = approximately 0.1) in comparison with more malignant medullary thyroid cancers, the incidence of which has not been found to be increased in irradiated subjects. Analysis of two cohorts of subjects exposed to therapeutic radiation for benign conditions and a review of the literature have revealed that: 1. There are major differences in background thyroid cancer rates among unirradiated individuals of different reported cohorts. Analysis suggests that these differences are related, at least in part, to life-style, although ascertainment may also play a critical role. 2. Females are roughly 3 times as susceptible to radiogenic, as well as nonradiogenic (background), thyroid cancer as males. Hence, relative-risk estimates do not differ significantly by sex. 3. The excess risk from radiation exposure is considerably greater among children who are exposed during the first 5 years of life than in those exposed later. On the basis of an examination of data from the Japanese adult health study, not yet published by the RERF, the Committee concludes that the risk of radiation-induced thyroid cancer in adults is only one half, or less, of that in children. 4. Although the data are best fit by an excess-risk model that includes allowance for cohort effects, latency, age at exposure, and sex, a relative risk model is preferred because of the strong dependence of the estimates on the background incidence of the particular cohort under consideration. The model, which is based on the risk in Israeli-born children who were exposed when they were more than 5 years of age, yields a relative risk at 1 Gy of 8.3 for both sexes. 5. The risk ratio for 131I/x rays has been estimated as 0.66, but the 95% confidence interval of the ratio is broad (0.14-3.15), since the risks from internally deposited radionuclides of iodine are not well understood. 6. The development of thyroid cancer from initiated cells is profoundly dependent on hormone balance. ESOPHAGUS In a recently completed registry-based study of the incidence of second primary cancers in women following radiation treatment for cervical cancer (Bo85), an overall relative risk for esophageal cancer of 1.5 (p < 0.05) was seen among patients with invasive cervical cancer who were treated with radiotherapy (40 observed cases versus 27 expected cases). The corresponding risks for women with invasive cervical cancer who did not receive radiation and women with in situ cervical cancer (the majority of whom only received surgical treatment) were 1.0 and 0.5, respectively. However, when attention was restricted to patients with invasive cervical cancer who were treated with radiation and followed for 10 or more years, the relative
RADIOGENIC CANCER AT SPECIFIC SITES 299 risk was reduced to 1.1, which was no longer statistically significant. The authors of the study included the esophagus among the organs estimated to receive small doses of radiation, that is, an estimated average dose of less than 0.5 Gy, and concluded that cancers at these sites were either not elevated or were probably increased because of other major risk factors, such as the use of cigarettes or alcohol. Esophageal cancer was not included in a recent case control analysis of these data (Bo88). Stronger evidence in support of the relationship between esophageal cancer and exposure to ionizing radiation is provided by the analysis of the updated mortality experience of the cohort of patients with ankylosing spondylitis (Da87). The authors reported that while a highly significant (p < 0.001) increase was observed for all neoplasms other than leukemia or colon cancer when considered collectively, the number of deaths observed 25 years or more after treatment tended to decline, closely approaching expected values (i.e., relative risk, 1.07). The main exception to this trend was esophageal cancer, which was significantly increased (p < 0.01) during both the intervals of 5-24.9 years and 25 or more years posttreatment, with relative risks of 2.05 and 2.41, respectively. Overall, the relative risk for x-ray-treated patients 5 years or more after treatment was 2.20 (28 observed versus 12.73 expected esophageal cancer deaths), which was highly significant (p <0.001). The estimated mean dose to the esophagus was quite high (over 4 Gy) (Le88). The 1950-1982 follow-up data for the combined Japanese survivor populations (Pr87a), based on T65 dosimetry, continues to support the hypothesis of radiation-induced cancer of the esophagus. In this report from the Radiation Effects Research Foundation, relative risks ranging from 0.65 in the 1-9-rad dose group to 2.03 in the >400 rad dose group were observed. Preston and colleagues (Pr87a) also reported a nonsignificant (p = 0.30), decreasing trend in the relative risk of esophageal cancer over time and a significant (p = 0.03) effect of sex on the relative risk of esophageal cancer (i.e., the estimated relative risk for exposures of 1 Gy for males and females were 1.09 and 2.23, respectively). If ethnic and other differences in potential risk factors are ignored, then this latter result suggests that the lack of a more impressive and sustained esophageal effect in the cohort of patients with cervical cancer may be related, at least in part, to the relatively low organ-specific levels of exposure that they received. The latest analysis of the Japanese cohorts (Sh88) using the new DS86 dosimetry data indicates that in terms of the kerma at a survivor's location the relative risk for esophageal cancer is estimated at 1.43/Gy (p < 0.05), with a corresponding excess risk of 0.34/104 PYGy. In terms of dose to the esophagus, the relative risk is 1.58/Gy; excess risk 0.45/104 PYGy (Sh88).
RADIOGENIC CANCER AT SPECIFIC SITES 300 Summary Carcinoma of the esophagus has been observed to occur with increased frequency in several irradiated human populations. The available dose- incidence information is sparse, but the data from the various studies are consistent with those from the A-bomb survivors, in whom the relative risk is estimated to approximate 1.58 per Gy (organ dose). SMALL INTESTINE Although carcinomas of the small intestine can be induced with a high frequency in mice and rats by intensive irradiation of the ileum or jejunum, as noted below, their induction by irradiation in humans has yet to be established. In comparison with cancers of the stomach and colon, however, cancer of the small intestine occurs infrequentlyâits annual incidence in humans approximates only 0.8 cases/100,000 people (Yo81). In addition, little is known about the factors that affect its occurrence in the general population (Li82). In 2,068 women treated with irradiation of the ovaries for excessive menstrual bleeding, an excess of mortality from cancer of the small intestine was observed; that is, there were 3 observed deaths, compared with only 0.4 expected deaths (Sm76). Similarly, in an international study of 82,616 women treated with radiation for carcinoma of the uterine cervix, a two-fold excess of cancer of the small intestine was observed; that is, there were 21 observed versus 9.5 expected cases (Bo85). The excess was evident, however, within the first year after treatment and did not significantly increase with time. Furthermore, a comparable excess was observed in women with invasive cervical cancer who received no radiotherapy; that is, there were 4 observed versus 0.9 expected cases. New case control analyses of these data yield a relative risk of 1 for cancer of the small intestine (Bo88). Hence, any causal relationship between the excess cases and radiation is questionable. No excess cases of the disease have been reported in Japanese A-bomb survivors, patients treated with radiation for ankylosing spondylitis, or other irradiated populations; but cancers of the small intestine have not been reported separately from cancers of the colon in most such studies (La86). Adenocarcinomas of the small intestine have been observed in more than 20% of LAF1 mice surviving midlethal doses (3.5 Gy) of whole-body neutron radiation, whereas such tumors are rare in nonirradiated controls or in mice surviving midlethal doses of x-irradiation (No59). Similarly, in rats, a high incidence (>50%) of such tumors has been observed after localized exposure of the ileum or jejunum to x rays (Os63, Ts73, and Co74) or deuterons (Bo52) at doses exceeding 15 Gy.
RADIOGENIC CANCER AT SPECIFIC SITES 301 Summary Although adenocarcinomas of the small intestine can be produced by intensive localized irradiation in laboratory animals, no carcinogenic effects of radiation on the small intestine have been evident in any of the irradiated human populations studied to date. Hence the risk of radiation carcinogenesis in the small intestine, although not quantifiable, appears to be low. COLON AND RECTUM Colon Irradiation has been observed to increase the risk of colon cancer in humans and laboratory animals. The strongest evidence of the carcinogenic effects of radiation on the human colon is provided by the dose-dependent excess of colon cancers observed in Japanese A-bomb survivors. At doses of 1 Gy or higher, a total of 25 deaths from colon cancer were observed between 1950 and 1982 in members of the Life Span Study cohort population versus 14.50 expected deaths; no such excess was evident during the first 15 years after irradiation (i.e., before 1959), nor has any excess been evident at doses below about 1.0 Gy (Pr87a). The relative risk per Gy (organ dose) in the DS86 subcohort was estimated to amount to 1.85 (1.39-2.45), which corresponds to an excess of 0.81 (0.40-1.30) deaths per 104 PYGy (Sh88). A comparable association between cancer of the colon and therapeutic irradiation has been observed in two series of women treated for benign gynecologic conditions. Four deaths from intestinal cancer were observed (versus one expected death) in a series of 297 women followed for an average of 16 years after irradiation of the ovaries for benign pelvic disease (Br69), and 24 deaths from colon cancer were observed (versus 13.86 expected deaths) more than 5 years after treatment in a series of 2,067 women treated with irradiation for metropathic hemorrhagica (Sm76). No significant excess deaths were observed, however, in two other series of women treated with radiation for similar disorders (Di69, Wa84). Likewise, in a large series of patients (82,616 women) treated with x rays for carcinoma of the uterine cervix, in whom the average dose to the colon was estimated to have exceeded 5 Gy, no consistent excess number of deaths was observed within the first three decades after irradiation (Bo85). The new case control analysis of these data yielded an insignificant excess risk of only 1.02 (Bo88). In 14,106 patients who were treated with x rays to the spine for ankylosing spondylitis during 1935-1954 and who were followed until 1985, a total of 47 deaths from colon cancer were observed, versus 36.11 expected (relative risk, 1.30) (Da87); however, the relative risk in this population
RADIOGENIC CANCER AT SPECIFIC SITES 302 was higher during the first 2-5 years after irradiation (ratio of observed to expected deaths 6/2.50 = 2.40), in keeping with the known associations between ankylosing spondylitis and ulcerative colitis and between ulcerative colitis and colon cancer. In view of the confounding influence of these associations, the excess deaths in this population have not been attributed to radiation per se (e.g., NRC80, Sm82). In laboratory rats, localized exposure of the colon to 45 Gy of collimated x rays has been observed to cause a high incidence (47%) of adenocarcinomas, with smaller increases at higher and lower dose levels (De78). Such neoplasms have also been induced in a large percentage of rats (75%) by localized beta irradiation from yttrium administered in the diet (Li47). Similarly, rats and dogs exposed to neutron beams or subjected to irradiation of the bowel by dietary polonium-210 or cerium-144 have been observed to develop benign and malignant tumors of the colon (Le73). Although whole-body gamma or x- irradiation at doses in the range of 5-10 Gy has been reported to cause only a small increase in the incidence of such tumors (5%) in rats (Br53, Wa86) and mice (Up69), a high incidence (27%) has been induced in mice (No59) by near- lethal whole-body fast neutron irradiation. Rectum Carcinoma of the rectum has been observed to be increased in frequency in humans (La86) and laboratory animals (Wa86) by intensive localized irradiation. In a large series of women (82,616) who were treated with radiation for carcinoma of the uterine cervix, and who were estimated to have received an average dose of more than 50 Gy to the rectum, no excess in the number of rectal cancers was observed within the first decade after irradiation, but a growing excess was observed at later intervals, with the relative risk after 30 years reaching 4.1 (p < 0.05) (Bo85). A similar excess, which also arose in the second decade after treatment, was observed in a smaller series of women treated with radiation for carcinoma of the uterine cervix (ratio of observed to expected = 20/8.8) (Kl82). Suggestive evidence for an excess of rectal cancers also has been reported in women treated with radiation for benign pelvic disease (Br69) but not in the Japanese A-bomb survivors (Pr87a). As yet, however, there is no evidence of such an excess in patients treated with radiation for ankylosing spondylitis (Da87). In ICR and CF1 mice, the incidence of rectal carcinoma was observed to be increased by intensive x-irradiation of the pelvis, rising from zero at a dose of 20 Gy to 95% in ICR mice exposed to 60 Gy delivered in three exposures and to 70% in CF1 mice exposed to 40 Gy delivered in two exposures (Hi77). In C57Bl mice, the induction of rectal carcinomas by intensive
RADIOGENIC CANCER AT SPECIFIC SITES 303 x-irradiation has been observed to be enhanced by the administration of the radiosensitizer midonidazole shortly before irradiation (Ro78). In rats, similar tumors have been reported to be induced by localized exposure to negative pi- mesons (Bl80). Summary The data imply that the risks of cancer of the colon and cancer of the rectum can be increased by intensive irradiation in humans and laboratory animals; however, the shapes of the dose-incidence curves and the risks per unit dose are highly uncertain. In the Japanese A-bomb survivors, the dose- dependent excess of colon cancers corresponds to a relative risk of 1.85/Gy, or 0.81 fatal cases per 104 PYGy, and was not evident until more than 15 years after irradiation. LIVER Introduction Evidence of radiation-induced liver cancer comes mainly from observations on human populations and laboratory animals with high intrahepatic concentrations of radionuclides. Human Studies Follow-up studies of patients in West Germany, Portugal, and Denmark have noted the occurrence of increased numbers of liver cancers, particularly angiosarcomas, bile duct carcinomas, and hepatic cell carcinomas, many years after intravascular injection of Thorotrast, an x-ray contrast medium containing colloidal 232ThO2. From the results of these studies, a linear lifetime risk coefficient of 300 liver cancers/104 person-Gy of alpha radiation was estimated by the BEIR III Committee (NRC80); however, the extent to which chemical toxicity of Thorotrast may have influenced the risk was not known. More recently, the data from patients who received Thorotrast, including those in West Germany, Portugal, Japan, Denmark, and the United States, were extensively re-reviewed and reanalyzed by the BEIR IV Committee (NRC88). The follow-up of the West German patients was the largest of these studies, involving 5,159 Thorotrast-exposed and 5151 control subjects (Va84) (NRC88), in which 347 cases of liver cancer (primarily cholangiocarcinomas and hemangiosarcomas) were observed in the Thorotrast-exposed group and 2 cases of liver cancer in the control group. Latency ranged from 16 to more than 40 years. The average alpha dose to
RADIOGENIC CANCER AT SPECIFIC SITES 304 the liver was calculated to range from 2 to 1.5 Gy. Based on an assumed latent period of 20 years, the lifetime cancer risk from alpha irradiation of the liver was estimated to be 300 cancers/104 person-Gy (NRC88). Similar lifetime risks were calculated on the basis of the Japanese and Danish studies. If a 10-year instead of a 20-year latent period had been assumed, the risk estimates would have been reduced by about one-third. In the BEIR IV report, it was noted that the risk estimates applied only to intravascularly administered Thorotrast. The same radionuclide administered by different routes, or other radionuclides, could cause different patterns of dose distribution and thus different risks of liver cancer. Experiments in animals have demonstrated that the chemical toxicity of Thorotrast contributes little to the induction of liver cancer (NRC88). The follow-up of Japanese A-bomb survivors covering the period 1950-1982 (Pr87a) is the first in which cancers of the liver, gall bladder and bile ducts were reported separately from those of other organ sites. The dose-trend test for liver cancer was suggestive of a significant response (p = 0.05), there being 59 deaths due to primary cancers of the liver and intrahepatic bile ducts, 19 of which occurred in the unexposed group and 40 of which occurred in the exposed group. The estimated relative risk in terms of the T65 dosimetry was 1.35 (90% confidence interval, 0.98-2.04), and the excess risk was 0.08 deaths/104 PYGy (90% confidence interval, 0.00-0.20). These results are based on death certificate diagnoses for which both poor detection and poor confirmation of liver cancer have been observed. A study of a smaller number of histologically diagnosed cases of liver cancer for the period 1950-1980 found no relationship between radiation dose and the incidence of primary liver cancer for persons in either Hiroshima or Nagasaki or for both cities combined (As82). Additional information on the occurrence of liver cancer in the Japanese A- bomb survivors for the years 1950-1985 has been reported by Shimizu et al. (Sh88) who discussed the questionable significance of an increase in mortality from liver cancer among survivors. On the basis of the 420 liver cancer deaths that were not otherwise specified, the relative risk using the DS86 dosimetry was estimated to be 1.26 (90% confidence interval, 1.05-1.53) at 1 Gy, and the excess risk 0.45/104 PYGy (90% confidence interval, 0.09, 0.88). Such estimates are complicated by the inclusion of metastases of other cancers to the liver. For the 77 cases of confirmed primary liver cancer, the relative risk of 1.12 was not statistically significant (90% confidence interval, 0.87-1.71). In 14,106 patients who received a single treatment course of x rays for ankylosing spondylitis and were followed through 1982, a total of 6 liver cancers were observed more than 5 years after exposure, with 2 cases between 5 and 25 years and the other 4 cases more than 25 years after exposure; the observation of the 6 liver cancers was not significantly
RADIOGENIC CANCER AT SPECIFIC SITES 305 different from the expected number, 5.44 (Da87). The x-ray dose to the liver in this study population was estimated to be 1.63 Â± 1.26 Gy (Le88). In a study of second cancers arising after radiation treatment of the pelvic region for cancer of the uterine cervix, Boice et al. (Bo85) (Bo88) found no evidence of radiation-induced liver cancer (ratio of observed to expected cancers 19/20). Animal Studies Much of our knowledge of the induction of liver cancer from intra-hepatic radionuclides is derived from studies in laboratory animals. As noted in Chapter 1, not all species have prolonged hepatic retention of actinide or lanthanide radionuclides. The BEIR IV report briefly discussed the prolonged retention of actinide radionuclides in the livers of Chinese hamsters, deer mice, grasshopper mice, and beagle dogs, compared with the shorter retention half- times seen in laboratory mice and rats (NRC88). Prolonged retention times increase the radiation doses received by the liver and increase the carcinogenic effects observed in some species. In a series of life-span studies in which beagles received a single inhalation exposure to monodisperse aerosols of 238PuO2, late-occurring cancers were prevalent findings in the skeleton, liver, and lungs (Gi88). Almost all of the cancers found in the liver and skeleton were considered to have been induced by the 238Pu that was absorbed after particle fractionation in the lung. These results demonstrate that inhaled as well as injected alpha-emitting radionuclides can cause liver cancers under appropriate conditions. In regard to low-LET irradiation, primary liver cancers, principally hemangiosarcomas, bile duct carcinomas, and hepatocellular carcinomas, were prominent long-term effects of chronic beta irradiation of the liver in dogs that had inhaled 144CeCl3 or had been injected intravenously with 137CsCl (Mu86). The estimated lifetime risk of liver cancer in these dogs was 90 liver cancers/104 dog Gy. On the basis of this value and other information on the induction of liver cancers by alpha-emitting radionuclides in humans and in dogs, Muggenburg et al. (Mu86) estimates the lifetime risk for humans exposed to internally deposited beta-emitters to be 30 liver cancers/104 person-Gy). Summary Follow-up studies of Thorotrast-exposed patients have provided conclusive evidence of carcinogenic effects on the liver from chronic alpha irradiation by internally deposited 232Th and its radioactive decay products. In laboratory animals, likewise, prolonged hepatic retention of actinide
RADIOGENIC CANCER AT SPECIFIC SITES 306 and lanthanide radionuclides has produced similar carcinogenic effects on the liver, through chronic irradiation by alpha-emitters and beta-emitters. Collectively, the data indicate that the lifetime risk of liver cancer from Thorotrast is about 300 liver cancers/104 person Gy and that the risk from chronic beta irradiation may be about 10 times lower. SKELETON Human Data Low-LET Irradiation Among 14,106 persons given a single treatment course of x rays for ankylosing spondylitis, four bone cancers were observed at times greater than 5 years after exposure (ratio of observed to expected bone cancers, 4/1.36), corresponding to a relative risk of 2.95 ( p < 0.05) (Da87). The mean doses received by various parts of the skeleton were estimated to be 9.44 Â± 6.05 Gy to the pelvis, 4.41 Â± 3.42 Gy to the ribs, 14.39 Â± 9.66 Gy to the spine and 0.48 Â± 0.61 Gy to other parts of the skeleton (Le88); however, the doses received by individual subjects are not available. Furthermore, each individual's total radiation dose was received in successive fractions over several weeks, with rather large dispersions in the numbers of fractions and numbers of weeks. Also, the dispersion of the dose at each site in each patient was usually quite large, as was the dispersion of doses among patients. The lack of information at this time on the doses to bone in each individual, as well as the small number of extra bone cancers seen in this study, precludes using the data to estimate a risk coefficient for bone cancer. In a long-term follow-up study on the occurrence of second cancers following radiation treatment of women for cancer of the uterine cervix, an ostensibly radiation-related distribution of bone cancers was observed, with 55% of the bone cancers in the exposed group occurring in the pelvis, compared with 15% in a control group (Bo88). The overall relative risk was 1.3, rising threefold for bone doses greater then 10 Gy. The possible induction of bone cancer from medically related x-or gamma- radiation was examined by Kim et al. (Ki78), who reported 27 cases of bone sarcoma that were judged to have been induced by radiation. The latent periods for the tumors that occurred in the irradiated field ranged from 4 to 27 years, with a median of 11 years. No bone sarcomas were seen after treatment doses of less than 30 Gy given over a period of 3 weeks. Similarly, Yoshizawa reviewed 262 cases of skeletal cancer attributed to therapeutic external irradiation (Yo77b); however, these cases do not lend themselves to analysis of the risks of bone cancer.
RADIOGENIC CANCER AT SPECIFIC SITES 307 In the most recent report on the Life Span Study population of A-bomb survivors, covering the period 1950-1985, bone cancer was reported to show no statistically significant increase with dose (Sh88). Likewise, in a long-term follow-up study of 339 British radiologists who began their practice prior to 1921, no statistically significant excess of deaths from bone cancer was found (Sm81). From the radiotherapy studies described above, it can be seen that large doses of acutely delivered x-or gamma-radiation can produce bone cancer; however, the uncertainties in dosimetry preclude the estimation of dose- response relationships for low doses of low-LET irradiation. Therefore, other approaches, including studies of the effects of internally deposited alpha emitters in human populations or studies of the comparative carcinogenic effects of alpha irradiation and beta irradiation in laboratory animals, are required. High-LET Irradiation With internally deposited 224Ra, 226Ra or 228Ra, the main long-term biological effect has been observed to be the induction of bone cancer, primarily osteosarcoma (NRC80, NRC88, UN86, Va86). The internal deposition of 226Ra in dial painters, chemists, and medical patients has resulted in life-long alpha irradiation of the bone volume, whereas with 224 Ra administered medically the alpha irradiation has been of relatively short duration (because of the short half-life of 224Ra) and delivered mainly to endosteal bone surfaces. Risk coefficients for alpha-radiation-induced bone cancer given in the BEIR III report (NRC80) included a linear function based on the 224Ra data, (i.e., 27 Ã 10â6 sarcomas/person rad) and a dose squared function based on the 226Ra and 228Ra data (i.e, 3.7 Ã 10â8 sarcoma/person-rad2). Recognizing the lack of available human data from which to estimate a risk coefficient for low-LET radiation-induced bone cancer, the BEIR III Committee divided the risk factors for high-LET radiation, by an estimated relative effectiveness factor of 20, to estimate the risk coefficient for low-LET irradiation of the skeleton. In this way, a lifetime linear risk coefficient of 1.4 Ã 10â6 bone sarcomas/person-rad and a dose-squared risk coefficient of 9.2 Ã 10â 11 bone sarcoma/person rad2 were derived. No direct evidence for the derivation of this relative effectiveness factor was cited, except that it corresponded to the ratio of currently used quality factors for alpha emissions, as compared with beta emissions, from internally deposited radionuclides. In discussing these risk coefficients, the BEIR III Committee noted that the shapes of the dose-response relationships for
RADIOGENIC CANCER AT SPECIFIC SITES 308 radiation-induced bone cancer were uncertain and that a quadratic dose- response function might be more appropriate than a linear function for low-LET radiation because of the sparsely ionizing nature of the radiation. The long-term follow up studies of persons with elevated body burdens of 224Ra, 226Ra, or 228Ra were examined again in detail in the BEIR IV report (NRC88). Because of the short (3.62-day) radioactive half-life of 224Ra, alpha radiation is confined primarily to the sites of initial deposition on bone surfaces. In 224Ra-injected subjects, bone cancers have been seen at times ranging from 3.5 to 25 years after initial exposure, with a peak occurrence at about 8 years. Several different dose-response functions for internally deposited 224Ra, and their associated uncertainties, were discussed by the BEIR IV Committee, and the lifetime risk of osteosarcoma was estimated to be about 2 Ã 10â2/person-Gy for a well-protracted exposure (NRC88). A range of intake-response or dose-response functions for internally deposited 226Ra and 228Ra was also examined by the BEIR IV Committee (NRC88). The lifelong presence of 226Ra in the skeleton after deposition affects both the doses that are received and the biological responses that are observed. In contrast to the results for 224Ra, the alpha dose from 226Ra continues to accumulate throughout life, and bone cancers have occurred over a much longer period of time after initial deposition of 226Ra (up to 63 years after the first exposure). Because of the long-continued alpha irradiation of the skeleton, the ongoing biological processes of remodeling of bone tissues, and the associated nonuniform local deposition and redeposition of 226Ra, the BEIR IV Committee recommended the use of intake-response instead of dose-response functions. No estimate of the lifetime risk as a function of the dose in person-Gy comparable to the estimate given above for 224Ra was given for 226Ra. Studies in Laboratory Animals Because of the sparseness of human data for the risks of bone cancer from low-LET irradiation, studies with laboratory animals provide another means of estimating these risks. Most of the currently available data on the long-term biological effects of low-LET irradiation in laboratory animals have been derived from chronic beta irradiation by internally deposited beta-emitting radionuclides such as 32P, 45Ca, and 90Sr (Go86a). Of these radionuclides, 90Sr has been studied in the greatest detail because of its relatively long-term persistence in fission product mixtures that may be released into the environment. A broad range of species of laboratory animals has been used in these studies. The most extensive of these have been the life-span studies of dogs exposed to 90Sr in relatively soluble form by intravenous injection, at the
RADIOGENIC CANCER AT SPECIFIC SITES 309 University of Utah, Salt Lake City; by ingestion, at the University of California at Davis; and by inhalation, at the Lovelace Inhalation Toxicology Research Institute, Albuquerque, New Mexico. Parallel studies of 226Ra injected 1 or 8 times into young adult beagle dogs have been conducted at the University of Utah and the University of California at Davis, respectively, to provide a direct link between the biological responses seen in dogs and the human data base (Go86a). Mays (Ma80) examined the relative effectiveness of chronic alpha and chronic beta radiations by comparing the average absorbed doses of alpha and beta radiations required to produce equal incidences of bone cancer in dogs. When the incidence of bone cancer induced by 226Ra was plotted against the average absorbed dose of alpha radiation, an approximately linear relationship was obtained, whereas the plot of 90Sr-induced tumor incidence was concave upward at higher doses. It was observed that the effectiveness of alpha irradiation relative to that of beta irradiation increased as the dose decreased, reaching a value of 26 at an incidence of 8.7%. A similar comparison of mice injected with 226Ra and with 90Sr gave a relative effectiveness factor of 25 at an incidence of 7.7%. The increase in relative effectiveness resulted primarily from the decreased response seen at lower doses in 90Sr-exposed dogs. No bone tumors were seen in dogs that received an average skeletal dose of 6 Gy or less from 90Sr. Experiments designed to study the long-term effects of 90Sr ingested daily in food, as compared with the effects of eight fortnightly injections of 226Ra (Go86b), also demonstrated the reduced carcinogenic response of the dog skeleton to chronic beta irradiation as compared with chronic alpha irradiation. Raabe et al. (Ra83), using a log-normal dose-response model, reported that the relative effectiveness of these two chronic exposure modalities was of about equal potency when the average skeletal dose rate was about 0.1 Gy/day, but that the relative effectiveness of 90Sr at lower dose rates decreased in comparison with that of 226Ra, eventually reaching a point at which 90Sr was only 1/30 as effective as 226Ra. The nonparallel nature of the dose-response relationships seen for 90Sr and 226Ra was consistent with similar observations made at the University of Utah (Ma80). A proportional hazards model also has been used to compare the life-span carcinogenic response from 90Sr inhaled by beagle dogs (Mc86, Gi87) with the carcinogenic responses from inhalation of 238PuO2 or intravenous injection of 90Sr, 226Ra or 239Pu (Me86). The results of this comparison show that the relative risk coefficients for bone cancer from 90Sr are the same in dogs that received one exposure to inhaled 90Sr or injected 90Sr and were about 5, 48, and 30 times lower than the relative risk coefficients for injected 226Ra, injected 239Pu, or inhaled 238PuO , respectively. Collectively, these studies demonstrate 2 that the risks per Gy of bone cancer from
RADIOGENIC CANCER AT SPECIFIC SITES 310 internally deposited 90Sr are appreciably less than those from internally deposited 226Ra (NRC88). Summary 1. Currently available information on persons who have received x ray or gamma radiation delivered therapeutically for medical purposes indicates that large doses of low-LET radiation can produce bone cancer. Skeletal dosimetry in these cases is too uncertain to provide precise information about the dose-response relationship. 2. The data currently available from the study of Japanese A-bomb survivors provide no evidence of an excess of bone cancer resulting from low-LET irradiation at levels in the 0 to 4 Gy range. 3. The most definitive dose-response relationships for radiation- induced bone cancer come from studies of persons with elevated body burdens of the alpha-emitting radionuclides 224Ra and 226Ra, in whom the lifetime risk of bone cancer from internally deposited 224Ra has been estimated to be about 2 Ã 10â2/person Gy. 4. Studies of the carcinogenic response of the skeleton to internally deposited 90Sr in beagle dogs have demonstrated a nonlinear, concave upward, dose-response relationship for chronic beta irradiation of the dog skeleton. Parallel studies with 226Ra in beagles dogs have demonstrated chronic beta irradiation from 90Sr to be less effective than chronic alpha irradiation from 226Ra, by a factor of up to 25, the carcinogenic effectiveness of chronic beta irradiation being greatly reduced at low doses and low dose rates. BRAIN AND NERVOUS SYSTEM Radiation has been observed to increase the incidence of tumors of the nervous system in humans and laboratory animals. The human data are derived from studies of populations exposed prenatally to diagnostic x radiation and populations exposed postnatally to therapeutic x radiation or A-bomb radiation (La86, Ku87). In a 1% sample of 734,000 children exposed to diagnostic x radiation in utero, MacMahon (Ma62) observed an excess mortality from cancer of the central nervous system, amounting to approximately 6.3 deaths/10 4 PYGy (80% confidence limits, 1.1-17.2) after adjustment for birth order, religion, maternal age, sex, and pay status of parents. A similar risk estimate (6.1 excess deaths/104 PYGy) was subsequently reported by Bithell and Stewart (Bi75), based on their finding of a history of antenatal irradiation in 1,332 British children dying of malignant central nervous system tumors before the age of 15. Although later studies of children exposed prenatally to x
RADIOGENIC CANCER AT SPECIFIC SITES 311 rays (Di73, Mo84) or A-bomb radiation (Ja70) failed to confirm an excess number of central nervous system tumors, the results of such studies were not statistically inconsistent with the previous risk estimates (La86). A smaller excess has been reported in persons given radiotherapy to the scalp for tinea capitis in childhood. In one series of such persons, including 2,215 patients who were followed for an average of 25 years after a dose to the brain that was estimated to average 1.4 Gy, 8 brain tumors (3 malignant) were observed, versus 1.4 expected (none were observed in 1,413 controls), corresponding to an excess of 1.0 Â± 0.4 cases/104 PYGy, or an excess relative risk of 3.4 Â± 1.5%/cGy (Sh76, La86). The tumors included glioblastomas as well as meningiomas. In another series, which included 10,842 persons followed for an average of 22.6 years after receiving an estimated brain dose of 1.21-1.39 Gy, 21 brain tumors (10 malignant) were observed, versus 6 (4 malignant) in an equal number of controls; 9 other central nervous system tumors (2 malignant) were also observed, versus none in the controls (Ro84a). From these observations, the excess of brain tumors has been estimated to approximate 0.71 Â± 0.20 cases/104 PYGy, and the total excess of all central nervous system tumors has been estimated to approximate 1.09 Â± 0.24 cases/104 PYGy (La86). Other patients in whom the risk of central nervous system tumors has been observed to be increased after therapeutic irradiation include a series of 3,108 persons who were followed for an average of 22 years after x-ray treatment of the head and neck during childhood, in whom 14 intracranial tumors (6 malignant) were observed (Co78). On the basis of an estimated average midbrain dose of 0.8 Gy, a latent period of 5 years, and an expectation of about 1.6 intracranial tumors, the excess of intracranial tumors in this series has been calculated to approximate 2.9 Â± 0.9 cases/104 PYGy, or 9.7 (Â± 2.9)%/cGy (La86). Also among 592 children treated with irradiation of the cranium for acute lymphatic leukemia (ALL), the relative risk of subsequent brain tumor was reported by Rimm et al. (Ri87) to approximate 20 (0.25 cases/104 person- year); and in a comparable series of 468 children, the relative risk was reported by Albo et al. (Al85) to approximate 226 (ratio of observed to expected cases, 9/0.0398). Similarly, 3 of 904 patients treated with radium implants in the nasopharynx and followed for an average of 25 years after treatment were observed to develop brain tumors, versus none in 2,021 controls (Sa82); on the basis of an estimated average dose to the brain of 0.15-0.4 Gy and an expectation of 0.57 brain tumors, the excess in this series has been calculated to range from 3.4 Â± 2.4 to 9.0 Â± 6.4 cases/10 4 PYGy for doses of 0.4 Gy and 0.15 Gy, respectively (La86). Likewise, an excess of intracranial tumors (22 observed after a latency of 5 years, versus 14.03 expected) has been observed in a series of 14,106 patients treated with spinal irradiation for ankylosing spondylitis and followed for up to 48 years after treatment (Da87); if the
RADIOGENIC CANCER AT SPECIFIC SITES 312 average dose to the affected part of the brain in such patients is assumed to have been less than 0.15 Gy, the excess of intracranial tumors can be calculated to exceed 5.8/104 PYGy. In atomic-bomb survivors no excess of intracranial or other central nervous system tumors has been evident thus far (La86, Pr87a). In pioneer radiologists who entered practice in the United States during the 1920s, mortality from brain cancer was about 3 times higher than that in other medical specialists (Ma75); however, the numbers that were exposed and the doses that they received are not known. Hence, the magnitude of the risk per unit dose cannot be estimated (La86). An association between intracranial meningiomas and previous medical or dental radiography has been suggested by the results of a case-control study of such tumors diagnosed during 1972-1975 in women of Los Angeles County (Pr80). The strongest association (relative risk, 4.0, p < 0.01) was with a history of exposure to full-mouth dental x-ray examinations at a young age (<20 years). Radiation dose estimates were not reported. Other series of patients in whom an excess of meningiomas occurred after previous localized irradiation have been reported by Soffer et al. (So83) and Rubinstein et al. (Ru84). In monkeys exposed acutely to x rays, neutrons, or protons, the incidence of glioblastoma multiforme has been observed to be increased at high doses; for example, two of four rhesus monkeys exposed to 15 Gy of x radiation delivered in a single exposure to the head alone developed such tumors during the subsequent decade (Wa82). Similarly, the incidence of such tumors was greatly increased in Macaca mulatto monkeys exposed to whole-body fission neutron irradiation (Br81) or whole-body 55-MeV proton irradiation delivered in a single exposure (Yo85). In the latter, the incidence rose from zero at a surface dose of 2 Gy or less to about 30% at a surface dose of 6 Gy and 33% at a surface dose of 8 Gy. In addition to glioblastomas, other types of intracranial tumors (ependymomas, meningiomas, and pituitary adenomas) also occurred with increased frequency, the total excess of intracranial tumors greatly exceeding that of all other radiation-induced neoplasms combined. The incidence of intracranial tumors in rats has been observed to be increased after acute whole-body x-irradiation at doses in the range of 1-2 Gy (Kn82). A brain tumor (oligodendroglioma) was also observed in 1 of 11 dogs exposed to 1.33 Gy of fast neutrons delivered to the head alone (Zo80). Summary Radiation has been observed to increase the incidence of nervous system tumors in human populations and laboratory animals. The tumors include malignant as well as benign growths of the brain, meninges, and
RADIOGENIC CANCER AT SPECIFIC SITES 313 peripheral nerves. The increase has been evident after irradiation in childhood at doses of less than 1-2 Gy. Although the dose-incidence relation is uncertain, the data indicate the brain to be relatively sensitive to the carcinogenic effects of radiation. OVARY There is a wealth of experimental data on the induction of ovarian tumors in mice by ionizing radiation. It has been demonstrated that all nonreproductive cells (i.e., all cells other than oocytes) in the ovary are susceptible to the carcinogenic effects of radiation; that even low levels of acute exposure (e.g., as little as 50 rad) can cause a significant increase in the rate of tumor induction; and that the dose-response curve for radiation-induced mouse ovarian tumors is generally S-shaped, depending on the total dose, dose rate, LET of the radiation, and the strain and age of the mice exposed (Up70 and Ul79). The induction of these growths is ascribed to abnormal gonadotrophic stimulation incident to the cessation of ovulatory cycles resulting from radiation-induced sterilization (Cl59). The strongest evidence of an association between radiation and cancer of the human ovary comes from the combined cohort of Japanese atomic bomb survivors in whom the relative risk at 1 Gy (using the new dosimetry) was estimated to be 2.33/Gy (Sh88) In women with cancer of the uterine cervix who were treated with radiation (Bo85), the relative risk of ovarian cancer increased significantly over time, reaching a (nonsignificant) high of 3.4 (3 observed versus 0.9 expected cases) in those whose first treatment had occurred 30 or more years previously. Overall, the risk of ovarian cancer was significantly decreased (p < 0.001) (relative risk, 0.7); however, when only those women who were followed for at least 10 years after radiotherapy were considered, the risk was similar to that seen in the general population (i.e., 70 cases observed versus 76 expected cases). The significance of the increasing trend over time is likely to be due in large part to the marked reduction in risk seen the first 10 years after radiotherapy, which, in turn, may have arisen because of the frequent incorporation of bilateral oophorectomies in the treatment regimen. The results from the latest update (Da87) of the Court-Brown and Doll cohort (Co65) of patients with ankylosing spondylitis offer little additional support for the presumed relationship between ovarian cancer and exposure to ionizing radiation. The overall mortality risk among patients 5 years or more after radiotherapy is slightly below the expectation based on the experience of the national population (i.e., 5 observed versus 5.37 expected deaths). As was noted above in the discussion of uterine cancer deaths
RADIOGENIC CANCER AT SPECIFIC SITES 314 in this cohort, the results need to be interpreted cautiously since they are based on such a small number of cases. UTERUS Epidemiologic evidence of an association between irradiation and uterine cancer has been based in the past primarily on the Tumor Registry data from A- bomb survivors of Hiroshima (but not those of Nagasaki), which gave some indication of a linear trend in cervical cancer cases in groups exposed to 10+ and 50+ rads (i.e., p values of 0.06 and 0.09, respectively), and on data from the Smith and Doll (Sm76) cohort of 2,068 women who had undergone pelvic irradiation for benign uterine bleeding. In the latter study, 16 deaths from uterine cancer were observed among the patients 5 or more years after treatment, while only 10.3 deaths were expected (p = 0.08). The most recent analysis of the mortality experience of the combined cohorts of A-bomb survivors from Hiroshima and Nagasaki (Sh87) provides the strongest evidence of a potential relationship between radiation exposure and uterine cancer. Using the new dosimetry calculations, Shimizu and coworkers estimated that the relative risk for cancer of the cervix uteri and uterus was 1.22, which represented a suggestive but not significantly increased risk (i.e., p = 0.07). In a study of the occurrence of second primary tumors among 182,040 women treated for cervical cancer (Bo85), an overall deficit in the number of cancers of the uterine corpus was observed among patients who received radiotherapy, in comparison with the expectated number based on general population rates. A total of 133 cases of cancer were observed in this subgroup, versus an expected value of 215, giving a significantly reduced (p < 0.001) relative risk of 0.6. However, when attention was restricted to those radiation- treated patients who were followed for 10 years or more, the relative risk returned to a nearly normal or baseline value (84 observed versus 86 expected cases) and was no longer significant. Indeed, the trend in the relative risk over 5- year intervals since the administration of treatment was significantly increased (p < 0.001), primarily because of the significant reduction in risk seen in the first 10 years after radiation exposure. Since many of the radiation-treated patients also had a hysterectomy as part of their treatment regimen, the authors speculated that use of general population rates to predict the expected number of uterine cancers may actually lead to an overestimation of risk. As a result, they concluded that "â¦an RR of nearly one in all women followed for more than 10 years probably corresponds to a substantial excess in those women with intact uteri and is likely associated with the prior radiotherapy." No excess was observed in recent case control analysis of these data (Bo88).
RADIOGENIC CANCER AT SPECIFIC SITES 315 Data from the recently updated study of long-term mortality in patients treated for ankylosing spondylitis with a single treatment course of x rays (Da87) provide little additional insight into the potential relationship between radiation exposure and uterine cancer. Among patients who died in less than 5 years, 5.0-24.9 years, and 25 or more years, respectively, after the first treatment, uterine cancer relative risks of 0.00 (0 observed versus 1.24 expected deaths), 1.15 (5 observed versus 4.35 expected deaths), and 0.65 (1 observed versus 1.54 expected deaths) were seen. The study's authors qualified these negative findings to some extent by noting that their cohort of patients with ankylosing spondylitis provided little information on radiation-based cancer of the uterus, since so few women were included in the study population. Summary Taken collectively, the new data that have been accumulated since the BEIR III report (NRC80) still do not resolve the question of the potential association between radiation exposure and uterine cancer. TESTIS Relatively little information on the possible association between ionizing radiation and testicular cancer is available, particularly with reference to potential human risk. In the BEIR III report (NRC80), testes were included as a site or tissue in which radiation-induced cancer has not been observed. Darby et al. (Da85) specifically addressed the question of radiation-induced testicular cancer in a study involving a parallel analysis of cancer mortality among the Japanese A-bomb survivors and patients with ankylosing spondylitis in the United Kingdom who had been given x-ray therapy. A comparison of observed and expected testicular cancer deaths among these two study populations revealed no observed deaths in either group. In another recent analysis (Sh88), of the Japanese A-bomb survivors, testicular cancer was not among the site- specific cancers that showed a significant increase in occurrence with dose (DS86 system). No excess cases of testicular cancer have been identified in other epidemiologic studies to date. Thus, the limited data available for humans suggest that the human testes may be relatively resistant to cancer induction by exposure to ionizing radiation. There are a number of studies in the experimental literature that indicate that there is some association between whole-body or site-specific radiation exposure and the induction of testicular cancer, particularly interstitial cell tumors in rats (see, e.g., Wa86). It has been postulated that these tumors occur in part because of a probable hormonal imbalance from
RADIOGENIC CANCER AT SPECIFIC SITES 316 radiation damage to the testes. At present, however, the experimental data have not been paralleled by epidemiologic evidence. Summary The existing data imply that the human testis is relatively insensitive to the carcinogenic effects of radiation. PROSTATE Introduction In the 1980 BEIR report (NRC80), the prostate was recorded as an organ with little or no sensitivity to the induction of cancer by radiation, since no epidemiologic evidence suggesting radiogenic prostate cancer was available at that time. For the same reason, the 1985 National Institutes of Health report (NIH85) on the probability of causation did not include prostate cancer as a radiogenic neoplasm. In the interim since these reports, data suggesting a weak association between prostatic cancer and radiation have been reported, as summarized below. Human Studies Japanese A-Bomb Survivors In the 1987 RERF report on the cancer experience of the Japanese A-bomb survivors (Pr87a), prostate cancer was considered separately for the first time. In the survivor population, mortality from carcinoma of the prostate was uncommon; only 51 deaths were reported. A causal association with radiation dose was not statistically significant, the average relative risk under the T65 dosimetry being 1.27 at 1 Gy. However, the combination of a moderate excess risk of 0.14 excess prostate cancer deaths/104 PYGy together with low background mortality rates yielded an attributable risk of 11.7%. The data suggested that the relative risk of death from prostate cancer may have increased with time. However, the time trend was not statistically significant, even though the average increase in the excess relative risk of 37.5%/year was one of the largest estimated among the cancers considered. Additional uncertainties were introduced due to inaccuracies of death certificate diagnoses of prostate cancer; confirmation rates were 39%, while the detection rate was only 21%. In the most recent Life Span Study cohort report (Sh88), prostate cancer mortality in the years 1950-85 showed no significant increase with increasing dose. In the DS86 subcohort, which included 75,991 exposed
RADIOGENIC CANCER AT SPECIFIC SITES 317 persons, there were 52 deaths from cancer of the prostate; the estimated relative risk at 1 Gy (shielded kerma) was 1.05; the absolute risk was 0.03 excess cancer deaths/104 PYGy; and the attributable risk was 1.95%. Ankylosing Spondylitis Patients Darby et al. (Da85) included prostate cancer patients in the category of those who were treated with x-irradiation to heavily irradiated sites. In the most recent follow-up on cancer mortality, to January 1, 1983, covering 11,772 men with ankylosing spondylitis given a single course of x ray treatment during the period 1935-1954 (152,979 person-years at risk), there were 21 observed versus 18.15 expected prostate cancers (ratio of observed to expected cases, 1.16) during the period â¤ 5 years after the first treatment (Da87). An early excess (ratio of observed to expected cancers 4.0/1.31 = 3.04, p < 0.5) was limited to the first 5 years after treatment. Thus, the relative risk for cancer of the prostate was the third highest in this series, and the risk was significantly increased; but the authors noted that this disease is often confused with ankylosing spondylitis, since it is frequently present with pain in the back due to metastases to the spine (Da87). The mean dose to the prostate was calculated to be 24 rad (0.24 Gy) (Le88). Based on this average organ dose, estimated by using the Monte Carlo method (Le88), the increase in relative risk of prostate cancer for the period 5.0-24.9 years after exposure (ratio of observed to expected cancers, 1.24) was estimated to be 0.66%/rad at 0.01 Gy (1 rad). Nuclear Workers Beral et al. (Be85) have recently reported an increase in the relative risk of mortality from prostate cancer among British nuclear workers, as discussed in Chapter 7 (SMR = 145 for those with 10 or more years of employment). In other groups of nuclear workers also, a nonsignificant elevation of risk has been reported (Sm86). United States Radiologists The early U.S. radiologists are estimated to have had lifetime exposures of 2 to 20 Gy. Their cancer mortality experience has been analyzed by Matanoski et al. (Ma84), who reported standardized mortality ratios (SMRs) for selected cancers among members of the Radiological Society of North America (RSNA), the American College of Physicians (ACP), and the American Academy of Ophthalmology and Otolaryngology (AAOO). During the period 1920-1939, SMRs for prostate cancer were 1.24 for members of RSNA, 1.03 for members of ACP, and 0.81 for members of AAOO. The excess of prostate cancer mortality in radiologists was not statistically significant (p < 0.05). During the period 1940-1969, the SMRs
RADIOGENIC CANCER AT SPECIFIC SITES 318 were 1.01 for members of RSNA and ACP, 0.98 for members of AAOO and OTOL, and 1.40 for members of OPH, values which were not significantly different from unity. Patients Receiving Iodine-131 Therapy In a study (Ho84b) of the incidence of malignant tumors in 4,557 Swedish patients treated with 131I for hyperthyroidism, prostate cancer was analyzed in 726 men, with 6,400 person-years at risk. With doses of less than 370 megabecquerels (MBq) (10 mCi), there were 11 observed, and 10.3 expected cases (relative risk, 1.07). For doses of greater than or equal to 370 MBq (10 mCi), there were 5 observed and 8.0 expected cases (relative risk, 0.63). For the overall doses combined, there were 16 observed, and 18.4 expected cases (relative risk, 0.87). Thus, none of the relative risks were considered significantly different from unity. Animal Data In rats, intensive x-irradiation has been observed to induce carcinoma of the prostate, but only at doses of 10 Gy or more (Wa86). Summary From the studies available thus far, the relative risk of radiation-induced prostate cancer appears to be small. Hence, although the data suggest that there may be a weak association between prostate cancer and radiation, the sensitivity of the prostate to the induction of cancer by irradiation appears to be comparatively low. URINARY TRACT In the 1980 BEIR report (NRC80) the urinary organs, particularly the kidney and urinary bladder, were included among those tissues with definite but low sensitivity to radiation carcinogenesis. Since then substantial new information in support of this conclusion has become available. Japanese A-Bomb Survivors The Radiation Effects Research Foundation (RERF) Life Span Study Report 10 on cancer mortality among A-bomb survivors in Hiroshima and Nagasaki (1950-1982), which was based on T65D dosimetry, indicated a significant dose-related increase in the number of cases of urinary bladder cancer (Pr87a). Among 91,231 exposed survivors with T65D dose estimates,
RADIOGENIC CANCER AT SPECIFIC SITES 319 there were 6,270 cancer deaths during 1950-1982. Death certificate diagnoses for cancers of the urinary tract are moderately accurate. Of the 131 deaths caused by cancers of the urinary tract (bladder, kidney, and other unspecified urinary organs), 95 were urinary bladder, 33 were kidney, and 3 were ureter. For all sites combined, the relative risk at 1 Gy (T65D) was 1.55. The highly significant radiation dose response (p = 0.006) occurred primarily among those who died from bladder cancer (p = 0.003); for those who died from kidney cancer, the positive radiation dose response was slight (p = 0.3). An analysis of radiation-related cancer mortality during 1950-1985 according to the new DS86 dosimetry systems has recently been made on data from 75,991 Hiroshima and Nagasaki survivors, a subgroup of the RERF Life Span Study cohort designated the DS86 subcohort (Sh87, Sh88). In terms of the kerma at the survivor's location, the overall relative risk of death from cancers of the bladder and other urinary organs based on 133 cases was 2.06/Gy (90% confidence interval: 1.46-2.90). This relative risk is 1.49 times the risk calculated according to the T65D dosimetry system on the same cohort (Sh87). There was a suggestive trend (p < 0.10) of an increase in relative risk with time from exposure (Sh88). For bladder and kidney cancer mortality individually, the relative risk values were 2.13 (90% confidence interval, 1.40-3.28) and 1.58 (90% confidence interval, 0.91-2.94) at 1 Gy kerma, respectively (Sh88). The corresponding absolute risks were 0.41 (90% confidence interval, 0.16-0.70) and 0.09 (90% confidence interval, â 0.02-0.26)/104 PYGy, respectively. Sex had little effect on the relative risk of bladder cancer mortality; the male relative risk/female relative risk was 0.9. However, the absolute risk in males was nearly twice that in females; this reflects the higher incidence of bladder cancer in Japanese males than in females that is unrelated to radiation exposure (Sh88). The dose-response relationship had a strong linear component. In terms of organ dose rather than kerma, the relative risk at 1 Gy was 2.27 for bladder cancer mortality (90% confidence interval, 1.53-3.37). Ankylosing Spondylitis Series Darby et al. (Da85) compared radiation-induced cancer mortality among the Japanese A-bomb survivors who received doses of â¥1 Gy with that of British patients who had received x-ray therapy for ankylosing spondylitis. The estimated mean organ doses to the bladder were 1.02 Gy, based on T65D dosimetry, in the Japanese studies and 0.31 Gy in the spondylitics (NRC80). Eleven cases of bladder cancer were observed in each group. The risk estimates were given, with 90% confidence interval given in parentheses. The relative risk for A-bomb survivors was 3.0
RADIOGENIC CANCER AT SPECIFIC SITES 320 (1.7-5.2); for those with ankylosing spondylitis it was 1.6 (0.9-2.7). The excess risk/105 person-years for A-bomb survivors was 5.1 (1.3-8.8), and for spondylitics it was 4.1 (0.6-11.2). Land's review (La86) of the analysis of patients with ankylosing spondylitis indicated that urinary tract cancers occurred in excess numbers among those patients given one course of x-ray treatment. Eight deaths as a result of bladder cancer were seen while 5.1 were expected (p = 0.14). During the first 9 years the mortality was 3 deaths from bladder cancer versus 2.6 expected. The observed and expected values more than 9 years after the beginning of treatment corresponded to an absolute risk of 1.7 excess deaths from bladder cancer/104 PYGy and a relative risk of 1.9/Gy. Radiotherapy for Benign Uterine Bleeding Smith and Doll (Sm76, Sm77) reviewed the experience of English women treated by x radiation for metropathia hemorrhagica (benign uterine bleeding); 3 deaths from bladder cancer were observed 5 years or more after treatment, and 2.15 deaths from bladder cancer were expected (13.5 years of mean follow-up). Wagoner (Wa84) found 10 cases of bladder cancer versus 5.1 expected (p = 0.026) among women who had received radiation therapy for benign gynecologic disorders in Connecticut between 1935 and 1966. Cervical Cancer Series An international study was recently performed on 150,000 women with uterine cervical cancer who had been treated at 1 of 20 oncologic clinics and/or who were reported to 1 of 19 population-based registries. A sample of 4,188 of these women who had received radiation therapy for cervical cancer and who had second cancers and 6,880 matched controls were selected for detailed study (Bo88). A dose of 30 to 60 Gy to the bladder was associated with a relative risk for bladder cancer of 4.0. Women who were <55 years of age at the time of treatment were at especially high risk (relative risk of 16.0). Risk increased with time after exposure; the relative risk was 8.7 among those who survived â¥20 years. Over the range of 30-60 Gy, the relative risk increased significantly with dose (p < 0.001). The effects of smoking and radiation exposure were independent; controlling for smoking did not appreciably alter the radiation risk estimates. Overall bladder cancer risks were greater in the U.S. registry data (Bo88). An analysis of 148 cases of kidney cancer and 285 matched controls revealed an overall radiation-related relative risk of 1.2 (Bo88). The relative risk increased to 3.5 (90% confidence interval, 1.3-9.2) among those who
RADIOGENIC CANCER AT SPECIFIC SITES 321 survived â¥ 15 years after exposure. Women in U.S. registry areas and those exposed when they were <55 years of age had the greatest radiogenic kidney cancer risks (Bo88). Finally, a relative risk of 2.9 was found for cancers of the renal pelvis (23 cases) and ureter (4 cases); the epithelia of these structures are similar to those of the bladder. Cancer after Iodine-131 Therapy A recent report on iodine-131 treatment for hyperthyroidism in Sweden (1951-1975) indicated no increase in bladder and kidney cancers during the 24 years after therapy despite the administration of relatively high doses to the urinary tract (Ho84b). The values for kidney and bladder cancers combined were as follows: (1) males, 9 observed versus 7.5 expected (relative risk, 1.20); (2) females, 17 observed versus 19.0 expected (relative risk 0.89); (3) both sexes, 26 observed, versus 26.5 expected (relative risk 0.98). Conclusions The epidemiologic evidence shows that radiation can cause cancer of the bladder and, to a lesser extent, of the kidneys and other urinary organs. For such effects, the observed dose-response relationship is consistent with a linear nonthreshold function over a broad range of doses, from a few Gy to 60 Gy. Women less than 55 years of age at the time of exposure appear to be at greater risk than older women, and risk appears to increase with time after exposure. The most recent analysis of the A-bomb survivor data using the new DS86 dosimetry indicates a relative risk of 2.3 (90% confidence interval, 1.5-2.4) urinary tract cancer deaths/Gy of absorbed DS86 dose, and an absolute risk of 0.7 urinary tract cancer deaths/104 PYGy (Sh88). If the incidence of urinary tract cancer is assumed to be 3-4 times the rate of mortality from such cancer, the relative risk can be estimated to be Ë6.8-9.1/Gy of absorbed dose. PARATHYROID GLANDS By the time of publication of the BEIR III report, (NRC80), the parathyroid glands were included among those tissues that are susceptible to radiogenic neoplasia. Of the 64 women and 36 men examined more than 25 years after radiotherapy for cervical tubercular adenitis, a total of 11 were found to have parathyroid abnormalities, including 7 with adenomas and 4 with diffuse hyperplasias (Ti77). Of these 11, 7 were hypercalcemic; that is, they had hyperparathyroidism (HPT). None of the 27 subjects who had received <3 Gy had parathyroid dysfunction, while 3/39 (8%) and 4/28
RADIOGENIC CANCER AT SPECIFIC SITES 322 (14%) who had received 3-6 Gy and 6-12 Gy, respectively, had HPT. Four of six (67%) of those who had received >12 Gy had parathyroid disease. The mean time from exposure to diagnosis was 38 years. In a health survey done in Stockholm, Sweden, 15,903 subjects were screened, of whom 58 (44 women, 14 men) had HPT with parathyroid adenomas. None of 58 matched eucalcemic control subjects had had radiotherapy to the parathyroid region, whereas 8 (14%) of the patients with HPT had been irradiated at a mean age of 8.1 years. The dose range was 2-5 Gy, and the mean time to diagnosis was 47 years (Ch78). A total of 17% of 130 patients with HPT at the Henry Ford Hospital had a history of radiation exposure at a mean age of 16 years, whereas only 3% of 400 ambulatory eucalcemic patients had been irradiated (p < 0.025) (Ra80). Similarly, 8/73 (11%) of patients with HPT in a Dutch study had received radiotherapy for benign disease (Ne83). The mean time to diagnosis of HPT was 36 years among those in the Henry Ford Hospital study (Ra80) and 34 years among those in The Netherlands study. Among 200 subjects in the Henry Ford Hospital and other series who were known to have a history of irradiation received during childhood, the prevalence of HPT was 5%, â¥30 fold the prevalence in the general population (p < 0.025) (Ra80). HPT is not always associated with parathyroid hyperplasia or adenoma. In a study of 23 patients who received surgery for nodular thyroid disease and who had no known HPT, five women and three men (35%) had either parathyroid adenoma or hyperplasia (Pr81). All 23 patients had received radiotherapy when they were an average of 16 years of age; the time from irradiation to surgery averaged 33 years. The incidence of thyroid disease among 42 patients with HPT who had a history of receiving irradiation was compared with that in 162 patients with HPT who had not been exposed to radiation (Ka83). Seventy-nine percent of the irradiated patients with HPT had thyroid abnormalities, including 38% with thyroid adenomas and 29% with cancer; in contrast, 43% of patients with HPT with no history of irradiation had thyroid disease and only 10% each had thyroid adenoma and carcinoma (Ka83). Although thyroid disease was not reported to accompany HPT or parathyroid tumors in the Stockholm study (Ch78), an association between radiation-induced diseases of these two glands is a common finding in other series. All 11 patients with parathyroid disease (7 with adenomas, 4 with hyperplasias) of 100 irradiated subjects had thyroid abnormalities, including 2 with thyroid carcinomas and 1 with adenoma (Ti77). Of 73 patients with HPT, 8 were found to have a history of irradiation; 5 had parathyroid adenomas and 3 had hyperplasias. Of these eight patients, six had concurrent nodular goiters and one had struma lymphocytica (Ne83). Possible radiation-related carcinomas of the parathyroid have been
RADIOGENIC CANCER AT SPECIFIC SITES 323 reported sporadically (Ir85). The low frequency of overt cancer may be related in part to the often very long latency of symptomatic parathyroid disease. Mean times from exposure to diagnosis of HPT varied from 30 to 50 years in the different series. In addition, 90% of parathyroid adenomas were accompanied by clinically important HPT and a high percentage of those with radiogenic parathyroid hyperplasia or adenoma had concurrent thyroid disorders. Both conditions commonly require surgical intervention and, hence, removal of the possibly premalignant parathyroid tissue. A review of experimental studies, including those involving mice, rats, guinea pigs, dogs, and monkeys, illustrates that the parathyroid glands do not acutely express radiation damage at the histological level at x-ray doses below 5 Gy (Be72). Doses between 5 and 25 Gy cause modest edema and hyperemia, and higher doses cause severe damage. Late changes include hyperplasia, cyst formation, adenomatous nodules, gross adenomas, and carcinomas. In one experiment, a cumulative total of 12 parathyroid tumors were noted among 80 rats of each sex that were exposed to 250-kVp x rays at 5 Gy when they were 100 days of age; no tumors occurred in 160 unirradiated controls. Ten animals of each sex in the irradiated and control groups were necropsied every 3 months for 24 months after exposure, so the cumulative reported incidence of 8% parathyroid tumors is an underestimate of the true 24-month (or life span) incidence (Be67). Parathyroid neoplasia was also observed to follow irradiation with radioiodide. In one experiment, 185 or 370 kBq of 131I was administered to neonatal rats. Sixty-one percent (28/46) of such animals that survived 15 months were found to have parathyroid adenomas (Tr77); adenomas were found in 31 untreated control rats. Some of the animals had HPT as evidenced by elevated serum calcium. In summary, both experimental and human studies confirm that HPT and parathyroid hyperplasia, parathyroid adenoma, and less frequently, parathyroid carcinoma are late sequelae of radiation exposure. Most parathyroid neoplasms are hyperfunctional, and radiogenic HPT is frequently accompanied by thyroid dysfunction, neoplasia, or both. In humans, the time from irradiation until the time of diagnosis is most commonly at least 30 years. Although the incidence of HPT and neoplasia appears to increase with dose (Ti77), the data are inadequate for quantitative risk estimation. It is clear, however, that parathyroid neoplasia may eventually follow doses in the range of 1 to 5 Gy after exposures that cause little or no acute histopathologic evidence of damage in the glands. The possibility of HPT and parathyroid neoplasia should be considered in those individuals with a history of irradiation of the head and neck, and particularly those with thyroid dysfunction or thyroid nodules.
RADIOGENIC CANCER AT SPECIFIC SITES 324 NASAL CAVITY AND SINUSES In the BEIR III report (NRC80), the induction of cancer of the paranasal sinuses and mastoid air cells by internally deposited 226Ra was described, but no information was presented on the induction of such cancer by low-LET irradiation. In the latest report from the Life Span Study cohort of Japanese A- bomb survivors, Shimizu et al. (Sh88) reported that a total of 44 cases of nasal cancer had been seen during the interval 1950-1985, without any evidence of a dose-response relationship. Similarly, no radiation-induced excess has been evident in the 14,106 patients who received a single course of x-ray treatment for ankylosing spondylitis (Da87), although the mean dose received by the nasal region in such patients was estimated to be 0.47 Â± 0.44 Gy (Le88). These results imply that the nasal cavity is not highly sensitive to low-LET radiation. Conversely, carcinomas of the paranasal sinuses and mastoid air cells have been observed in radium-dial painters and other people exposed to internally deposited 226Ra. The occurrence of these cancers and the underlying radiation etiology are discussed in detail in the BEIR IV report (NRC88). The carcinomas are thought to arise as a result of alpha irradiation of the epithelium from 222Rn gas and radon progeny in the air above the epithelium and from emissions, primarily beta and gamma radiations, from 226Ra and its progeny in the underlying bone. Thirty-five carcinomas of the paranasal sinuses and mastoid air cells have occurred in the 4,775 226,228Ra-exposed subjects, of which there has been at least one determination of vital status (NRC88). The observed latent periods for these cancers have been quite long, ranging from 19 to 52 years (NRC80). In the BEIR III report (NRC80), the lifetime risk of 226Ra-induced paranasal sinus and mastoid carcinomas was estimated to be 64 carcinomas/106 person-rad. The response of the paranasal sinuses to radiation was also demonstrated in patients who received Thorotrast (232Th), an x-ray contrast medium, by antral injection into their sinuses. Fabrikant (Fa64) et al. reported on 10 patients with maxillary sinus carcinomas after maxillary sinus instillations, and Rankow et al. (Ra74b) reported that 13 of 14 patients who received Thorotrast by this route developed cancers of the maxillary and adjacent sinuses. Studies of inhaled or intravenously injected beta-emitting radionuclides in beagle dogs have shown that relatively high local dose rates can occur in the nasal cavity because of patterns of radionuclide deposition and retention (Be79, Bo86). These local accumulations, which result from deposition and retention of the inhaled material or from subsequent translocation of radionuclide to the underlying bone, persist for long periods of time, resulting in the accumulation of high local doses. Such irradiation from
RADIOGENIC CANCER AT SPECIFIC SITES 325 inhaled 144CeCl3, 91YCl3, or 90SrCl2 or from intravenously injected 137CsCl has been observed to induce nasal cavity cancers in dogs (Be79, Bo86a). In summary, there are currently no human dose-response data on cancers of the nasal cavity or cranial sinuses from low-LET irradiation. The only data on the induction of such tumors in human populations pertain to the internally deposited alpha-emitters 226Ra or 232Th and their decay products. The latency for such cancers has been at least 10 years. The induction of nasal cavity cancers in dogs by intensive irradiation from beta-emitting radionuclides implies that such a response might occur in humans under appropriate conditions of low-LET irradiation, but also that the risk would be vanishingly small. SKIN In pioneer radiation workers, carcinomas of the epidermis arising in areas of chronic radiodermatitis were the first radiation-induced neoplasms to be recognized as such (Br36, Ca48, He50). The early literature, consisting largely of case reports, affords no adequate basis for assessing the dose-incidence relationship (Al86). Although some epidemiologic studies of irradiated cohorts have provided dose-incidence data in recent years, such studies have been complicated by the fact that skin cancerâunlike cancer of other sitesâcarries a low mortality and is grossly underreported. The result is that its ascertainment is difficult and uncertain. These limitations notwithstanding, the results of several studies (summarized below) imply that the skin has a higher susceptibility to radiation carcinogenesis than has generally been suspected. Perhaps the most extensive study of radiation-induced skin cancer is an investigation of 2,226 persons who were treated in childhood with epilating doses of 100 kVp x rays to the scalp for tinea capitis and who have since been followed for an average of more than 25 years (Sh84a, Ha83b). The absorbed dose to the scalp in such persons averaged 4.5 Gy (3.3-6.0 Gy), while the dose at the margins of the scalp averaged 2.4 Gy and the dose to the face and neck averaged 0.1-0.5 Gy. In 41 of the 1,680 white members of the cohort, 80 basal cell carcinomas of the skin had appeared, whereas none had appeared in the 546 nonwhite members and only 3 have appeared in a control group of 1,387 nonirradiated white tinea cases (Sh84a). The tumors began to appear about 20 years after exposure and were not limited to the most heavily irradiated parts of the scalp but tended to occur more commonly at the margins of the scalp and in neighboring areas of skin that were not covered by hair or clothing; an excess has been detected even on the cheek and the neck, where the dose is estimated to have been only 0.12 and 0.09 Gy, respectively (Ha83b). The distribution of tumors in relation to the dose suggests, therefore, that the carcinogenic effects of x-irradiation
RADIOGENIC CANCER AT SPECIFIC SITES 326 were enhanced by exposure to ultraviolet (UV) radiation (Ha83b). The cumulative excess increased with the dose of x rays in a manner consistent with linearity, amounting to about 3.3 Ã 10â5 cases/cm2 Person Gy in areas that were exposed to both x radiation and UV radiation, as compared with about 7.1 Ã 10â 6 cases/cm2 Person Gy in areas that were exposed to x rays alone. The average follow-up period for these observations was 25.7 years (Sh84a). Other populations for which risk estimates have been derived include 2,653 persons given x-ray therapy to the chest for enlargement of the thymus gland in infancy, in whom 8 skin cancers were observed to develop later in the irradiated area, versus 3 skin cancers in the corresponding area among 4,791 controls; the 8 cancers included 6 basal cell carcinomas and 2 malignant melanomas (E. Woodward and L. Hempelmann, personal communication). The average dose to the irradiated skin was estimated to approximate 3.3 Gy, and the excess relative risk of cancer in the irradiated area between 10 and 49 years postirradiation was interpreted to amount to 4.8 (Al86), giving an average excess relative risk of about 1.5/Gy. The absolute risk has been estimated to range from 0.66/104 PYGy at doses of less than 4 Gy to 0.32/104 PYGy at doses exceeding 4 Gy (Al86). An excess of skin cancer, primarily basal cell carcinomas of the face, has been observed also in Czechoslovakian uranium miners (Se78). On the basis of an estimated relative risk of 4.5 in this population and a cumulative dose to the affected skin from alpha radiation of approximately 1-2 Gy (20-40 Sv), the relative risk may be calculated to approximate 15%/Sv and the absolute risk 0.95/104 PYSv (Al86). As has been noted previously, however, the excess may not be attributable entirely to radiation, in view of the possible causal contribution that may have been made by arsenic in the uranium ore dust. No excess in numbers of skin cancers has been observed thus far in a number of other irradiated populations that have been studied epidemiologically (Al86). The failure to detect an excess in such studies may be attributable, however, to underascertainment of skin cancers, for the reasons cited above. Radiation carcinogenesis in the skin has been studied experimentally in several species of laboratory animals (UN77). In the rat, a variety of different types of skin tumors occur in response to irradiation, including tumors of hair follicles; in total numbers, the tumors induced by a given dose in the rat exceed those induced by the same dose in the mouse, a species in which the tumors are composed predominantly of squamous cell carcinomas (Bu86). The incidence of tumors in the rat increases as a linear-quadratic function of the dose and reaches a peak at 20-30 Gy of low-LET radiation or 9-10 Gy of high-LET (125 keV/Âµ) radiation. For maximal tumorigenic effectiveness per unit dose, the full thickness of the epidermis must be irradiated in the rat, including the entire hair follicle
RADIOGENIC CANCER AT SPECIFIC SITES 327 (Bu76). Fractionation or protraction of the dose to the skin of rat reduces the cumulative incidence per unit dose with low-LET radiation, but not with high- LET radiation (Bu80). For a given total dose, moreover, the yield of tumors may be increased by exposure to ultraviolet radiation, tumor-promoting agents, or other factors, depending on the particular experimental conditions in question (UN82, Fr86). Summary The risks of basal cell and squamous cell carcinomas of the skin have been observed to be increased by occupational and therapeutic radiation exposure. Although the data do not suffice to define the dose-incidence relationship precisely, the cumulative 30-year excess of basal cell carcinomas in fair- skinned persons treated with x rays to the scalp for tinea capitis in childhood has been observed to increase over a 25-year period with dose in a manner consistent with linearity, corresponding to 7.1 Ã 10â6 excess cases/cm2 Person Gy in areas of skin not exposed to sunlight and 3.3 Ã 10â5 excess cases/cm2 Person Gy in areas of skin exposed to sunlight as well as x rays. LYMPHOMA AND MULTIPLE MYELOMA An increase in the frequency of some forms of lymphoma has been associated with irradiation in humans and laboratory animals (UN77). In humans, the forms include multiple myeloma, in which the tumor cells proliferate primarily in the bone marrow, and non-Hodgkin's lymphoma, in which the tumor cells proliferate primarily in the lymph nodes. Multiple myeloma and non-Hodgkin's lymphoma, like chronic lymphocytic leukemia, are malignancies of B lymphocytes. Of the three diseases, however, only multiple myeloma and non-Hodgkin's lymphoma have been observed to increase in frequency after irradiation in humans. Multiple Myeloma Multiple myeloma has been observed to be increased in frequency by irradiation more consistently than that of any other human lymphoma. A review of the literature by Cuzick (Cu81) showed such an increase in 12 of the 17 irradiated populations analyzed (Table 5-6). In the cohorts tabulated, the pooled excess corresponded to a relative risk of 2.25 (ratio of observed to expected, 50/22.21), with the largest excesses occurring in those exposed to internal- emitters (14 observed versus 3.24 expected cases); however, a deficit of multiple myeloma was reported (3 observed versus 10.17 expected
RADIOGENIC CANCER AT SPECIFIC SITES 328 In A-bomb survivors, mortality from multiple myeloma has been observed at doses as low as 0.5-0.99 Gy (Pr87a), and the relative risk at 1 Gy is estimated to approximate 3.29 (1.67-6.31), corresponding to 0.26 excess fatal cases/104 PYGy (Sh87). For persons exposed to radiation in both Hiroshima and Nagasaki, the relative risk increased with dose in males and females aged 20-59 at the time of bombing but did not become evident until 20 years after exposure (Ic79). As noted elsewhere (Mi86), the data from A-bomb survivors, as well as from other populations, imply that for multiple myeloma the minimal latent period is appreciably longer, the relative risk smaller, and the age distribution later than for leukemia. Although mortality from multiple myeloma has been observed to be
RADIOGENIC CANCER AT SPECIFIC SITES 329 comparably increased (ratio of observed to expected mortality, 9/4.6 = 1.72) in 14,106 patients who were followed for up to 25+ years after radiation therapy for ankylosing spondylitis (Da87), no excess has been evident in 150,000 women who were followed for more than 15 years after radiation therapy for carcinoma of the uterine cervix (relative risk, 0.26) (Bo88). In Hanford nuclear plant workers, mortality from multiple myeloma was observed to be elevated in the 1970s (Gi79) and has been found to remain elevated in a more recent, expanded analysis of the same population (To83). A similar excess has since been reported in workers at two other nuclear installations (Be85, Sm86). No excess, however, has been evident in an early cohort of 27,011 Chinese x-ray workers, in whom the ratio of observed to expected cases is 0/0.5 (relative risk, 0) (Wa88b). Malignant Lymphoma For Hodgkin's disease, the data are reasonably consistent in showing no excess in irradiated populations. For other lymphomas, however, the data are inconsistent. As concerns non-Hodgkin's lymphoma, mortality from this disease has not been increased in A-bomb survivors (Sh87), notwithstanding a previous suggestion to the contrary (An64). Patients treated with radiation for ankylosing spondylitis, however, continue to show increased mortality from the disease (ratio of observed to expected mortality 16/7.14 = 2.24) (Da87). An excess of the disease has also been observed in women who were treated with radiation for benign gynecologic disorders (Wa84) and in women who were treated with radiation for carcinoma of the uterine cervix (relative risk, 2.51; 90% confidence interval, 0.8-7.6) (Bo88). Mortality from lymphosarcoma has been observed to be increased in American radiologists who entered practice in the 1920s and 1930s, when the average occupational radiation levels were higher than they are today. Although such early cohorts showed an increased standardized mortality ratio (2.73) for lymphosarcoma, no excess of this disease or of other lymphomas has been evident in American radiologists of more recent cohorts (Ma81b) or in pioneer Chinese x-ray workers (Wa88b). In laboratory animals, a variety of lymphoid neoplasms can be induced by irradiation (UN77). The best studied of these growths is the thymic lymphoma of the mouse, which, as discussed above (see the Section on parathyroid glands), often terminates as a lymphatic leukemia (Yo86). The dose- incidence curves for experimentally induced lymphomas vary markedly with the lymphoma in question, as well as with species, sex, age at exposure, conditions of irradiation, and other variables (UN86). Paradoxically, the incidence of one such neoplasm, a reticulum cell sarcoma of the mouse,
RADIOGENIC CANCER AT SPECIFIC SITES 330 typically decreases with increasing dose of whole-body radiation (UN77, UN86). Summary The incidence of multiple myeloma has been observed to be elevated after widespread irradiation of the bone marrow in the majority of populations studied to date. In A-bomb survivors, although the excess did not become detectable until 20 years after irradiation, it is now evident at doses as low as 0.05-0.99 Gy and corresponds to a relative risk of 3.29/Gy or to 0.26 fatal cases/104 PYGy. No other form of lymphoma has been consistently observed to be increased in frequency in irradiated human populations. PHARYNX, HYPOPHARYNX, AND LARYNX The review of radiation-induced cancers of the pharynx, hypopharynx and larynx by the BEIR III Committee (NRC80) was based primarily on several small studies of the late effects of therapeutic irradiation of adjacent tissues, such as the esophagus, larynx, thyroid and spine. In the cases of cancer reported, the mean latent periods ranged from 23 to 27 years, and the radiation doses involved were high (fractionated doses of 3,000-6,000 rad delivered over 3 to 6 weeks). In patients with ankylosing spondylitis treated with radiation and observed through January 1, 1970, 3 deaths from cancer of the larynx were observed versus 1.29 expected (ratio of observed to expected cancer deaths, 2.33) and 3 deaths from cancer of the larynx were observed versus 2.25 expected (ratio of observed to expected cancer deaths, 1.33) (Sm82). Neither the excess for the pharynx nor that for the larynx was significant at p < 0.05. Similarly, Darby and colleagues later found no significant excess in deaths from cancer of the pharynx or larynx in ankylosing spondylitis patients and in Japanese A-bomb survivors (Da85). In an update of the Japanese A-bomb survivor data for the period 1950-1985, Shimizu et al. likewise found no excess mortality from cancer of the pharynx, hypopharynx, or larynx, reporting a total of 23 cancers of the pharynx and 46 cancers of the larynx (Sh88). Summary Although cancers of the pharynx and larynx have been observed to arise as a late complication of therapeutic irradiation, after doses in the range of 30-60 Gy, no significant excess of such cancers has been found in the Japanese A- bomb survivors or other populations exposed to doses in
RADIOGENIC CANCER AT SPECIFIC SITES 331 the range below 1 Gy. The sensitivity of the pharynx, hypopharynx, and larynx to radiation carcinogenesis thus appears to be relatively low. SALIVARY GLANDS The incidence of salivary gland tumors has been observed to be increased in patients treated with irradiation for diseases of the head and neck, in Japanese A-bomb survivors, and in persons exposed to diagnostic x radiation. The therapeutically irradiated populations fall primarily into three groups: (1) those treated with x rays to the head and neck during childhood or infancy, in whom the dose to the salivary glands has usually exceeded 1 Gy (Sa60, Ja71, He75, Sc78, Ma81); (2) those treated with x rays to the scalp for tinea capitis in childhood, in whom the dose to the salivary gland is estimated to have averaged about 0.4 Gy (Mo74, Sh76); and (3) women treated with iodine-131 during later middle age, in whom the dose to the thyroid gland is estimated to have averaged about 5.3 Gy (Ho82, La86). The data from the three types of studies are remarkably consistent in yielding an average excess of 0.26 Â± 0.06 malignant tumors/104 PYGy, or an average increase in relative risk of 6.9 (Â±5.5)%/rad, excluding the first 5 years after irradiation (chi-square of 12.7 on 12 degrees of freedom; p = 0.39). For benign tumors, an average excess of 0.44 Â± 0.11/104 PYGy, or 3.6 (Â±2.1)%/rad (chi-square of 12.9 on 10 degrees of freedom; p = 0.23) (La86). In Japanese A-bomb survivors, although mortality from salivary gland tumors has not been detectably affected, the incidence of such tumors has shown a dose-dependent increase (Table 5-7). The increase is smaller than that in radiotherapy patients, however, possibly because of differences in ascertainment or case reporting. No marked variation of susceptibility with age at the time of irradiation has been evident in the A-bomb survivors. Persons exposed to diagnostic x radiation of the head and neck also have been reported to show an increase in the risk of cancer of the parotid gland, the risk being highest in those receiving full-mouth or panoramic dental radiography or some other type of major diagnostic examination of the head before the age of 20 (Pr88b). In laboratory animals, irradiation has been observed to induce cancer of the salivary gland infrequently (G162, Ta75), indicating that the susceptibility of the salivary gland to radiation carcinogenesis is relatively low in comparison with that of other organs.
TABLE 5-7 Incidence of Salivary Gland Tumors Among A-Bomb Survivors, Hiroshima, 1953-1971, Open City Population Exposure Distance (m) Trend Test, No. of Cases/106 P Estimated Risk (%/ 0-5,000 m (Pvalue) rad rad)a <1.500 1.501-2,000 2,001-5,000 >5,000 Person years 322,768 366,887 800,486 4,333,814 Average kerma (rad)a 124.3 9.9 0 0 No. malignant tumors 7 4 6 5 0.030 0.056 Â± 0.036 0.69 Â± 0.57 No. benign tumors 7 2 5 25 0.010 0.063 Â± 0.035 1.10 Â± 0.86 Total no. tumors 14 6 11 30 0.002 0.120 Â± 0.050 0.86 Â± 0.49 aBased on T65D dose estimates (see La86). SOURCE: From La86. based on data from Ta76 and Oh78. RADIOGENIC CANCER AT SPECIFIC SITES 332
RADIOGENIC CANCER AT SPECIFIC SITES 333 Summary The incidence of salivary gland tumors has been observed to be increased by irradiation in A-bomb survivors, patients treated with x rays to the head and neck in childhood, and women treated with iodine-131 in middle age. The excess relative risk of salivary gland cancer averages 550% per Gy, or 0.26 cases/104 PYGy. In patients treated for tinea capitis, the excess was evident at an estimated average dose of only 0.4 Gy, indicating that the susceptibility of the salivary gland to radiation carcinogenesis is relatively high. PANCREAS Cancer of the pancreas is the fourth leading type of fatal cancer in the United States (Yo81), although it is difficult to diagnose clinically and is verified histologically in only a small percentage of cases (Ma82). Excess mortality from the disease has been observed inconsistently in irradiated human populations and has borne no clear relationship to dose or time after irradiation. One of the first populations in which an excess of the disease was observed is the well-studied series of 14,106 patients who were treated with radiation to the spine for ankylosing spondylitis, in whom 27 deaths from the disease have been reported versus 22.39 expected (Da87); however, the relative risk in this population was increased significantly only within the first 5 years after treatment (6 observed versus 1.85 expected deaths). Because cancer of the pancreas frequently causes pain in the back and is thus prone to be confused with ankylosing spondylitis, it is conceivable that the disease was present before irradiation in some of the observed cases (Da87). Other therapeutically irradiated patients in whom an excess has been reported include a series of men and women treated for lymphoma (Jo76) and a series of 82,616 women treated for cervical cancer (Bo85). In the latterâas in the patients with ankylosing spondylitisâthe relative risk was increased maximally soon (1-4 years) after irradiation and not consistently thereafter. Furthermore, a comparable excess (ratio of observed to expected cases = 34/25 = 1.4) was observed in a companion series of women with in situ carcinoma of the cervix who received no therapeutic radiation (Bo85). A case control analysis of these data also yielded a null result (Bo88). In Japanese A-bomb survivors, no dose-or time-dependent excess in mortality from cancer of the pancreas has been observed; the relative risk at 1 Gy (T65DR shielded kerma) is estimated to approximate 0.9974 Â± 0.1069 (Pr87a). Although data from the Nagasaki Tumor Registry for 1959-1978 suggested an increase with dose (P value for trend test, 0.0740),
RADIOGENIC CANCER AT SPECIFIC SITES 334 corresponding to an excess of 1.15 Â± 0.66 cases/104 PYGy (Wa83, La86), no dose-dependent excess was evident in the concurrent data (then incomplete from the Hiroshima Tumor Registry (Be78, La86). Among occupationally exposed persons, an excess in the number of deaths from the disease was reported among British radiologists who entered the practice of radiology before 1921 (6 deaths versus 1.9 expected through 1976) but was not evident in later cohorts (Sm81) nor in U.S. radiologists who entered practice after 1920 (Ma75). Among radiation workers at the Hanford Plant, a dose-related excess number of deaths from pancreatic cancer was reported a number of years ago (Ma77, Ma78b, Gi79), but the excess has not been confirmed by more recent follow-up (To83). Summary An association between cancer of the pancreas and previous irradiation, suggested by several reports in the past, has not been confirmed in more recent and thorough studies of irradiated human populations. The pancreas appears, therefore, to be relatively insensitive to radiation carcinogenesis. REFERENCES Al85 Albo, V., D. Miller, S. Leiken, N. Satlok, and D. Hammond. 1985. Nine brain tumors (BT) as a late effect in children ''cured" of acute lymphoblastic leukemia (ALL) from a single protocol study. Proc. Am. Soc. Clin. Oncol. 4:172. Al86 Albert, R. E., and R. E. Shore. 1986. Carcinogenic effects of radiation on the human skin. Pp. 335-345 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. J. Burns, and R. E. Shore, eds. New York: Elsevier. An64 Anderson, R. E., and K. Ishida. 1964. Malignant lymphoma in survivors of the atomic bomb in Hiroshima. Am. Inst. Med. 61:853-862. As82 Asano, M., K. Yoshimoto, S. Seyama, H. Itakura, T. Hamada, and S. Iijuma. 1982. Primary liver carcinoma and liver cirrhosis in atomic bomb survivors, Hiroshima and Nagasaki, 1961-75 with special reference to hepatitis B surface antigen . J. Natl. Cancer Inst. 69:1221-1227. Ba78 Barendsen, G. W. 1978. Fundamental aspects of cancer induction in relation to the effectiveness of small doses of radiation. Pp. 263-276 in Late Biological Effects of Ionizing Radiation, Vol. II. Vienna: International Atomic Energy Agency. Ba86 Bair, W. J. 1986. Experimental carcinogenesis in the respiratory tract. Pp. 151-167 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. J. Burns, and R. E. Shore, eds. New York: Elsevier. Be67 Berdjis, C. C. 1967. Pathogenesis of radiation-induced endocrine tumors. Oncology 21:49-60. Be78 Beebe, G. W., H. Kato, and C. E. Land. 1978. Studies of the mortality of A-bomb survivors. 6. Mortality and radiation dose, 1950-74. Radiat. Res. 75:138-201.
RADIOGENIC CANCER AT SPECIFIC SITES 335 Be79 Benjamin, S. A., B. B. Boecker, R. G. Cuddihy, and R. O. McClellan. 1979. Nasal carcinomas in beagles after inhalation of relatively soluble forms of beta-emitting radionuclides. J. Natl. Cancer Inst. 63:133-139. Be85 Beral, V., H. Inskip, P. Fraser, et al. 1985. Mortality of employees of the United Kingdom Atomic Energy Authority, 1946-79. Br. Med. J. 29:440. Be72 Berdjis, C. C. 1972. Parathyroid diseases and irradiation. Strahlentherapie 143:48-62. Bi75 Bithell, J. F., and A. M. Stewart. 1975. Pre-natal irradiation and childhood malignancy: A review of British data from the Oxford survey. Br. J. Cancer 31:271-287. Bl80 Black, W. C., L. S. Gomez, J. M. Yuhas, and M. M. Kligerman. 1980. Quantitation of the late effects of x-radiation on the large intestine. Cancer 45:444-451. Bl84 Blot, W. J., S. Akiba, and H. Katon. Ionizing radiation and lung cancer: A review including preliminary results from a case-control study among a-bomb survivors. In Atomic Bomb Survivor Data: Utilization and Analysis, R. L. Prentice, and D. J. Thompson, eds. Philadelphia: Society for Industrial and Applied Mathematics. Bo52 Bond, V. P., M. N. Swift, C. A. Tobias, and G. Brecher. 1952. Bowel lesions following single deuteron irradiation. Fed. Proc. 11:408-409. Bo60 Bond, V. P., E. P. Cronkite, S. W. Lippincott, and C. J. Shellabarger. 1960. Studies on radiation-induced mammary gland neoplasms in the rat. III. Relation of the neoplastic response to dose of total body radiation. Radiat. Res. 12:276-285. Bo79 Boice, J. D., Jr. 1979. Multiple chest fluoroscopies and the risk of breast cancer. Pp. 147-156 in Advances in Medical Oncology Research and Education, Vol. 1. Oxford: Pergamon. Bo84 Boice, J., and Fraumeni, Jr., eds. 1984. Radiation Carcinogenesis: Epidemiology and Biological Significance. New York: Raven Press. Bo85 Boice, J. D., Jr., N. E. Day, A. Andersen, L. A. Brinton, R. Brown, N. W. Choi, E. A. Clarke, M. P. Coleman, R. E. Curtis, J. T. Flannery, M. Hakama, T. Hakulinen, G. R. Howe, O. M. Jensen, R. A. Kleinerman, D. Magnin, K. Magnus, K. Makela, B. Malker, A. B. Miller, N. Nelson, C. C. Patternson, F. Petterssen, V. Pompe-Kirn, M. Primic-Zakelj, P. Prior, R. Ravnihar, R. G. Skeet, J. E. Skjerven, P. G. Smith, M. Sok, R. F. Spengler, H. H. Storm, M. Stovall, G. H. O. Thomkins, and C. Wall. 1985. Second Cancers Following Radiation Treatment for Cervical Cancer. An international collaboration among cancer registries. J. Natl. Cancer Inst. 74:955-975. Bo86 Boice, J. D., Jr., and R. H. Kleinerman. 1986. Meeting highlights: Radiation studies of women treated for benign gynecologic disease. J. Natl. Cancer Inst. 76:549-551. Bo86a Boecker, B. B., F. F. Hahn, R. G. Cuddihy, M. B. Snipes, and R. O. McClellan. 1986. Is the human nasal cavity at risk from inhaled radionuclides? Pp. 564-576 in Life-Span Radiation Effects Studies in Animals: What Can They Tell Us? U.S. Department of Energy Report CONF 830951. Springfield, Va.: National Technical Information Service. Bo87 Boice, J. D., Jr., M. Blettner, R. A. Kleinerman, M. Stovall, W. C. Moloney, G. Engholm, D. F. Austin, A. Bosch, D. L. Cookfair, E. T. Krementz, H. B. Latourette, L. J. Peters, M. D. Schulz, M. Lundell, F. Pattersson, H. H. Storm, C. M. J. Bell, M. P. Coleman, P. Fraser, M. Palmer, P. Prior, N. W. Choi, T.G. Hislop, M. Koch, D. Robb, D. Robson, R. F. Sprengler, D. von Fournler,
RADIOGENIC CANCER AT SPECIFIC SITES 336 R. Frischkorn, H. Lochmuller, V. Pompe-Kirn, A. Rimpels, K. Kjorstad, M. H. Pejovic, K. Sigurdsson, P. Pisani, H. Kucera, and G. B. Hutchison. 1987. Radiation Dose and Leukemia Risk in Patients Treated for Cancer of the Cervix. JNCI 79:1295-1311 Bo88 Boice, J. D., Jr., G. Engholm, R. A. Kleinerman, M. Blettner, M. Stovall, H. Lisco, W. C. Moloney, D. F. Austin, A. Bosch, D. L. Cookfair, E. T. Krementz, H. B. Latourette, J. A. Merrill, L. J. Peters, M. D. Schulz, H. H. Storm, E. Bjorkholm, F. Pettersson, C. M. J. Bell, M. P. Coleman, P. Fraser, F. E. Neal, P. Prior, N. W. Choi, T. G. Hislop, M. Koch, N. Kreiger, D. Robb, D. Tobson, D. H. Thomson, H. Lochmuller, D. V. Fournier, R. Frischkorn, K. E. Kjorstad, A. Rimpela, M. H. Pejovic, V. P. Kirn, H. Stankusova, F. Berrino, K. Soigurdsson, G. B. Hutchison, and B. MacMahon. 1988. Radiation dose and second cancer risk in patients treated for cancer of the cervix. Radiat. Res. 116:3-55. Bo88b Boecker, B. B., F. F. Hahn, B. A. Muggenburg, R. A. Guilmette, W. C. Griffith, and R. O. McClellan. 1988. The relative effectiveness of inhaled alpha-and beta-emitting radionuclides in producing lung cancer. Pp. 1059-1062 in Radiation Protection Practice. Sydney: Pergamon Press. Br53 Brecher, G., E. P. Cronkite, and J. H. Peers. 1953. Neoplasms in rats protected against lethal doses of irradiation by parabiosis or para aminopropriophenone. J. Natl. Cancer Inst. 14:159-175. Br69 Brinkley, D., and J. L. Haybittle. 1969. The late effects of artificial menopause by x-radiation. Br. J. Radiol. 42:519-521. Br81 Broerse, J. J., C. F. Hollander, and M. J. Van Zwieten. 1981. Tumor induction in Rhesus monkeys after total body irradiation with X-rays and fission neutrons. Int. J. Radiat. Biol. 40:671-676. Br85 Broerse, J. J., L. A. Hermen, and M. J. van Zwieten. 1985. Radiation carcinogenesis in experimental animals and its implications for radiation protection. Int. J. Radiat. Biol. 48:167-187. Br36 Brown, P. 1936. American Martyrs to Science through the Roentgen Ray. Springfield, III.: Charles C Thomas. Bu80 Burns, F. J., and R. E. Albert. 1980. Dose-response for rat skin tumors induced by single and split doses of argon ions. In Biological and Medical Research with Accelerated Heavy Ions at the Bevalac. Berkeley: University of California. Bu86 Burns, F. J., and R. E. Albert. 1986. Radiation carcinogenesis in rat skin. Pp. 199-214 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. J. Burns, and R. E. Shore, eds. New York: Elsevier. Bu76 Burns, F. J., I. P. Sinclair, R. E. Albert, and M. Vanderlaan. 1976. Tumor induction and hair follicle damage for different electron penetrations in rat skin. Radiat. Res. 67:474-481. By85 Byers, T., S. Graham, T. Rzepka, and J. Marshall. 1985. Lactation and breast cancer: Evidence for a negative association in premenopausal women. Am. J. Epidemiol. 121:664-674. Ca48 Cade, S. 1948. Malignant Disease and Its Treatment by Radium, Vol. 1 . Baltimore: Williams & Wilkins. Ch78 Christensson, T. 1978. Hyperparathyroidism and radiation therapy. Ann. Intern. Med. 89:216-217. Cl74 Clayman, C. B., W. H. Kruskal, J. W. J. Carpenter, and W. L. Palmer. 1974. The neoplastic potential of gastric irradiation. In Gastric Irradiation in Peptic Ulcer, W. L. Palmer, ed. Cl59 Clifton, K. H. 1959. Problems in experimental tumorigenesis of the pituitary
RADIOGENIC CANCER AT SPECIFIC SITES 337 gland, gonads, adrenal cortices, and mammary glands: a review. Cancer Res. 19:2-22. Cl75 Clifton, K. H., and B. N. Sridharan. 1975. Endocrine factors and tumor growth. Pp. 249-285 in Cancer: A Comprehensive Treatise, Vol. 3. F. F. Becker, ed. New York: Plenum. Cl77 Clifton, K. H. 1977. The physiology of endocrine therapy. Pp. 573-597 in Cancer: A Comprehensive Treatise, Vol. 5. New York: Plenum Press. Cl78 Clifton, K. H., and J. Crowley. 1978. Effects of radiation type and role of glucocorticoids, gonadectomy and thyroidectomy in mammary tumor induction in MtT-grafted rats. Cancer Res. 38:1507-1513. Cl79 Clifton, K. H. 1979. Animal models of breast cancer. Pp. 1-20 in Endocrinology of Cancer, D. P. Rose, ed. Boca Raton, Fla.: CRC Press. Cl85a Clifton, K. H., and M. N. Gould. 1985. Clonogen transplantation assay of mammary and thyroid epithelial cells. Pp. 128-138 in Cell Clones: Manual of Mammalian Cell Techniques, C. S. Potten and J. H. Hendry, eds. Edinburgh: Churchill Livingstone. Cl85b Clifton, K. H., K. Kamiya, R. T. Mulcahy, and M. N. Gould. 1985. Radiogenic neoplasia in the thyroid and mammary clonogens: Progress, problems and possibilities. Pp. 329-342 in Assessment of Risk From Low-Level Exposure to Radiation and Chemicals: A Critical Overview. A. D. Woodhead, C. J. Shellabarger, V. Pond, and A. Hollaender, eds. New York: Plenum. Cl85c Clifton, K. H., J. Yasukawa-Barnes, M. A. Tanner, and R. V. Haning, Jr. 1985. Irradiation and prolactin effects on rat mammary carcinogenesis: Intrasplenic pituitary and estrone capsule implants . J. Natl. Cancer Inst. 75:167-175. Cl86a Clifton, K. H., M. A. Tanner, and M. N. Gould. 1986. Assessment of radiogenic cancer initiation frequency per clonogenic rat mammary cell in vivo. Cancer Res. 46:2390-2395. Cl86b Clifton, K. H. 1986. Thyroid cancer: Reevaluation of an experimental model for radiogenic carcinogenesis. Pp. 181-198 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. J. Burns, and R. E. Shore, eds. New York: Elsevier-North-Holland. Cl86c Clifton, K. H. 1986. Thyroid and mammary radiobiology: Radiogenic damage to glandular tissue. Br. J. Cancer 53(Suppl. VII):237-250. Co85 Coggle, J. E., D. M. Peel, and J. D. Tarling. 1985. Lung tumor induction in mice after uniform and nonuniform external thoracic x-irradiation. Int. J. Radiat. Biol. 48:95-106. Co76 Colman, M., L. R. Simpson, L. K. Patterson, and L. Cohen. 1976. Thyroid cancer associated with radiation exposure. Pp. 285-288 in Biological and Environmental Effects of Low Level Radiation, Vol. II. Vienna: International Atomic Energy Agency. Co78 Colman, M., M. Kirsch, and M. Creditor. Radiation induced tumors. Pp.167-180 in Late Biological Effects of Ionizing Radiation, Vol. 1. Vienna: International Atomic Energy Agency. Co80 Conard, R. A., et al. 1980. Review of Medical Findings in a Marshallese Population Twenty- six Years after Accidental Exposure to Radioactive Fallout. Report BNL 5126 (Biology and Medical TID-4500). Upton, N.Y.: Brookhaven National Laboratory. Co84 Conard, R. A. 1984. Late radiation effects in Marshall Islanders exposed to fallout 28 years ago. Pp. 57-71 in Radiation Carcinogenesis: Epidemiology and Biological Significance, J. D. Boice, Jr. and J. F. Fraumeni, Jr., eds. New York: Raven.
RADIOGENIC CANCER AT SPECIFIC SITES 338 Co74 Coop, K. L., J. G. Sharp, J. W. Osborne, and G. R. Zimmerman. 1974. An animal model for the study of small-bowel tumors. Cancer Res. 34:1487-1494. Co58 Court, W. M., and R. Doll. 1958. Expectation of life and mortality from cancer among British radiologists. Br. Med. J. 181-187. Co65 Court-Brown, W. M., and R. Doll. 1965. Mortality from cancer and other causes after radiotherapy for ankylosing spondylitis. Br. Med. J. 2:1327-1332. Co74a Cox, D. R., and D. V. Hinkley. 1974. Theoretical Statistics. London: Chapman and Hall. Cu81 Cuzick, J. 1981. Radiation-induced myelomatosis. N. Engl. J. Med. 304:204-210. Cu84 Curtis, R., B. Hankey, M. Myers, et al. 1984. Risk of leukemia associated with a first course of cancer treatment: An analysis of Surveillance, Epidemiology and End Results Program experience. J. Natl. Cancer Inst. 72:531-544. Da85 Darby, S. C., E. Nakashima, and H. Kato. 1985. A parallel analysis of cancer mortality among atomic bomb survivors and patients with ankylosing spondylitis. J. Natl. Cancer Inst. 75:1-21. Da87 Darby, S. C., R. Doll, S. K. Gill, and P. G. Smith. 1987. Long-term mortality after a single treatment course with X-rays in patients treated for ankylosing spondylitis. Br. J. Cancer 55:179-190. Da78 da Silva Horta, J., M. E. da Silva Horta, L. C. da Motta, and M. H. Tavares. 1978. Malignancies in Portuguese Thorotrast Patients. Health Phys. 35:137-151. De65 DeGroot, L. J. 1965. Current views of formation of thyroid hormones. N. Engl. J. Med. 272:243-250, 297-303, 355-362. De78 Denman, D. L., F. R. Kirchner, and J. W. Osborne. 1978. Induction of colonic adenocarcinoma in the rat by x-irradiation. Cancer Res. 38:1899-1905. Di73 Diamond, E. L., H. Schmerler, and A. M. Lilienfeld. 1973. The relationship of intrauterine radiation to subsequent mortality and development of leukemia in children. Am. J. Epidemiol. 97:283-313. Di69 Dickson, R. J. 1969. The late results of radium treatment for benign uterine hemorrhage. Br. J. Radiol. 42:582-594. Do74 Dobyns, B. M., G. E. Sheline, J. B. Workman, E. A. Tompkins, W. M. McConahey, and D. V. Becker. 1974. Malignant and benign neoplasms of the thyroid in patients treated for hyperthyroidism: A report of the cooperative thyrotoxicosis therapy follow-up study. J. Clin. Endocrinol. 38:976-998. Do76 Dolphin, G. W. 1976. A comparison of the observed and the expected cancers of the haematopoietic and lymphatic systems among workers at Windscale. London: Her Majesty's Stationery Office. Do63 Doniach, I. 1963. Effects including carcinogenesis of 131I and x-rays on the thyroid of experimental animals: A review. Health Phys. 9:1357-1362. Do67 Doniach, I. 1967. Damaging effect of x-irradiation of less than 1000 rads on goitrogenic capacity of rat thyroid gland. In Thyroid Neoplasia, S. Young and D. R. Inman, eds. New York: Academic Press. Do77 Doniach, I. 1977. Pathology of irradiation thyroid damage. In Radiation-Associated Thyroid Carcinoma, L. J. DeGroot, L. A. Grohman, E. L. Kaplan, and S. Refetoff, eds. New York: Grune and Stratton. Du50 Duffy, B. J., Jr., and P. J. Fitzgerald. 1950. Cancer of the thyroid in children: A report of 28 cases. J. Clin. Endocrinol. 10:1296-1308. Du80 Dumont, J. E., J. F. Malone, and A. J. Van Herle. 1980. Irradiation and Thyroid Disease: Dosimetric, Clinical and Carcinogenic Aspects. Report EUR 6713ER. Luxembourg: Commission of the European Communities. Ev86 Evans, J. S., J. E. Wennberg, and B. J. McNeil. 1986. The influence of
RADIOGENIC CANCER AT SPECIFIC SITES 339 diagnostic radiography on the incidence of breast cancer and leukemia. N. Engl. J. Med. 315:800-815. Fa78 Faber, M. 1978. Malignancies in Danish Thorotrast patients. Health Phys. 35:153-158. Fa64 Fabrikant, J. I., R. J. Dickson, and B. F. Fetter. 1964. Mechanism of radiation carcinogenesis at the clinical level. Br. J. Cancer 18:459-477. Fe87 Feola, J. M., Y. Maruyama, A. Pattarasumunt, and R. M. Kryscio. 1987. Cf-252 leukemogenesis in the C57BL mouse. Inst. J. Radiat. Oncol. Biol. Phys. 13:69-74. Fr73 Fritz, T. E., W. P. Norris, and D. V. Tolle. 1973. Myelogenous leukemia and related myeloproliferative disorders in beagles continuously exposed to 60Co y-radiation . Pp. 170-188 in Unifying Concepts of Leukemia, R. M. Dutcher and L. Chieco-Bianchi, eds. Bibliography of Haematology, No. 39. Basel: Karger. Fr86 Fry, R. J. M., J. B. Storer, and F. J. Burns. 1986. Radiation induction of cancer of the skin. Br. J. Radiol. 19(Suppl.):58-60. Fu36a Furth, J., and J. S. Butterworth. 1936. Neoplastic diseases occuring among mice subjected to general irradiation with x-rays. II. Ovarian tumors and associated lesions. Am. J. Cancer 28:66-95. Fu36b Furth, J., and O. B. Furth. 1936. Neoplastic diseases produced in mice by general irradiation with x-rays. I. Incidence and types of neoplasms. Am. J. Cancer 28:54-65. Fu75 Furth, J. 1975. Hormones as etiological agents in neoplasia. In Cancer: A Comprehensive Treatise, Vol. 1. F. F. Becker, ed. New York: Plenum. Ga63 Garner, R. J. 1963. Comparative early and late effects of single and prolonged exposure to radioiodine in young and adults of various animal speciesâa review. Health Phys. 9:1333-1339. Gi72 Gibson, R., S. Graham, A. Lilienfeld, et al. 1972. Irradiation in the epidemiology of leukemia among adults. J. Natl. Cancer Inst. 48:301. Gi79 Gilbert, E. S., and S. Marks. 1979. An analysis of the mortality of workers in a nuclear facility. Radiat. Res. 79:122-128. Gi87 Gillett, N. A., B. A. Muggenburg, B. B. Boecker, W. C. Griffith, F. F. Hahn, and R. O. McClellan. 1987. Single inhalation exposure to 90SrCl2 in the beagle dog: Late biological effects. J. Natl. Cancer Inst. 79:359-376. Gi88 Gillett, N. A., B. A. Muggenburg, J. A. Mewhinney, F. F. Hahn, F. A. Seiler, B. B. Boecker, and R. O. McClellan. Primary liver tumors in beagle dogs exposed by inhalation to aerosols of plutonium-238 dioxide. Am. J. Pathol. (submitted). G162 Glucksman, A., and C. P. Cherry. 1962. The induction of adenoma by the irradiation of salivary glands of rats. Radiat. Res. 17:186-202. Go86a Goldman, M. 1986. Experimental carcinogenesis in the skeleton. Pp. 215-331 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. J. Burns, and R. E. Shore, eds. New York: Elsevier. Go86b Goldman, M., L. S. Rosenblatt, and S. A. Book. 1986. Lifetime radiation effects research in animals: An overview of the status and philosophy of studies at University of California- Davis Laboratory for Energy Related Health Research. Pp. 53-65 in Life-Span Radiation Effects Studies in Animals: What Can They Tell Us?, R. C. Thompson and J. A. Mahaffey, eds. U.S. Department of Energy Report CONF-830951. Springfield, Va.: National Technical Information Service. Gr65 Gray, L. H. 1965. Radiation biology and cancer. Pp. 18-19 in Cellular Radiation Biology. Baltimore: Williams & Wilkins.
RADIOGENIC CANCER AT SPECIFIC SITES 340 Gr84 Griem, M. L., J. Justman, and L. Weiss. 1984. The neoplastic potential of gastric irradiation. IV. Am. J. Clin. Oncol. 7:675-677. Gr87 Griffith, W. C., B. B. Boecker, R. C. Cuddihy, R. A. Guilmette, F. F. Hahn, R. O. McClellan, B. A. Muggenburg, and M. B. Snipes. 1987. Preliminary Radiation Risk Estimates of Carcinoma Incidence in the Lung as a Function of Cumulative Radiation Dose Using Proportional Tumor Incidence Rates. Pp. 196-204 in 1986-87 Inhalation Toxicology Research Institute Report, J. D. Sun and J. A. Mewhinney, eds. U.S. Department of Energy Report LMF-120. Springfield, Va.: National Technical Information Service. Gu64 Gunz, F., and H. Atkinson. 1964. Medical radiation and leukemia: A retrospective survey. Br. Med. J. 1:389. Ha83a Hahn, F. F., B. B. Boecker, R. G. Cuddihy, C. H. Hobbs, R. O. McCellan, and M. B. Snipes. 1983. Influence of Radiation Dose Patterns on Lung Tumor Incidence in Dogs That Inhaled Beta Emitters: A Preliminary Report. Radiat. Res. 96:505-517. Ha83b Harley, N., A. B. Kolber, R. E. Shore, R. E. Albert, S. M. Altman, and B. Pasternack. 1983. The skin dose and response for the head and neck in patients irradiated with x-ray for tinea capitis: implications for environmental radioactivity. Pp. 125-142 in Epidemiology Related to Health Physics. Proceedings of the 16th Midyear Topical Meeting of the Health Physics Society . CONF-83011. Springfield, Va.: National Technical Information Service. Ha87 Hamilton, T. E., G. van Belle, and J. P. Lo Gerfo. 1987. Thyroid neoplasia in Marshall Islanders exposed to nuclear fallout. J. Am. Med. Assoc. 258:629-636. He75 Hempelmann, L. H., W. J. Hall, M. Phillips, R. A. Cooper, and W. R. Ames. 1975. Neoplasms in persons treated with x-rays in infancy: Fourth survey in 20 years. J. Natl. Cancer Inst. 55:519-530. He85 Henderson, B. E., L. N. Kolonel, R. Dworsky, D. Kerford, E. Mori, K. Singh, and H. Thevenot. 1985. Cancer incidence in the islands of the Pacific. Natl. Cancer Inst. Monogr. 69:73-81. He88 Henderson, B.E., Ross, R. and Bernstein, L. 1988. Estrogens as a cause of human cancer Cancer Res. 48:246-253. He50 Henry, S. A. 1950. Cutaneous cancer in relation to occupation. Ann. R. Coll. Surg. Engl. 7:425. Hi69 Hirose, F. 1969. Experimental induction of carcinoma in the glandular stomach by localized x- irradiation of gastric region, Pp. 75-113 in Experimental Carcinoma of the Glandular Stomach. Japanese Cancer Association GANN Monograph No. 8. Tokyo. Hi77 Hirose, F., K. Fukazawa, H. Watanabe, Y. Terada, I. Fujii, and S. Ootuska. 1977. Induction of rectal carcinoma in mice by local x-irradiation. GANN 68:669-680. Ho83 Hoel, D. G., T. Wakabayashi, and M. C. Pike. 1983. Secular trends in the distributions of the breast cancer risk factorsâmenarche, first birth, menopause and weightâin Hiroshima and Nagasaki, Japan. Am. J. Epidemiol. 118:78-89. Ho82 Hoffman, D. A., W. M. McConahey, and L. T. Kurland. 1982. Cancer incidence following treatment for hyperthyroidism. Int. J. Epidemiol. 11:218-224. Ho84a Hoffman, D. A. Late effects of I-131 therapy in the United States. 1984. Pp. 273-280 in Radiation Carcinogenesis: Epidemiology and Biological Significance , J. D. Boice, Jr., and J. F. Fraumeni, Jr., eds. New York: Raven. Ho63 Hollingsworth, D. R., H. B. Hamilton, H. Tamagaki, and G. W. Beebe. 1963. Thyroid disease: A study in Hiroshima, Japan. Medicine 42:47-71.
RADIOGENIC CANCER AT SPECIFIC SITES 341 Ho80a Holm, L.-E., I. Dahlqvist, A. Israelsson, and G. Lundell. 1980. Malignant thyroid tumors after iodine-131 therapy. N. Engl. J. Med. 303:188-191. Ho80b Holm, L.-E., G. Eklund, and G. Lundell. 1980. Incidence of malignant thyroid tumors in humans after exposure to diagnostic doses of iodine-131. II. Estimation of thyroid gland size, thyroid radiation dose, and predicted versus observed number of malignant thyroid tumors. J. Natl. Cancer Inst. 65:1221-1224. Ho80c Holm, L.-E., G. Lundell, and G. Walinder. 1980. Incidence of malignant thyroid tumors in humans after exposure to diagnostic doses of iodine-131. I. Restrospective cohort study. J. Natl. Cancer Inst. 64:1055-1059. Ho84b Holm, L.-E. 1984. Malignant disease following iodine-131 therapy in Sweden. Pp. 263-271 in Radiation Carcinogenesis: Epidemiology and Biological Significance. J. D. Boice, Jr., and J. F. Fraumeni, Jr., eds. New York: Raven Press. Ho88 Holm, L.-E., K. E. Wicklund, G. E. Lundell, J. D. Boice, N. A. Bergman, G. Bjelkengren, E. S. Cederquist, U.-B. C. Ericsson, L.-G. Larsson, M. E. Lidberg, R. S. Lindberg, and H. V. Wicklund. 1988. Thyroid cancer after diagnostic doses of iodine-131: A retrospective study. J. Natl. Cancer Inst. 80:1132-1136. Ho70 Howard, E. B., and W. J. Clarke. 1970. Strontium-90-induced hematopoietic neoplasms in miniature swine. Pp. 379-401 in Myeloproliferative Disorders of Animals and Man. U.S.A.E.C. Div. Tech. Info., W. J. Clarke, E. B. Howard, and P. L. Hackett, eds. Hr89 Hrubec, Z., J. Boice, R. Monson, and M. Rosenstein. 1989. Breast cancer after multiple chest fluoroscopies: Second follow-up of Massachusetts women with tuberculosis. Cancer Res. 49:229-234. Hu63 Huggins, C., and R. Fukunishi. 1963. Cancer in the rat after single exposures to irradiation or hydrocarbons. Radiat. Res. 20:493-503. Hu87 Humphreys, E. R., J. F. Loutit, and V. A. Stones. 1987. The induction by 239 Pu of myeloid leukemia and osteosarcoma in female CBA mice. Int. J. Radiat. Biol. 51:331-339. Ic79 Ichimaru, M., T. Ishimaru, M. Mikami, and M. Matsunaga. 1979. Multiple myeloma among atomic bomb survivors, Hiroshima and Nagasaki, 1950-1976. Technical Report 9-79. Hiroshima: Radiation Effects Research Foundation. ICRP80 International Commission on Radiological Protection (ICRP). 1980. Pp. 1-108 in Biological Effects of Inhaled Radionuclides. ICRP Publication 31. Oxford: Pergamon. ICRP87 International Commission on Radiological Protection (ICRP). 1987. Pp. 1-60 in Lung Cancer Risk from Indoor Exposures to Radon Daughters. ICRP Publication 50. Oxford: Pergamon. Ir85 Ireland, J. P., S. J. Fleming, D. A. Levison, W. R. Cattell, and L. R. I. Baker. 1985. Parathyroid carcinoma associated with chronic renal failure and previous radiotherapy to the neck. J. Clin. Pathol. 38:1114-1118. Ja70 Jablon, S., and H. Kato. 1970. Childhood cancer in relation to prenatal exposure to atomic- bomb radiation. Lancet ii:1000-1003. Ja71 Janower, M. L., and O. S. Miettinen. Neoplasms after childhood irradiation of the thymus gland. J. Am. Med. Assoc. 215:753-756. Ka82 Kato, H., and W. J. Schull. 1982. Studies of the mortality of a-bomb survivors. 7. Mortality 1950-1978: part 1. Cancer mortality. Radiat. Res. 90: 395-432. Ka83 Katz, A., and G. D. Braunstein. 1983. Clinical, biochemical, and pathologic features of radiation-associated hyperparathyroidism. Arch. Intern. Med. 143:79-82.
RADIOGENIC CANCER AT SPECIFIC SITES 342 Ka85 Kamiya, K., A. Inoh, Y. Fujii, K. Kanda, T. Kobayashi, and K. Yokoro. 1985. High mammary carcinogenicity of neutron irradiation in rats and its promotion by prolactin. Jpn. J. Cancer Res. 76:449-456. Ke78 Kennedy, A. R., and J. B. Little. 1978. Radiation carcinogenesis in the respiratory tract. Pp. 189-261 in Pathogenesis and Therapy of Lung Cancer, C. C. Harris, ed. A Monograph in the series "Lung Biology in Health and Disease." C. Lenfant, ed. New York: Marcel Dekker. Ki78 Kim, J. H., F. C. Chu, H.Q. Woodward, M. R. Melamed, A. Huvos, and J. Cantin. 1978. Radiation-Induced Soft-Tissue and Bone Sarcoma. Kl87 Kleinberg, D. L. 1987. Prolactin and breast cancer (editorial). N. Engl. J. Med. 316:269-271. Kl82 Kleinerman, R. A., R. E., Curtis, J. D. Boice, Jr., J. T. Flannery, and J. F. Fraumeni, Jr. 1982. Second cancers following radiotherapy for cervical cancer. J. Natl. Cancer Inst. 69:1027-1033. Kn82 Knowles, J. F. 1982. Radiation-induced nervous system tumours in the rat. Int. J. Radiat. Biol. 41:79-84. Ko86 Kopecky, K. J., E. Nakashima, T. Yamamoto, and H. Kato. 1986. Lung Cancer, Radiation and Smoking Among A-Bomb Survivors. RERF TR 13-86. Ku87 Kumar, P. R., R. R. Good, F. M. Skultety, L. G. Liebrock, and G. S. Severson. 1987. Radiation-induced neoplasms of the brain. Cancer 59:1274-1282. La87 Laird, N. M. 1987. Thyroid cancer risk from exposure to ionizing radiation: A case study in the comparative potency model. Risk Anal. 7:299-309. La89 Lafuma, J., D. Chmelevsky, J. Chameaud, M. Morin, R. Masse, and A. M. Kellerer. 1989. Lung carcinomas in Sprague-Dawley rats after exposure to low doses of radon daughters, fission neutrons, or gamma rays. Radiat. Res. 118:230-245. La80 Land, C. E., J. D. Boice, Jr., R. E. Shore, J. E. Norman, and M. Tokunaga. 1980. Breast cancer risk from low-dose exposures to ionizing radiation: Results of parallel analysis of three exposed populations of women. J. Natl. Cancer Inst. 65:353-365. La86 Land, C. E. 1986. Carcinogenic effects of radiation on the human digestive tract and other organs. Pp. 347-378 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. J. Burns, and R. E. Shore, eds. New York: Elsevier. Le63 Lewis, E. B. 1963. Leukemia, multiple myeloma, and aplastic anemia in American radiologists. Science 142:1492-1494. Le73 Lebedeva, G. A. 1973. Intestinal polyps arising under the influence of various kinds of ionizing radiations. Vop. Onkol. 19:47-51. Le79 Lee, W., B. Schlein, N. C. Telles, and R. P. Chiacchierini. 1979. An accurate method of 131I dosimetry in the rat thyroid. Radiat. Res. 79:55-62. Le82 Lee, W., R. P. Chiacchierini, B. Shlein, and N. C. Telles. 1982. Thyroid tumors following I-131 or localized x-irradiation to the thyroid and the pituitary glands in rats. Radiat. Res. 92:307-319. Le88 Lewis, C. A., P. G. Smith, I. M. Stratton, S. C. Darby, and R. Doll. 1988. Estimated Radiation Doses to Different Organs Among Patient Treated for Ankylosing Spondylitis with a Single Course of X-rays. Br. J. Radiol. 61:212-220. Li82 Lightdale, C. J., T. D. Koepsell, and P. Sherlock. 1982. Small intestine. In Cancer Epidemiology and Prevention, D. Schottenfeld and J. F. Fraumeni, Jr., eds. Philadelphia: W. B. Saunders. Li63 Lindsay, S., and I. L. Chaikoff. 1963. The effects of irradiation on the thyroid
RADIOGENIC CANCER AT SPECIFIC SITES 343 gland with particular reference to the induction of thyroid neoplasms: A review. Cancer Res. 24:1099-1107. Li80 Linos, A., J. Gray, and A. Orvis. 1980. Low dose radiation and leukemia. N. Engl. J. Med. 302:1101. Li47 Lisco, H., M. P. Finkel, and A. M. Brues. 1947. Carcinogenic properties of radioactive fission products and of plutonium. Radiology 49:61-63. Lu87a Lundgren, D. L., F. F. Hahn, W. C. Griffin, R. G. Cuddihy, P. J. Haley, and B. B. Boecker. 1987. Effects of relatively low-level exposure of rats to inhaled 144CeO2. III. Pp. 308-312 in 1986-87 Inhalation Toxicology Research Institute Annual Report, J. D. Sun, and J. A. Mewhinney, eds. U.S. Department of Energy Report LMF-120. Springfield, Va.: National Technical Information Service. Lu87b Ludgren, D. L., F. F. Hahn, W. C. Griffin, R. G. Cuddihy, F. A. Seiler and B. B. Boecker. 1987. Effects of relatively low-level thoracic or whole-body exposure of rats to x-rays. I. Pp. 313-317 in 1986-87 Inhalation Toxicology Research Institute Annual Report, J. D. Sun, and J. A. Mewhinney, eds. U.S. Department of Energy Report LMF-120. Springfield, Va.: National Information Service. Ma82 Mack, T. M. 1982. Pancreas. In Cancer Epidemiology and Prevention, D. Schottenfeld and J. F. Fraumeni, Jr., eds. Philadelphia: W. B. Saunders. Ma65 Mackenzie, I. 1965. Breast cancer following multiple fluoroscopies. Br. J. Cancer 19:1-9. Ma62 MacMahon, B. 1962. Prenatal x-ray exposure and childhood cancer. J. Natl. Cancer Inst. 28:1173-1191. Ma73 MacMahon, B., P. Cole, and J. Brown. 1973. Etiology of human breast cancer: A review. J. Natl. Cancer Inst. 50:21-42. Ma78 Major, I. R., and R. H. Mole. 1978. Myeloid Leukemia in X-ray Irradiated CBA Mice. Nature 272, 455-456. Ma77 Mancuso, R. F., A. Stewart, and G. Kneale. 1977. Radiation exposures of Hanford workers dying from cancer and other causes. Health Phys. 33:369-385. Ma78b Marks, S., E. S. Gilbert, and B. D. Breitenstein. 1978. Cancer mortality in Hanford workers. Pp. 369-386 in Late Biological Effects of Ionizing Radiation, Vol. I. Vienna: International Atomic Energy Agency. Ma75 Matanoski, G. M. 1975. The current mortality rates of radiologists and other physician specialists: Specific causes of death. Am. J. Epidemiol. 101:199-210. Ma81b Matanoski, G. M. 1981. Risk of cancer associated with occupational exposure in radiologists and other radiation workers. Pp. 241-254 in Cancer Achievements, Challenges and Prospects for the 1980's, Vol. 1, J. H. Burchenal and H. F. Oettgen, eds. New York: Grune & Stratton. Ma84 Matanoski, G. M., P. Sartwell, E. Elliott, J. Tonascia, A. Sternberg. 1984. Cancer risks in radiologists and radiation workers. Pp. 83-96 in Radiation Carcinogenesis: Epidemiology and Biological Significance, J. D. Boice, Jr., J. F. Fraumeni, Jr., eds. New York: Raven. Ma81 Maxon, H. R., E. L. Saenger, C. R. Buncher, J. G. Kereiakes, S. R. Thomas, M. L. Shafer, and C. A. McLaughlin. 1981. Radiation-associated carcinoma of the salivary glands. A controlled study. Ann. Otol. 90:107-108. Ma80 Mays, C. W., and M. P. Finkel. 1980. RBE of Alpha-Particles vs Beta-Particles in Bone Sarcoma Induction. Pp. 401-405 in Proceedings of the 6th Congress of the International Radiation Protection Association. Berlin: International Radiation Protection Association. Ma88 Mays, C. W. Personal communication.
RADIOGENIC CANCER AT SPECIFIC SITES 344 Mc86 McClellan, R. O., B. B. Boecker, F. F. Hahn, and B. A. Muggenburg. 1986. Lovelace ITRI Studies on the Toxicity of Inhaled Radionuclides in Beagle Dogs. Pp. 74-96 in Life-Span Radiation Effects Studies in Animals: What Can They Tell Us?, R. E. Thompson and J. A. Mahaffey, eds. U.S. Department of Energy Report CONF-830951. Mc86a McTiernan, A., and D. B. Thomas. 1986. Evidence for a protective effect of lactation on risk of breast cancer in young women. Am. J. Epidemiol. Me88 Metivier, H., R. Masse, G. Rateau, D. Nolibe, and J. Lafuma. 1988. New Data on the Toxicity of 239PuO2 in Baboons. To be published in the Proceedings of the CEC/CEA/DOE- Sponsored Workshop on Biological Assessment of Occupational Exposure to Actinides, Versailles, France, May 30-June 2. Me86 Mewhinney, J. A., F. F. Hahn, M. B. Snipes, W. C. Griffith, B. B. Boecker, and R. O. McClellan. 1986. Incidence of 90SrCl2 or 238PuO 2; Implications for Estimation of Risk in Humans. Pp. 535-555 in Life Span Radiation Effects Studies in Animals: What Can They Tell Us?, R. C. Thompson and J. A. Mahaffey, eds. U.S. Department of Energy Report CONF-83051. Springfield, Va.: National Technical Information Service. Mi86 Miller, R. W., and G. W. Beebe. 1986. Leukemia, lymphoma, and multiple myeloma. Pp. 245-260 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. Burns, and R. E. Shore, eds. New York: Elsevier. Mi89 Miller, A. B., G. R. Howe, G. J. Sherman, J. P. Lindsay, M. J. Yaffe, P. Dinner, H. A. Risch, and D. C. Preston. 1989. Breast cancer mortality following irradiation in a cohort of Canadian tuberculosis patients. N. Engl. J. Med. (in press). Mo74 Modan, B., D. Baidatz, H. Mart, R. Steinitz, and S. G. Levin. 1974. Radiation-induced head and neck tumours. Lancet i:277-279. Mo82 Mole, R. H., and J. A. G. Davids. 1982. Induction of myeloid leukemia and other tumors in mice by irradiation with fission neutrons. Pp. 31-43 in Neutron Carcinogenesis, J. J. Broerse and G. B. Gerber, eds. Luxembourg: Commission of the European Communities. Mo83a Mole, R. H., and J. R. Major. 1983. Myeloid leukemia frequency after protraced exposure to ionizing radiation: experimental confirmation of the flat dose-response found in ankylosing spondylitis after a single treatment course with x-rays. Leukemia Res. 7:295-300. Mo83b Mole, R. H., D. G. Papworth, and M. J. Corp. 1983. The dose response for x-ray induction of myeloid leukemia in male CBA.H mice. Br. J. Cancer 47:285-291. Mo84 Monson, R. R., and B. MacMahon. 1984. Prenatal x-ray exposure and cancers in children. Pp. 97-105 in Radiation Carcinogenesis: Epidemiology and Biological Significance, J. D. Boice, Jr., and J. F. Fraumeni, Jr., eds. New York: Raven. Mo77 Montour, J. L., R. C. Hard, Jr., and R. E. Flora. 1977. Mammary neoplasia in the rat following high-energy neutron irradiation. Cancer Res. 37:2619-3623. MRC56 Medical Research Council. 1956. The Hazards to Man of Nuclear and Allied Radiations. London: Her Majesty's Stationery Office. Mu86 Muggenburg, B. A., B. B. Boecker, F. F. Hahn, W. C. Griffith, and R. O. McClellan. 1986. The Risk of Liver Tumors in Dogs and Man from Radioactive Aerosols. Pp. 556-563 in Life-Span Radiation Effects Studies in Animals: What Can They Tell Us?, R. C. Thompson and J. A. Mahaffey, eds. U.S. Department of Energy Report CONF-830951. Springfield, Va: National Technical Information Service.
RADIOGENIC CANCER AT SPECIFIC SITES 345 Mu87 Muirhead, C. R., and S. C. Darby. 1987. Modelling the relative and absolute risks of radiation-induced cancers. J. R. Statist. Soc. A 150(part 2):83-118. Mu80 Mulcahy, R. T., M. N. Gould, and K. H. Clifton. 1980. The survival of thyroid cells: In vivo irradiation and in situ repair. Radiat. Res. 84:523-528. Mu80a Mulcahy, R. T., M. N. Gould, and K. H. Clifton. 1980 The survival of thyroid cells: In vivo irradiation in situ repair. Radiat. Res. 84:523-528. Mu80b Mulcahy, R. T., D. P. Rose, J. M. Mitchen, and K. H. Clifton. 1980. Hormonal effects on the quantitative transplantation of monodispersed rat thyroid cells. Endocrinology 106:1769-1775. Mu84 Mulcahy, R. T., M. N. Gould, and K. H. Clifton. 1984. Radiation initiation of thyroid cancer: A common cellular event. Int. J. Radiat. Biol. 45:419-426. NRC80 National Research Council, Committee on the Biological Effects of Ionizing Radiations. The Effects on Populations of Exposure to Low Levels of Ionizing Radiation (BEIR III). Washington, D.C.: National Academy Press. Pp 524. NRC88 National Research Council, Committee on the Biological Effects of Ionizing Radiations. Health Risks of Radon and Other Internally Deposited Alpha-Emitters (BEIR IV). Washington, D.C.: National Academy Press. Pp 602. NCI81 National Cancer Institute. 1981. Surveillance, Epidemiology and End Results: Incidence and Mortality Data, 1973-77 (SEER Report). Natl. Cancer Inst. Monogr. 57:794-797. NCRP84 National Council on Radiation Protection and Measurements (NCRP). 1984. Evaluation of Occupational and Environmental Exposures to Radon and Radon Daughters in the United States. Report No. 78. Bethesda, Md.: National Council on Radiation Protection and Measurements. NCRP85 National Council on Radiation Protection and Measurements (NCRP). 1985. Induction of Thyroid Cancer By Ionizing Radiation. NCRP Report No. 80. Bethesda, Md.: National Council on Radiation Protection and Measurement. Ne83 Netelenbos, C., P. Lips, and C. van der Meer. 1983. Hyperparathyroidism following irradiation of benign diseases of the head and neck. Cancer. 52:458-461. NIH85 Report of the National Institutes of Health Ad Hoc Working Group to Develop Radioepidemiological Tables. 1985. NIH publication 85-2748. Washington, D.C.: U.S. Government Printing Office. No59 Nowell, P. C., and L. G. Cole. 1959. Late effects of fast neutrons versus x-rays in mice: Nephrosclerosis, tumors, longevity. Radiat. Res. 11:545-556. Oh78 Ohkita, T., H. Takebashi, N. Takeichi, and F. Hirose. 1978. Prevalence of leukemia and salivary gland tumors among Hiroshima atomic bomb survivors. Pp. 71-81 in Late Biological Effects of Ionizing Radiation, Vol. 1. Vienna: International Atomic Energy Agency. Os63 Osborne, J. W., D. P. Nicholson, and K. N. Prasad. 1963. Induction of intestinal carcinoma in the rat by x-irradiation of the small intestine. Radiat. Res. 18:76-85. Pa86 Park, J. F., G. E. Dangle, H. A. Ragan, R. E. Weller, and D. L. Stevens. 1986. Current Status of Life-Span Studies with Inhaled Plutonium in Beagles at Pacific Northwest Laboratory. Pp. 445-470 in Life-Span Radiation Effects Studies in Animals: What Can They Tell Us?, R. E. Thompson and J. A. Mahaffey, eds. U.S. Department of Energy Report CONF-830951. Springfield, Va.: National Technical Information Service. Pi83 Pike, M. C., M. D. Krailo, B. E. Henderson, J. T. Casagrande, and D. G. Hoel. 1983. ''Hormonal" risk factors, "breast tissue age" and the age-incidence of breast cancer. Nature 303:767-770.
RADIOGENIC CANCER AT SPECIFIC SITES 346 Po78 Polednak, A. P., A. F. Stehney, and R. E. Rowland. 1978. Mortality among women first employed before 1930 in the U.S. radium dial painting industry. A group ascertained from employment lists. Am. J. Epidemiol. 107:179-195. Pr80 Preston-Martin S., A. Paganini-Hill, B. E. Henderson, M. Pike, and C. Wood 1980. Case control study of intracranial meningiomas in women in Los Angeles County, California. J. Natl. Cancer Inst. 65:67-73. Pr81 Prinz, R. A., E. Paloyan, A. M. Lawrence, A. L. Barbato, S. S. Braithwaite, and M. H. Brooks. 1981. Unexpected parathyroid disease discovered at thyroidectomy in irradiated patients. Am. J. Surg. 142:355-357. Pr82 Prentice, R. L., H. Kato, K. Yoshimoto, and M. Mason. 1982. Radiation exposure and thyroid cancer incidence among Hiroshima and Nagasaki residents. Natl. Cancer Inst. Monogr. 62:207-212. Pr83 Prentice, R. L., Y. Yoshimoto, and M. W. Mason. 1983. Relationship of cigarette smoking and radiation exposure to cancer mortality in Hiroshima and Nagasaki. J. Natl. Cancer Inst. 70:611-622. Pr87a Preston, D. L., H. Kato, K. J. Kopecky, and S. Fugita. 1987. Life Span Study Report 10. Part 1. Cancer Mortality among A-Bomb Survivors in Hiroshima and Nagasaki, 1950-82. Technical Report RERF TR 1-86. Hiroshima: Radiation Effects Research Foundation. Pr87b Preston, D. L., H. Kato, K. J. Kopecky, and S. Fujita. 1987. Life Span Study Report 10. Part 1. Cancer mortality among a-bomb survivors in Hiroshima and Nagasaki, 1950-82. Radiat. Res. 111:151-178. Pr88 Preston D., and D. Pierce, 1988. The effect of changes in dosimetry on cancer mortality risk estimates in atomic bomb survivors. Radiat Res. 114:437-466. Pr88b Preston-Martin S., D.C. Thomas, S. C. White, D. Cohen. 1988. Prior exposure to medical and dental x-rays related to tumors of the parotid gland . J. Natl. Cancer Inst. 80:943-949. Ra83 Raabe, O. G., S. A. Book, and N. J. Parks. 1983. Lifetime bone cancer dose-response relationships in beagles and people from skeletal burdens of 226Ra and 90Sr. Health Phys. 44(Suppl. 1):33-48. Ra74 Rallison, M. L., B. M. Dobyns, F. R. Keating, Jr., J. E. Rall, and F. H. Tyler. 1974. Thyroid disease in children. A survey of subjects potentially exposed to fallout radiation. Am. J. Med. 56:457-463. Ra75 Rallison, M. L., B. M. Dobyns, F. R. Keating, Jr., J. E. Rall, and F. H. Tyler. 1975. Thyroid nodularity in children. J. Am. Med. Assoc. 233:1069-1072. Ra74b Rankow, R. M., J. Conley, and P. Fodor. 1974. Carcinoma of the maxillary sinus following Thorotrast instillation. J. Max. Fac. Surg. 2:119-126. Ra80 Rao, S. D., B. Frame, M. J. Miller, M. Kleerekoper, M. A. Block, and A. M. Parfitt. 1980. Hyperparathyroidism following head and neck irradiation. Arch. Intern. Med. 140:201-207. Ri87 Rimm, I. J., F. C. Li, N. J. Tarbell, K. R. Winston, and S. E. Sallan. 1987. Brain tumors after cranial irradiation for childhood acute lymphoblastic leukemia. Cancer 59:1506-1508. Ro78b Robinson, C. V., and A. C. Upton. 1978. Competing risk analysis of leukemia and nonleukemia mortality in x-irradiated, male RF mice. J. Natl. Cancer Inst. 60:995-1007. Ro87 Roesch, W. C., ed. 1987. U.S.-Japan Joint Reassessment of Atomic Bomb Radiation Dosimetry in Hiroshima and Nagasaki, Vol. 1. Hiroshima: Radiation Effects Research Foundation. Ro84a Ron, E., and B. Modan. 1984. Thyroid and other neoplasms following childhood scalp irradiation. Pp. 139-151 in Radiation Carcinogenesis: Epidemiology
RADIOGENIC CANCER AT SPECIFIC SITES 347 and Biological Significance, J. D. Boice, Jr., and J. F. Fraumeni, Jr., eds. New York: Raven. Ro79 Rose, D. P. 1979. Endogenous hormones in the etiology and clinical course of breast cancer. Pp. 21-60 in Endocrinology of Cancer, Vol. I, D. P. Rose, ed. Boca Raton, Fla.: CRC Press. Ro78 Rostom, A. Y., S. L. Kauffman, and G. G. Steel. 1978. Influence of midonidazole on the incidence of radiation-induced intestinal tumors in mice. Br. J. Cancer 38:530-536. Ro84b Rothman, R. J. 1984. Significance of studies of low-dose radiation fallout in the western United States. Pp. 73-82 in Radiation Carcinogenesis: Epidemiology and Biological Significance, J. D. Boice, Jr., and J. F. Fraumeni, Jr., eds. New York: Raven. Ro77 Roudebush, C. P., and L. J. DeGroot. 1977. The Natural History of Radiation-Associated Thyroid Cancer. L. J. DeGroot, L. A. Grohman, E. L. Kaplan, and S. Refetoff, eds. New York: Grune and Stratton. Ru84 Rubinstein, A. B., M. N. Shalit, M. L. Cohen, U. Zandbank, and E. Reichenthal. 1984. Radiation-induced cerebral meningioma: A recognizable entity. J. Neurosurg. 61:966-971. Ru82 Russo, J., L. K. Tay, and L. H. Russo. 1982. Differentiation of the mammary gland and susceptibility to carcinogenesis. Breast Cancer Res. Treat. 2:5-73. Sa60 Saenger, E. L., F. N. Silverman, T. D. Sterling, and M. E. Turner. 1960. Neoplasia following therapeutic irradiation for benign conditions in childhood. Radiology 74:889-904. Sa82 Sandler, D. P., G. N. Cornstock, and G. M. Matanoski. 1982. Neoplasms following childhood irradiation of the nasopharynx. J. Natl. Cancer Inst. 68:3-8. Sa88 Sanders, C. L., K. E. Lauhala, J. A. Mahaffey, and K. E. McDonald. 1988. Low-Level 239PuO2 Lifespan Studies. Pp. 31-34 in Pacific Northwest Laboratory Annual Report for 1987 to the DOE Office of Energy Research, Part 1, Biomedical Sciences. U.S. Department of Energy PNL-6500 Pt. 1. Springfield, Va.: National Technical Information Service. Sc78 Schneider, A. B., M. J. Favus, M. E. Stachura, M. J. Arnold, and L. A. Frohman. 1978. Salivary gland neoplasms as a late consequence of head and neck irradiation. Ann. Int. Med. 87:160-164. Sc85 Schneider, A. B., E. Shore-Freedman, U. Y. Ryo, C. Bekerman, M. Favus, and S. Pinsky. 1985. Radiation-induced tumors of the head and neck following childhood irradiation. Medicine 64:1-15. Se78 Sevcoca, M., J. Sevc, and J. Thomas. 1978. Alpha irradiation of the skin and the possibility of late effects. Health Phys. 35:803-806. Sh57 Shellabarger, C. J., E. P. Cronkite, V. P. Bond, and S. W. Lippincott. 1957. The occurrence of mammary tumors in the rat after sublethal whole-body irradiation. Radiat. Res. 6:501-512. Sh66 Shellabarger, C. J., V. P. Bond, G. E. Aponte, and E. P. Cronkite. 1966. Results of fractionation and protraction of total-body radiation on rat mammary neoplasia. Cancer Res. 26:509-513. Sh71 Shellabarger, C. J. 1971. Induction of mammary neoplasia after in vitro exposure to x-rays. Proc. Soc. Exp. Biol. Med. 136:1103-1106. Sh80 Shellabarger, C. J., D. Chmelevsky, and A. M. Kellerer. 1980. Induction of mammary neoplasms in the Sprague-Dawley rat by 430-keV neutrons and x-rays. J. Natl. Cancer Inst. 64:821-833. Sh82 Shellabarger, C. J., D. Chmelevsky, A. M. Kellerer, J. P. Stone, and S.
RADIOGENIC CANCER AT SPECIFIC SITES 348 Holtzman. 1982. Induction of mammary neoplasms in the ACI rat by 430-keV neutrons, x- rays and diethylstilbestrol. J. Natl. Cancer Inst. 69:1135-1146. Sh86a Shellabarger, C. J., J. P. Stone, and S. Holtzman. 1986. Experimental carcinogenesis in the breast. Pp. 169-180 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. B. Burns, and R. E. Shore, eds. New York: Elsevier. Sh87 Shimizu, Y., H. Kato, W. J. Schull, D. L. Preston, S. Fujita, and D. A. Pierce. Life Span Study Report 11. Part 1. Comparison of Risk Coefficients for Site-Specific Cancer Mortality Based on the DS86 and T65DR Shielded Kerma and Organ Doses. Technical Report RERF TR 12-87. Hiroshima: Radiation Effects Research Foundation. Sh88 Shimizu, Y., H. Kato, and W. J. Schull. Life Span Study Report 11. Part II. Cancer Mortality in the Years 1950-85 Based on the Recently Revised Doses. Technical Report. Hiroshima: Radiation Effects Research Foundation. Sh76 Shore, R. E., R. E. Albert, and B. S. Pasternack. 1976. Follow-up study of patients treated by x-ray epilation for tinea capitis: Resurvey of post-treatment illness and mortality experience. Arch. Environ. Health 31:17-24. Sh84a Shore, R. E., R. E. Albert, M. Reed, N. Harley, and B. S. Pasternack. 1984. Skin cancer incidence among children irradiated for ringworm of the scalp. Rad. Res. 100:192-204. Sh84b Shore, R. E., E. D. Woodward, and L. H. Hempelmann. 1984. Radiation-induced thyroid cancer. Pp. 131-138 in Radiation Carcinogenesis: Epidemiology and Biological Significance, J. D. Boice, Jr., and J. F. Fraumeni, Jr., eds. New York: Raven. Sh85 Shore, R. E., E. Woodward, N. Hildreth, P. Dvoretsky, L. Hempelmann, and B. Pasternack. 1985. Thyroid tumors following thymus irradiation. J. Natl. Cancer Inst. 74:1177-1184. Sh86b Shore, R., N. Hildreth, E. Woodward, P. Dvoretsky, L. Hempelmann, and B. Pasternack. 1986. Breast cancer among women given x-ray therapy for acute postpartum mastitis. J. Natl. Cancer Inst. 77:689-696. Si81 Sikov, M. R. 1981. Carcinogenesis following prenatal exposure to radiation. Biol. Res. Pregnancy 2:159-167. Si55 Simpson, C. L., L. H. Hempelmann, and L. M. Fuller. 1955. Neoplasia in children treated with x-rays in infancy for thymic enlargement. Radiology 64:840-845. Sm76 Smith, P.G., and R. Doll. 1976. Late effects of x-irradiation in patients treated for metropathia haemorrhagica. Br. J. Radiol. 49:224-232. Sm77 Smith, P. G. 1977. Leukemia and other cancers following radiation treatment of pelvic disease. Cancer 39:1901-1905. Sm81 Smith, P.G., and R. Doll. 1981. Mortality from cancer and all causes among British radiologists. Br. J. Radiol. 54:187-194. Sm82 Smith, P. G., and R. Doll. 1982. Mortality among patients with ankylosing spondylitis after a single treatment course with X-rays. Br. Med. J. 284:449-460. Sm86 Smith, P. G., and A. J. Douglas. 1986. Mortality of workers at the Sellafield plant of British Nuclear Fuels. Br. Med. J. 293:845-854. So63 Socolow, E. L., A. Hashizume, S. Neriishi, and R. Niitani. 1963. Thyroid carcinoma in man after exposure to ionizing radiation. N. Engl. J. Med. 268:406-410. So83 Soffer, D., S. Pittaluga, M. Feiner, and A. J. Beller. 1983. Intracranial meningiomas following low-dose irradiation to the head. J. Neurosurg. 59:1048-1053.
RADIOGENIC CANCER AT SPECIFIC SITES 349 Sp58 Spiess, H., A. Gerspach, and C. W. Mays. 1958. Soft-tissue effects following Ra injections into humans. Health Phys. 35:61-81. St73 Stewart, A. 1973. The carcinogenic effects of low level radiation. A re-appraisal of epidemiologic methods and observations. Health Phys. 24:223-240. St62 Stewart, A., W. Pennypacker, and R. Barber. 1962. Adult leukemia and diagnostic x-rays. Br. Med. J. 2:882. Ta75 Takeichi, N. 1975. Induction of salivary gland tumors following x-ray examination. II. Development of salivary gland tumors in long-term experiments. Med. J. Hiroshima Univ. 23:391-411. Ta76 Takeichi, N., F. Hirose, and H. Yamamoto. 1976. Salivary gland tumors in atomic bomb survivors, Hiroshima, Japan. I. Epidemiology observations. Cancer 38:2462-2468. Th86 Thompson, R. C., and J. A. Mahaffey, eds. 1986. Life-Span Radiation Effects Studies in Animals: What Can They Tell Us? U.S. Department of Energy Report No. CONF-830951. Springfield, Va.: National Technical Information Service. Ti77 Tisell, L. E., G. Hansson, S. Lindberg, and I. Ragnhult. 1977. Hyperparathyroidism in persons treated with x-rays for tuberculous cervical adenitis. Cancer 40:846-854. To87 Tokunaga, M., C. E. Land, T. Yamamoto, M. Asano, S. Tokuoka, H. Ezaki, and I. Nishimori. 1987. Incidence of female breast cancer among atomic bomb survivors, Hiroshima and Nagasaki, 1950-1980. Radiat. Res. 112:243-272. To83 Tolley, H. D., S. Marks, J. A. Buchanan, and E. S. Gilbert. 1983. A further update of the analysis of mortality of workers in a nuclear facility. Radiat. Res. 95:211-213. Tr77 Triggs, S. M., and E. D. Williams. 1977. Irradiation of the thyroid as a cause of parathyroid adenoma. Lancet i:293-294. Ts73 Tsubouchi, S., and T. Matsuzawa. 1973. Nodular formations in rat small intestine after local abdominal x-irradiation. Cancer Res. 33:3155-3158. Ul79 Ullrich, R. L., and J. B. Storer. 1979. Influence of irradiation on the development of neoplastic disease in mice. II. Solid tumors . Radiat. Res. 80:317-324. Ul84 Ullrich, R. L. 1984. Tumor induction in BALB/c mice after fractionated or protracted exposures to fission spectrum neutrons. Radiat. Res. 97:587-597. Ul86 Ullrich, R. L. 1986. The rate of progression of radiation-transformed mammary epithelial cells is enhanced after low-dose-rate neutron irradiation. Radiat. Res. 105:68-75. Ul87 Ullrich, R. L., and R. J. Preston. 1987. Myeloid leukemia incidence in male RFM mice following irradiation with x-rays or fission spectrum neutrons. Radiat. Res. 109:165-170. Ul87b Ullrich, R. L., M. C. Jernigan, L. C. Satterfield, and N. D. Bowles. 1987. Radiation carcinogenesis: Time-dose relationships. Radiat. Res. 111:179-184. UN77 United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR). 1977. Sources and Effects of Ionizing Radiation. Report E. 77. IX. 1. New York: United Nations. Pp. 725. UN82 United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR). 1982. Ionizing Radiation: Sources and Biological Effects. Report E, 82, IX, 8. New York: United Nations. Pp. 773. UN86 United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR). 1986. Genetic and Somatic Effects of Ionizing Radiation, Report. New York: United Nations. Pp. 366.
RADIOGENIC CANCER AT SPECIFIC SITES 350 UN88 United Nations Scientific Committee on the Effects of Atomic Radiation (NSCEAR). 1988. Sources, Effects and Risks of Ionizing Radiation. Report E.88.IX.7. New York: United Nations. Up64 Upton, A. C., V. K. Jenkins, and J. W. Conklin. 1964. Myeloid leukemia in the mouse. Ann. N.Y. Acad. Sci. 114:189-201. Up66 Upton, A. C., V. K. Jenkins, H. E. Walburg, Jr., R. L. Tyndall, J. W. Conklin, and N. Wald. 1966. Observations on viral, chemical, and radiation-induced myeloid and lymphoid leukemias in RF mice. Natl. Cancer Inst. Monogr. 22:329-347. Up70 Upton, A. C., M. L. Randolph, and J. W. Conklin (with the collaboration of M. A. Kastenbaum, M. Slater, G. S. Melville, Jr., F. P. Conte, and J. A. Sproul, Jr.). 1970. Late effects of fast neutrons and gamma-rays in mice as influenced by the dose rate of irradiation: Induction of neoplasia. Radiat. Res. 41: 467-491. Va78 van Kaick, G., D. Lorenz, H. Muth, and A. Kaul. 1978. Malignancies in German thorotrast patients and estimated tissue dose. Health Phys. 35:127-136. Va84 van Kaick, G. H. Muth, and A. Kaul. 1984. The German Thorotrast Study. Results of Epidemiological, Clinical and Biophysical Examinations on Radiation-Induced Late Effects in Man Caused by Incorporated Colloidal Thorium Dioxide (Thorotrast). Report No. EUR 9504 EN. Luxembourg: Commission of the European Communities. Va86 Vaughn, J. Carcinogenic effects of radiation on the human skeleton and supporting tissues. Pp. 311-334 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. J. Burns, and R. E. Shore, eds. New York: Elsevier. Vo72 Vogel, H. H., and R. Valdivar. Neutron-induced mammary neoplasms in the rat. Cancer Res. 32:933-938. Wa64 Wagoner, J. K., V. E. Archer, V. E. Carroll, D. A. Holaday, and P. A. Lawrence . 1964. Cancer mortality patterns among U.S. uranium miners and millers, 1950 through 1962. J. Natl. Cancer Inst. 32:787-801. Wa84 Wagoner, J. K. 1984. Leukemia and other malignancies following radiation therapy for gynecological disorders. Pp. 153-159 in Radiation Carcinogenesis: Epidemiology and Biological Significance, J. D. Boice, Jr., and J. F. Fraumeni, Jr., eds. New York: Raven. Wa82 Wakisaka, S., T. L. Kemper, H. Nakagaki, and R. R. O'Neill. 1982. Brain tumors induced by radiation in Rhesus monkeys. Fukuoka Acta Med. 73:585. Wa83 Wakabayashi, T., H. Kato, T. Ikeda, and W. J. Schull. Studies of the mortality of A-bomb survivors. Report 7, Part III. Incidence of cancer in 1959-1978, based on the tumor registry, Nagasaki. Radiat. Res. 93:112-146. Wa68 Wanebo, C. K., K. G. Johnson, K. Sato, and T. W. Thorslund. 1968. Breast cancer after exposure to the atomic bombings of Hiroshima and Nagasaki. N. Engl. J. Med. 279:667-671. Wa88b Wang, J.-X., J. D. Boice, Jr., B.-X. Li, J.-Y. Zhang, and J. F. Fraumeni Jr.. 1988. Cancer among medical diagnostic x-ray workers in China. J. Natl. Cancer Res. 80:344-350. Wa86 Watanabe, H., A. Ito, and F. Hirose. 1986. Experimental carcinogenesis in the digestive and genitourinary tracts. Pp. 233-244 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. J. Burns, and R. E. Shore, eds. New York: Elsevier. Wa88 Watanabe, H., M. A. Tanner, F. E. Domann, M. N. Gould, and K. H. Clifton. 1988. Inhibition of carcinoma formation and of vascular invasion in grafts of radiation-initiated thyroid clonogens by unirradiated thyroid cells. Carcinogenesis 9:1329-1335.
RADIOGENIC CANCER AT SPECIFIC SITES 351 Yo85 Yochmowitz, M.G., D. H. Wood, and Y. L. Salmon. 1985. Seventeen-year mortality experience of proton radiation in Macaca mulatta. Radiat. Res. 102:14. Yo77 Yokoro, K., M. Nakano, A. Ito, K. Nagao, and Y. Kodama. 1977. Role of prolactin in rat mammary carcinogenesis: Detection of carcinogenicity of low-dose carcinogens and of persisting dormant cancer cells. J. Natl. Cancer Inst. 58:1777-1783. Yo78 Yokoro, K., C. Sumi, A. Ito, K. Hamada, K. Kanda, and T. Kobayashi. 1978. Mammary carcinogenic effect of low-dose fission radiation in Wistar/Furth rats and its dependency on prolactin. J. Natl. Cancer Inst. 64:1459-1466. Yo86 Yokoro, K. 1986. Experimental radiation leukemogenesis in mice. Pp. 137-150 in Radiation Carcinogenesis, A. C. Upton, R. E. Albert, F. J. Burns, and R. E. Shore, eds. New York: Elsevier. Yo77b Yoshizawa, Y., T. Kusama, and K. Morimoto. 1977. Search for the lowest irradiation dose from literature on radiation-induced bone tumors. Nippon Acta Radiol. 37:377-386. Yo81 Young, J. L., C. L. Percy, and A. J. Asire. 1981. Surveillance, Epidemiology, and End Results. Incidence and Mortality Data. 1973-1977. National Cancer Institute Monograph 57. NIH Publication No. 81-2330. Washington, D.C.: U.S. Government Printing Office. Yu88 Yuan, J. M., M. C. Yu, R. K. Ross, Y. T. Gao, et al. 1988. Risk factors for breast cancer in Chinese women in Shanghai. Cancer Res. 48:1949-1953. Zo80 Zook, B. C., E. W. Bradley, G. W. Casarett, and C. C. Rogers. 1980. Pathologic findings in canine brain irradiated with fractionated fast neutrons or photons. Radiat. Res. 84:562-578.