Risks of second primary cancer among patients with major histological types of lung cancers in both men and women.

BACKGROUND: Patterns of second primary cancers (SPCs) following first primary lung cancers (FPLCs) may provide aetiological insights into FPLC. METHODS: Cases of FPLCs in 13 cancer registries in Europe, Australia, Canada, and Singapore were followed up from the date of FPLC diagnosis to the date of SPC diagnosis, date of death, or end of follow-up. Standardised incidence ratios (SIRs) were calculated to estimate the magnitude of SPC development following squamous cell carcinoma (SCC), small cell lung carcinoma (SCLC), and adenocarcinoma (ADC). RESULTS: Among SCC patients, male SIR=1.58 (95% confidence interval (CI)=1.50-1.66) and female SIR=2.31 (1.94-2.72) for smoking-related SPC. Among SCLC patients, the respective ratios were 1.39 (1.20-1.60) and 2.28 (1.73-2.95), and among ADC patients, they were 1.73 (1.57-1.90) and 2.24 (1.91-2.61). We also observed associations between first primary lung ADC and second primary breast cancer in women (SIR=1.25, 95% CI=1.05-1.48) and prostate cancer (1.56, 1.39-1.79) in men. CONCLUSION: The FPLC patients carried excess risks of smoking-related SPCs. An association between first primary lung ADC and second primary breast and ovarian cancer in women at younger age and prostate cancers in men may reflect an aetiological role of hormones in lung ADC.

Risk of second malignant neoplasms after childhood central nervous system malignant tumours: an international study.

PURPOSE: The aim of this study was to assess the risk of second malignant neoplasms (SMNs) other than central nervous system (CNS) neoplasms after childhood CNS cancer in an international multicentre study. METHODS: Individual data on cases of CNS cancer in children (0-14 years) and on subsequent SMNs were obtained from 13 population-based cancer registries contributing data for different time periods in 1943-2000. Standardised incidence ratios (SIRs) with 95% confidence intervals (CI), absolute excess risk and cumulative incidence of SMNs were computed. RESULTS: We observed 43 SMNs in 8431 CNS cancer survivors. The SIR was 10.6 (4.85-20.1) for thyroid cancer (nine cases), 2.75 (1.01-5.99) for leukaemia (six cases) and 2.47 (0.90-5.37) for lymphoma (six cases). The SIRs were highest in the first 10 years after CNS cancer diagnosis. The cumulative incidence of non-CNS SMNs was 3.30% (0.95-5.65%) within 45 years after a CNS cancer diagnosis. Within 15 years, the cumulative incidence was highest for cases diagnosed after 1980 (0.56%, 95% CI: 0.29-0.82%). CONCLUSION: This population-based study indicates that about one every 180 survivors of a childhood CNS cancer will develop a non-CNS SMN within the following 15 years. The excess is higher after glioma and embryonal malignant tumour than after another CNS tumour.

Quality, comparability and methods of analysis of data on childhood cancer in Europe (1978-1997): report from the Automated Childhood Cancer Information System project.

In collaboration with 62 population-based cancer registries contributing to the Automated Childhood Cancer Information System (ACCIS), we built a database to study incidence and survival of children and adolescents with cancer in Europe. We describe the methods and evaluate the quality and internal comparability of the database, by geographical region, period of registration, type of registry and other characteristics. Data on 88,465 childhood and 15,369 adolescent tumours registered during 1978-1997 were available. Geographical differences in incidence are caused partly by differences in definition of eligible cases. The observed increase in incidence rates cannot be explained by biases due to the selection of datasets for analyses, and only partially by the registration of non-malignant or multiple primary tumours. Part of the observed differences in survival between the regions may be due to variable completeness of follow-up, but most is probably explained by resource availability and organisation of care. Further standardisation of data and collection of additional variables are required so that this study may continue to yield valuable results with reliable interpretation.

Second primary cancers among 109 000 cases of non-Hodgkin's lymphoma.

An analysis of other primary cancers in individuals with non-Hodgkin's lymphoma (NHL) can help to elucidate this cancer aetiology. In all, 109 451 first primary NHL were included in a pooled analysis of 13 cancer registries. The observed numbers of second cancers were compared to the expected numbers derived from the age-, sex-, calendar period- and registry-specific incidence rates. We also calculated the standardised incidence ratios for NHL as a second primary after other cancers. There was a 47% (95% confidence interval 43-51%) overall increase in the risk of a primary cancer after NHL. A strongly significant (P<0.001) increase was observed for cancers of the lip, tongue, oropharynx*, stomach, small intestine, colon*, liver, nasal cavity*, lung, soft tissues*, skin melanoma*, nonmelanoma skin*, bladder*, kidney*, thyroid*, Hodgkin's lymphoma*, lymphoid leukaemia* and myeloid leukaemia. Non-Hodgkin's lymphoma as a second primary was increased after cancers marked with an asterisk. Patterns of risk indicate a treatment effect for lung, bladder, stomach, Hodgkin's lymphoma and myeloid leukaemia. Common risk factors may be involved for cancers of the lung, bladder, nasal cavity and for soft tissues, such as pesticides. Bidirectional effects for several cancer sites of potential viral origin argue strongly for a role for immune suppression in NHL.

Second primary malignancies in patients with male breast cancer.

An international multicentre study of first and second primary neoplasms associated with male breast cancer was carried out by pooling data from 13 cancer registries. Among a total of 3409 men with primary breast cancer, 426 (12.5%) developed a second neoplasia; other than breast cancer, a 34% overall excess risk of second primary neoplasia, affecting the small intestine (standardised incidence ratio, 4.95, 95% confidence interval, 1.35-12.7), rectum (1.78, 1.20-2.54), pancreas (1.93, 1.14-3.05), skin (nonmelanoma, 1.65, 1.16-2.29), prostate (1.61, 1.34-1.93) and lymphohaematopoietic system (1.63, 1.12-2.29). A total of 225 male breast cancers was recorded after cancers other than breast cancer, but an increase was found only after lymphohaematopoietic neoplasms. BRCA2 (and to some extent BRCA1) mutations may explain the findings for pancreatic and prostate cancers. Increases at other sites may be related to unknown factors or to chance. This large study shows that the risks for second discordant tumours after male breast cancer pose only a moderate excess risk.

European health systems and cancer care.

INTRODUCTION: Data on the survival of all incident cases collected by population-based cancer registries make it possible to evaluate the overall performance of diagnostic and therapeutic actions on cancer in those populations. EUROCARE-3 is the third round of the EUROCARE project, the largest cancer registry population based collaborative study on survival in European cancer patients. The EUROCARE-3 study analysed the survival of cancer patients diagnosed from 1990 to 1994 and followed-up to 1999. Sixty-seven cancer registries of 22 European countries characterised by differing health systems participated in the study. This paper includes essays providing brief overviews of the state and evolution of the health systems of the considered countries and comments on the relation between cancer survival in Europe and some European macro-economic and health system indicators, in the 1990s. OVERVIEW OF THE EUROPEAN HEALTH SYSTEMS: The European health systems underwent a great deal of reorganisation in the last decade; a general tendency being to facilitate expanding involvement of the private sector in health care, a process which occurred mainly in the eastern countries (i.e. the Czech Republic, Estonia, Poland, Slovakia and Slovenia). In contrast, organisational changes in the northern European countries (i.e. Denmark, Iceland, Finland and Sweden) tended to confirm the established public sector systems. Other countries, including the UK and some southern European countries (i.e. England, Scotland, Wales, Malta and Italy) have reduced the public role while the systems remain basically public, at least at present. Our findings clearly suggest that cancer survival (all cancer combined) is related to macro-economic variables such as the gross domestic product (GDP), the total national (public and private) expenditure on health (TNEH) and the total public expenditure on health (TPEH). We found, however, that survival is related to wealth (GDP), but only up to a certain level, after which survival continues to be related to the level of health investment (both TNEH and TPEH). According to the Organisation for Economic Co-operation and Development (OECD), the TNEH increased during the 1990s in all EUROCARE-3 countries, while the ratio of TPEH to TNEH reduced in all countries except Portugal. CONCLUSIONS: Cancer survival depends on the widespread application of effective diagnosis and treatment modalities, but our enquiry suggests that the availability of these depends on macro-economic determinants, including health and public health investment. Analysis of the relationship between health system organisation and cancer outcome is complicated and requires more information than is at present available. To describe cancer and cancer management in Europe, the European Cancer Health Indicator Project (EUROCHIP) has proposed a list of indicators that have to be adopted to evaluate the effects on outcome of proposed health system modifications.

Cancer prevalence in Central Europe: the EUROPREVAL Study.

BACKGROUND: Information on cancer prevalence is either absent or largely unavailable for central European countries. MATERIALS AND METHODS: Austria, Germany, The Netherlands, Poland, Slovakia, Slovenia and Switzerland cover a population of 13 million inhabitants. Cancer registries in these countries supplied incidence and survival data for 465 000 cases of cancer. The prevalence of stomach, colon, rectum, lung, breast, cervix uteri, corpus uteri and prostate cancer, as well as skin melanoma, Hodgkin's disease, leukaemia and all malignant neoplasms combined was estimated for the end of 1992. RESULTS: A large heterogeneity was observed within central European countries. For all cancers combined, estimates ranged from 730 per 100 000 in Poland (men) to 3350 per 100 000 in Germany (women). Overall cancer prevalence was the highest in Germany and Switzerland, and the lowest in Poland and Slovenia. In Slovakia, prevalence was higher than average for men and lower than average for women. This was observed for almost all ages. As shown by incidence data, breast cancer was the most frequent malignancy among women in all countries. Among men, prostate cancer was the leading malignancy in Germany, Austria and Switzerland, and lung cancer was the major cancer in Slovenia, Slovakia and Poland. The Netherlands had a high prevalence of both prostate and lung cancer. Time-related magnitude of prevalence within each country and the variability of such proportions across the countries has been estimated and cancer prevalence is given by time since diagnosis (1 year, 1-5 years, 5-10 years, >10 years) for each site. The weight of 1-year prevalence (248 per 100 000 among men and 253 per 100 000 among women) was <15% of total prevalence. Prevalent cases between 1 and 5 years since diagnosis represented between 22% and 34% of the total prevalence. Prevalent cases diagnosed from 5 to 10 years before (335 per 100 000 for men and 505 per 100 000 for women) represented between 17% and 23% of prevalent cancers. Finally, long-term cancer prevalence (diagnosed >10 years before), reflecting long-term survival, and number of people considered as cured from cancer were 490 per 100 000 for men and 1028 per 100 000 for women, with a range between 26% (The Netherlands, men) and 50% (Slovakia, women). CONCLUSION: It is clear from observing countries in Central Europe, that high cancer prevalence is associated with well-developed economies. This burden of cancer could be interpreted as a paradoxical effect of better treatments and thereby survival. It could also be taken as a sign for not being satisfied with the advances in treating patients diagnosed with cancer, and for supporting more primary prevention.

Second primary cancers in laryngeal cancer patients in Slovenia, 1961-1996.

We analysed the incidence of second primary cancers (SPC) in male laryngeal cancer patients in Slovenia and their survival for the period of 1961-1996. Data were taken from the population-based Cancer Registry of Slovenia. The person-years approach was used and the risk for SPC was expressed as a standardised incidence ratio (SIR). Survival analysis was carried out using the Kaplan-Meier method. Of 2275 male patients, 369 developed SPC (16.2%, total SIR 2.83), most commonly in the head and neck region (SIR 6.07-15.97), lung (SIR 4.15), oesophagus (SIR 4.66), and bladder (SIR 3.0), which points to an important role of common risk factors of smoking and alcohol. SPC were diagnosed in significant excess up to 20 years after the diagnosis of laryngeal cancer. The median survival time from the diagnosis of laryngeal cancer was 3.25 years for patients without a SPC and 6.47 years for patients who developed a SPC. However, the median survival time from the diagnosis of a SPC was only 0.84 years. Patients with laryngeal cancer in Slovenia have a higher risk of developing a SPC than was reported in similar studies in Europe and the USA. This high risk is partly responsible for their relatively poor survival.

Mortality from cardiovascular diseases and exposure to inorganic mercury.

OBJECTIVE: To study the mortality from cardiovascular and other chronic non-neoplastic diseases after long term exposure to inorganic mercury. Limited information is available on the effect of chronic exposure to mercury on the cardiovascular system. METHODS: The mortality was studied among 6784 male and 265 female workers from four mercury mines and mills in Spain, Slovenia, Italy, and the Ukraine. Workers were employed between 1900 and 1990; the follow up period lasted from the 1950s to the 1990s. The mortality of the workers was compared with national reference rates. RESULTS: Among men, there was a slight increase in overall mortality (standardised mortality ratio (SMR) 1.08, 95% confidence interval (95% CI) 1.04 to 1.12). An increased mortality was found from hypertension (SMR 1.46, 95% CI 1.08 to 1.93), heart diseases other than ischaemic (SMR 1.36, 95% CI 1.20 to 1.53), pneumoconiosis (SMR 27.1, 95% CI 23.1 to 31.6), and nephritis and nephrosis (SMR 1.55, 95% CI 1.13 to 2.06). The increase in mortality from cardiovascular diseases was not consistent among countries. Mortality from hypertension and other heart diseases increased with estimated cumulative exposure to mercury; mortality from ischaemic heart disease and cerebrovascular diseases increased with duration of employment, but not with estimated exposure to mercury. Results among women were hampered by few deaths. CONCLUSION: Despite limited quantitative data on exposure, possible confounding, and likely misclassification of disease, the study suggests a possible association between employment in mercury mining and refining and risk in some groups of cardiovascular diseases.

Future trends in breast, cervical, lung, mouth and pharyngeal cancer incidence in Slovenia.

OBJECTIVES: Breast, cervical, lung, mouth and pharyngeal cancers are important public health problems in Slovenia, and in many other Central and South European countries. The aim of this study was to predict the incidence of these cancers in Slovenia up to the year 2009, based on the data of the Cancer Registry of Slovenia for the period 1965-1994 and on the official national population projections for the Republic of Slovenia. METHODS: Age-period-cohort models were applied. In the case of data heterogeneity in lung as well as in mouth and oropharyngeal cancer in males, an additional parameter indicating differences in lifestyle was introduced in the model. RESULTS: After accounting for major site-specific risk factors, we predict in females a steady increase in breast and lung cancer, but no major changes in cervical cancer case-load. In males a steady decrease in the lung cancer case-load is expected throughout the predicted period, while for mouth and pharyngeal cancer a moderate decrease is expected only after the year 2000. CONCLUSION: Despite some uncertainties inherent in cancer incidence predictions, the obtained results are important in setting priorities for national cancer control strategies in Slovenia, especially in further efforts towards primary prevention of lung, mouth and pharyngeal cancer, and in more efficient early detection of breast, cervical, mouth and pharyngeal cancer.

Population density and childhood leukaemia: results of the EUROCLUS Study.

The EUROCLUS study assembled incidence data for 13,551 cases of childhood leukaemia (CL) diagnosed between 1980 and 1989 in 17 countries (or regions of countries). These were referenced by location at diagnosis to small census areas of which there were 25,723 in the study area. Population counts, surface area and, hence, population density were available for all these small areas. Previous analyses have shown limited extra-Poisson variation (EPV) of case counts within small areas; this is most pronounced in areas of intermediate population density (150-499 persons/km2). In this study, the data set was examined in more detail for evidence that variations in incidence and EPV of CL are associated with population density. Incidence showed a curvilinear association with population density and was highest in areas which were somewhat more densely populated (500-750 persons/km2), where the incidence rate ratio relative to areas having > or = 1000 persons/km2 was 1.16 (95% confidence interval 1.07-1.26) and the P value for quadratic trend across eight strata of population density was 0.02. Incidence in these areas is uniformly elevated and showed no evidence of heterogeneity (i.e. EPV). Statistically significant evidence of EPV was evident amongst some of the areas previously classified as intermediate density areas (specifically, those with a density of 250-499 persons/km2, P < 0.001 for CL). These results were interpreted in terms of the current aetiological hypotheses for CL which propose that exposure to localised epidemics of one or more common infectious agent may contribute to the development of leukaemia. They suggest that such epidemics arise regularly in moderately densely populated areas and also sporadically in areas which are somewhat less densely populated. Although other interpretations are possible, these results may assist in the identification of characteristics which infectious agents must possess if direct or indirect causes of CL.

Spatial clustering of childhood leukaemia: summary results from the EUROCLUS project.

The interpretation of reports of clusters of childhood leukaemia is difficult, first because little is known about the causes of the disease, and second because there is insufficient information on whether cases show a generalized tendency to cluster geographically. The EUROCLUS project is a European collaborative study whose primary objective is to determine whether the residence locations of cases at diagnosis show a general tendency towards spatial clustering. The second objective is to interpret any patterns observed and, in particular, to see if clustering can be explained in terms of either infectious agents or environmental hazards as aetiological agents. The spatial distribution of 13351 cases of childhood leukaemia diagnosed in 17 countries between 1980 and 1989 has been analysed using the Potthoff-Whittinghill method. The overall results show statistically significant evidence of clustering of total childhood leukaemia within small census areas (P=0.03) but the magnitude of the clustering is small (extra-Poisson component of variance (%) = 1.7 with 90% confidence interval 0.2-3.1). The clustering is most marked in areas that have intermediate population density (150-499 persons km[-2]). It cannot be attributed to any specific age group at diagnosis or cell type and involves spatial aggregation of cases of different ages and cell types. The results indicate that intense clusters are a rare phenomenon that merit careful investigation, although aetiological insights are more likely to come from investigation of large numbers of cases. We present a method for detecting clustering that is simple and readily available to cancer registries and similar groups.

Cancer occurrence among European mercury miners.

OBJECTIVES: To study the carcinogenicity of inorganic mercury in humans. METHODS: We studied the mortality from cancer among 6784 male and 265 female workers of four mercury mines and mills in Spain, Slovenia, Italy and the Ukraine. Workers were employed between the beginning of the century and 1990; the follow-up period lasted from the 1950s to the 1990s. We compared the mortality of the workers with national reference rates. RESULTS: Among men, there was no overall excess cancer mortality; an increase was observed in mortality from lung cancer (standardized mortality ratio [SMR] 1.19, 95 percent confidence interval [CI] 1.03-1.38) and liver cancer (SMR 1.64, CI 1.18-2.22). The increase in lung cancer risk was restricted to workers from Slovenia and the Ukraine: no relationship was found with duration of employment or estimated mercu ry exposure. The increase in liver cancer risk was present both among miners and millers and was stronger in workers from Italy and Slovenia: there was a trend with estimated cumulative exposure but not with duration of employment, and the excess was not present in a parallel analysis of cancer incidence among workers from Slovenia. No increase was observed for other types of cancer, including brain and kidney tumours. Among female workers (Ukraine only), three deaths occurred from ovarian cancer, likely representing an excess. CONCLUSIONS: Exposure to inorganic mercury in mines and mills does not seem strongly associated with cancer risk, with the possible exception of liver cancer; the increase in lung cancer may be explained by co-exposure to crystalline silica and radon.

Second primary cancers in breast cancer patients in Slovenia.

Data from the Cancer Registry of Slovenia were used in a cohort study to determine whether the incidence of second primary cancers in patients with first primary breast cancer differs from the incidence expected in the general population. Special interest was given to long-term survivors. The expected numbers of second primary cancers were calculated by multiplying the number of appropriate person-years at risk by the corresponding age- and calendar-period-specific cancer incidence rates for women in Slovenia. The risk of a second primary cancer was expressed as the standardized incidence ratio (SIR). Of the 8,917 patients newly diagnosed in the period 1961-85 and followed-up to the end of 1994, 547 (6.2 percent) developed second primary cancers, whereas 410 (4.7 percent) were expected (SIR = 1.3, 95 percent confidence interval [CI] = 1.2-1.4). The risk was higher among younger patients. In long-term survivors, the risk was increased significantly for second primary cancer of the breast (SIR = 1.4, CI = 1.1-1.7), lung cancer (SIR = 1.6, CI = 1.1-2.3), melanoma (SIR = 2.7, CI = 1.5-4.4) and non-melanoma skin cancers(SIR = 2.0, CI = 1.6-2.4), corpus uteri cancer(SIR = 1.6, CI = 1.2-2.1), ovarian cancer(SIR = 2.3, CI = 1.7-3.0), and thyroid cancer (SIR = 2.5, CI = 1.2-4.6). Our results confirm the findings of several cohort studies carried out in Europe, the United States, and Japan, indicating that breast cancer patients should be monitored carefully for the occurrence of second primary cancers.

Second cancers occurring after cancers of the mouth and pharynx: data from three population-based registries in Australia, Scotland and Slovenia.

Data over at least 20 years from three large population-based registries in Europe and Australasia have been used to assess the risk of second primary tumours occurring after a cancer of the mouth or pharynx. These patients have previously been shown in clinical series to be at a particularly high risk of subsequent tumours, while data from cancer registries have shown conflicting results on the magnitude of the risk. In this study, patients were found to have between a 2-fold (Scotland and New South Wales) and 4-fold (Slovenia) increase in risk of a subsequent tumour over that in the population, although the actual risk in each centre was similar (between 2.8 and 3.1 per 100 person years). The risk remained for 10 years after diagnosis of the original tumour and was primarily in the upper aerodigestive tract. The most elevated risks (approximately 10-fold) were for tumours in the oral cavity and oesophagus. These data provide higher estimates of risk than previously reported from European cancer registries for second primary tumours and emphasize the need for close follow-up of patients who may represent an appropriate population in which to assess possible new chemopreventive agents.

Intestinal metaplasia types and the risk of gastric cancer: a cohort study in Slovenia.

Between 1967 and 1976, 1,525 Slovenian patients with a histological diagnosis of intestinal metaplasia (IM) were classified according to subtype of IM based on morphology and mucin staining; 518 cases were diagnosed with type I, 197 with type II and 275 with type III, but in 291 the diagnosis of IM was not confirmed. Patients who had developed cancer or died up to 1986 were identified by record linkage at the Slovenia Cancer Registry and the Central Population Registry in Slovenia. A total of 34 incident cases of gastric cancer occurring at least 6 months after the diagnosis of IM were identified. The standardised incidence ratio (SIR) for stomach cancer was 2.23 in the whole cohort. It was highest for IM type III, followed by type II and IM-unconfirmed, but not increased for type I. The relative risk (RR) of developing gastric cancer based on Cox's proportional hazards model was 2.14 for type II and 4.58 for type III, compared with type I. The RR was especially increased for a subgroup of type III secreting sulphomucins in their goblet cells in comparison with types I-II negative to sulphomucins. Our results confirm that subtyping of IM is useful for identifying individuals at high risk for gastric cancer.

Diffuse-type gastric cancer in a young adult presenting with neurological symptoms and liver rupture.

A 27-yr-old man presented with neurological symptoms and subsequent painful liver enlargement. Cranial computerized tomographic scans and gastroscopy were negative. Ultrasound examination revealed nonhomogeneous echo pattern of the enlarged liver; a guided biopsy specimen suggested hepatoma. The patient died of hemorrhagic shock 5 days after liver biopsy and 4 wk after the initial presentation. Autopsy revealed diffuse-type gastric carcinoma as the primary lesion, liver ruptures at sites distant from the biopsy, thrombosis of the sagittal sinus, and widespread permeation of blood and lymphatic vessels with anaplastic carcinoma cells. The incidence of gastric cancer in young adults from the Cancer Registry of Slovenia is presented for the period from 1979 to 1987. Our case corroborates the importance of considering this malignancy early in the evaluation of young symptomatic patients.

Solving the problem of small-population-based areas for the analysis of rare diseases by clustering with constraints methods.

The aim of this study was to find a sensible fusion of small geographical areas into, as far as possible, homogeneous larger regions with the necessary minimal population size according to 14 indicators of socioeconomic development, which is known to be indirectly related to cancer incidence. The starting point was the minimal population size which could still provide an estimation of a statistically significantly lower rate relative to the national average. Being aware of the heterogeneity and complexity of cancer etiology, the problem was studied step by step: regionalization was obtained according to selected socioeconomic indicators with different numbers of regions (from 60 to 32). With the best-obtained regionalization into 32 regions by clustering with constraints methods, zero values were reduced from 112 to 6, while almost the same variance of most cancers was retained.

Radiation dose and leukemia risk in patients treated for cancer of the cervix.

To quantify the risk of radiation-induced leukemia and provide further information on the nature of the relationship between dose and response, a case-control study was undertaken in a cohort of over 150,000 women with invasive cancer of the uterine cervix. The cases either were reported to one of 17 population-based cancer registries or were treated in any of 16 oncologic clinics in Canada, Europe, and the United States. Four controls were individually matched to each of 195 cases of leukemia on the basis of age and calendar year when diagnosed with cervical cancer and survival time. Leukemia diagnoses were verified by one hematologist. Radiation dose to active bone marrow was estimated by medical physicists on the basis of the original radiotherapy records of study subjects. The risk of chronic lymphocytic leukemia, one of the few malignancies without evidence for an association with ionizing radiation, was not increased [relative risk (RR) = 1.03; n = 52]. However, for all other forms of leukemia taken together (n = 143), a twofold risk was evident (RR = 2.0; 90% confidence interval = 1.0-4.2). Risk increased with increasing radiation dose until average doses of about 400 rad (4 Gy) were reached and then decreased at higher doses. This pattern is consistent with experimental data for which the down-turn in risk at high doses has been interpreted as due to killing of potentially leukemic cells. The dose-response information was modeled with various RR functions, accounting for the nonhomogeneous distribution of radiation dose during radiotherapy. The local radiation doses to each of 14 bone marrow compartments for each patient were incorporated in the models, and the corresponding risks were summed. A good fit to the observed data was obtained with a linear-exponential function, which included a positive linear induction term and a negative exponential term. The estimate of the excess RR per rad was 0.9%, and the estimated RR at 100 rad (1 Gy) was 1.7. The model proposed in this study of risk proportional to mass exposed and of risk to an individual given by the sum of incremental risks to anatomic sites appears to be applicable to a wide range of dose distributions. Furthermore, the pattern of leukemia incidence associated with different levels of radiation dose is consistent with a model postulating increasing risk with increasing exposure, modified at high doses by increased frequency of cell death, which reduces risk.