Childhood cancer research in Oxford I: the Oxford Survey of Childhood Cancers.

BACKGROUND: Significant research on the epidemiology and natural history of childhood cancer took place in the Universities of Oxford and Birmingham over sixty years. This is the first of three papers recording this work and describes the Oxford Survey of Childhood Cancers (OSCC), the largest case-control survey of childhood cancer ever undertaken. METHODS: The OSCC studied deaths in Britain from 1953 to 1981. Parents were interviewed and medical records from ante-natal clinics and treatment centres were followed up and abstracted. The survey left Oxford in 1975 and was run subsequently from Birmingham. The data are now being documented and archived to make them available for future study. RESULTS: Many papers have resulted from this survey, most notably those relating to the association first reported therein between childhood cancer and ante-natal X-raying. This paper is a historical review of the OSCC. CONCLUSIONS: In spite of many analyses of the study, this historic data set has continuing value because of the large number of examples of some very rare tumours and the detailed clinical and family history data that are available; and also because of the possibility of carrying out new analyses to investigate emerging research issues.

Second and subsequent tumours among 1927 retinoblastoma patients diagnosed in Britain 1951-2004.

BACKGROUND: Retinoblastoma is an eye tumour of childhood that occurs in heritable and non-heritable forms. In the heritable form, there is a predisposition to the development of non-ocular subsequent primary tumours (SPTs). METHODS: This study included 1927 retinoblastoma patients diagnosed in Britain from 1951 to 2004. Ascertainment was through the (UK) National Registry of Childhood Tumours; cases were followed-up for the occurrence of SPTs. Standardised incidence ratios (SIRs) were calculated. RESULTS: We identified 169 SPTs in 152 patients. The SIR analysis included 145 SPTs with cancer registrations from the years 1971 to 2009. These tumours occurred in 132 patients: 112 of the 781 heritable and 20 of the 1075 (presumed) non-heritable cases under surveillance at the start of this period developed at least one registered SPT. The SIRs for all tumours combined were 13.7 (95% confidence interval 11.3-16.5) in heritable cases and 1.5 (0.9-2.3) in non-heritable cases. The main types of SPT in the heritable cases were leiomyosarcoma, (31 cases; SIR 1018.7 (692.2-1446.0)), osteosarcoma (26 cases; SIR 444.6 (290.4-651.4)), and skin melanoma (12 cases; SIR 18.6 (9.6-32.4)). CONCLUSION: The risk of SPTs in heritable retinoblastoma is extremely high. This has important implications for the clinical follow-up and counselling of survivors and their families.

Childhood cancer and magnetic fields from high-voltage power lines in England and Wales: a case-control study.

BACKGROUND: Epidemiological evidence suggests that chronic low-intensity extremely-low-frequency magnetic-field exposure is associated with increased risk of childhood leukaemia; it is not certain the association is causal. METHODS: We report a national case-control study relating childhood cancer risk to the average magnetic field from high-voltage overhead power lines at the child's home address at birth during the year of birth, estimated using National Grid records. From the National Registry of Childhood Tumours, we obtained records of 28,968 children born in England and Wales during 1962-1995 and diagnosed in Britain under age 15. We selected controls from birth registers, matching individually by sex, period of birth, and birth registration district. No participation by cases or controls was required. RESULTS: The estimated relative risk for each 0.2 µT increase in magnetic field was 1.14 (95% confidence interval 0.57 to 2.32) for leukaemia, 0.80 (0.43-1.51) for CNS/brain tumours, and 1.34 (0.84-2.15) for other cancers. CONCLUSION: Although not statistically significant, the estimate for childhood leukaemia resembles results of comparable studies. Assuming causality, the estimated attributable risk is below one case per year. Magnetic-field exposure during the year of birth is unlikely to be the whole cause of the association with distance from overhead power lines that we previously reported.

Cancer in the offspring of female radiation workers: a record linkage study.

This study uses record linkage between the National Registry of Childhood Tumours (NRCT) and the National Registry for Radiation Workers to re-assess our earlier finding that the offspring of women radiation workers exposed to ionising radiation before the child's conception may be at an increased risk of childhood cancer. An additional 16,964 childhood cancer patients taken from the NRCT, together with the same number of matched controls, are included. Pooled analyses, based on the new and original datasets, include 52,612 cases and their matched controls. Relative risks (RRs) for maternal employment as a radiation worker, maternal exposure or not during the relevant pregnancy and pattern of employment relative to conception and diagnosis dates were calculated.The new data provide no evidence of an increased risk of childhood cancer associated with maternal preconception radiation work and thus do not support our earlier finding of a raised risk in the offspring of female radiation workers. Considering the pooled data, a weak association was found between maternal radiation work during pregnancy and childhood cancer in offspring although the evidence is limited by the small numbers of linked cases and controls.

Retinoblastoma incidence and survival in European children (1978-1997). Report from the Automated Childhood Cancer Information System project.

Based on 2283 cases of retinoblastoma diagnosed in children aged 0-14 years, incidence and survival in Europe during the period 1978-1997 are described. Data were provided to the Automated Childhood Cancer Information System (ACCIS) from 60 paediatric and general cancer registries. During 1988-1997, the cumulative incidence of retinoblastoma in the ACCIS regions was found to be between 44.2 and 67.9 per million births. The highest incidence was seen in the first year of life. The age-standardised (World standard) incidence rate for the age-range 0-14 years was 4.1 per million. Approximately one-third of cases had bilateral tumours. Overall incidence increased over the period 1978-1997 by 1% per year, as derived from a model adjusted for sex, age group and type of registry (general or paediatric). The 5-year survival rate improved from 89% to 95% during the period covered by the study. This improvement was seen in both unilateral and bilateral cases but was significant only for the unilateral tumours. Survival was lower in the East region, although smaller differences were also observed between the other four regions (British Isles, North, South and West). Availability and quality of registration data on retinoblastoma need to be improved for effective evaluation of incidence and survival.

Calculation of sample size in trials of screening for early diagnosis of disease.

The calculation of the sample sizes required for trials of screening for disease with the aim of reducing mortality involves the estimation of both the mortality in the control group at intervals after the start of the trial, and the potential reduction due to screening. Since at the start of a screening trial the population is selected to be free of the disease in question, the mortality rate in the control group will differ from that in the general population. A method of estimating this mortality rate using published incidence and survival data is described. The expected reduction in mortality due to screening depends both on the number of, and intervals between, screens and on parameters concerning the natural history of the disease; the means by which these parameters can be estimated are discussed. A trial of screening for colorectal cancer by a faecal occult blood test is used to illustrate these calculations.

Apparent association between benzene and childhood leukaemia: methodological doubts concerning a report by Knox.

A recent study by Knox concludes that cases and "clusters" of two or more cases of childhood leukaemia and non-Hodgkin's lymphoma occur closer to many kinds of industrial installation than to supposedly comparable control locations. It is argued that these findings could be largely or entirely artefactual, the apparent differences arising out of the inappropriateness of the control data. Knox used randomly selected postcode units as controls, a procedure that leads to the comparison of individuals located in areas with typically quite different population densities from those for the cases. The resulting potential for bias is explored and the arguments are exemplified by analysing household data based on postcodes.

Occupations of fathers of children dying from neoplasms.

It has been suggested in a number of recent reports that there is a possible relationship between parental occupation and malignant disease in children. A proportional mortality analysis relating deaths among children in England and Wales in 1959-63 and 1970-72 to occupation of father as stated on the child's death certificate has not shown any convincing evidence for such associations. Earlier papers published on the subject are reviewed. Although there is some slight evidence for associations between childhood tumours and certain parental occupations there is little consistency between the results reported by different authors. A previously reported association between higher social class and deaths from neoplasms was found also in this study. The explanation for this finding is unknown, and it remains possible that it is an artefact.

The Inter-Regional Epidemiological Study of Childhood Cancer (IRESCC): past medical history in children with cancer.

The Inter-Regional Epidemiological Study of Childhood Cancer (IRESCC) collected interview and medical information relating to the child's past medical experiences from parents of 555 children diagnosed with cancer and parents of 1110 unaffected matched controls. No significant associations emerged overall for ante-natal care, place and mode of delivery, length of gestation, birth weight, condition at birth, special care, neonatal procedures or breast-feeding. Few risk factors relating to previous illnesses and medication were found, although increasing numbers of illnesses appeared to be associated with an increased risk of childhood cancer, particularly acute lymphoblastic leukaemia. A highly significant excess of case children had not been immunised (p = 0.005). In general, these results indicate that past medical experiences have little influence on the development of cancer in children.

Retinoblastoma in Great Britain 1969-80: incidence, treatment, and survival.

Patients with retinoblastoma diagnosed from 1969 to 1980 have been followed up for periods of up to 17 years. Data from a previous study of patients diagnosed from 1962 to 1968 have been included for analysis of incidence and second primary tumours, and for study of trends in treatment. The registration rate in Britain (which may be about 10% less than the true incidence) is about one in 23,000 live births, approximately 40% of cases being known to be genetic. There is no apparent trend in incidence during the period covered by these two studies. The three-year survival rate in 88%. Patients with bilateral tumours have a better survival rate than those with unilateral tumours for the first few years, but their long-term survival rate is worse because of later deaths from ectopic intracranial retinoblastoma or second primary neoplasms. Older children tend to have a worse prognosis, which is related to the fact that their tumours are diagnosed at a more advanced stage. There is a significantly higher survival rate for boys than for girls; this is partly accounted for by difference in age and stage at diagnosis between the sexes. Children referred to units specialising in the treatment of retinoblastoma have a higher three-year survival rate than those treated at other hospitals. Comparing methods of treatment between the periods 1962-8 and 1969-80, we find there has been a trend towards more conservative treatment. The use of chemotherapy is now usually reserved for recurrences and metastases and for palliative treatment in terminal retinoblastoma.

An overview of reports and current research concerning childhood leukaemia and cancer around nuclear installations in the UK.

Studies around UK nuclear installations suggest that it is possible that the incidence of leukaemia in young persons may be increased in the vicinity of such installations. However, the results of these studies are inconsistent and the cause or causes of the apparent increases are unknown; they may be at least in part attributable to chance variations in incidence, to selection factors, or to variations in completeness of recording. The cluster in Seascale, the village closest to the Sellafield reprocessing plant, may be the result of germ cell mutations in the fathers of the affected children.

Distribution of childhood leukaemias and non-Hodgkin's lymphomas near nuclear installations in England and Wales.

OBJECTIVE: To examine the relation between the risk of childhood leukaemia and non-Hodgkin's lymphoma and proximity of residence to nuclear installations in England and Wales. DESIGN: Observed and expected numbers of cases were calculated and analysed by standard methods based on ratios of observed to expected counts and by a new statistical test, the linear risk score test, based on ranks and designed to be sensitive to excess incidence in close proximity to a putative source of risk. SETTING: Electoral wards within 25 km of 23 nuclear installations and six control sites that had been investigated for suitability for generating stations but never used. SUBJECTS: Children below age 15 in England and Wales, 1966-87. MAIN OUTCOME MEASURE: Registration of any leukaemia or non-Hodgkin's lymphoma. RESULTS: In none of the 25 km circles around the installations was the incidence ratio significantly greater than 1.0. The only significant results for the linear risk score test were for Sellafield (P = 0.00002) and Burghfield (P = 0.031). The circles for Aldermaston and Burghfield overlap; the incidence ratio was 1.10 in each. One of the control sites gave a significant linear risk score test result (P = 0.020). All the tests carried out were one sided with P values estimated by simulation. CONCLUSION: There is no evidence of a general increase of childhood leukaemia or non-Hodgkin's lymphoma around nuclear installations. Apart from Sellafield, the evidence for distance related risk is very weak.

Cancer in Cumbria and in the vicinity of the Sellafield nuclear installation, 1963-90.

OBJECTIVE: To reappraise the epidemiological findings reported by the Black Advisory Group concerning a possible excess of malignant disease, particularly of childhood acute lymphoid leukaemia and non-Hodgkin lymphomas, in the vicinity of the Sellafield nuclear installation, and to determine whether any excess of malignant disease had occurred among people aged 0-24 years in the area in the years after the Black report--that is, from 1984 to 1990. DESIGN: Calculation of incidence of cancer using data from population based cancer registries and special surveys. SETTING: England and Wales; county of Cumbria; county districts Allerdale and Copeland within Cumbria; Seascale ward within Copeland. SUBJECTS: All residents under the age of 75 years in the above areas, but with particular reference to those aged 0-24 years. MAIN OUTCOME MEASURES: Numbers of cases and incidence particularly of lymphoid leukaemia and non-Hodgkin lymphomas in those aged 0-24 years, but including other cancers and age groups. RESULTS: Previous reports of an increased incidence of cancer, especially of leukaemia, among those aged 0-24 years in Seascale during the period up to and including 1983 are confirmed. During 1984-90 there was an excess of total cancer among those aged 0-24 years. This was based on four cases including two cases of non-Hodgkin lymphoma but none of leukaemia. There was an increased, but nonsignificant, incidence of other cancers, based on two cases (one pinealoma and one Hodgkin's disease) occurring among those aged 15-24 years during 1984-90. This was not observed in the younger age group or in previous years. For the immediately surrounding area--that is, the county districts of Allerdale and Copeland excluding Seascale and in the remainder of Cumbria--there was no evidence of an increased incidence of cancer among those aged 0-24 years in either period. CONCLUSIONS: During 1963-83 and 1984-90 the incidence of malignant disease, particularly lymphoid leukaemia and non-Hodgkin lymphomas, in young people aged 0-24 in Seascale was higher than would be expected on the basis of either national rates or those for the surrounding areas. Although this increased risk is unlikely to be due to chance, the reasons for it are still unknown.

Cancer in the offspring of radiation workers: a record linkage study.

OBJECTIVES: To test the "Gardner hypothesis" that childhood leukaemia and non-Hodgkin lymphoma can be caused by fathers' exposure to ionising radiation before the conception of the child, and, more generally, to investigate whether such radiation exposure of either parent is a cause of childhood cancer. DESIGN: Case-control study. SETTING: Great Britain. SUBJECTS: 35,949 children diagnosed as having cancer, together with matched controls. MAIN OUTCOME MEASURES: Parental employment as radiation worker as defined by inclusion in the National Registry for Radiation Workers and being monitored for external radiation before conception of child; cumulative dose of external ionising radiation for various periods of employment before conception; dose during pregnancy. RESULTS: After cases studied by Gardner and colleagues were excluded, fathers of children with leukaemia or non-Hodgkin lymphoma were significantly more likely than fathers of controls to have been radiation workers (relative risk 1.77, 95% confidence interval 1.05 to 3.03) but there was no dose-response relation for any of the exposure periods studied; indeed, the association was greatest for those with doses below the level of detection. No increased risk was found for fathers with a lifetime preconception dose of 100 mSv or more, or with a dose in the 6 months before conception of 10 mSv or more. There was no increased risk for the group of other childhood cancers. Mothers' radiation work was associated with a significant increase of childhood cancer (relative risk 5.00, 1.42 to 26.94; based on 15 cases and 3 controls). Only four of the case mothers and no controls were radiation workers during pregnancy. CONCLUSIONS: These results do not support the hypothesis that paternal preconception irradiation is a cause of childhood leukaemia and non-Hodgkin lymphoma; the observed associations may be chance findings or results from exposure to infective or other agents. If there is any increased risk for the children of fathers who are radiation workers, it is small in absolute terms: in Britain the average risk by age 15 years is 6.5 per 10,000; our best estimate, using all available data, is that the increase is 5.4 per 10,000. For mothers, the numbers are too small for reliable estimates of the risk, if any, to be made.

The analysis of registry data in relation to various different types of hypothesis regarding the geographical distribution of disease.

Disease registries will often contain the addresses of cases included in the registry. If the registry includes information on all cases, or deaths, occurring in a defined geographical area and time period and if there is a postcode/zip code or map reference for each case it is possible to carry out a variety of different types of geographical analysis that may give clues to the aetiology of the disease. For such analyses it will usually also be necessary to have population data for the region covered by the registry and for separate sub-regions within it. In this paper we review types of analysis that may be applied to such data and give references to examples of applications and the statistical methods used. These include, first, methods of presenting incidence rates, and particularly the use of maps; of particular concern is the development of methods for presenting data that take into account the problems of rates calculated for small populations and which may therefore happen to be high or low simply by chance. Secondly, we consider, the analysis of "clustering" and "clusters" of cases of disease. These problems have been the subject of considerable methodological development in recent years. Analyses of clustering address the question of whether there is a general tendency for there to be aggregations of cases or areas of high incidence the analysis of clusters is concerned with problems of detecting specific locations where there are unusual aggregations of cases. The third type of problem considered here is whether there are, within the registry region, aetiological factors that vary geographically with consequent variations in disease incidence in different sub-regions. Where there is geographical variation it may be possible to use regression analysis to relate such variation to factors such as socio-economic status or levels of some environmental hazard. Finally we consider the problem of determining whether disease rates in certain areas may be related to distance from the source of some potential causative agent.