Oncogene (2004) 23, 6429–6444. doi:10.1038/sj.onc.1207717

Epidemiology and genetics of childhood cancer

Charles A Stiller1

1Childhood Cancer Research Group, Department of Paediatrics, University of Oxford, 57 Woodstock Road, Oxford OX2 6HJ, UK

Correspondence: CA Stiller, E-mail:



Childhood cancer is rare everywhere in the world, with age-standardized annual incidence usually between 70 and 160 per million at age 0–14 years. Greater variation is seen between populations for some specific tumour types. Some of the largest variations are geographical and are attributable to environmental factors, whereas variation mainly on ethnic lines seems likely to be a marker of genetic predisposition. A wide range of familial and genetic syndromes is associated with an increased risk of childhood cancer. Virtually all the excess risk of cancer among first-degree relatives of children with cancer can be accounted for by known hereditary syndromes. Studies of weak predisposition and gene–environment interaction have so far shown limited consistency. There are very few established environmental or exogenous risk factors and most of these are infective agents. Many putative risk factors can hardly ever be investigated epidemiologically except by interview or questionnaire studies. Some recent examples illustrate the continuing problems of participation bias and recall bias.


childhood cancer, incidence, familial syndromes, genetic associations, risk factors

Cancer is very rare among children everywhere in the world. In industrialized countries, only about 0.5% of all cancers occur among children aged under 15 years. Whereas most adult cancers are carcinomas, childhood cancers are histologically very diverse. It is therefore more appropriate for childhood tumours to be classified principally according to histology rather than primary site. The standard scheme is the International Classification of Childhood Cancer (Kramárová and Stiller, 1996), in which diagnostic groups are defined according to codes for morphology as well as topography in the second edition of the International Classification of Diseases for Oncology (ICD-O). The 12 major groups are leukaemias, lymphomas, brain and spinal tumours, sympathetic nervous system tumours, retinoblastoma, kidney tumours, liver tumours, bone tumours, soft-tissue sarcomas, gonadal and germ-cell tumours, epithelial tumours, and other and unspecified malignant neoplasms. A new version of the classification based on the third edition of ICD-O is in preparation.

The present review begins with a brief account of the incidence of childhood cancer, including international and ethnic variations that relate to known or postulated genetic or environmental elements in its aetiology. This is followed by a review of the genetics of childhood cancer from an epidemiological standpoint and a brief discussion of selected risk factors that have been the subject of recent reports or illustrate methodological difficulties.



Unless otherwise stated, the data on incidence are taken from the IARC study International Incidence of Childhood Cancer (Parkin et al., 1988, 1998). The total incidence rate is usually in the range 70–160 per million children. There are marked variations between populations in the incidence of particular types of childhood cancer and these can provide valuable pointers to aetiology. Table 1 summarizes international variations in incidence from population-based registries.

Among predominantly white populations of Europe, the Americas and Oceania, and in much of eastern Asia, about a third of all childhood cancers are leukaemias, with an age-standardized rate (ASR) of 35–50 per million. Incidence is somewhat lower, usually below 30 per million, in South Asia and the Middle East and among Black children in the USA and sub-Saharan Africa.

This international variation is largely accounted for by acute lymphoblastic leukaemia (ALL), which comprises around 80% of the total in many populations. High incidence of ALL, up to 40 per million, occurs in the predominantly white populations of North America, Western Europe and Oceania, and also in the Chinese of Hong Kong and Singapore. Somewhat lower rates, around 20–30 per million, have tended to be found in the former socialist countries of Eastern Europe, and in Japan, much of Latin America, and parts of the Middle East. Incidence is still lower, 15–20 per million among US Blacks and in India, and lowest of all in much of sub-Saharan Africa. In general, incidence is correlated with levels of socioeconomic development. The most notable exception to this pattern is that the highest incidence has been recorded among Hispanic populations not only in California (Glazer et al., 1999) and Florida (Wilkinson et al., 2001) but also in Costa Rica. The already high incidence recorded in Costa Rica in the early 1980s was tentatively linked with high levels of pesticide use (Linet and Devesa, 1991). After increasing throughout the 1980s, incidence declined again during 1991–1996, but the reasons for this are unclear (Monge et al., 2002).

The common (early pre-B) immunophenotype accounts for around 70% of childhood ALL in high-incidence populations, and the early peak in total incidence is due to the especially marked peak for this subtype (Greaves et al., 1993). In lower-incidence populations such as India (Rajalekshmy et al., 1994) and Nigeria (Greaves et al., 1993), common ALL accounts for only a third of cases, and T-cell ALL, whose incidence is relatively constant throughout childhood, has a higher relative frequency.

The strong association of higher incidence of common ALL in early childhood with higher levels of socioeconomic development suggests that environmental factors linked to affluence are important in its aetiology. Further evidence is provided by the emergence of the early childhood peak in Britain and among US Whites during the first half of the twentieth century (Court Brown and Doll, 1961). The more moderate peak among US Blacks had appeared by the 1970s. In the Czech Republic, incidence of ALL at age 1–4 years increased with improving socioeconomic conditions during the 1990s, resulting in a more marked early peak similar to that in Western Europe (Hrusák et al., 2002).

Acute myeloid leukaemia (AML) usually has an incidence rate of 4–10 per million. There is little systematic variation between regions of the world or ethnic groups, and a high relative frequency of AML is generally a consequence of lower incidence of ALL rather than a raised risk of AML. High incidence rates, 11–14 per million, have been recorded among two indigenous Pacific island populations, the Maori of New Zealand and Hawaiians of Hawaii. The reasons for this are unknown but the fact that the high incidence in each case occurs in one ethnic group of a multiethnic country suggests that genetic predisposition is important.

The highest incidence of childhood Hodgkin's disease, commonly above 7 per million, is found in Western Asia and North Africa. Moderately high rates have also been found in some parts of India and Latin America. In all those regions, the age-incidence curve rises fairly gently from early childhood, sometimes with a peak in the 5–9 years age group, and substantial numbers of cases are of mixed cellularity subtype. Among predominantly white, Western populations, by contrast, total incidence is usually below 7 per million, there is a steep rise throughout childhood and nodular sclerosis accounts for the majority of cases. These patterns are consistent with the long-standing suggestion of accelerated onset in poor socioeconomic environments. Epstein–Barr virus (EBV) is implicated in more than half of all cases worldwide. EBV positivity in the malignant cells is most frequent in mixed cellularity Hodgkin's disease and in cases from developing countries (Glaser et al., 1997). Around 60% of childhood Hodgkin's disease in industrialized countries and 80% in developing countries may be a rare response to EBV infection, together with an unidentified cofactor related to the socioeconomic level of the population (Parkin et al., 1999). Incidence is generally below 3 per million in populations of East Asian ethnic origin, in both Asia and Hawaii, suggesting that genetic predisposition is also important.

Burkitt's lymphoma (BL) has a very high incidence in parts of tropical Africa and in Papua New Guinea, where it is the commonest childhood cancer. Nearly all cases in this high incidence zone are associated with EBV, with intense malaria infection as the cofactor (Crawford, 2001). Elsewhere in the world, it is difficult to distinguish the patterns of incidence of BL from those for other non-Hodgkin lymphoma (NHL), despite the characteristic histology and cytogenetics of BL. BL everywhere can be identified by breakages and translocations in chromosome 8, but the precise breakpoint on chromosome 8 varies geographically (Shiramizu et al., 1991; Gutiérrez et al., 1992). EBV is associated with 50–70% of BL in North Africa and South America and around 20% in Europe and North America (Crawford, 2001). In the latter regions, at least, the cofactor is very unlikely to be malaria and is as yet unknown. Outside the endemic BL zone, the highest incidence of total NHL in childhood occurs in Mediterranean and Middle Eastern countries and in parts of Latin America. The Israeli-born offspring of migrants to Israel from elsewhere in Asia and North Africa retain their high incidence, indicating a likely role for genetic susceptibility (Iscovich and Parkin, 1997).

Brain and spinal tumours are collectively outnumbered only by leukaemia among children of industrialized countries, with incidence nearly always between 25 and 40 per million. Recorded rates in developing countries are much lower, usually below 15 per million but it is impossible to tell how much of the deficit represents reduced risk rather than underdiagnosis and underascertainment. In the US, Black children have lower incidence than Whites, and in California the incidence was lower in Hispanic than non-Hispanic Whites (Glazer et al., 1999). Incidence is lower among non-Jews than Jews in Israel and lower among Malays than Chinese in Singapore. These differences are far from infallible pointers to differential genetic predisposition, as the lower rates in economically disadvantaged populations may reflect poorer access to medical care. In New Zealand, however, Maori and non-Maori children have similar total incidence, while in Britain low incidence among South Asian compared with White children seems to be due to a marked deficit among those of Indian rather than Pakistani origin, despite the lower socioeconomic status of the latter group (Powell et al., 1995). The most frequent childhood CNS tumours are astrocytomas, followed by primitive neuroectodermal tumours (PNETs). Patterns of incidence for astrocytomas tend to be similar to those for all brain and spinal tumours combined. The highest rates for PNET have been found in New Zealand Maori and the Hawaiian ethnic group in Hawaii, the two ethnic groups that also have the highest incidence of childhood AML. Incidence of childhood CNS tumours has increased at upwards of 1% per year in several countries. It has been suggested that a rather sudden increase in the US in the 1980s, especially for low-grade gliomas, was an artefact of diagnosis following the widespread introduction of MRI scanning (Smith et al., 1998). Steady increases in the incidence of pilocytic astrocytoma and PNET over a 45-year period in North West England are less readily explained by changes in diagnostic or reporting practice and are consistent with increasing exposure to environmental risk factors (McNally et al., 2001).

Nearly all childhood cancers of the sympathetic nervous system are neuroblastomas. Recorded incidence tends to be lowest in developing countries, probably as a result of under diagnosis and underascertainment. Neuroblastoma is the only childhood cancer for which mass screening has been shown to be feasible. Screening has been offered nationally in Japan, and has been the subject of population studies in several other countries. It has invariably led to a substantial increase in incidence in the first year of life but little or no decrease among older children and little effect on mortality (Ajiki et al., 1998; Schilling et al., 2002; Woods et al., 2002). Many neuroblastomas, especially in very young children, regress spontaneously and the increase in infants with screening is clearly due to over diagnosis. Even without mass screening, neuroblastoma can be readily detected incidentally at routine examinations or during investigation of an unrelated disorder (Powell et al., 1998). In a European comparative study, the lower total incidence in the UK than in France and Germany was the result of a substantial deficit of early-stage disease among infants, partially offset by higher incidence of metastatic disease among older children (Powell et al., 1998). It was suggested that in the UK, diagnosis was delayed and some over diagnosis avoided because of less rigorous routine child health checks.

Retinoblastoma essentially occurs in two forms, heritable and nonheritable. The heritable form includes all cases of bilateral disease and a small number of children with unilateral tumours; the incidence of bilateral retinoblastoma is a close minimum estimate for that of the heritable form. Total incidence is generally higher in developing countries, especially in sub-Saharan Africa; Black children have a higher incidence than Whites in the US. Incidence of bilateral retinoblastoma varies little between regions of the world or ethnic groups and the higher total incidence in many developing countries reflects a higher risk of developing unilateral, predominantly nonheritable disease. These variations therefore presumably result from environmental factors operating after conception rather than from parental germ-cell mutation. Retinoblastoma incidence has been correlated with ultraviolet radiation in international data (Hooper, 1999), but the association was not present in the US and became nonsignificant internationally after adjusting for ethnic group and tropical climate (Jemal et al., 2000). The only notable variation in the incidence of bilateral retinoblastoma is in the US, where a higher total incidence among Blacks compared with Whites is due to an excess of bilateral disease. The causes of this are unknown but it seems unlikely that an environmental factor could cause increased risk of new-mutation bilateral disease but not of unilateral disease in one ethnic group.

The highest incidence of Wilms' tumour, the most common childhood renal tumour, occurs in some Black populations in Africa and the USA, with rates commonly in the range 9–12 per million. Among White populations, the incidence is generally somewhat lower, typically 6–10 per million. Consistently lower rates, under 5 per million, are found throughout East Asia, and ethnic East Asian children in California and Hawaii also have lower incidence. The variation in Wilms' tumour largely along ethnic rather than geographical lines suggests a strong element of genetic predisposition in its aetiology. Further evidence for this is provided by the finding of genetic differences between Wilms' tumours in Japanese and White children (Nakadate et al., 2001).

Malignant liver tumours in children are mostly hepatoblastoma or hepatocellular carcinoma (HCC). There is little geographic variation in incidence of hepatoblastoma. HCC is rare in childhood in Europe and North America and most common in those regions where there are also high rates in adults, namely sub-Saharan Africa, East and South-East Asia, and Melanesia. Nearly all childhood cases in these areas occur in chronic carriers of hepatitis B (Wu et al., 1987). Mass immunization against hepatitis B in Taiwan has been followed by a reduction in incidence of HCC in childhood (Chang et al., 1997).

Almost all malignant bone tumours in children are osteosarcoma or Ewing's sarcoma. There is rather little geographic or ethnic variation in the incidence of osteosarcoma. Ewing's sarcoma is especially rare among Black and East Asian populations, suggesting the existence of predisposing or protective genetic factors. Genetic differences have been observed between Ewing's tumours from Japanese and Caucasian patients (Ozaki et al., 2002).

The most frequent soft-tissue sarcoma of childhood in most regions of the world is rhabdomyosarcoma. Incidence rates are around 4–6 per million in Europe, the Americas, and Oceania. In most of Asia, the incidence of rhabdomyosarcoma tends to be somewhat lower. In India it is only 2–3 per million and, while the deficit could be partially accounted for by underascertainment, the lower incidence among ethnic South Asian children in Britain suggests a lower risk, most likely due to lower genetic susceptibility (Powell et al., 1995). In Africa, the incidence of rhabdomyosarcoma is unremarkable, but in many sub-Saharan countries Kaposi's sarcoma has long been the commonest soft-tissue sarcoma among children. Before the AIDS epidemic, it had an incidence in childhood of 2–2.5 per million in Uganda and Zimbabwe, attributable to the endemic form of the disease. Since then, the incidence rose to over 50 per million in Kampala, Uganda (Wabinga et al., 2000) and 10 per million among Africans in Harare, Zimbabwe. These increases are clearly related to the AIDS epidemic, which has been exceptionally severe in East and Central Africa. Much of the increase may be explained by heightened susceptibility to HHV8 among HIV-positive children. In a Ugandan case–control study, HIV infection was associated with an odds ratio of 95 for Kaposi's sarcoma (Newton et al., 2001). In another study in the same country, however, only 78% of children with Kaposi's sarcoma were HIV positive (Ziegler and Katongole-Mbidde, 1996), indicating that a raised prevalence of HHV8 has also contributed to the increase.

Germ-cell tumours generally account for under 4% of all childhood cancers. There is a well-known excess of intracranial germ-cell tumours in Japan but similar incidence is found in Singapore Chinese. The highest incidence of malignant germ-cell tumours of gonadal and other sites has also been recorded in East Asia.

Adrenocortical carcinoma (ACC) almost always has an incidence among children of less than 0.5 per million. It is, however, much more common in southern Brazil, with an incidence of at least 1.5 per million in Sao Paulo and Parana (Sandrini et al., 1997).

The commonest site for carcinomas in children in many world regions is the thyroid, although incidence seldom exceeds 1.5 per million. Differentiated carcinomas, predominantly papillary, are most frequent. An enormous increase in thyroid carcinoma was observed in areas of Belarus, Ukraine, and Russia most heavily contaminated by fallout from Chernobyl. Some of the excess may well have been due to intensified case-finding but the scale of the increase and the unusually aggressive histology of many of the tumours strongly suggest a substantial excess risk attributable to radiation exposure (Moysich et al., 2002). Further support for an association with short-lived radioactive fallout is provided by the fact that incidence returned to pre-Chernobyl levels among children conceived after the explosion of the nuclear reactor (Shibata et al., 2001).

The highest incidence, up to 3 per million, of nasopharyngeal carcinoma in childhood is found in North Africa, a region of intermediate risk for adults. Among Jewish children and adolescents born in Israel, the highest incidence is among those whose parents were born in North Africa (Steinitz et al., 1990). East Asian populations, who have the highest incidence among adults, have moderately high incidence in childhood, up to 1 per million. Predominantly white populations in industrialized countries hardly ever have an incidence above 0.4 per million. In France, the incidence was 0.8 per million but it is not known what proportion of cases occurred in children of North African ethnic origin. In the USA, a similar incidence was found in black children, five times that in whites. Virtually all nasopharyngeal carcinoma in regions of moderate to high incidence is associated with EBV, as is a high proportion of cases in low-incidence regions (Parkin et al., 1999). The cofactors are unknown, but presumably vary in type or prevalence between populations with different incidence rates.

By far, the highest incidence of skin carcinoma in childhood occurs in Tunisia, where both squamous cell and basal cell carcinomas are seen in association with xeroderma pigmentosum.

Malignant melanoma has an incidence of about 1 per million in most childhood populations. The highest rates have been found in Oceania. While some registrations may have been cases that were in fact nonmalignant, the incidence in New Zealand of malignant melanoma confirmed by pathology review was still double that in white populations of the northern hemisphere (Becroft et al., 1999). The relatively high incidence in the region, which also has the highest incidence among adults, is presumably the result of heavy exposure to sunlight in predominantly fair-skinned populations.



It is a truism that cancer is a result of genetic alteration. The age-incidence patterns and cell types of origin for many childhood cancers point to an origin at latest in utero. Chromosome translocations involved in many childhood leukaemias have been shown to originate during fetal haematopoiesis on the basis of studies of monozygotic twins concordant for leukaemia (Greaves et al., 2003a) and by detection in neonatal blood spots (Greaves and Wiemels, 2003b). The genetic mechanisms underlying childhood cancer have been reviewed elsewhere (e.g. Anderson and Pritchard-Jones, 2004). The present review will concentrate on familial aggregations of childhood cancers and associations with specific genetic syndromes, with particular attention, where possible, to risk estimates.

Familial aggregations have long proved a fruitful field for research into genetic aetiology. Pairs of siblings with childhood cancer certainly occur more often than would be expected by chance, and the risk of childhood cancer in a sibling of an affected child without a known family history is approximately doubled (Draper et al., 1977). When cancer does occur in a second sibling, however, this will often be the defining event for the existence of a known syndrome in the family concerned. Population-based studies in the Nordic countries have shown that there is virtually no excess risk of cancer in the siblings, parents, or offspring of children with cancer that cannot be accounted for by known hereditary syndromes (Olsen et al., 1995; Sankila et al., 1998; Winther et al., 2001). The contribution of any as yet undefined familial syndromes is therefore likely to be very small, although this need not make their discovery be of little scientific importance.

Table 2 lists the principal familial neoplastic syndromes that give rise to an increased risk of childhood cancer.

Retinoblastoma is the classic example of a cancer resulting from an inherited genetic abnormality. There are numerous families with more than one affected member, often in more than one generation. In most familial cases, both eyes are affected. Heritable retinoblastoma is usually defined as any case with bilateral tumours or a family history. Retinoblastoma results from two mutations in the RB1 tumour suppressor gene. In heritable cases, the first mutation is prezygotic, which either is inherited or is a rare germ-cell mutation; in nonheritable cases, both mutations are postzygotic. The pattern of inheritance is autosomal dominant, with penetrance about 90% where an affected parent has bilateral tumours and probably somewhat lower in the rare cases of unilateral heritable retinoblastoma. About 2% of sporadic unilateral cases in fact carry the retinoblastoma gene and the probability that a child of a sporadic unilateral case will be affected is about 1% (Draper et al., 1992). When that happens, of course, the family is redefined as having heritable retinoblastoma. The survival rate from retinoblastoma has been very high for many decades, and several families with the disease in multiple generations had already been reported half a century ago. Survivors have a remarkably high risk of developing further primary tumours, many of which are definitely unrelated to treatment received for retinoblastoma. The incidence of osteosarcoma, occurring mainly in the first three decades, is hundreds of times that in the general population (Draper et al., 1986; Wong et al., 1997). The elevated risk of developing subsequent primaries continues in later life, although the specific types of cancer change with advancing age. The excess risk may be especially high for malignant melanoma and for carcinoma of the lung and bladder (Sanders et al., 1989; Kleinerman et al., 2000).

Familial aggregations of the other characteristic embryonal tumours of childhood are comparatively rare. In the US, only 1.5% of 6209 children with Wilms' tumour in the National Wilms' Tumor Study (NWTS) had a family history of the disease (Breslow et al., 1996). Not only does bilateral disease represent a much smaller proportion of total cases in Wilms' tumour than in retinoblastoma, around 7%, family history of the same tumour is also much rarer – 3.3% in the NWTS (Breslow et al., 1996). While the proportion of familial cases is expected to rise following increased survival, it will always be small compared with the proportion of retinoblastoma with family history (Li et al., 1988b). There are at least three familial Wilms' tumour susceptibility genes, and possibly more (Rapley et al., 2000).

Familial aggregations of neuroblastoma are even more infrequent. While several chromosomal regions have been investigated, no candidate region or gene has yet been linked to familial neuroblastoma (Perri et al., 2002).

The Li–Fraumeni cancer family syndrome (LFS) was first identified with children who had rhabdomyosarcoma as probands. The formal definition of LFS is a proband with sarcoma diagnosed before the age of 45 years, a first-degree relative with any cancer also before the age of 45 years, and another first or second-degree relative with sarcoma at any age or any cancer before the age of 45 years (Li et al., 1988a). The cancers that are characteristic of LFS include soft-tissue sarcomas, adrenocortical carcinoma, premenopausal breast carcinoma, central nervous system tumours, and osteosarcoma. The relative risk of childhood cancer is about 20. Subsequently, Li–Fraumeni-like syndrome (LFL) was defined as a proband with any tumour before the age of 45 years, or a sarcoma, brain tumour or adrenocortical tumour before the age of 45 years, together with a first- or second-degree relative with a typical LFS tumour at any age, and another first- or second-degree relative with any cancer before the age of 60 years (Birch et al., 1994). In the largest single study, germline mutations of the TP53 tumour suppressor gene have been detected in 77% of LFS and 40% of LFL families (Varley, 2003). Germline mutations of TP53 are present in over 80% of unselected children with adrenocortical carcinoma (Varley et al., 1999; Ribeiro et al., 2001) and were detected in 10% of a series of children with rhabdomyosarcoma and no family history (Diller et al., 1995). An inherited mutation of TP53 that was found in 35 of 36 children with ACC from the high-incidence area of southern Brazil appears to be specific to the geographical area (Ribeiro et al., 2001). The causes are unknown but seem likely to be environmental. Among family members of a large series of children with soft-tissue sarcoma, the risk of lung cancer in carriers of a TP53 mutation was 38.5 times that in noncarriers (Hwang et al., 2003b).

In LFS families where no TP53 mutation has been found, this could of course be due to chance or failure of detection. Some LFS or LFL families have been found to have germline mutations of other genes, however, including CHK2 (Bell et al., 1999) and SNF5 (Sévenet et al., 1999). Mutations of SNF5 have been particularly associated with choroid plexus tumours and atypical teratoid rhabdoid tumours of the brain (Biegel et al., 1999; Taylor et al., 2000).

Turcot's syndrome encompasses two distinct groups of patients with brain tumours and colorectal polyposis or cancer (Paraf et al., 1997). The first consists of children and adolescents with gliomas, usually high-grade astrocytomas, and colorectal adenomas without polyposis, and their siblings with glioma and/or colorectal adenoma. The risk of a brain tumour in members of families with hereditary nonpolyposis colorectal cancer is about five times that in the general population (Vasen et al., 1996). The risk of brain tumours is higher in MSH2 mutation carriers than in MLH1 mutation carriers (Vasen et al., 2001). In the second group, the brain tumours are usually medulloblastomas and occur in members of familial adenomatous polyposis (FAP) families.

Hepatoblastoma is about 100 times commoner in FAP families than in the general population (Garber et al., 1988).

The proportion of childhood medulloblastoma due to Gorlin syndrome is 2–3% (Cowan et al., 1997). The tumours are always of the desmoplastic subtype (Cowan et al., 1997; Amlashi et al., 2003).

Neurofibromatosis type 1 (NF1) carries an increased risk of several types of childhood cancer, of which CNS tumours, especially astrocytoma, are the most frequent. The risk of optic nerve glioma (astrocytoma) by age 15 years among children with NF1 is 4–5% (McGaughran et al., 1999). The relative risk of death from malignant brain tumours among US children and adolescents with NF1 was 4 at age 0–9 years and 11 at age 10–19 (Rasmussen et al., 2001). The relative risk of childhood CNS tumours in NF1 is over 40 (Narod et al., 1991). There is a similar relative risk for soft-tissue sarcomas, with the excess accounted for by malignant peripheral nerve sheath tumours and rhabdomyosarcoma (Narod et al., 1991). The relative risk is highest, about 200, for juvenile myelomonocytic leukaemia (Stiller et al., 1994). Relative risks of around 5 were found for ALL and NHL in children with NF1 (Stiller et al., 1994). It now seems likely that some of these children in fact had the NF1-like syndrome of multiple cafe au lait spots that has been reported in several families with germline mutations of the DNA mismatch repair genes MLH1 (Ricciardone et al., 1999; Wang et al., 1999) or MSH2 (Whiteside et al., 2002; Bougeard et al., 2003).

Neurofibromatosis type 2 (NF2) has only a tenth of the frequency of NF1 and the most commonly associated CNS tumours are meningiomas; it must therefore account for a much lower proportion of childhood cancer than NF1. In many instances where childhood meningioma is the first symptom, NF2 is only diagnosed in adulthood. Among 22 cases of meningioma in the population-based Manchester Children's Tumour Registry, at least three went on to develop classic features of NF2 (Evans et al., 1999).

Half of the children in Britain with medullary thyroid carcinoma diagnosed during 1971–1983 were from families with multiple endocrine neoplasia type 2 (MEN2) (Narod et al., 1991), but the proportion will probably have decreased since then with widespread prophylactic thyroidectomy in MEN2 families.

Several inherited immune deficiency syndromes carry an increased risk of childhood cancer, mainly lymphomas and leukaemias (Table 3). Owing to their rarity, they account together for less than 0.1% of all childhood cancers. Numerically, the most important is ataxia telangiectasia; more than 10% of affected children develop lymphoma or leukaemia before 15 years (Morrell et al., 1986). There is a similar incidence of NHL among children with Wiskott–Aldrich syndrome (Sullivan et al., 1994). Bloom syndrome, common variable immunodeficiency syndrome, X-linked agammaglobulinaemia, IgA deficiency, severe combined immunodeficiency (SCID), Duncan's disease, and Nijmegen breakage syndrome are all associated with a high risk of NHL and leukaemia but these extremely rare syndromes between them only account for a handful of new childhood cancers worldwide (Mueller and Pizzo, 1995; German, 1997; The International Nijmegen Breakage Syndrome Study Group, 2000). Other childhood cancers reported more than once in Bloom syndrome are Wilms' tumour and osteosarcoma (German, 1997). In a trial of gene therapy for SCID, two of 10 children developed T-cell ALL, thought to be the result of insertional oncogenesis (Kohn et al., 2003).

The most common inherited bone marrow failure syndromes associated with childhood cancer (Table 3) are Fanconi anaemia (FA), Diamond–Blackfan anaemia (DBA) and Shwachman–Diamond syndrome (SDS). The risk of AML or myelodysplasia by age 15 years in FA is in the range 5–15% (Kutler et al., 2003; Rosenberg et al., 2003). HCC is also seen, almost always following androgen therapy for aplastic anaemia (Alter, 2003). No risk estimates appear to have been published, but children with DBA probably have an excess risk of AML and osteosarcoma (Lipton et al., 2001). The risk of AML in SDS varies widely between series but seems likely to be around 10% in the first 15 years of life (Dror and Freedman, 2002).

FA is one of a group of autosomal recessive DNA repair disorders associated with childhood cancer, which also includes ataxia telangiectasia, Bloom syndrome, and Nijmegen breakage syndrome, all mentioned above, together with xeroderma pigmentosum (XP) and Rothmund–Thomson syndrome (RTS) (Table 4). In the US XP Registry, carcinomas and melanomas of the skin and cancers of the anterior eye and of the tongue all occurred at age 0–19 years at least 1000 times as frequently as in the general population (Kraemer et al., 1994); basal cell carcinoma was roughly twice as common as squamous cell. In the largest reported series of RTS patients, 32% developed osteosarcoma at a median age of 11.5 years (Wang et al., 2001).

Miscellaneous non-neoplastic genetic syndromes that are associated with an increased risk of childhood cancer are listed in Table 4.

Two syndromes involve the Wilms' tumour suppressor gene WT1. Over 30% of children with WAGR (Wilms' tumour, aniridia, genitourinary abnormalities, mental retardation) syndrome develop Wilms' tumour (Coppes et al., 1994). A Danish record linkage study gave a relative risk of 67 for Wilms' tumour in children with sporadic aniridia (Grønskov et al., 2001). Denys–Drash syndrome carries a much higher risk of Wilms' tumour, around 90% (Coppes et al., 1994).

The most common overgrowth syndrome associated with childhood cancer is Beckwith–Wiedemann syndrome (BWS), with a cumulative cancer risk of 10% by age 4 years (DeBaun and Tucker, 1998). Wilms' tumour is the most frequent cancer in BWS, and there is also a raised risk of hepatoblastoma, neuroblastoma, and pancreatoblastoma (Drut and Jones, 1988; DeBaun and Tucker, 1998). Hemihypertrophy, alone or as part of BWS, is associated with Wilms' tumour, hepatoblastoma, and adrenocortical carcinoma (Sotelo-Avila et al., 1980). The risk of childhood cancer in isolated hemihypertrophy has been estimated as 6% (Hoyme et al., 1998). In Costello syndrome, which can resemble BWS in neonates, the risk of rhabdomyosarcoma may be as high as 10%, and there also appears to be an unusually high risk of bladder carcinoma at very young ages (Gripp et al., 2002). Simpson–Golabi–Behmel syndrome is phenotypically similar to BWS; among 18 affected boys, two had Wilms' tumour, including one who had originally been given a diagnosis of BWS (Lindsay et al., 1997). The even rarer Perlman syndrome, which also overlaps phenotypically with BWS, again has a raised risk of Wilms' tumour (Henneveld et al., 1999). A wide variety of childhood cancers has been associated with Sotos syndrome, with the risk estimated as one in 41 (Hersh et al., 1992), but this should be treated with caution because of past difficulty in distinguishing from other disorders such as Weaver syndrome (Opitz et al., 1998).

Among patients with either form of tuberous sclerosis, 5–14% develop brain tumours in childhood (Webb et al., 1996; Lindor and Greene, 1998). About 0.5% of all childhood CNS tumours are associated with tuberous sclerosis (Narod et al., 1991).

In a multicentre review of children with tyrosinaemia who survived to at least 2 years, a third developed HCC, with the peak age at diagnosis lying between 4 and 5 years of age (Weinberg et al., 1976). The incidence of HCC in tyrosinaemia should have decreased in recent years with the adoption of prophylactic liver transplantation.

Table 5 shows numerical chromosome abnormalities associated with childhood cancer. Of these, Down syndrome accounts for by far the largest number of cases. Very high relative risks of leukaemia have been repeatedly found (Hasle et al., 2000; Hermon et al., 2001), about 50-fold in the first 5 years of life and 10-fold in the next 10 years. About 60% of the leukaemias are ALL and 40% are AML; 2–3% of children with leukaemia have Down syndrome. There is strong evidence of an excess of germ-cell tumours (Hasle et al., 2000; Hasle, 2001; Satgé et al., 2003), although their much smaller numbers make it hard to estimate the risk. With the possible exception of lymphoma and retinoblastoma (Satgé et al., 2003), however, other solid tumours occur less frequently than expected in children with Down syndrome. In particular there is a complete absence of neuroblastoma (Satgé et al., 1998; Satgé et al., 2003) and Wilms' tumour (Olson et al., 1995; Satgé et al., 2003), raising the possibility that there are genes on chromosome 21 that protect against a wide range of cancers.

Wilms' tumour is associated more frequently than expected with trisomy 18 and the tumours are diagnosed at unusually high ages (Olson et al., 1995). In a cohort study of Turner syndrome patients, there was no excess of any type of cancer (Swerdlow et al., 2001), although three girls in a series of 394 with Turner syndrome developed neuroblastoma (Blatt et al., 1997) and an excess of Turner syndrome has been found in Wilms' tumour (Olson et al., 1995). Adolescents and young men with Klinefelter syndrome in a record-based cohort study had a relative risk of 67 for mediastinal germ-cell tumours (Hasle et al., 1995).

In addition to their contribution to the understanding of carcinogenesis, studies of genetic syndromes associated with childhood cancer can also be important for the care of long-term survivors. The increased risk of lung cancer in survivors of heritable retinoblastoma indicates that it is unusually important for them to avoid smoking. While the relative risk of lung cancer associated with cigarette smoking may be lower in germline TP53 mutation carriers than in the general population (Hwang et al., 2003a), the absolute risk is still much higher, implying that it is especially important for this group also not to smoke.


Weak predisposition and gene–environment interaction

Possible associations of childhood ALL with HLA haplotypes have been studied for more than 30 years. Recent studies have investigated associations with HLA-DP, -DQ, and -DR. A preliminary study in north west England found a positive association between common ALL and the HLA-DPB1 locus allele *0201 (Taylor et al., 1995). A subsequent, much larger UK national study confirmed this finding, and additionally found a positive association between T-cell ALL and *0201 and a decreased risk of common ALL associated with DPB1*0101 (Taylor et al., 2002). The earlier study also found positive associations with DQB1*05 (Dearden et al., 1996) and, for boys only, with DQA1*0101/*0104 and DQB1*0501 (Taylor et al., 1998); these results have yet to be confirmed from the national study. Two studies, in Wales (Dorak et al., 1999) and Turkey (Dorak et al., 2002), found significant associations of ALL with HLA-DRB4 in boys only. All these results support an infective aetiology for childhood ALL, although it has also been pointed out that increased parental HLA-DR compatibility is associated with a history of spontaneous abortion, which has emerged as a risk factor in several case–control studies of ALL (Dorak et al., 1999).

The most comprehensive investigation so far of genetic variation in metabolic pathways in relation to a childhood cancer is a study of ALL among the white French-Canadian population of Quebec (Krajinovic et al., 2001). The CYP1A1*2A and GSTM1 null genotypes were both associated with an increased risk of ALL (Krajinovic et al., 1999). CYP1A1 is important in the activation of polycyclic aromatic hydrocarbons, which are present in cigarette smoke. While there was no association overall of ALL with reported parental smoking, there was some evidence for an increased risk with maternal smoking in pregnancy and paternal smoking postnatally among children with the CYP1A1*2A allele (Infante-Rivard et al., 2000). Among children with the CYP1A1m1 or CYP1A1m2 mutation, there was some evidence for an increased risk with maternal use of pesticides during pregnancy or the child's exposure to them postnatally, although the numbers were small and there was only limited consistency in the results (Infante-Rivard et al., 1999).

Carriers of CYP2E1*5 had a raised risk of ALL, as did carriers of NQO1*2 and *3; a gene–gene interaction between the GSTM1-null genotype and NQO1 mutant alleles was also found (Krajinovic et al., 2002). Overall, there was a protective effect of maternal alcohol consumption during pregnancy, tentatively attributed to the effects of flavonoids, but there were significantly raised interaction odds ratios for the GSTM1-null genotype and third-trimester consumption and for CYP2E1*5 and consumption during the nursing period (Infante-Rivard et al., 2002b). Exposure to trihalomethanes, byproducts of disinfection, in drinking water was a risk factor during pregnancy for CYP2E1*5 and postnatally for the GSTT1-null genotype (Infante-Rivard et al., 2002a).

Slow NAT2 acetylation genotype had a raised odds ratio for ALL, and the risk was further increased for NAT2 slow acetlylators who also had the GSTM1-null or CYP1A1*2A genotypes (Krajinovic et al., 2000).

Finally, while there was little evidence that variants in the mismatch repair genes MLH1 and MSH3 were independent risk factors for ALL, MLH1 variant Val-219 increased the risk of ALL associated with GSTM1, CYP1A1*2A, and CYP2E1*5 (Mathonnet et al., 2003), and MLH1 (ex 8) was associated with a reduction of risk from postnatal X-ray exposure among girls (Infante-Rivard, 2003).

Some of these polymorphisms have been investigated in other studies of childhood leukaemia. In two US studies, one at St Jude Children's Research Hospital (Chen et al., 1997) and the other in the multi-institution Children's Cancer Group (Davies et al., 2002), frequencies of GSTM1 null and GSTT1 null were similar for cases and controls among white children. There were nonsignificant excesses of GSTM1 null among black cases in both studies and no consistent effect for GSTT1 null among blacks. A small Portuguese study found a raised risk for GSTM1 null but not for GSTT1 null (Alves et al., 2002). In Turkey, no significant association was found between ALL and the GSTM1-null, GSTT1-null, or CYP1A1*2A genotypes (Balta et al., 2003). Two studies, in the UK and US, have found an association of low or null function NQO1 genotypes and leukaemia with MLL rearrangements (Wiemels et al., 1999; Smith et al., 2002).

Results have not been published from the Quebec study on polymorphisms of methylenetetrahydrofolate reductase (MTHFR), which is involved in folate metabolism, but these have been investigated in two other studies of childhood leukaemia. In the larger study, the 677CT and TT genotypes were associated with reduced risk of leukaemia with MLL rearrangements and the 1298CC genotype with a reduced risk of hyperdiploid ALL (Wiemels et al., 2001). In the other study, presence of the 677T allele was associated with a reduced risk of ALL (Franco et al., 2001). An inverse association between maternal consumption of folate supplements in pregnancy and childhood ALL has been found in one study (Thompson et al., 2001), but this was based on fewer than 100 cases, and MTHFR polymorphisms and ALL cytogenetics were not evaluated.

Weak predisposition to childhood AML has been less extensively investigated. In the largest study, in the US, the GSTM1-null genotype had a significantly raised odds ratio of 2.0, while the odds ratio of 1.6 for GSTT1 null was not statistically significant (Davies et al., 2000). A much smaller Turkish study, by contrast, found a significant protective effect for GSTT1 null and no association with GSTM1 (Balta et al., 2003).

There appears so far to be only one published study of weak predisposition and any other childhood cancer. This was an investigation of myeloperoxidase (MPO) polymorphism in relation to hepatoblastoma (Pakakasama et al., 2003), in which the A allele was found to have a significant protective effect.

Hitherto, there has been only limited consistency between studies of disease susceptibility in relation to genetic variants (Ioannidis et al., 2001; Hirschhorn et al., 2002). Childhood leukaemia is no exception to this pattern and, furthermore, studies of genetic polymorphisms as determinants of susceptibility have tended to be based on small numbers of cases. Nevertheless, the results of some studies could help to account for previous findings of associations with various exogenous factors.

Overall, human genetic diversity is overwhelmingly accounted for by differences between individuals within a population rather than by differences between ethnic groups (Barbujani et al., 1997; Rosenberg et al., 2002). Frequencies of many polymorphisms do show substantial variation between populations (Burchard et al., 2003), however, and these include some that have been linked to childhood leukaemia. In particular, GSTT1 null, CYP1A1*2A and *2B, and CYP2E1*5 are all more frequent in East Asians than Whites (Garte et al., 2001), and NAT2 slow acetylator mutations are also more frequent in East Asians (Lin et al., 1994). These patterns do not appear to produce corresponding variations in the risk of ALL between the ethnic groups as incidence is similar among Asian and White populations at similar levels of socioeconomic development. The MTHFR 677T allele, which has been suggested as protective against ALL, appears to have a particularly high frequency in Hispanic populations of the US (Wilcken et al., 2003), who also have some of the highest reported incidence rates of ALL.

If further studies confirm the association of specific polymorphisms with childhood cancer, this could have important consequences for follow-up of survivors. For example, some CYP variants that have been associated with ALL may also, alone or in combination with GSTM1 deficiency, increase susceptibility to tobacco-related cancers in adulthood (Bartsch et al., 2000), providing an additional reason to discourage smoking among childhood cancer survivors.


Risk factors that need not to be exclusively genetic or environmental in origin

A wide range of congenital anomalies has been reported among children with cancer in the apparent absence of any known genetic syndrome. In the largest population-based study, 4.4% of children with solid tumours had a congenital anomaly, as against 2.6% of those with leukaemia or lymphoma (Narod et al., 1997). The highest rates, above 6%, were in Wilms' tumour, hepatoblastoma, Ewing's sarcoma, and germ-cell tumours. Some of these associations may turn out to have a genetic aetiology, perhaps especially where multiple anomalies occur in the same child. Others, such as an excess of neural tube defects in children with CNS tumours, may more plausibly have a shared environmental cause. All four studies that have looked at inguinal hernia in relation to Ewing's sarcoma have found an excess in cases compared with controls (Holly et al., 1992; Winn et al., 1992; Cope et al., 2000; Valery et al., 2003). As hereditary, congenital and environmental factors all influence the risk of hernia, a common cause for hernia and Ewing's sarcoma could equally be an in utero environmental exposure or a genetic factor (Valery et al., 2003).

A series of case reports of cancer in children conceived with assisted reproduction technology (ART) gave rise to anxiety over the possible neoplastic risk of ART. Table 6 shows the observed and expected numbers of cases in the four large cohort studies that included all forms of childhood cancer. Details of study design and the range of ART techniques varied between the studies but there is little sign of a marked increase in risk of all childhood cancers combined. One further study, not included in Table 6, reported a cohort of only 332 births in Israel (Lerner-Geva et al., 2000). Unsurprisingly, no cancers were found, but this result is impossible to evaluate as the expected number of cancers was given as 1.7, implying an implausibly high cumulative incidence of 1/195 in the child population. Most recently, it was reported that five children diagnosed with retinoblastoma in the Netherlands and born within a 7-year period were conceived by in vitro fertilization (IVF), compared with an estimated 0.69 expected (Moll et al., 2003). The expected number of cases was based on a lower rate of IVF births than in the Dutch cohort study (Klip et al., 2001) and so the sevenfold relative risk may be an overestimate. Also, the 17 cancers in the Swedish, Australian, and Dutch cohort studies included no cases of retinoblastoma (cancer types were not reported for the two cases in the UK study). Nevertheless, the increasing number of children born by ART and the low statistical power of studies to date weigh strongly in favour of the study of cancer incidence among further, larger cohorts born following ART, both to improve the precision of risk estimates for all cancers combined and to assess the risk of specific types of childhood cancer.

Parental age has been investigated in numerous studies of childhood cancer. There is a well established increasing trend in the risk of ALL with increasing maternal and paternal age despite the tendency for risk to be lower with higher birth order (Dockerty et al., 2001). The status of very young maternal age as a risk factor in childhood leukaemia is more controversial. In a German case–control study, there was a statistically significant odds ratio of 1.9 for acute leukaemia with maternal age at delivery of less than 20 years compared with age 20–34 years (Schüz et al., 1999). The 1.4% prevalence of mothers aged under 20 years in the control group born during 1986–1994 was low in comparison not only with the 3.3% in the leukaemia cases but also with the 2.9% in the general population of Germany for the same birth years (Schüz, 2003) and it was concluded that the high leukaemia risk was likely to be an artefact of nonparticipation bias. Support was given by a recent study from California which was record based and in which the issue of nonparticipation therefore did not arise (Reynolds et al., 2002); the nonsignificant odds ratios for maternal age under 20 years compared with age 20–34 years were 0.84 for ALL and 0.56 for AML. In most case–control studies, participation basis has not been investigated, indeed with some commonly used methods of control selection, such as random digit dialling, it would be impossible to assess. In the UK Childhood Cancer Study (UKCCS), however, where controls were selected from registration lists of geographically defined health authorities, control parents who agreed to participate tended to live in more affluent areas than initially selected controls (Law et al., 2002). While participation bias appears to account for the high odds ratio for young maternal age in the German study, however, there is no universal agreement in other record-based studies, shown in Table 7, that could not suffer from this bias. In case–control analyses of both lymphoid and myeloid leukaemia in Sweden and in a cohort analysis of myeloid leukaemia in Denmark, the youngest maternal age group was associated with the highest risk.


Environmental and exogenous risk factors

A huge range of environmental or exogenous exposures have been suggested as risk factors for childhood cancer. The IARC monograph Epidemiology of Childhood Cancer (Little, 1999) includes a comprehensive review covering most relevant publications up to 1997. More recent reviews have covered the literature since then relating to all childhood cancers (e.g. Stiller, 2004) or to particular diagnostic groups or specific putative risk factors.

Very few exposures are firmly established as risk factors. The relation between antenatal obstetric irradiation and subsequent cancer in the child was discovered 45 years ago (Stewart et al., 1958). At that time, diagnostic X-rays in pregnancy may have caused 5% of all childhood cancers but reductions in the use of X-rays since then will have reduced the attributable fraction very substantially.

The only established transplacental carcinogen is the hormone diethylstilboestrol (DES), formerly given to pregnant women with threatened abortion. Exposure in utero caused clear cell adenocarcinoma of the vagina or cervix mainly in young women, although occasionally in girls under 15 (Giusti et al., 1995). As the use of DES ceased about 30 years ago, no further DES-related childhood cancers should occur.

Of specific infections, EBV, hepatitis B, HIV, and HHV8 are responsible for some of the international variations in incidence of childhood lymphomas, nasopharyngeal carcinoma, hepatic carcinoma, and Kaposi's sarcoma.

Simian virus 40 (SV40) contaminated large quantities of polio vaccine between 1955 and 1963 and some, although not all, laboratory studies have indicated that SV40 is associated with some cancers that occur in children, including brain tumours and osteosarcoma (Vilchez et al., 2003). Cohort studies of exposed populations, however, have failed to find evidence of increased risk. Some of these studies have been criticized on the grounds that not all vaccine was contaminated and that the extent of use was poorly documented; however, a large national cohort study in Denmark was not subject to these limitations and it, also, found no evidence of increased cancer incidence with SV40 exposure (Engels et al., 2003). A case–control study of medulloblastoma in Connecticut found that mothers of cases had a significantly higher rate of polio vaccination in pregnancy during the period of SV40 contamination than mothers of controls (Farwell et al., 1984). When this result is cited in support of an association, however, it is sometimes not mentioned that only 62 cases were born during the relevant period and vaccination status was unknown for 40% of case and control mothers.

A putative risk factor for childhood cancer that has been the subject of greater public concern over the past decade, but for which epidemiological study has been hampered by poor quality data, is intramuscular vitamin K given neonatally to prevent haemorrhagic disease of the newborn. In a pooled analysis of data from six record-based case control studies, including the one that originally gave rise to the concern (Golding et al., 1992), there was little evidence that intramuscular vitamin K is associated with childhood leukaemia or other childhood cancers (Roman et al., 2002). Overall, there was no record of whether or not vitamin K was given for 40% of cases and 35% of controls. In four of the six individual studies, including the original one, the odds ratios for no record of whether vitamin K was given compared with vitamin K definitely not given was higher than that for intramuscular vitamin K. Similar results have since been found in the current UK national case–control study of childhood cancer (Fear et al., 2003), with over 20% of both case and control children lacking a record of whether vitamin K was given.

Especially if record-based case–control studies are carried out by computerized record linkage, or if abstraction from records is carried our without knowing whether each subject is a case or control, they should be free of bias in the proportion of subjects with incorrect or missing information. Studies based on interviews or postal questionnaires are subject to recall bias and this is well illustrated by a recent report from a German national study of parental occupational exposures (Schüz et al., 2003). The odds ratios for leukaemia were above unity for 77% of self-reported exposures analysed. Among fathers of cases, but not of controls, there was a consistent decline in the proportion reporting exposures before the index child's birth with increasing time since that birth at interview. It seemed likely that this reflected a tendency for case fathers to recall insignificant or infrequent exposures if they had occurred relatively recently. The use of job titles as proxy for exposures was felt to be unsatisfactory, however, as generally low proportions of respondents reported exposures that would be expected on the basis of those job titles. With the exception of a very few exposures, notably ionizing radiation, for which independent records exist, studies of childhood cancer in relation to parental occupational exposures will inevitably rely on data derived from questionnaires. Therefore, improvements are needed in future studies including probing questions and better interview techniques (Schüz et al., 2003). These considerations of course also apply to many other putative risk factors for childhood cancer that cannot be retrospectively verified, such as the diet of mother and child.



  1. Ajiki W, Tsukuma H, Oshima A and Kawa K. (1998). Cancer Causes Control, 9, 631–636.
  2. Alter BP. (2003). Cancer, 97, 425–440. | Article | PubMed |
  3. Alves S, Amorim A, Ferreira F, Norton L and Prata MJ. (2002). Leukemia, 16, 1565–1567. | Article |
  4. Amlashi SFA, Riffaud L, Brassier G and Morandi X. (2003). Cancer, 98, 618–624.
  5. Anderson J and Pritchard-Jones K. (2004). Pinkerton CR, Plowman PN and Pieters R (eds) Paediatric Oncology, 3rd edn. Arnold: London, pp. 25–51.
  6. Balta G, Yuksek N, Ozyurek E, Ertem U, Hicsonmez G, Altay C and Gurgey A. (2003). Am. J. Hematol., 73, 154–160.
  7. Barbujani G, Magagni A, Minch E and Cavalli-Sforza LL. (1997). Proc. Natl. Acad. Sci. USA, 94, 4516–4519. | Article | PubMed | ChemPort |
  8. Bartsch H, Nair U, Risch A, Rojas M, Wikman H and Alexandrov K. (2000). Cancer Epidem. Biomark. Prev., 9, 3–28.
  9. Becroft DMO, Dockerty JD, Berkeley BB, Chan Y-F, Lewis ME, Skeen JE, Synek BJL and Teague LR. (1999). Pathology, 31, 83–89.
  10. Bell DW, Varley JM, Szydlo TE, Kang DH, Wahrer DCR, Shannon KE, Lubratovich M, Verselis SJ, Isselbacher KJ, Fraumeni JF, Birch JM, Li FP, Garber JE and Haber DA. (1999). Science, 286, 2528–2531. | Article | PubMed | ISI | ChemPort |
  11. Bergh T, Ericson A, Hillensjö T, Nygren K-G and Wennerholm U-B. (1999). Lancet, 354, 1579–1585. | Article | PubMed | ISI | ChemPort |
  12. Biegel JA, Zhou J-Y, Rorke LB, Stenstrom C, Wainwright LM and Fogelgren B. (1999). Cancer Res., 59, 74–79. | PubMed | ISI | ChemPort |
  13. Birch JM, Hartley AL, Tricker KJ, Prosser J, Condie A, Kelsey AM, Harris M, Morris Jones PH, Binchy A, Crowther D, Craft AW, Eden OB, Evans DGR, Thompson E, Mann JR, Martin J, Mitchell ELD and Santibáñez-Koref MF. (1994). Cancer Res., 54, 1298–1304. | PubMed | ISI | ChemPort |
  14. Blatt J, Olshan AF, Lee PA and Ross JL. (1997). J. Pediatr., 131, 666–670.
  15. Bougeard G, Charbonnier F, Moerman A, Martin C, Ruchoux MM, Drouot N and Frébourg T. (2003). Am. J. Hum. Genet., 72, 213–216. | Article | PubMed | ISI | ChemPort |
  16. Breslow NE, Olson J, Moksness J, Beckwith JB and Grundy P. (1996). Med. Pediatr. Oncol., 27, 398–403.
  17. Bruinsma F, Venn A, Lancaster P, Speirs A and Healy D. (2000). Hum. Reprod., 15, 604–607. | Article | PubMed | ISI | ChemPort |
  18. Burchard EG, Ziv E, Coyle N, Gomez SL, Tang H, Karter AJ, Mountain JL, Pérez-Stable EJ, Sheppard D and Risch N. (2003). N. Eng. J. Med., 348, 1170–1175. | Article |
  19. Chang M-H, Chen C-J, Lai M-S, Hsu H-M, Wu T-C, Kong M-S, Liang D-C, Shau W-Y, Chen D-S and for the Taiwan Childhood Hepatoma Study Group. (1997). N. Eng. J. Med., 336, 1855–1859.
  20. Chen C-L, Liu Q, Pui C-H, Rivera GK, Sandlund JT, Ribeiro R, Evans WE and Relling MV. (1997). Blood, 89, 1701–1707. | PubMed | ISI | ChemPort |
  21. Cnattingius S, Zack M, Ekbom A, Gunnarskog J, Linet M and Adami HO. (1995a). Cancer Epidem. Biomark. Prev., 4, 441–445.
  22. Cnattingius S, Zack MM, Ekbom A, Gunnarskog J, Kreuger A, Linet M and Adami H-O. (1995b). J. Natl. Cancer Inst., 87, 908–914. | Article | PubMed | ChemPort |
  23. Cope JU, Tsokos M, Helman LJ, Gridley G and Tucker MA. (2000). Med. Pediatr. Oncol., 34, 195–199.
  24. Coppes MJ, Haber DA and Grundy PE. (1994). N. Eng. J. Med., 331, 586–590.
  25. Court Brown WM and Doll R. (1961). Br. Med. J., I, 981–988.
  26. Cowan R, Hoban P, Kelsey A, Birch JM, Gattamaneni R and Evans DGR. (1997). Br. J. Cancer, 76, 141–145. | PubMed | ISI | ChemPort |
  27. Crawford DH. (2001). Philos. Trans. R. Soc. Lond. B. Biol. Sci., 356, 461–473. | PubMed | ISI | ChemPort |
  28. Davies SM, Bhatia S, Ross JA, Kiffmeyer WR, Gaynon PS, Radloff GA, Robison LL and Perentesis JP. (2002). Blood, 100, 67–71. | Article | PubMed | ISI | ChemPort |
  29. Davies SM, Robison LL, Buckley JD, Radloff GA, Ross JA and Perentesis JP. (2000). Cancer Epidem. Biomark. Prev., 9, 563–566.
  30. Dearden SP, Taylor GM, Gokhale DA, Robinson MD, Thompson W, Ollier W, Binchy A, Birch JM, Stevens RF, Carr T and Bardsley WG. (1996). Br. J. Cancer, 73, 603–609.
  31. DeBaun MR and Tucker MA. (1998). J. Pediatr., 132, 398–400. | Article | PubMed | ISI | ChemPort |
  32. Diller L, Sexsmith E, Gottlieb A, Li FP and Malkin D. (1995). J. Clin. Invest., 95, 1606–1611. | PubMed | ISI | ChemPort |
  33. Dockerty JD, Draper GJ, Vincent TJ, Rowan SD and Bunch KJ. (2001). Int. J. Epidemiol., 30, 1428–1437. | Article | PubMed | ISI | ChemPort |
  34. Dorak MT, Lawson T, Machulla HKG, Darke C, Mills KI and Burnett AK. (1999). Blood, 94, 694–700. | PubMed | ISI | ChemPort |
  35. Dorak MT, Oguz FS, Yalman N, Diler AS, Kalayoglu S, Anak S, Sargin D and Carin M. (2002). Leuk. Res., 26, 651–656.
  36. Doyle P, Bunch KJ, Beral V and Draper GJ. (1998). Lancet, 352, 452–453. | Article | PubMed | ISI | ChemPort |
  37. Draper GJ, Heaf MM and Kinnier Wilson LM. (1977). J. Med. Genet., 14, 81–90. | PubMed | ChemPort |
  38. Draper GJ, Sanders BM, Brownbill PA and Hawkins MM. (1992). Br. J. Cancer, 66, 211–219.
  39. Draper GJ, Sanders BM and Kingston JE. (1986). Br. J. Cancer, 53, 661–671. | PubMed | ISI | ChemPort |
  40. Dror Y and Freedman MH. (2002). Br. J. Haematol., 118, 701–713. | Article | PubMed | ISI |
  41. Drut R and Jones MC. (1988). Pediatr. Pathol., 8, 331–339. | PubMed | ChemPort |
  42. Engels EA, Katki HA, Nielsen NM, Winther JF, Hjalgrim H, Gjerris F, Rosenberg PS and Frisch M. (2003). J. Natl. Cancer Inst., 95, 532–539.
  43. Evans DRG, Birch JM and Ramsden RT. (1999). Arch. Dis. Child, 81, 496–499. | PubMed | ISI | ChemPort |
  44. Farwell JR, Dohrmann GJ and Flannery JT. (1984). J. Neurosurg., 61, 657–664. | PubMed | ISI | ChemPort |
  45. Fear NT, Roman E, Ansell P, Simpson J, Day N, Eden OB and on behalf of the United Kingdom Childhood Cancer Study Investigators. (2003). Br. J. Cancer, 89, 1228–1231. | Article |
  46. Franco RF, Simões BP, Tone LG, Gabellini SM, Zago MA and Falcão RP. (2001). Br. J. Haematol., 115, 616–618. | Article | PubMed | ChemPort |
  47. Garber JE, Li FP, Kingston JE, Krush AJ, Strong LC, Finegold MJ, Bertario L, Bülow S, Filippone A, Gedde-Dahl T and Järvinen HJ. (1988). J. Natl. Cancer Inst., 80, 1626–1628.
  48. Garte S, Gaspari L, Alexandrie A-K, Ambrosone C, Autrup H and Autrup JL et al. (2001). Cancer Epidem. Biomark. Prev., 10, 1239–1248.
  49. German J. (1997). Cancer Genet. Cytogenet., 93, 100–106. | Article | PubMed | ChemPort |
  50. Giusti RM, Iwamoto K and Hatch EE. (1995). Ann. Intern. Med., 122, 778–788. | PubMed | ChemPort |
  51. Glaser SL, Lin RJ, Stewart SL, Ambinder RF, Jarrett RF, Brousset P, Pallesen G, Gulley ML, Khan G, O'Grady J, Hummel M, Preciado MV, Knecht H, Chan JKC and Claviez A. (1997). Int. J. Cancer, 70, 375–382. | Article | PubMed | ISI | ChemPort |
  52. Glazer ER, Perkins CI, Young Jr JL, Schlag RD, Campleman SL and Wright WE. (1999). Cancer, 86, 1070–1079.
  53. Golding J, Greenwood R, Birmingham K and Mott M. (1992). Br. Med. J., 305, 341–346. | ISI | ChemPort |
  54. Greaves MF, Colman SM, Beard MEJ, Bradstock K, Cabrera ME, Chen P-M, Jacobs P, Lam-Po-Tang PRL, MacDougall LG, Williams CKO and Alexander FE. (1993). Leukemia, 7, 27–34. | PubMed | ISI | ChemPort |
  55. Greaves MF, Maia AT, Wiemels JL and Ford AM. (2003a). Blood, 102, 2321–2333. | Article | PubMed | ISI | ChemPort |
  56. Greaves MF and Wiemels J. (2003b). Nat. Rev. Cancer, 3, 639–649. | Article | PubMed | ISI | ChemPort |
  57. Gripp KW, Scott CI, Nicholson L, McDonald-McGinn DM, Ozeran JD, Jones MC, Lin AE and Zackai EH. (2002). Am. J. Med. Genet., 108, 80–87. | Article | PubMed | ISI |
  58. Grønskov K, Olsen JH, Sand A, Pedersen W, Carlsen N, Jylling AMB, Lyngbye T, Brøndum-Nielson K and Rosenberg T. (2001). Hum. Genet., 109, 11–19.
  59. Gutiérrez MI, Bhatia K, Barriga F, Diez B, Muriel FS, de Andreas ML, Epelman S, Risueño C and Magrath IT. (1992). Blood, 79, 3261–3266. | PubMed | ChemPort |
  60. Hasle H. (2001). Lancet Oncol., 2, 429–436. | Article | PubMed | ChemPort |
  61. Hasle H, Clemmensen IH and Mikkelsen M. (2000). Lancet, 355, 165–169. | Article | PubMed | ISI | ChemPort |
  62. Hasle H, Mellemgaard A, Nielsen J and Hansen J. (1995). Br. J. Cancer, 71, 416–420. | PubMed | ISI | ChemPort |
  63. Hemminki K, Kyyrönen P and Vaittinen P. (1999). Epidemiology, 10, 271–275. | PubMed | ISI | ChemPort |
  64. Henneveld HT, van Lingen RA, Hamel BCJ, Stolte-Dijkstra I and van Essen AJ. (1999). Am. J. Med. Genet., 86, 439–446. | Article | PubMed | ISI | ChemPort |
  65. Hermon C, Alberman E, Beral V, Swerdlow AJ and for the Collaborative Study Group of Genetic Disorders. (2001). Ann. Hum. Genet., 65, 167–176.
  66. Hersh JH, Cole TR, Bloom AS, Bertolone SJ and Hughes HE. (1992). J. Pediatr., 120, 572–574. | PubMed | ISI | ChemPort |
  67. Hirschhorn JN, Lohmueller K, Byrne E and Hirschhorn K. (2002). Genet. Med., 4, 45–61. | Article | PubMed | ISI | ChemPort |
  68. Holly EA, Aston DA, Ahn DK and Kristiansen JJ. (1992). Am. J. Epidemiol., 135, 122–129.
  69. Hooper ML. (1999). Br. J. Cancer, 79, 1273–1276. | Article |
  70. Hoyme HE, Seaver LH, Jones KL, Procopio F, Crooks W and Feingold M. (1998). Am. J. Med. Genet., 79, 274–278. | Article | PubMed | ChemPort |
  71. Hrusák O, Trka J, Zuna J, Poloucková A, Kalina T and Starý J. (2002). Leukemia, 16, 720–725. | Article |
  72. Hwang S-J, Cheng LS-C, Lozano G, Amos CI, Gu X and Strong LC. (2003a). Hum. Genet., 113, 238–243. | Article | PubMed | ISI | ChemPort |
  73. Hwang S-J, Lozano G, Amos CI and Strong LC. (2003b). Am. J. Hum. Genet., 72, 975–983. | Article | PubMed | ISI | ChemPort |
  74. Infante-Rivard C. (2003). Health Phys., 85, 60–64.
  75. Infante-Rivard C, Amre D and Sinnett D. (2002a). Environ. Health Perspect., 110, 591–593.
  76. Infante-Rivard C, Krajinovic M, Labuda D and Sinnett D. (2000). Cancer Causes Control, 11, 547–553.
  77. Infante-Rivard C, Krajinovic M, Labuda D and Sinnett D. (2002b). Epidemiology, 13, 277–281. | PubMed |
  78. Infante-Rivard C, Labuda D, Krajinovic M and Sinnett D. (1999). Epidemiology, 10, 481–487. | Article | PubMed | ISI | ChemPort |
  79. Ioannidis JPA, Ntzani EE, Trikalinos TA and Contopoulos-Ioannidis DG. (2001). Nat. Gen., 29, 306–309. | Article | ChemPort |
  80. Iscovich J and Parkin DM. (1997). Int. J. Cancer, 70, 649–653. | Article | PubMed |
  81. Jemal A, Devesa SS, Fears TR and Fraumeni JF. (2000). Br. J. Cancer, 82, 1875–1878. | Article |
  82. Kerr B, Mucchielli ML, Sigaudy S, Fabre M, Saunier P, Voelckel MA, Howard E, Elles R, Eden TOB, Black GC and Philip N. (2003). J. Med. Genet., 40, 469–471.
  83. Kleinerman RA, Tarone RE, Abramson DH, Seddon JM, Li FP and Tucker MA. (2000). J. Natl. Cancer Inst., 92, 2037–2039. | Article | PubMed |
  84. Klip H, Burger CW, de Kraker J, Van Leeuwen FE and for the OMEGA-project group. (2001). Hum. Reprod., 16, 2451–2458. | PubMed | ISI | ChemPort |
  85. Kohn DB, Sadelain M and Glorioso JC. (2003). Nat. Rev. Cancer, 3, 477–488. | Article | PubMed | ISI | ChemPort |
  86. Kraemer KH, Lee M-M, Andrews AD and Lambert WC. (1994). Arch. Dermatol., 130, 1018–1021. | Article | PubMed | ISI | ChemPort |
  87. Krajinovic M, Labuda D, Richer C, Karimi S and Sinnett D. (1999). Blood, 93, 1496–1501. | PubMed | ISI | ChemPort |
  88. Krajinovic M, Labuda D and Sinnett D. (2001). Rev. Environ. Health, 16, 263–279. | PubMed |
  89. Krajinovic M, Richer C, Sinnett H, Labuda D and Sinnett D. (2000). Cancer Epidem. Biomark. Prev., 9, 557–562.
  90. Krajinovic M, Sinnett H, Richer C, Labuda D and Sinnett D. (2002). Int. J. Cancer, 97, 230–236. | Article | PubMed | ISI | ChemPort |
  91. Kramárová E and Stiller CA. (1996). Int. J. Cancer, 68, 759–765. | Article | PubMed | ISI | ChemPort |
  92. Kutler DI, Singh B, Satagopan J, Batish SD, Berwick M, Giampietro PF, Hanenberg H and Auerbach AD. (2003). Blood, 101, 1249–1256. | Article | PubMed | ISI | ChemPort |
  93. Law GR, Smith AG, Roman E and United Kingdom Childhood Cancer Study Investigators. (2002). Br. J. Cancer, 86, 350–355. | Article | PubMed | ChemPort |
  94. Lerner-Geva L, Toren A, Chetrit A, Modan B, Mandel M, Rechavi G and Dor J. (2000). Cancer, 88, 2845–2847. | Article | PubMed | ISI | ChemPort |
  95. Li FP, Fraumeni JF, Mulvihill JJ, Blattner WA, Dreyfus MG, Tucker MA and Miller RW. (1988a). Cancer Res., 48, 5358–5362. | PubMed | ISI | ChemPort |
  96. Li FP, Williams WR, Gimbrere K, Flamant F, Green DM and Meadows AT. (1988b). Pediatrics, 81, 147–149.
  97. Lin HJ, Han C-Y, Lin BK and Hardy S. (1994). Pharmacogenetics, 4, 125–134. | PubMed | ChemPort |
  98. Lindor NM, Greene MH and The Mayo Familial Cancer Program. (1998). J. Natl. Cancer Inst., 90, 1039–1071. | Article | PubMed | ChemPort |
  99. Lindsay S, Ireland M, O'Brien O, Clayton-Smith J, Hurst JA, Mann J, Cole T, Sampson J, Slaney S, Schlessinger D, Burn J and Pilia G. (1997). J. Med. Genet., 34, 480–483. | PubMed | ISI | ChemPort |
  100. Linet MS and Devesa SS. (1991). Br. J. Cancer, 63, 424–429. | PubMed | ChemPort |
  101. Lipton JM, Federman N, Khabbaze Y, Schwartz CL, Hilliard LM, Clark JI and Vlachos A. (2001). J. Pediatr. Hematol. Oncol., 23, 39–44.
  102. Little J. (1999). Epidemiology of Childhood Cancer (IARC Scientific Publications No 149). International Agency for Research on Cancer: Lyon.
  103. Mathonnet G, Krajinovic M, Labuda D and Sinnett D. (2003). Br. J. Haematol., 123, 45–48. | Article | PubMed | ISI | ChemPort |
  104. McGaughran JM, Harris DI, Donnai D, Teare D, Macleod R, Westerbeek R, Kingston H, Super M and Evans DGR. (1999). J. Med. Genet., 36, 197–203.
  105. McNally RJQ, Kelsey AM, Cairns DP, Taylor GM, Eden OB and Birch JM. (2001). Cancer, 92, 1967–1976. | Article | PubMed | ChemPort |
  106. Moll AC, Imhof SM, Cruysberg RM, Schouten-Van Meeteren AYN, Boers M and Van Leeuwen FE. (2003). Lancet, 361, 309–310. | Article | PubMed | ISI |
  107. Monge P, Wesseling C, Rodríguez AC, Cantor KP, Weiderpass E, Reutfors J, Ahlbom A and Partanen T. (2002). Paediatr. Perinat. Epidemiol., 16, 210–218.
  108. Morrell D, Cromartie E and Swift M. (1986). J. Natl. Cancer Inst., 77, 89–92. | PubMed | ChemPort |
  109. Moysich KB, Menezes RJ and Michalek AM. (2002). Lancet Oncol., 3, 269–279. | Article | PubMed | ISI |
  110. Mueller BU and Pizzo PA. (1995). J. Pediatr., 126, 1–10.
  111. Nakadate H, Yokomori K, Watanabe N, Tsuchiya T, Namiki T, Kobayshi H, Suita S, Tsunematsu Y, Horikoshi Y, Hatae Y, Endo M, Komada Y, Eguchi H, Toyoda Y, Kikuta A, Kobayashi R and Kaneko Y. (2001). Int. J. Cancer, 94, 396–400. | Article | PubMed | ChemPort |
  112. Narod SA, Hawkins MM, Robertson CM and Stiller CA. (1997). Am. J. Hum. Genet., 60, 474–485.
  113. Narod SA, Stiller C and Lenoir GM. (1991). Br. J. Cancer, 63, 993–999. | PubMed | ISI | ChemPort |
  114. Newton R, Ziegler J, Beral V, Mbidde E, Carpenter L, Wabinga H, Mbulaiteye S, Appleby P, Reeves G, Jaffe H and Uganda Kaposi's Sarcoma Study Group. (2001). Int. J. Cancer, 92, 622–627. | Article | PubMed | ISI | ChemPort |
  115. Olsen JH, Boice JD, Seersholm N, Bautz A and Fraumeni JF. (1995). N. Eng. J. Med., 333, 1594–1599.
  116. Olson JM, Hamilton A and Breslow NE. (1995). Med. Pediatr. Oncol., 24, 305–309.
  117. Opitz JM, Weaver DW and Reynolds JF. (1998). Am. J. Med. Genet., 79, 294–304. | Article | PubMed | ISI | ChemPort |
  118. Ozaki T, Schaefer K-L, Wai D, Yokoyama R, Ahrens S, Diallo R, Hasegawa T, Shimoda T, Hirohashi S, Kawai A, Naito N, Morimoto Y, Inoue H, Boecker W, Juergens H, Winkelmann W, Dockhorn-Dworniczak B and Poremba C. (2002). Ann. Oncol., 13, 1656–1664.
  119. Pakakasama S, Chen TT-L, Frawley W, Muller C, Douglass EC and Tomlinson GE. (2003). Int. J. Cancer, 106, 205–207.
  120. Paraf F, Jothy S and Van Meir EG. (1997). J. Clin. Oncol., 15, 2744–2758.
  121. Parkin DM, Kramárová E, Draper GJ, Masuyer E, Michaelis J, Neglia J, Qureshi S and Stiller CA (eds) (1998). International Incidence of Childhood Cancer, Vol. 2 (IARC Scientific Publications No 144). International Agency for Research on Cancer: Lyon.
  122. Parkin DM, Pisani P, Muñoz N and Ferlay J. (1999). Cancer Surv., 33, 5–33. | ISI |
  123. Parkin DM, Stiller CA, Draper GJ, Bieber CA, Terracini B and Young JL (eds) (1988). International Incidence of Childhood Cancer (IARC Scientific Publications No. 87). International Agency for Research on Cancer: Lyon.
  124. Perri P, Longo L, McConville C, Cusano R, Rees SA, Seri M, Conte M, Romeo G, Devoto M and Tonini GP. (2002). Ann. N. Y. Acad. Sci., 963, 74–84.
  125. Powell JE, Esteve J, Mann JR, Parker L, Frappaz D, Michaelis J, Kerbl R, Mutz ID and Stiller CA. (1998). Lancet, 352, 682–687. | Article | PubMed | ChemPort |
  126. Powell JE, Kelly AM, Parkes SE, Cole TRP and Mann JR. (1995). Br. J. Cancer, 72, 1563–1569.
  127. Rajalekshmy KR, Abitha AR, Pramila R, Gnanasagar T, Maitreyan V and Shanta V. (1994). Leuk. Res., 18, 183–190.
  128. Rapley EA, Barfoot R, Bonaïti-Pellié C, Chompret A, Foulkes W, Perusinghe N, Reeve A, Royer-Pokora B, Schumacher V, Shelling A, Skeen J, de Toureil S, Weirich A, Pritchard-Jones K, Stratton MR and Rahman N. (2000). Br. J. Cancer, 83, 177–183. | Article |
  129. Rasmussen SA, Yang Q and Friedman JM. (2001). Am. J. Hum. Genet., 68, 1110–1118. | Article | PubMed | ISI | ChemPort |
  130. Reynolds P, Von Behren J and Elkin EP. (2002). Am. J. Epidemiol., 155, 603–613. | Article | PubMed |
  131. Ribeiro RC, Sandrini F, Figueiredo B, Zambetti GP, Michalkiewicz E, Lafferty AR, DeLacerda L, Rabin M, Cadwell C, Sampaio G, Cat I, Stratakis CA and Sandrini R. (2001). Proc. Natl. Acad. Sci. USA, 98, 9330–9335. | Article | PubMed | ChemPort |
  132. Ricciardone MD, Özçelik T, Cevher B, Özdag H, Tuncer M, Gürgey A, Uzunalimoglu Ö, Çetinkaya H, Tanyeli A, Erken E and Öztürk M. (1999). Cancer Res., 59, 290–293. | PubMed | ISI | ChemPort |
  133. Roman E, Fear NT, Ansell P, Bull D, Draper G, McKinney P, Michaelis J, Passmore SJ and von Kries R. (2002). Br. J. Cancer, 86, 63–69. | Article | PubMed | ISI | ChemPort |
  134. Rosenberg NA, Pritchard JK, Weber JL, Cann HM, Kidd KK, Zhivotovsky LA and Feldman MW. (2002). Science, 298, 2381–2385. | Article | PubMed | ISI | ChemPort |
  135. Rosenberg PS, Greene MH and Alter BP. (2003). Blood, 101, 822–826. | Article | PubMed | ISI | ChemPort |
  136. Sanders BM, Jay M, Draper GJ and Roberts EM. (1989). Br. J. Cancer, 60, 358–365. | PubMed | ISI | ChemPort |
  137. Sandrini R, Ribeiro RC and DeLacerda L. (1997). J. Clin. Endocrinol. Metab., 82, 2027–2031.
  138. Sankila R, Olsen JH, Anderson H, Garwicz S, Glattre E, Hertz H, Langmark F, Lanning M, Moller T, Tulinius H and for the Association of the Nordic Cancer Registries the Nordic Society of Paediatric Haematology Oncology. (1998). N. Eng. J. Med., 338, 1339–1344.
  139. Satgé D, Sasco AJ, Carlsen NLT, Stiller CA, Rubie H, Hero B, De Bernardi B, de Kraker J, Coze C, Kogner P, Langmark F, Hakvoort-Cammel FGAJ, Beck D, von der Weid N, Parkes S, Hartmann O, Lippens RJJ, Kamps WA and Sommelet D. (1998). Cancer Res., 58, 448–452. | PubMed | ISI | ChemPort |
  140. Satgé D, Sasco AJ and Lacour B. (2003). Int. J. Cancer, 106, 297–298.
  141. Schilling FH, Spix C, Berthold F, Erttmann R, Fehse N, Hero B, Klein G, Sander J, Schwarz K, Treuner J, Zorn U and Michaelis J. (2002). N. Eng. J. Med., 346, 1047–1053.
  142. Schüz J. (2003). Paediatr. Perinat. Epidemiol., 17, 106–112.
  143. Schüz J, Kaatsch P, Kaletsch U, Meinert R and Michaelis J. (1999). Int. J. Epidemiol., 28, 631–639. | Article | PubMed | ISI | ChemPort |
  144. Schüz J, Spector LG and Ross JA. (2003). Am. J. Epidemiol., 158, 710–716.
  145. Sévenet N, Sheridan E, Amram D, Schneider P, Handgretinger R and Delattre O. (1999). Am. J. Hum. Genet., 65, 1342–1348. | Article | PubMed | ISI | ChemPort |
  146. Shibata Y, Yamashita S, Masayakin VB, Panasyuk GD and Nagataki S. (2001). Lancet, 358, 1965–1966. | Article | PubMed | ChemPort |
  147. Shiramizu B, Barriga F, Neequaye J, Jafri A, Dalla-Favera R, Neri A, Guttierez M, Levine P and Magrath I. (1991). Blood, 77, 1516–1526. | PubMed | ChemPort |
  148. Smith MA, Freidlin B, Gloeckler Ries LA and Simon R. (1998). J. Natl. Cancer Inst., 90, 1269–1277. | Article | PubMed | ChemPort |
  149. Smith MT, Wang Y, Skibola CF, Slater DJ, Lo Nigro L, Nowell PC, Lange BJ and Felix CA. (2002). Blood, 100, 4590–4593. | Article | PubMed | ISI | ChemPort |
  150. Sotelo-Avila C, Gonzalez-Crussi F and Fowler JW. (1980). J. Pediatr., 96, 47–50. | PubMed | ChemPort |
  151. Steinitz R, Iscovich JM and Katz L. (1990). Cancer Detect. Prev., 14, 547–553.
  152. Stewart A, Webb J and Hewitt D. (1958). Br. Med. J., i, 1495–1508.
  153. Stiller CA. (2004). In Pinkerton CR, Plowman PN and Pieters R (eds) Paediatric Oncology, 3rd. edn.. Arnold: London, pp. 3–24.
  154. Stiller CA, Chessells JM and Fitchett M. (1994). Br. J. Cancer, 70, 969–972. | PubMed | ChemPort |
  155. Sullivan KE, Mullen CA, Blaese RM and Winkelstein JA. (1994). J. Pediatr., 125, 876–885. | PubMed | ISI | ChemPort |
  156. Swerdlow AJ, Hermon C, Jacobs PA, Alberman E, Beral V, Daker M, Fordyce A and Youings S. (2001). Ann. Hum. Genet., 65, 177–188. | Article | PubMed | ISI | ChemPort |
  157. Taylor GM, Dearden S, Payne N, Ayres M, Gokhale DA, Birch JM, Blair V, Stevens RF, Will AM and Eden OB. (1998). Br. J. Cancer, 78, 561–565. | PubMed |
  158. Taylor GM, Dearden S, Ravetto P, Ayres M, Watson P, Hussain A, Greaves M, Alexander F, Eden OB and UKCCS Investigators. (2002). Hum. Mol. Genet., 11, 1585–1597. | Article | PubMed | ISI | ChemPort |
  159. Taylor GM, Robinson MD, Binchy A, Birch JM, Stevens RF, Jones PM, Carr T, Dearden S and Gokhale DA. (1995). Leukemia, 9, 440–443.
  160. Taylor MD, Gokgoz N, Andrulis IL, Mainprize TG, Drake JM and Rutka JT. (2000). Am. J. Hum. Genet., 66, 1403–1406. | Article | PubMed | ISI | ChemPort |
  161. The International Nijmegen Breakage Syndrome Study Group (2000). Arch. Dis. Child., 82, 400–406. | Article | PubMed | ISI |
  162. Thompson JR, Fitz Gerald P, Willoughby MLN and Armstrong BK. (2001). Lancet, 358, 1935–1940. | Article | PubMed | ISI | ChemPort |
  163. Valery PC, McWhirter W, Sleigh A, Williams G and Bain C. (2003). Int. J. Cancer, 105, 825–830.
  164. Varley JM. (2003). Hum. Mutat., 21, 313–320. | Article | PubMed | ISI | ChemPort |
  165. Varley JM, McGown G, Thorncroft M, James LA, Margison GP, Forster G, Evans DGR, Harris M, Kelsey AM and Birch JM. (1999). Am. J. Hum. Genet., 65, 995–1006. | Article | PubMed | ISI | ChemPort |
  166. Vasen HFA, Sanders EACM, Taal BG, Nagengast FM, Griffioen G, Menko FH, Kleibeuker JH, Houwing-Duistermaat JJ and Meera Khan P. (1996). Int. J. Cancer, 65, 422–425. | Article | PubMed | ISI | ChemPort |
  167. Vasen HFA, Stormorken A, Menko FH, Nagengast FM, Kleibeuker JH, Griffioen G, Taal BG, Moller P and Wijnen JT. (2001). J. Clin. Oncol., 19, 4074–4080. | PubMed | ISI | ChemPort |
  168. Vilchez RA, Kozinetz CA, Arrington AS, Madden CR and Butel JS. (2003). Am. J. Med., 114, 675–684. | Article | PubMed | ISI |
  169. Wabinga HR, Parkin DM, Wabwire-Mangen F and Nambooze S. (2000). Br. J. Cancer, 82, 1585–1592. | Article | PubMed | ISI | ChemPort |
  170. Wang LL, Levy ML, Lewis RA, Chintagumpala MM, Lev D, Rogers M and Plon SE. (2001). Am. J. Med. Genet., 102, 11–17. | Article | PubMed | ISI | ChemPort |
  171. Wang Q, Lasset C, Desseigne F, Frappaz D, Bergeron C, Navarro C, Ruano E and Puisieux A. (1999). Cancer Res., 59, 294–297. | PubMed | ISI | ChemPort |
  172. Webb D, Fryer AE and Osborne JP. (1996). Dev. Med. Child Neurol., 38, 146–155.
  173. Weinberg AG, Mize CE and Worthen HG. (1976). J. Pediatr., 88, 434–438.
  174. Westergaard T, Andersen PK, Pedersen JB, Olsen JH, Frisch M, Sorensen HT, Wohlfahrt J and Melbye M. (1997). J. Natl. Cancer Inst., 89, 939–947. | Article | PubMed | ChemPort |
  175. Whiteside D, McLeod R, Graham G, Steckley JL, Booth K, Somerville MJ and Andrew SE. (2002). Cancer Res., 62, 359–362. | PubMed | ISI | ChemPort |
  176. Wiemels JL, Pagnamenta A, Taylor GM, Eden OB, Alexander FE, Greaves MF and United Kingdom Childhood Cancer Study Investigators. (1999). Cancer Res., 59, 4095–4099. | PubMed | ISI | ChemPort |
  177. Wiemels JL, Smith RN, Taylor GM, Eden OB, Alexander FE, Greaves MF and United Kingdom Childhood Cancer Study Investigators. (2001). Proc Natl. Acad. Sci. USA, 98, 4004–4009. | Article | PubMed | ChemPort |
  178. Wilcken B, Bamforth F, Li Z, Zhu H, Ritvanen A, Redlund M, Stoll C, Alembik Y, Dott B, Czeizel AE, Gelman-Kohan Z, Scarano G, Bianca S, Ettore G, Tenconi R, Bellato S, Scala I, Mutchinick OM, López MA, de Walle H, Hofstra R, Joutchenko L, Kavteladze L, Bermejo E, Martínez-Frías ML, Gallagher M, Erickson JD, Vollset SE, Mastroiacovo P, Andria G and Botto LD. (2003). J. Med. Genet., 40, 619–625. | Article | PubMed | ISI | ChemPort |
  179. Wilkinson JD, Fleming LE, MacKinnon J, Voti L, Wohler-Torres B, Peace S and Trapido E. (2001). Cancer, 91, 1402–1408.
  180. Winn DM, Li FP, Robison LL, Mulvihill JJ, Daigle AE and Fraumeni Jr JF. (1992). Cancer Epidem. Biomark. Prev., 1, 525–532.
  181. Winther JF, Sankila R, Boice JD, Tulinius H, Bautz A, Barlow L, Glattre E, Langmark F, Møller TR, Mulvihill JJ, Olafsdottir GH, Ritvanen A and Olsen JH. (2001). Lancet, 358, 711–717. | Article | PubMed | ISI | ChemPort |
  182. Wong FL, Boice Jr JD, Abramson DH, Tarone RE, Kleinerman RA, Stovall M, Goldman MB, Seddon JM, Tarbell N, Fraumeni Jr JF and Li FP. (1997). JAMA, 278, 1262–1267. | Article | PubMed | ISI | ChemPort |
  183. Woods WG, Gao R-N, Shuster JJ, Robison LL, Bernstein M, Weitzman S, Bunin G, Levy I, Brossard J, Dougherty G, Tuchman M and Lemieux B. (2002). N. Eng. J. Med., 346, 1041–1046.
  184. Wu TC, Tong MJ, Hwang B, Lee SD and Hu MM. (1987). Hepatology, 7, 46–48. | PubMed |
  185. Ziegler JL and Katongole-Mbidde E. (1996). Int. J. Cancer, 65, 200–203. | PubMed |


I am grateful to Janette Wallis for secretarial assistance. This work was undertaken by the Childhood Cancer Research Group, which receives funding from the Department of Health and the Scottish Ministers. The views expressed in the publication are those of the authors and not necessarily those of the Department of Health and the Scottish Ministers.



These links to content published by NPG are automatically generated


Cancer: Malignant tumours generated by recessive mutations

Nature News and Views (16 Oct 1986)