Second primary cancers after childhood cancer

BMJ 1996; 312 doi: http://dx.doi.org/10.1136/bmj.312.7035.861 (Published 06 April 1996) Cite this as: BMJ 1996;312:861
  1. Leslie L Robison
  1. Professor of pediatrics Division of Pediatric Epidemiology and Clinical Research, University of Minnesota Cancer Center, Minneapolis, MN 55455, USA

    Low absolute risk, but prevention and monitoring are high priorities

    Over the past two decades, impressive advances have been made in the treatment and survival of children with cancer. Population based data from the United States document a 68% five year survival for children aged 0-15 years who were diagnosed with cancer in 1985.1 In the vast majority of cases, five year survival represents cure. Accordingly, the number of children and young adults in the population who have been exposed to cancer treatments, which often include radiation and chemotherapy, continues to increase. These children, cured of their cancer, are at risk from adverse effects of the treatment they received.

    Little is known about the long term consequences of cancer treatment. Rapid improvements in the treatment and survival of children with cancer were first realised in the early 1970s, and therefore the first large cohort of survivors of childhood cancer is now entering adulthood, a decade or more after being treated. A retrospective cohort of children in Britain diagnosed with cancer since 1940 has provided an opportunity to investigate the long term risk in a large population.2 3 The recent report by Hawkins and colleagues from the Childhood Cancer Research Group in Oxford describes the risk of bone cancer in over 13000 children who survived three years after treatment.4

    Hawkins et al emphasise the need to put the risks in perspective. Review of the literature gives a picture of massive excesses in risk, with observed to expected ratios in excess of 50—that is, 50 times the expected number of cases occurring in children with previous cancers.5 But from the patient's perspective, the relevant statistic is the cumulative probability or the absolute risk of developing a second cancer. Reports on large populations of patients have shown that the cumulative probability of second cancer can be quite low.6 Despite this low absolute risk, high priority should be given to characterising and preventing second neoplasms because they are often associated with high levels of morbidity and mortality.

    Several questions remain regarding the future risk of cancer in child survivors. Firstly, what are the latencies between specific cancer treatments and subsequent cancers? While the risk of leukaemia associated with alkylating agents seems to be restricted to the first decade after exposure,7 data suggest that radiation will play a substantial role in the risk of non-leukaemic malignancies two and more decades later.7 8

    Secondly, what interactions exist between treatments? The possibility of interactions between radiation and some chemotherapeutic agents has been suggested.9 However, testing for such interactions often requires large study populations, which up until now have been scarce.

    Thirdly, what interactions exist between treatment and the patient's genetic characteristics? As the report by Hawkins et al confirms,4 patients with hereditary retinoblastoma who have received radiation exposure are at massively increased risk of developing a subsequent bone tumour.10 Other genetic characteristics that may interact in this way include p53, ATM, and GST null phenotype.

    Fourthly, what interactions exist between treatment and lifestyle? It is clear that lifestyle factors, including tobacco, alcohol, diet, and occupation, can influence the risk of cancer in the general population. It is reasonable to speculate that exposure to therapeutic agents during childhood may modify the impact of these lifestyle factors, resulting in a multiplicative increase in cancer risk. Conversely, it may be that healthy lifestyle behaviours result in a lower treatment related risk.

    Studying survivors of childhood cancer can also provide insights into the cause of primary malignancies. For example patients treated with epipodophyllotoxins have an excess risk of biologically distinct secondary acute myeloid leukemia.11 This suggests that environmental exposures (such as diet and drugs) that inhibit topoisomerase-II may be important in the aetiology of leukemia in infants with MLL (mixed lineage leukaemia) gene abnormalities.12

    Treatment protocols will continue to evolve, with new combinations of drugs and different dose intensity and sequencing, all of which may modify the risk of a second cancer. The ultimate goal is clear—to successfully treat and cure children with cancer while maximising their chances of a long and healthy life. This will need both primary prevention strategies (such as changes in treatment protocols) and secondary prevention programmes (such as screening). Efforts like those of Hawkins and colleagues to carefully evaluate the experience of previous and future patient cohorts provide data essential for achieving this goal.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.