Long-term population-based risks of second malignant neoplasms after childhood cancer in Britain

Br J Cancer. 2004 Nov 29;91(11):1905-10. doi: 10.1038/sj.bjc.6602226.

Abstract

In a population-based, retrospective cohort study of 16 541 3-year survivors of childhood cancer treated in Britain up to the end of 1987, 278 second malignant neoplasms (SMNs) were identified against 39.4 expected giving a standardised incidence ratio (SIR) of 6.2. The overall cumulative risk of an SMN by 25 years from 3-year survival from childhood cancer was 4.2%. Analysis of the cohort of nonretinoblastoma childhood cancers combined revealed a significant decline in SIR of SMN with increasing duration of follow-up. There was a greater risk of developing a SMN, particularly secondary acute myeloid leukaemia, in those diagnosed with childhood cancer from 1980 onwards. However, on multivariate modeling, this was not an independent risk factor. There was significant heterogeneity (P<0.001) in SIR of SMN across different treatment groups, the greatest risk observed in the group exposed to both radiotherapy and chemotherapy. The risks of SMN observed were comparable with those in other population-based studies. While the decline in SIR with duration of follow-up and the small excess numbers of cancers observed over later decades after diagnosis are reassuring, the high excess risk, particularly of leukaemia, associated with recent more intense therapy is of concern.

Publication types

  • Comparative Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Acute Disease
  • Adolescent
  • Child
  • Child, Preschool
  • Cohort Studies
  • Female
  • Follow-Up Studies
  • Genetics, Population*
  • Humans
  • Infant
  • Infant, Newborn
  • Leukemia, Myeloid / etiology*
  • Male
  • Neoplasms, Second Primary / epidemiology*
  • Retrospective Studies
  • Risk Factors
  • Survivors
  • Time Factors
  • United Kingdom / epidemiology