There has been considerable progress in the past decade in elucidating the molecular basis of malignant transformation of retinoblastoma, Wilms' tumour and neuroblastoma. For retinoblastoma, the story is relatively simple and the laboratory focus should be on rapid detection of germline mutations, identifying additional genetic changes associated with tumour progression and the prevention of second primary cancers, if possible. In the case of Wilms' tumour, the locus (or loci) responsible for hereditary predisposition is still unknown. Furthermore, the WT2 locus has not been cloned and there is no consistent evidence of oncogene activation. Clearly, there is still much to be done before we fully understand the genetic basis of this disease. Finally, substantial insights have been gained from the molecular analysis of neuroblastomas, but this tumour is perhaps the least well understood. Two sites of allelic loss have been identified (1p36 and 14q32), but the presumptive suppressor genes that are the targets of these changes have yet to be identified. Furthermore, it is not clear if either of these loci is responsible for a genetic predisposition to develop this tumour. Although N-myc amplification is a powerful prognostic marker of aggressive tumours, no other oncogene has been shown to be activated in tumours lacking amplification. Finally, the NGFR pathway may have an important role in regulating differentiation and programmed cell death in these cells, but other NGFR family pathways or unrelated genes may be involved as well. Hopefully, the next decade will provide us with answers to many of these open questions.