The availability of murine models with precisely defined genetic lesions has greatly increased our understanding of the genetic control of cell death, with functional dependence established for a wide range of genes including (amongst others) the p53 and Bcl-2 gene family members, the mismatch repair (MMR) genes and the methyl binding domain family member Mbd4. These studies raised the attractive hypotheses that tumour predisposition may be explained in terms of failed cell death, and also that tumour regression may be initiated through activation of an apoptotic programme. The studies that have addressed these notions have revealed complex consequences of a failed death programme, such that these simple hypotheses have not always been supported. Remarkably, however, some tissues show more predictable responses than others, most apparent in the contrast between the intestine and the haematopoietic system. This review will focus upon a discussion of these relationships, and will also consider the relevance of some of these findings to tumour predisposition and regression.