Because of important roles of cytochromes P450 in the metabolic activation of many precarcinogens, extensive research in the past has focused on the relationship between the distribution of polymorphic variants of different isozymes of P450 and cancer susceptibility. In this respect three isozymes in particular have been studied, CYP1A1, CYP2D6, and CYP2E1. Both CYP1A1 and CYP2E1 participate in the metabolism of many suspected as well as established carcinogens, whereas essentially only one carcinogenic substrate has been identified for CYP2D6. Polymorphic sites for the three CYP genes have been identified both in the open reading frame as well as in introns and the regulatory 5' region. In the present contribution we summarize the molecular epidemiological research relating CYP polymorphism to cancer susceptibility and in some cases to toxicity. An interesting polymorphism has been described on the phenotypic level for the inducibility of CYP1A1, a polymorphism that in some studies has been related to a mutation in the 3' flanking region of the CYP1A1 gene. However, the genetic basis for this polymorphism might be inherited in the genes coding for proteins responsible for the induction of CYP1A1, ie, the Ah receptor or the ARNT protein. Data on lung cancer and CYP1A1 gene polymorphism indicate that carriers of genotypes associated with CYP1A1 inducibility are at higher risk for cancer, but that, at least for Caucasians, the recognized mutations probably identify only a fraction of the inducible individuals. The amount of DNA adducts correlates well in some studies to the individual activity registered for CYP1A1. CYP2D6 phenotype and genotype have mainly been related to the incidence of lung cancer, but results from 13 different studies now show an absence of any significant correlation between these parameters. In the case of CYP2E1, some studies indicate a relationship between lung cancer and the occurrence of a rare allele, although future research is needed in order to establish a significant relationship. It is concluded that, at the present stage, none of the polymorphic sites determined in the CYP genes can yet be used as markers for increased lung cancer risk. Future research in this field might be focused on the establishment of new polymorphic sites in the CYP genes, affecting inducibility or function, and on the molecular basis for the interesting differences in CYP1A1 inducibility.