Abstract Recently, several molecular genetic bases of polymorphic enzyme activities involved in drug activation and detoxification have been elucidated. Many molecular epidemiology studies based on these premises have sought to gather information on the association of genetically determined metabolic variants with different risks of environmentally induced cancer. While rare alterations of tumor suppressor genes dramatically raise cancer risk for the single affected subjects, far more common and less dramatic differences in genes encoding for drug metabolism enzymes can be responsible for a relatively small, but rather frequent increase of cancer risk at the population level. This increase could be especially important in specific cases of occupational, pharmacological or environmental exposure. Examination of the current literature reveals that the most extensively investigated metabolic polymorphisms are those of P450 1A1 and P450 2D6 cytochromes, glutathione S-transferases (GSTs; M1 and, to a lesser extent, M3, P1 and T1) and N-acetyltransferases (NATs; NAT1 and NAT2). Making reference to these enzymes, we have assayed the current knowledge on the relations among polymorphisms of human xenobiotic-metabolizing enzymes and cancer susceptibilities. We have found intriguing models of susceptibility toward different types of cancer. We have reviewed and commented these models on light of the complex balance among different enzyme activities that, in each individual, determines the degree of each cancer susceptibility. Moreover, we have found techniques of molecular genetic analysis, more suitable than previous ones on phenotypic expression, now allowing better means to detect individuals at risk of cancer. According to the models presently available, a systematic screening of individuals at risk seems to make sense only in situations of well defined carcinogenic exposures and when performed by the polymorphism analysis of coordinated enzyme activities concurring to the metabolism of the carcinogen(s) in question. Genetic polymorphism analysis can allow for the detection of patients more prone to some types of specific cancers, or to the adverse effects of specific pharmaceutical agents. Considering the increasingly confirmed double-edged sword nature of metabolism polymorphism (both wild-type and variant alleles can predispose to cancer, albeit in different situations of exposure), individual susceptibility to cancer should be monitored as a function of the nature, and mechanism of action, of the carcinogen(s) to which the individual under study is known to be exposed, and with reference to the main target organ of the considered type of exposure.