There is considerable evidence that consumption of fruits and vegetables may contribute to the prevention of cancer. It is however remarkable that evidence for such a preventive action arising from mechanistic studies is becoming stronger, whereas results of some recent prospective studies are less convincing. This apparent discrepancy may be overcome, or at least understood, by introducing molecular markers in future epidemiological studies, taking modulation of molecular processes as well as genetic variability in human populations into account. Both human and animal studies demonstrated that vegetable intake modulates gene expression in the gastrointestinal tract of many genes involved in biological pathways in favor of cancer risk prevention. Gene sets identified in this type of studies can be further evaluated, linked to the biological effects of phytochemicals and developed into biomarkers for larger human studies. Human dietary intervention studies have demonstrated that, apart from target tissues, also peripheral lymphocytes can be used for biomonitoring of chemopreventive effects. Transcriptomic responses and metabolite profiling may link phenotypic markers of preventive effects to specific molecular processes. The use of genomics techniques appears to be a promising approach to establish mechanistic pathways involved in chemoprevention by phytochemicals, particularly when genetic variability is taken into account.