Damage to DNA by oxygen radicals and other reactive oxygen/nitrogen/chlorine species occurs in vivo despite the presence of multiple antioxidant defence and repair systems. Such damage is thought to make a significant contribution to the age-related development of cancer. Modulation of oxidative DNA damage by diet thus constitutes a "biomarker" putatively predictive of the effect of diet on cancer incidence, provided that DNA damage can be accurately quantitated by validated methods. Current issues addressed in this article include the problems of artifactual DNA oxidation during isolation and analysis, the relative merits of different analytical methods, the advantages and disadvantages of relying on measurement of 8-hydroxydeoxyguanosine (8OHdG, 8-oxodG) as an index of oxidative DNA damage, and the limited data that are so far available on how diet can affect "steady-state" levels of oxidative DNA damage in humans. It appears that such damage can be modulated by vegetable intake, although the effects of vegetables may be mediated by components different from the "classical" antioxidants vitamin C, alpha-tocopherol and beta-carotene.