Prevention is an important and effective measure for reducing death caused by cancer. Thus information on individual susceptibility to cancer is valuable in suggesting high risk individuals to avoid intake of carcinogenic substances and receive frequent physical screening. To this end, polymorphisms found within cytochrome P450 (CYP) genes implicated in the metabolism of procarcinogens are expected to be good genetic targets in assessing human cancer susceptibility. We have found polymorphisms in the CYP2E1 and CYP1A1 genes associated with lung cancer susceptibility, though there were some discrepancies from observations made by other investigators. Discrepancies among investigators from different regions, however, are very common in these pharmacogenetic studies. We present an explanation for these discrepancies, difficulties associated with prediction of relative risk of individuals, and future directions.