2953

Background: During the last few decades, tremendous effort has been put forth to identify the sources of genetic susceptibility to cancer. Continuing advances in genotyping technologies and the set up of observational studies with DNA sample collections have resulted in a large and increasing number of genetic association studies. This leads us to the question about the overall contribution of these studies to our current understanding of the genetic susceptibility to cancer. In an attempt to synthesize previous association studies and overcome small sample size issues of individual studies, meta- and pooled analyses have been used as a systematic approach of integrating results from a number of independent studies to identify a trend in the data. Methods: To evaluate the overall progress of observational studies on genetic variants and cancer risk, we systematically examined the published results of meta- and pooled analyses for genetic polymorphisms and risk of cancer. For this analysis, we used the keyword combinations of “cancer + meta + gene” and “cancer + pooled + gene” to search the PubMed database for meta- and pooled analyses that evaluated the association between genetic polymorphisms and cancer risk in humans. We also used the keyword combinations of “gene + cancer” and “genetic + cancer” restricted to publication type “meta-analysis” to identify additional studies that may have been initially missed. The search included studies published before November 2005. While we are focusing here on main effects of genes we are currently expanding our systematic review to gene-gene and gene-environmental interactions. Results: Sixty-four published meta- and pooled analyses were identified, encompassing 17 cancer sites and 45 different genes. These 64 meta- and pooled analyses evaluated 150 gene-cancer associations and included on average eight studies per investigated association. Most analyses have been conducted for breast (n=28), colorectal (n=23), lung (n=23), and prostate (n=21) while very few analyses have evaluated genetic associations in stomach, endometrial and ovarian cancer. Among the 150 gene-cancer associations evaluated, the summary OR for 54 (36%) associations was significant. Of these 54 significant summary ORs, 11 reported ORs below 1.0 with a mean OR of 0.60 (range 0.36-0.76) and 43 reported ORs above 1.0 with a mean OR of 1.80 (range 1.08-12.8). Deletion of glutathione S-transferases (GSTs) phase II enzymes, GSTM1 and GSTT1, accounted for a large proportion of both significant and non-significant associations reported, followed by polymorphisms in genes of the cytochrome P450 family and DNA repair. Conclusions: About one third of all meta- and pooled analyses resulted overall in significant gene-cancer associations. Results from meta- and pooled analyses can be helpful in directing future genetic association studies towards areas that still require confirmation and away from those that do not.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]