Abstract 4740: Identification of new genes of clinical interest for colorectal cancer: Study design considerations for association studies

Abstract

Abstract Introduction. About 70% of the familial aggregation of colorectal cancer (CRC) cannot be explained by known defects on the predisposing MisMatch Repair (MMR) genes. Recent Genome-Wide Association Studies (GWAS) have detected only a few variants which collectively account for a marginal proportion of this otherwise unexplained excess familial risk. The emerging paradigm is that genetic variants of clinical interest for familial CRC are rarer than those tested by GWAS, and therefore are not explored. Using data simulation and mathematical modelling, we tried to identify appropriate study designs in this setting. Methods. Using the mathematical properties of multilocus exchangeable models, we first selected combinations of allele frequencies, genotypic risks, number of loci, and degree of dominance of unknown biallelic susceptibility genes that were both compatible with the genetic epidemiology of CRC and of clinical interest. Then, using a simulation program for familial data, we assessed the sample sizes required to reach 90% power when comparing ascertainment strategies for association studies. Specifically, we investigated the effect of case selection according to familial criteria, and compared both population- and familial-based approaches. Our simulation program took into account real population data including demographics, CRC incidence by age and global mortality. Results. As expected, selecting cases on familial criteria considerably increased the power of association studies under the hypothesis of rare causal alleles of moderate to high effect. For example, for the case-control study design, if the mode of inheritance of causal variants was additive, with an allele frequency of 0.001 and a genotype relative risk of 10, the sample size required could be divided by 3 if all CRC cases had at least one affected first degree relative. When using the transmission disequilibrium test with the same genetic parameters in the familial-based approach, similar results were obtained in terms of numbers of parents-affected child triads. However, compared with the case-control design, the relative gains in power were lower for more frequent alleles. Finally, we found that unaffected relatives of familial CRC cases had a frequency of at-risk alleles similar to that of sporadic CRC cases, and therefore could be added to CRC cases in case-control studies. Conclusion. Although testing rare alleles is difficult because few individuals are exposed in a given population, ascertainment strategies based on cases selected from multiple CRC families, including both affected and unaffected individuals, could dramatically reduce the sample size for association studies. Our results could be useful to plan future studies, and highlight the potential of data collected in cancer genetics clinics. Note: This abstract was not presented at the AACR 101st Annual Meeting 2010 because the presenter was unable to attend. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 4740.