O-017 Can single nucleotide variants in TGFBR1 and SMAD7 modify colorectal screening recommendations?

Abstract

Introduction: Standard recommendations for screening for colorectal cancer (CRC) specify a colonoscopy at age 50 and if normal a repeat after 10 years. However, the yield of the follow-on colonoscopy in this setting is poor. The yield of screening could be improved with a preliminary test to identify subjects with a higher risk of CRC.We are interested in the role of blood-based genetic markers in CRC pre-screening because of their low cost and relative non-intrusiveness. Methods: Blood samples were taken from 187 CRC patients and 94 healthy controls from a European Caucasian population. After gDNA extraction, selected sequences were amplified by PCR followed by melting curve analysis. The SNP status for TGFBR1 (rs334348) and SMAD7 (rs4939827) was determined for each subject. The association between allele frequency for the various SNPs and CRC status was evaluated by logistic regression. Both additive and dominance effects were considered. Information from both markers for each subject was condensed into an index which could be used to distinguish high-risk subjects from low-risk ones, so that the high-risk group would become the focus of further screening and early detection of cancer. Furthermore, various thresholds of the index could be defined to optimise the economic performance of the test for various relative costs of false positives or false negatives. Results: For TGFBR1, there was a highly significant additive association between the G allele and colorectal cancer. The AG (heterozygosity) and GG (homozygosity) genotypes were associated with progressively greater risk relative to the AA genotype (Odds ratio = 3.43, p < 0.00005, highly significant). There was no further significant effect for dominance (OR = 0.70, p = 0.216, not significant). Concerning SMAD7, even though there was no significant additive effect, i.e. the risks associated with the CC and TT genotypes were similar (OR = 1.00, p = 0.987, not significant), there was a significant dominance effect, and the CT heterozygote was associated with a lower risk (OR = 0.50, p = 0.014, significant). Sensitivity was calculated as the ratio of test-positive cases to total cases, and specificity as the ratio of test-negative controls to total controls. The trade-off between specificity and sensitivity was explored by ROC (receiver operating characteristic) analysis. For example, when sensitivity was 0.33, specificity was 0.96; when sensitivity was 0.65, specificity was 0.65 also; and when sensitivity was 0.87, specificity was 0.33. Conclusion: Presently the yield of mass screening for colorectal cancer has by its nature a large number of persons of whom the majority will not have cancer. Thus repeat screening after a normal first colonoscopy has a low yield and nevertheless misses a certain number of intermediary cancers. By being able to predict a clearly higher risk group, the repeat colonoscopy might have to be done earlier than stated by guidelines for these individuals. Furthermore, this inexpensive germline test, needing to be done only once, can be carried out at any age, including much earlier than is recommended by current guidelines.