Content area
Full Text
Large scale genetic studies which aim to identify the moderate and low penetrance loci involved in many genetic but non-Mendelian diseases appear now almost to be a commonplace. Notable success in this field of research includes the recognition of a relationship between the ApoE genotype and Alzheimer's disease risk 1 and in the field of infectious diseases where variation in TNF and HLA have been shown to be associated with substantially different risks of TB and malaria. 2-4 This is not to say that all these studies are near completion and in some cases, such as multiple sclerosis, localisation of predisposition genes (let alone gene identification) is proving very difficult. 5 6 While it is easy to underestimate the difficulties inherent in studying diseases like diabetes, asthma, and rheumatoid arthritis, and indeed to underestimate the remaining problems, it is nevertheless surprising that the study of cancer has lagged behind that of other complex genetic diseases. Almost all cancer susceptibility alleles identified so far are rare and highly penetrant (for example, APC, BRCA1, BRCA2, MSH2, LMH1, PTEN, CDNK2A ). 7 They may cause a substantial proportion of cancers at young ages, but they are unlikely to be responsible for a high proportion of all cancers, leaving a considerable potential contribution from less penetrant genes. It is possible to gain an insight into the potential impact of such genes on cancer incidence given that the relative risk of disease in first degree relatives is only of the order of 1.5-2.5 over all ages for most common cancers. 7 Under a dominant low penetrance model, the ratio of cancer risk in susceptibles to that in the general population cannot exceed about 5, if 50% of all cancers occur in susceptibles, when about 10% of the population must be at increased risk. A dominant gene carried by 2% of the population will cause a risk 11 times that in the general population and will cause 22% of all cancers. Within this range, such genes will rarely produce striking multiple case families, except possibly for breast cancer, where an 11-fold increase in risk would correspond to a penetrance of 43% by the age of 70. In the case of colon cancer, for example, where the cumulative risk in the...