A second and more problematic category of molecular diagnosis has become possible in the last decade; several genes have been identified which, when defective in the germline, predispose to a high penetrance form of one of the common forms of cancer. Such patients lack the characteristic syndromic features that allow easy targeting of molecular genetic expertise. The breast/ovarian cancer predisposition genes BRCA1 and BRCA2 and members of the mismatch repair gene family, such as MSH2 and MLH1 that predispose to colorectal cancer and endometrial cancer among others, are the classic examples of this category.

These are large genes in which many hundreds of pathological changes are possible. Distinguishing causative mutations from harmless population variants in families with multiple affected members is a major challenge which is being met through international cooperation. Methods of rapid analysis of such genes are becoming available. In the foreseeable future it is likely that all breast, ovarian, colorectal and endometrial cancers will be checked for germline mutations in these and similar genes so as to better characterise the cancer and allow more effective therapy. In the meantime, limited diagnostic resources must be targeted at those individuals most likely to generate a result of diagnostic value.