9 May 2005
The definition of hereditary nonpolyposis colorectal cancer (HNPCC) has been uncertain since the discovery that some, but not all families with clinically-defined HNPCC have germ-line mutations in one of six genes involved in the repair of mismatched base pairs in DNA; of these, mutations in the MSH2 or MLH1 genes are by far the most common. Syngal et al have attempted to assess the sensitivity and specificity of four different sets of clinical criteria for HNPCC for predicting the presence of an MSH2 or MLH1 mutation [Syngal, S. et al (2000) J Med Genet 37, 641-645 (Abstract)]. The criteria were, in decreasing order of stringency, the Amsterdam criteria, the Amsterdam II and Modified Amsterdam criteria, and the Bethesda criteria. 70 families were enrolled in the study on the basis of early age of onset of disease, multiple affected family members, and familial association of colorectal cancer with other HNPCC-associated tumours. An affected person with early onset disease from each family was tested for germ-line mutations in the MSH2 or MLH1 genes. The sensitivities of the four sets of criteria for predicting mutations were 61%, 78%, 72% and 94% respectively, and the specificities were 67%, 61%, 50% and 25%. Using just the first three of the Bethesda criteria did not affect the sensitivity but increased the specificity to 49%; the authors suggest that these criteria should be used to assess whether genetic testing is appropriate in a family.
Comment: This study is the most recent of several that have attempted to define clinical criteria for offering genetic testing for mismatch repair gene mutations (e.g. see Ponz de Leon et al in June 99 newsletter; Park et al in August 99 newsletter; Vasen et al in August 99 newsletter). The proportion of clinically ascertained families in which mutations can be detected varies widely in different reports. The Boston group represented by Syngal et al has consistently found higher percentages of mutation-positive families than most other groups. The reason for the difference is not known but may reflect differences in the populations from which the families are drawn, or the way families participating in the study are chosen. Perhaps the usefulness of genetic testing in HNPCC will need to be independently assessed in different populations.