IDENTIFICATION OF HIGH-RISK FAMILIES
Efficient screening for Lynch syndrome requires that high-risk individuals are targeted for mutation testing. Approximately 10 % of patients with CRC have an affected first-degree relative, and 2 % have two affected first-degree relatives. However, the frequency of families that fulfill the “Amsterdam Criteria” (AC) for Lynch syndrome is less than 5 % of all cases (Stephenson et al. 1991; Peel et al. 1997). The Bethesda criteria are more sensitive but less specific (Umar et al. 2004). Most population-based mutation analysis studies start with individuals with microsatellite unstable cancers and then look for germline mutations in these cases only: these studies have found that between 1 and 4 % of all CRC is due to germline mutations in either MLH1 or MSH2 (Peel et al. 2000; Salovaara et al. 2000) (Table 5.5). Some patients with early- onset CRC with no family history still have germline mutations in the mismatch repair genes (Dunlop et al. 1997), and these may account for a substantial proportion of CRC in patients diagnosed below 35 years, but only a minority of older cases. Persons with CRC caused by germline MSH6 mutations often do not fulfill either AC1 or AC2 (Sjursen et al. 2010).
Table 1 Frequency of germline MLH1 and MSH2 mutation carriers in early-onset CRC
|MLH1 mutation carriers||MSH2 mutation carriers|
|Age range (years)||Number of index cases||Number||%||Number||%|
In families not fulfilling the Amsterdam or Bethesda criteria for Lynch syndrome (Tables 11.4, 11.5, 11.6), pathological Tumours analysis by immunohistochemical stains (IHC) gives a high specificity for the diagnosis of Lynch syndrome and indicates the gene which may be mutated in the germline, although MSH2 and MSH6 staining may both be lost in Tumours in people with germline mutations in MSH2. Measurement of MSI in Tumours is less specific because approximately 15 % of all CRCs may be MSI positive and does not indicate which gene is likely to be involved in the germline. Some cases, especially older women with MSI-high cancers, may have somatic MLH1 inactivation. The specificity of MSI for Lynch syndrome is increased considerably if two Tumours from the same individual demonstrate MSI positivity (Vasen 2007; Lubbe et al. 2009; EGAPP Recommendation Statement 2009; Tresallet et al. 2012; Parsons et al. 2012).
In terms of identifying individuals at risk, it should be remembered that some patients with early-onset CRC without a family history of Lynch syndrome have germline mutations in mismatch repair genes (Dunlop et al. 1997; Hampel et al. 2005). If cases of CRC unselected for family history are studied for mismatch repair gene mutation analysis, the mutation detection rate is dependent on the age at diagnosis of CRC in the case (Mitchell et al. 2002; see Table 1). Familial clustering of CRC that does not fulfill the AC for a diagnosis Lynch syndrome is less likely to be due to germline MLH1 or MSH2 mutations. In early studies, only about 8 % of families not meeting the rather stringent AC1 had detectable mutations although some may have deletions or other gene rearrangements (Wijnen et al. 1997). This may be an underestimate, and detailed searches can find mutations in many more, as long as MSI-H is present in the CRC of the subject of the mutation analysis (Wagner et al. 2003), and would be considered to be very likely if MSH2 protein is absent, even if the family clustering in no way fulfills AC. In this context, looking for BRAF mutations (Davies et al. 2002) in the colorectal Tumours may be a good step following MSI and IHC, as among MSI-H cancers that do not express MLH1, the presence of a BRAF mutation makes a germline MLH1 mutation much less likely. Interestingly, loss of expression of MSH2 is rarely if ever associated with BRAF mutations (Wang et al. 2003; Koinuma et al. 2004; Domingo et al. 2004). More recent publications have presented improved algorithms for testing Tumours in newly diagnosed cases of CRC, stratified by age at diagnosis and family history, and in general MSI analysis is performed initially, as it is more sensitive but less specific than IHC, and then IHC performed to indicate the most likely gene involved. IHC can be the first step in cases conforming to the Bethesda criteria for Lynch syndrome (Lubbe et al. 2009; Halbert et al. 2004; Hampel 2010; Hall 2010; Rodriguez-Bigas et al. 1997; Parsons et al. 2012; Steinhagen et al. 2012; Tresallet et al. 2012).