TUMOURS OF THE COLON AND RECTUM
Malignant disease of the large bowel is one of the commonest causes of cancer death, with an incidence of 32 per 100,000 population in the UK, but a much lower incidence in parts of Africa (2.5 per 100,000 in Nigeria) and in Asia. Almost all (98 %) large bowel cancers are adenocarcinomas. Adenomas of the colon are thought to have the potential to develop into malignancy and coexist in 75 % of cases in which more than one primary carcinoma is
present in the colon; one or more adenomas are present in about a third of cases of colon carcinoma. It is thus probable that most carcinomas develop from adenomas (Morson 1966). The incidence of solitary colonic polyps in the general population is age-related, reaching 34 % in the sixth decade and 75 % in those over the age of 75 years (Lanspa et al. 1990; Dunlop 2002; Fletcher 2008). It has been calculated that the risk of invasive cancer in a single polyp is approximately 0.25 % per year, but for larger and/or villous polyps, the risks are up to 50 times higher (Eide 1986), indicating that screening guidelines need to be modified by the pathological findings at each colonoscopy. Inflammation is probably also a predisposing factor for bowel cancer, since ulcerative colitis and Crohn’s disease are associated with an increased risk of colon carcinoma (Judge et al. 2002). The risk of colorectal cancer in these conditions is similar for a similar extent and duration of colonic involvement, and surveillance guidelines have been developed for individuals with inflammatory bowel disease (Eaden and Mayberry 2002; Cairns et al. 2010). Individuals with acromegaly also have an increased prevalence of colorectal adenomas and cancers, RR 7.4, prevalence 3.7 % overall (Jenkins and Besser 2001).
Genetic factors are important in the pathogenesis of cancer, and a comprehensive family history should be part of the assessment of all patients with CRC. Early age at diagnosis in the affected relative is an important predictor of risk. Empiric risk estimates based on the number of relatives affected by (and age at diagnosis of) CRC are given in Table 1 and indicate that risk increases with earlier age at diagnosis and increased numbers of affected relatives. Twin studies indicate that 35 % of CRCs are partly due to inherited factors (Lichtenstein et al. 2000). Identification of individuals at increased genetic risk of colon cancer is important because screening can be offered to high-risk individuals, and evidence is accumulating that this is likely to be effective in reducing morbidity and mortality from CRC.
Table 1 Family history and risk of death from colon cancer: empiric risk estimates for counseling
|Affected relatives||Relative risk of CRC||Lifetime risk of CRC|
|General population||1 in 50|
|One first-degree relative||×3 (OR 1.8)||1 in 17|
|One first degree aged <45 years||×5 (OR 3.7)||1 in 10|
|One first and one second degree||1 in 12|
|Both parents||1 in 8.5|
|Two first-degree relatives||(OR 5.7)||1 in 6|
|Three first-degree relatives||1 in 3|
Odds ratio (OR) figures from St John et al. (1993); lifetime risks from Houlston et al. (1990)
It is estimated that about 5 % of CRCs occur in individuals with a dominantly inherited predisposition. Genetic conditions associated with colonic polyps carry an increased risk of CRC (Table 2). These include familial adenomatous polyposis, juvenile polyposis, and Peutz–Jeghers syndrome. Famililal adenomatous polyposis is characterized by the development of colorectal adenomas, whereas the latter conditions are characterized by gastrointestinal hamartomas, which are potentially premalignant although adenomas have more malignant potential. A rare familial syndrome of familial giant hyperplastic polyposis coli has been described which predisposes to CRC (Jeevaratnam et al. 1996; Sheikholeslami et al. 2004). More recently, the importance of some types of hyperplastic polyps, particularly as precursors of right-sided microsatellite unstable colon cancers, has been stressed (Jass et al. 2002). Conditions predisposing to gastrointestinal polyps are listed above (Ngeow et al. 2013).
Germline PTEN mutations are also associated with hamartomatous gastrointestinal polyps, but unlike the genes listed above, mutation carriers (affected with Cowden or Ruvalcaba–Myhre–Smith syndromes) do not seem to have a greatly increased risk of CRC, but there may be some risk in specific families.
Familial adenomatous polyposis is the most common of these disorders but accounts for fewer than 1 % of all cases of colon cancer. A larger proportion of colon cancer (probably 2–3 %) is accounted for by Lynch syndrome (HNPCC) (Evans et al. 1997; Salovaara et al. 2000), an autosomal dominant inherited predisposition to CRC, in which bowel cancers occur with high frequency at an early age (on average two decades earlier than sporadic cases). Mutations in one of several genes involved in the repair of DNA mismatch errors (MLH1, MSH2, MSH6, and PMS2) have been found to cause Lynch syndrome, and microsatellite instability (MSI) is seen, where multiple allelic changes are demonstrable in Tumours DNA when compared to the constitutional DNA of the patient. The condition predisposes to early- onset Tumours with increased proportion of right-sided colon cancers (65 % versus 25 % in sporadic cases), and a risk of multiple primary CRC. The risk of extracolonic cancers is higher in individuals with MSH2 than MLH1 mutations (Vasen et al. 2001; Bonadona et al. 2011), and MSH6 mutations particularly predispose to endometrial cancer (Wijnen et al. 1999). These extracolonic malignancies comprise mainly endometrial, ovarian, pancreatic, gastric, and urinary tract cancers (Wijnen et al. 1999). Full details of Lynch syndrome are given on page 249. Muir–Torre syndrome is Lynch syndrome associated with sebaceous adenomas and other characteristic skin lesions, and they are allelic (i.e., due to mutations in the same gene), but Muir–Torre syndrome has generally been considered to be due to mutations in MSH2 rather than MLH1 (Lucci-Cordisco et al. 2003). Biallelic MMR mutation carriers are at significantly increased risk of very early-onset colorectal cancer (Herkert et al. 2011).
Germline mutations in MLH3 and EXO1 may account for a small proportion of CRC cases with MSI-positive Tumours (Wu et al. 2001; Niessen et al. 2009; Liu et al. 2003; Laiho et al. 2002; Brassett et al. 1996; Sutter et al. 2004). Clinically important mutations in these genes are exceptionally rare.
Germline deletions of the last exons of EPCAM, the gene directly upstream of MSH2, result in loss of MSH2 expression by methylation of the promoter, and thus testing for EPCAM deletions is essential if IHC shows loss of MSH2, and yet no coding sequence or MLPA-identified mutation is found in MSH2 (Kuiper et al. 2011). Interestingly, such mutations are associated with a lower risk of endometrial cancer (Ligtenberg et al. 2013; Charbonnier et al. 2002).
The estimated cumulative risks of colorectal cancer by 70 years of age in individuals with Lynch syndrome have recently been reported as 41 % (95 % confidence intervals [CI], 25–70 %) for MLH1 mutation carriers, 48 % (95 % CI, 30–77 %) for MSH2, and 12 % (95 % CI, 8–22 %) for MSH6 (Hampel et al. 2006). For endometrial cancer, corresponding risks were 54 % (95 % CI, 20–80 %), 21 % (95 % CI, 8–77 %), and 16 % (95 % CI, 8–32 %). For ovarian cancer, they were 20 % (95 % CI, 1–65 %), 24 % (95 % CI, 3–52 %), and 1 % (95 % CI, 0–3 %). The estimated cumulative risks by age 40 years did not exceed 2 % (95 % CI, 0–7 %) for endometrial cancer or 1 % (95 % CI, 0–3 %) for ovarian cancer. The estimated lifetime risks for other Tumours types did not exceed 3 % with any of the gene mutations (Bonadona et al. 2011). These risks are lower than those published previously.
The term hereditary non-polyposis colon cancer (HNPCC) is something of a misnomer (Umar et al. 2004) since colonic adenomas do occur in patients with this condition; thus, the name Lynch syndrome has generally replaced HNPCC in common usage (Vasen et al. 1999). Although colonic adenomas are not more common in Lynch syndrome than in the general population, the adenomas progress more rapidly through the adenoma-carcinoma sequence than in normal individuals (Jass 1995a, b). This has implications for the frequency with which colonoscopic surveillance should be performed in these individuals. It is very rare for more than 50 polyps to develop in this condition, which distinguishes it from classical familial adenomatous polyposis in which there is a minimum of 100 colonic adenomas. There are, however, families with intermediate numbers of polyps that cannot easily be classified as Lynch syndrome or familial adenomatous polyposis (sometimes known as attenuated familial adenomatous polyposis, AFAP, or AAPC), and affected individuals in some such families may be demonstrated to have germline mutations in exons 3 or 4 of the APC gene (Spirio et al. 1993; Gismondi et al. 2002). Others may have MUTYH-associated polyposis, an autosomal recessive condition characterized by variable numbers of colonic adenomas and relatively early-onset CRC, with usually 30–100 adenomas (Halford et al. 2003). Very recently, rare mutations in POLE and POLD1 have been identified in families with multiple colorectal adenomas and associated carcinomas. The families may resemble AFAP families in some respects (Palles et al. 2013). A subgroup of Lynch syndrome-related CRC may be distinguished by the occurrence of flat adenomas in the right hemicolon (Lynch et al. 1993), but it seems likely that this is merely another variant of Lynch syndrome. However, serrated adenomas are common in HMPS (hereditary mixed polyposis syndrome), characterized by colonic adenomas and hamartomas and early-onset CRC, caused by a single germline deletion in GREM1 (Jaeger et al. 2012).
The prevalence of pathogenic APC and biallelic MUTYH mutations was 80 % and 2 %, respectively, among individuals with 1,000 or more adenomas, 56 % and 7 % among those with 100 to 999 adenomas, 10 % and 7 % among those with 20 to 99 adenomas, and 5 % and 4 % among those with 10 to 19 adenomas in a recent study, so APC mutations predominated in patients with classic polyposis, whereas prevalence of APC and MUTYH mutations was similar in attenuated polyposis (Grover et al. 2012).
A polymorphism in the APC gene, a T-A mutation germline mutation (predicted to result in a change from isoleucine to lysine position 1,307 of the protein; see familial adenomatous polyposis section), confers an approximately twofold increase in the risk of CRC. The mechanism for this at a molecular level appears to be that the polymorphism, which converts an AAATAAAA sequence to (A)8, predisposes to the development of somatic mutations in the APC gene (Laken et al. 1997). This mutation is common (about 6 %) in individuals of Ashkenazi Jewish origin, but is very rare in other ethnic groups. The relative risk or odds ratio for CRC in association with this mutation is about 1.5–1.8 (Gryfe et al. 1999), which is generally thought to be insufficient to warrant colonoscopic surveillance. The mutation is at least 2,200 years old (Niell et al. 2003) and has probably become frequent in this population by genetic drift.
There is possibly a slightly increased relative risk of colon cancer in individuals with germline BRCA1 mutations, but this is accompanied by a decrease in risk for rectal cancer, and this increased risk has not been reported in men (Thompson and Easton 2002). Overall, the risks are probably not clinically important. In support of the absence of an effect, there is no
apparent increased risk for CRC for Ashkenazi Jewish carriers of BRCA1:185delAG or BRCA1:5382inC (Ford et al. 1995; Struewing et al. 1997; Kirchhoff et al. 2004; Niell et al. 2004).
NAT2 (which alters the ability to acetylate N-acetyl transferase) and other metabolic gene variants may alter susceptibility to colonic adenomas and cancer in the general population and alter polyp density in FAP (Crabtree et al. 2002; Jass 2004), but are unlikely to cause a sufficiently increased CRC risk to be clinically important on their own. It may be that multiple variants in different genes including MLH1, MSH2, APC, AXIN1, and CTNNB1 (beta catenin) can contribute to susceptibility to colonic adenomas and cancer (Fearnhead et al. 2004; Lammi et al. 2004), but none of these loci are clinically useful.