FAMILIAL ADENOMATOUS POLYPOSIS
This is the most common of the hereditary polyposis syndromes, with a prevalence of about 1 in 8,000. It is inherited as an autosomal dominant trait with high penetrance but variable age at onset, and the mutation rate is high (15–20 % of cases are considered to be new mutations, although this may be an overestimate, and some single generation cases may be due to biallelic germline mutations in MUTYH) (Sieber et al. 2003a, b). Some sporadic cases (about 15 %) are mosaic for the APC mutation. Polyps usually develop in the teens, and penetrance is almost complete by age of 40 years in classical cases (Fig. 1). Progression to malignancy is inevitable, and colorectal carcinoma develops in untreated cases by the fourth to fifth decade, or even in childhood, 20–30 years earlier than in nonfamilial colon cancer.
Histologically, single crypt adenomas are a characteristic feature. Polyps also occur elsewhere in the GI tract. Gastric polyps in FAP are of two types: benign hyperplastic fundic gland polyps occur in most patients, and adenomas may also occur, usually in the pyloric region of the stomach, but at a much lower frequency. Gastric cancer may develop, even in young patients. Adenomatous duodenal polyps occur in most patients with FAP (over 80 %), are most numerous around the ampulla of Vater, and are associated with a significant risk of malignant transformation. The severity of duodenal polyposis is assessed using the Spigelman scale. Duodenal cancer is now the leading cause of death in this condition if colorectal cancer is prevented and occurs in about 5 % of cases (Spigelman et al. 1995; Brosens et al. 2005; Groves et al. 2002). Carcinoma of the gallbladder and bile ducts may also occur.
Fig. 1 Age-related carrier risk (larger squares) for relatives at 50 % prior risk of FAP, that is an affected parent, and a negative bowel examination at that age. The age-related risk was derived from the age-related penetrance data (smaller squares) (Adapted from Burn et al. 1991, with permission from J. Burn)
Extra-intestinal lesions develop in most patients with FAP and may be apparent before the bowel lesions. Epidermoid cysts may occur in two-thirds of patients, and although they can develop anywhere on the body, they are most noticeable on the scalp. These are very rare in normal children before puberty. Osteomas of the mandible may be detected in more than 90 % of patients using orthopantomograms and are uncommon in the general population (4 %). A third of patients may have impacted teeth, and dentigerous cysts and supernumerary and unerupted teeth may occur.
Exostoses may develop in the skull, digits, and long bones, and cortical thickening. The most common extraintestinal manifestation of FAP is multiple areas of retinal pigmentation, called congenital hypertrophy of the retinal pigment epithelium (CHRPE). These are found in about three-quarters of affected individuals. These are discrete, darkly pigmented, rounded lesions, 50–200 mm in diameter, and may have depigmentation around them (Fig. 2). Smaller, solitary, unilateral lesions may be seen in normal people, but it is rare for normal individuals to have more than three lesions. In FAP four or more CHRPEs are often present, characteristically large, oval, pigmented lesions with surrounding halo (type A lesions) which are specific for FAP; CHRPEs in controls are usually small dots (type B) (Olschwang et al. 1993). The larger lesions are probably congenital and appear to be choristomas of myelinated axons; the small lesions may show enlarged retinal pigment epithelial cells with increased pigment (Parker et al. 1990). There are interfamilial differences in predisposition to CHRPE in FAP, related to the position of the mutation in the APC gene, mutations distal to exon 9 being more prone to be associated with such lesions (Bunyan et al. 1995). However, there are exceptions to this (Pack et al. 1996), and there is intrafamilial variability in numbers of CHRPEs (Hodgson et al. 1994). Nevertheless, it has been estimated that in families in which CHRPEs appear to be a feature of the disease, the finding of fewer than three lesions in an individual at 50 % risk reduces the carrier risk, whereas the presence of more than three lesions conveys a very high probability that the individual is affected (Burn et al. 1991).
Fig. 2 FAP: (a) pigmented CHRPE, (b) depigmented CHRPE, and (c) florid colonic polyposis (Reproduced with permission from Cambridge University Press)
Frequency of Complications in FAP:
Congenital hypertrophy of the retinal pigment epithelium (70–80 %) Thyroid cancer (2–3 %)
Epidermoid cysts (50 %) Brain Tumours (1 %) Osteoma (50–90 %)
Hepatoblastoma (1 %)
Desmoid Tumours (10–15 %) Supernumerary teeth (11–27 %) Adrenal gland adenomas (7–13 %)
(Adapted from Vasen et al. 2009)
Desmoid disease is common in FAP, occurring in about 10–15 % of cases. It is nearly twice as common in females as in males and occurs at an earlier age in females (Brosens et al. 2005). Desmoids are more common in some families than in others, and this is related to the position of the germline APC mutation (mutations beyond codon 1444 being more likely to cause the disease). Histologically, they are composed of very vascular, fibrous tissue and may be diffuse or encapsulated. They occur predominantly in the small bowel mesentery, peritoneum, or abdominal wall, although they occasionally develop at extra-abdominal sites and often develop after a surgical procedure or pregnancy; 10 % resolve; 50 % may remain stable for prolonged periods; 30 % fluctuate, and 10 % grow rapidly. Desmoid Tumours do not metastasize but do infiltrate locally and can cause major morbidity and death, being one of the most common causes of death in FAP. They tend to recur after surgery, so treatment with nonsteroidal anti-inflammatory drugs (sulindac), antiestrogens, or, in resistant cases, cytotoxic chemotherapy or radiotherapy with computerized tomography monitoring is preferable.
Extraintestinal cancers associated with FAP include papillary carcinoma of the thyroid, brain Tumours (medulloblastoma, astrocytoma), and hepatoblastoma. Papillary carcinoma of the thyroid, characteristically the cribriform variant, appears to occur at increased frequency in young women (under 35 years of age) with FAP and in individuals with specific mutations in the APC gene, particularly at codon 1061 (Fenton et al. 2001). Brain Tumours (especially astrocytomas) are rare overall, but the relative risk is high in individuals with FAP: 23 to age of 29 years for all brain Tumours (7 for all ages) and 99 to age of 29 years for cerebellar medulloblastomas. Some cases of Turcot syndrome are a variant of FAP. A number of cases of hepatoblastoma have been described in children with FAP, and although the absolute risk is small, the relative risk is high (<500), and screening might be considered in high-risk families, although not usually performed (Hirschman et al. 2005).
Classically, the diagnosis of FAP is made clinically by the finding of
more than 100 adenomatous polyps in the colon and rectum, with histological evidence of single crypt adenomas. However, there is a milder form of the disease, attenuated FAP (AFAP), characterized by the presence of fewer adenomas and later onset of disease, which is seen in about 8 % of cases. The differential diagnosis is from other causes of intestinal polyposis, most importantly MUTYH-associated polyposis and other causes of inherited bowel cancer, such as Lynch syndrome and individuals with germline mutations in POLD1 and POLE (Palles et al. 2013). Individuals with AFAP usually have fewer than 100 adenomas, and such a phenotype may be caused by mutations in the first 4 exons of the APC gene (Olschwang et al. 1993; Knudsen et al. 2010) or by biallelic mutations in MUTYH. Biallelic germline MUTYH mutations have been assessed as causing 30 % of cases of >15 and <100 and 8 % classical polyposis with no APC mutation (Sieber et al. 2003a, b).
The FAP gene (APC) was localized to chromosome 5q21 following the report of a mentally retarded man who had Gardner syndrome and a constitutional deletion of chromosome 5q (Herera et al. 1986).
Characterization of the mutation in the APC gene in an affected individual in the family (possible in >90 % cases) allows predictive testing to be made available to at-risk relatives. The gene is large and has 15 exons, of which exon 15 is the largest. Most mutations in FAP patients are frameshift (2/3) or nonsense (1/3) mutations which result in the production of a truncated protein (Nagase and Nakamura 1993). Many different mutations have been described, but there are common ones, at codons 1061, 1450, and 1309 in exon 15. The APC gene product functions as a Tumours suppressor with subcellular location and interaction with catenins. It contains a number of coiled-coil heptad repeats at the 5′ end that promote oligomerization; the central part of the gene contains beta-catenin-binding domains, involved in cell–cell interaction, and Armadillo repeats, and the 3′ end contains tubulin- binding domains, with properties of binding to microtubules (Ilyas and Tomlinson 1997; Fodde et al. 2001).
Phenotype-genotype correlations are apparent (Stormorken et al. 2007), with the common mutations (1309 and 1061) being associated with a severe phenotype (Nugent et al. 1994), mutations before exon 9 usually being associated with a lack of CHRPE, and mutations in the first 6 exons being associated with a more variable and often milder phenotype, including “attenuated FAP” (Foulkes 1995). APC mutations found in “attenuated FAP” most commonly occur in exon 4, but many individuals with this “attenuated” form do not appear to have APC mutations; some may have MUTYH- associated polyposis. The observation that very short variant proteins may result in a less severe phenotype tends to support the “dominant negative” theory, but large deletions of the gene may be found in individuals with a severe phenotype, so that other factors including the loss of the mild APC mutant allele (Spirio et al. 1998) or the effects of other polymorphic alleles such at the NAT1 and NAT2 genes (Crabtree et al. 2004) may also be involved in the pathogenesis of the disease process.
Usually, there is reasonable consistency with regard to severity of adenomatous disease in different individuals with FAP within a single family, but some exceptions to this have been reported, with very variable age at onset of polyps in different cases, some presenting decades later than their affected relatives (Evans et al. 1993). The frequency of extracolonic features in FAP is associated with mutations in specific regions of the APC gene (Fig. 3).
Fig. 3 The APC gene and the mutation spectrum The frequency of extracolonic features in FAP is associated with mutations in specific regions of the APC gene (With permission from Ian Tomlinson) AAPC attenuated FAP, T thyroid cancer.
The phenotypic effects of a germline APC mutation may be affected by FAP modifier genes which include MOM in mice and genes encoding DNA methyltransferase and COX-2 (Fearon 1997; Crabtree et al. 2002, 2003).
A polymorphism in the APC gene was detected in an individual with a family history of colorectal cancer but not polyposis. Subsequent analysis suggested that this germline T–A mutation, predicted to result in a change from isoleucine to lysine position 1307 of the protein, predisposed to the development of somatic mutations of the APC gene and that gene carriers were therefore predisposed to colorectal cancer without florid polyposis, with a twofold increase in risk of bowel cancer in mutation carriers (Rozen et al. 2002). 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. This mutation is almost entirely restricted to the Ashkenazim (about 6 % of the Ashkenazi population probably carry this variant) (Laken et al. 1997), and there has been much debate about whether population screening for this mutation should be advocated in Ashkenazim, with colonoscopic surveillance for gene carriers. Currently the increase in risk is not thought to be sufficient for such a measure to be appropriate. Other germline variants such as E1317Q may be associated with an attenuated phenotype (Lamlum et al. 2000), but the evidence for this is less clear (Hahnloser et al. 2003), and recent data suggest that there is no increased colorectal neoplasia risk in carriers of this variant (Theodoratou et al. 2008).
Surveillance of the colorectum in patients at risk of FAP should begin before the age of 20 years; the risk of developing colorectal cancer before this age is very low, but some affected individuals do develop severe polyposis before the age of 10 years, so endoscopic surveillance should be commenced between 11 and 15 years, earlier if symptoms related to colonic problems occur beforehand (Vasen et al. 2008). Flexible sigmoidoscopy is probably sufficient before the age of 20 years as the polyps almost invariably develop in the distal colon initially, but colonoscopy should be performed annually from 20 years until florid polyposis develops.
Patients with FAP should be offered total colectomy with ileorectal anastomosis or proctocolectomy with restorative ileoanal anastomosis once colonic polyps have developed (Church 2006). The risk of colorectal cancer is significant once polyps have begun to develop, irrespective of polyp density (Phillips and Spigelman 1996). After surgery conserving the rectal stump, subsequent management should include lifelong surveillance of the rectal stump by yearly sigmoidoscopy as there is a 3.5 % risk of colorectal cancer in it after 5 years, rising to 10 % at 10 years. Proctocolectomy removes this risk and may be done after 50 years of age. In addition, since upper GI cancer is reported to occur in 5 % of patients, with an estimated prevalence of duodenal dysplasia of up to 90 %, initial surveillance by means of upper GI endoscopy with a side-viewing endoscope to allow detailed examination of the papilla, or a forward-viewing endoscope in early Spigelman stage cases, recommended to start at about 25 years of age, to give a baseline for subsequent follow-up (Arvanitis et al. 1990; Debinski et al. 1995; Theodoratu et al. 2008) (Fig. 11.4). If adenomas are detected, they should be biopsied; larger adenomas may be removed and further surveillance continued with 1–5-yearly duodenoscopy, depending on the Spigelman severity score (Dunlop 2002; Vasen et al. 2008). Duodenal cancer in FAP patients below 30 years of age is extremely rare, and the overall risk of duodenal cancer is small (5 %), but with a Spigelman stage lll–lV, the risk of cancer is 7–36 %.
Palpation of the thyroid gland and possibly ultrasound in young women with FAP have been suggested, but the rarity of death from thyroid cancer in this disease makes this of questionable benefit unless there is a family history of thyroid cancer, and this is not usually performed.
Relatives of affected individuals should be ascertained with the help of a genetic register, and those at risk of inheriting the disease should be offered screening and genetic testing if possible. Screening of at-risk relatives is commenced between the ages of 11 and 13 years by annual sigmoidoscopy because the rectum is involved by adenomas at an early stage, and polyps rarely develop before 11 years of age. However, if symptoms arise, it may be necessary to arrange endoscopy earlier because there are case reports of colonic polyps developing in young children (Distante et al. 1996). Where the APC mutation has been characterized, predictive testing may be offered to children in the family, and once a child has been found to carry the APC mutation in the family, it is appropriate to consider colonoscopy in their early teens to establish the extent of polyposis. In an at-risk child where a predictive genetic test is not available, if polyps are found, these are biopsied (to confirm they are adenomatous and to exclude malignancy) and colonoscopy is arranged. Although small numbers of polyps can be managed by endoscopic resection, because of the inevitability of florid polyposis developing and the risk of malignant change, definitive surgery is usually arranged at this stage. In the absence of positive findings, sigmoidoscopy is continued annually, with colonoscopies with dye spray after 20 years of age, to exclude more proximal polyposis. This is continued to at least the age of 40 years, by which time the risk to an individual with an affected parent has fallen below 1 % (see Fig. 11.1). However, since very variable age at onset has been described in some families, it is advisable to continue screening well beyond this age (Evans et al. 1993). Endoscopic surveillance should also be offered to individuals with attenuated polyposis and their close relatives (Heiskanen et al. 2000), with colonoscopic surveillance 2-yearly from age 18–20 years, and upper GI surveillance as for classical FAP (Debinsky et al. 1995; Vasen et al. 2008). If colectomy is necessary, ileorectal anastomosis should be offered; since the risk of duodenal cancer is also present, upper GI surveillance should also be recommended as above.
Predictive testing based on mutation analysis may be available, and individuals who have not inherited the mutation can be discharged from follow-up. A more cautious approach to management is indicated where the predictive test was based upon linkage analysis, surveillance being indicated until the estimated chance that the individual is affected is calculated as below 0.1 % on Bayesian risk analysis combining clinical and molecular data. Prenatal diagnosis is available by mutation or linkage analysis in informative families, although the uptake rate is low. The acceptability of such testing in a disorder such as FAP is a very personal matter, and the decision about whether to opt for such testing should be one made by the family after nondirective genetic counseling. Many parents wish their children to be tested for FAP at a very young age, but there are arguments for delaying such testing until clinical screening would normally be instituted, since, until this time, clinical management would not be affected (Hyer and Fell 2001).
The effectiveness of intervention strategies, such as the administration of aspirin, are being evaluated (Mathers et al. 2003), but are not sufficient to obviate the need for colectomy in affected individuals once polyps have developed. In FAP, the CAPP1 study showed that aspirin had a modest effect on polyp progression and significantly reduced the size of the largest polyps in 200 FAP patients, but no effect on polyp number (Burn et al. 2011). There has been some evidence that the administration of oral sulindac (a nonsteroidal anti-inflammatory drug) or cyclo-oxygenase-2 (COX-2) inhibitors in general may cause regression of duodenal adenomas (Phillips et al. 2002). Carcinoma of the rectal stump may occur (in about 10 % of cases) after surgical removal of the colon and ileorectal anastomosis, and COX-2 inhibitors may also be effective in reducing polyp recurrence in the colorectum (Steinbach et al. 2000), but these have been shown to increase cardiovascular morbidity in the VICTOR and APPROVE research studies and are therefore not recommended – but there is a delayed effect on colorectal cancer incidence, as seen in the CAPP2 study. Fish oils have long been thought to be protective against GI cancers, and a special formulation of eicosapentaenoic acid (EPA) was able to reduce polyp number and size (West et al. 2010).