July 2005

Section 5 Gastrointestinal Tract and Abdomen

Finally, there are a few other, less common, inherited hamartomatous polyposis syndromes, such as Cowden disease and Ruvalcaba-Myhre-Smith syndrome. At present, these syndromes appear to be associated with an exceedingly low risk of CRC; accordingly, prophylactic surgery is not indicated.¹

FAP is caused by mutations in the tumor suppressor gene APC, located at 5q21. Nearly 80% of FAP patients belong to known FAP kindreds; 10% to 30% have new mutations.¹ More than 300 distinct mutations have been identified within the APC gene locus in persons manifesting the FAP phenotype. More than half of the known germline mutations associated with the classic FAP phenotype are concentrated in the 5Á region of exon 15.¹ Genotype-phenotype correlative studies have revealed a wide range of phenotypic heterogeneity, ranging from the relatively mild presentation associated with attenuated FAP, which is caused by mutations in the 3Á and 5Á ends of the APC gene,² to the severe presentation associated with mutations downstream from codon 1250, particularly those in codon 1309. It has been reported that as many as 7.5% of patients with a classic FAP phenotype and no demonstrable APC mutation may have biallelic germline mutations in the base excision repair gene MYH.³

FAP, which accounts for less than 1% of the annual CRC burden, is characterized by the presence of more than 100 adenomatous polyps of the colorectum, virtually 100% penetrance, and a nearly 100% risk of CRC by the age of 40 if prophylactic colectomy is not performed.^1,4 Extracolonic manifestations are common and include desmoid tumors, osteomas, odontomas, sebaceous and epidermoid cysts, congenital hypertrophy of the retinal pigment epithelium, and periampullary neoplasms.¹

The polyps, which develop by the age of 20 years in 75% of cases, are typically less than 1 cm in size. In severe FAP, they may carpet the entire surface of the colorectal epithelium or, alternatively, may spare portions of the epithelial lining (e.g., the rectum). Adenomas may be either pedunculated or sessile and may have tubular, villous, or tubulovillous histology. Microscopic evaluation may reveal innumerable microadenomas within grossly normal-appearing colorectal mucosa. Foci of carcinoma in situ and invasive carcinoma may be found within larger polyps, and the incidence of invasive cancer is proportional to the extent of polyposis. Unlike CRC in the setting of HNPCC, CRC in the setting of FAP is more commonly located on the left side.¹

Screening and Surveillance

Screening (genetic testing or annual or biennial flexible sigmoidoscopy) for at-risk family members should begin around puberty (i.e., at 10 to 12 years of age) [see Table 1]. In families with a demonstrated APC mutation, informative genetic testing can be carried out with the protein truncation test [see Table 2]. This test, which detects foreshortened proteins resulting from truncating APC mutations, is approximately 80% sensitive⁵; however, the test results are commonly misinterpreted, even by physicians.⁶ Patients with normal protein truncation test results and a previously identified mutation in the family may be discharged from further screening with a nearly 100% certainty that the mutation is absent, but they should still undergo CRC screening starting at the age of 50 years, as is recommended for average-risk persons. When an APC mutation has not previously been identified in the family of an affected person, the patient should be tested first to identify the causative mutation. In families in which the protein truncation test fails to provide conclusive information on carrier status, at-risk individuals should continue with the recommended endoscopic surveillance program. Other options for detecting APC mutations include linkage analysis, single-stranded confirmation polymorphism, and direct sequence analysis.¹

Genetic counseling is an essential component of the evaluation of patients for FAP. Patients who have a positive genotype or who have adenomatous polyps on sigmoidoscopy should undergo full colonoscopy to establish the extent of polyposis.

A number of nonsteroidal anti-inflammatory drugs, including sulindac, celecoxib, and the sulindac metabolite exisulind, have been shown to reduce the number and size of polyps in FAP patients.^7–10 However, long-term use of chemopreventive agents for primary treatment of FAP is not recommended.¹¹ In a randomized, placebo-controlled, double-blind study of genotype-positive, phenotype-negative patients, the use of sulindac had no effect on the subsequent development of colorectal polyposis.¹² Furthermore, the development of rectal cancer has been reported in patients whose rectal polyps were effectively controlled with sulindac.¹⁰ Finally, these medications necessitate continued compliance⁹ and may be associated with significant side effects. Chemopreventive agents may be useful for reducing polyp load and facilitating endoscopic management of polyps in patients who have an ileal pouch or in patients who have an iliorectal anastomosis, are at high risk for polyp development, and refuse proctectomy. In such cases, however, it is still necessary to perform careful surveillance of the residual rectum or the ileoanal pouch every 6 months.¹¹

Surgical Therapy

The timing of surgical treatment depends to some degree on the extent of polyposis, in that the risk of CRC is partially dependent on the number of polyps present.¹³ Patients with mild polyposis (and thus a lower cancer risk) can undergo surgery in their midteens.¹¹ Practically speaking, the best time is usually the summer between high school and college. Patients with severe polyposis, dysplasia, adenomas larger than 5 mm, and significant symptoms should undergo surgery as soon after diagnosis as is practical.¹¹

There are three basic surgical options for treating FAP: (1) total proctocolectomy (TPC) with permanent ileostomy, (2) total abdominal colectomy with ileorectal anastomosis (IRA), and (3) proctocolectomy with ileal pouch—anal anastomosis (IPAA) [see 5:33 Procedures for Ulcerative Colitis]. The optimal procedure for a given patient is determined on the basis of a number of factors, including disease characteristics, differences in postoperative functional outcome, preoperative anal sphincter status, and patient preference.

TPC TPC with permanent ileostomy is rarely chosen as a primary procedure. More commonly, it is considered as an option for patients in whom a proctectomy is required but an IPAA is contraindicated (e.g., those with rectal tumors involving the sphincters or the levator complex or those with poor baseline sphincter function) or for patients in whom an IPAA is not technically feasible (e.g., those with desmoid disease and foreshortening of the small bowel mesentery). Occasionally, however, TPC is chosen as a primary procedure in patients whose lifestyle would be compromised by frequent bowel movements.

IPAA versus IRA The choice between IPAA and IRA is generally more challenging. The main considerations to be taken into account are the risk of rectal cancer development if the rectum is left in situ and the differences in functional outcome (and associated quality of life) between procedures.

It has been estimated that the risk of rectal cancer after IRA may be as high as 4% to 8% at 10 years and 26% to 32% at 25 years.^14,15 The true risk, however, may be somewhat lower. Most of the studies from which these figures were derived were completed before IPAA became available; thus, patients and physicians might have been more likely to choose IRA even in the setting of more extensive rectal disease, given that TPC and permanent ileostomy was the only other option at the time. The magnitude of risk in an individual patient is related to the overall extent of colorectal polyposis. IRA may be an option for patients with fewer than 1,000 colorectal polyps (including those with attenuated FAP) and fewer than 20 rectal adenomas, because these patients appear to be at relatively low risk for rectal cancer.^11,13,16 Ideally, patients with severe rectal (> 20 adenomas) or colonic (> 1,000 adenomas) polyposis, an adenoma larger than 3 cm, or an adenoma with severe dysplasia should be treated with IPAA.^11,13

The risk of secondary rectal excision as a consequence of uncontrollable rectal polyposis or rectal cancer may be estimated on the basis of the specific location of the causative APC mutation.^15–17 In a study of 87 FAP patients with an identified APC mutation who underwent IRA, those with a mutation located downstream from codon 1250 had an approximately threefold higher incidence of secondary rectal resection than those with a mutation located upstream of codon 1250.¹⁴ Furthermore, patients with a mutation located between codons 1250 and 1464 had a 6.2-fold higher risk of rectal cancer than those with a mutation before codon 1250 or after codon 1464.¹⁵

The risk of polyp and cancer development after index surgery is not limited to patients undergoing IRA. In patients undergoing IPAA, the pouch-anal anastomosis may be either handsewn after complete anal mucosectomy or stapled to a 1 to 2 cm anal transition zone. Neoplasia may occur at the site of the anastomosis, and the incidence appears to be higher after stapled anastomosis (28% to 31%) than after mucosectomy and handsewn anastomosis (10% to 14%).^18,19 Function, however, may be better after stapled anastomosis.¹⁹ In the case of anal transition zone neoplasia after stapled anastomosis, transanal mucosectomy can often be performed, followed by advancement of the pouch to the dentate line. Of additional concern is the development of adenomatous polyps in the ileal pouch itself, which occurs in 35% to 42% of patients at 7 to 10 years.^20–22

With respect to postoperative bowel function and associated quality of life, IPAA has been associated with a higher frequency of both daytime and nocturnal bowel movements, a higher incidence of passive incontinence and incidental soiling, and higher postoperative morbidity than IRA.²³ Accordingly, some authors recommend IRA for patients with mild rectal polyposis. Other authors, however, have found the two approaches to be equivalent in terms of functional results²⁴ and quality of life²⁵ and therefore recommend IPAA for most patients because of the risk of rectal cancer associated with IRA.

Regardless of which procedure is performed, however, lifetime surveillance of the rectal remnant (after IRA) or the ileal pouch (after IPAA) is required.¹¹ Endoscopic surveillance of the bowel at intervals of 6 months to 1 year after index surgery is recommended.^5,11 After IRA, small (< 5 mm) adenomas may be safely observed, with biopsy performed to rule out severe dysplasia. If adenomas increase in number, the frequency of surveillance should be increased, and polyps larger than 5 mm should be removed. When fulguration and polypectomy are repeated over a period of many years, subsequent polypectomy may become difficult, rectal compliance may be reduced, and flat cancers may be hard to identify against a background of scar tissue. The development of severe dysplasia or a villous adenoma larger than 1 cm is an indication for proctectomy.¹¹

Extracolonic Disease

After total abdominal colectomy with IRA and regular surveillance, the risk of death appears to be three times higher for FAP patients than for an age- and sex-matched control population.²⁶ The main causes of death after IRA are desmoid disease and upper gastrointestinal malignancy.

Desmoid disease Desmoids are histologically benign tumors that arise from fibroaponeurotic tissue and occur in 12% to 17% of FAP patients.^11,27,28 Unlike those in the general population, desmoids in FAP patients tend to be intra-abdominal (up to 80% of cases) and mainly occur after abdominal surgical procedures.^27,28 Patients with APC mutations located between codons 1310 and 2011 are at increased risk for these tumors.²⁹ Desmoids often involve the small bowel mesentery (> 50% of cases),²⁸ making complete resection difficult or impossible, and they may also involve the ureters.²⁷ Not uncommonly, patients present with small bowel obstruction.^27,28 Morbidity after attempted resection, which often involves removal of a significant length of small bowel, is substantial. The recurrence rate after attempted resection is also high, and the recurrent disease is often more aggressive than the initial desmoid.^27,28

Intra-abdominal desmoid formation may be more common after IRA than after IPAA, and the disease may be more severe after IRA as well.^28,30 When desmoid tumors involve the small bowel mesentery, the mesentery may become foreshortened and thereby render IPAA impracticable, especially in patients undergoing a subsequent completion proctectomy after an initial IRA.¹¹ This possibility should be considered in making the choice between IRA and IPAA as the initial procedure for FAP.

Medical therapy. When desmoid tumors are clinically inert, they may be treated with sulindac.¹¹ Tamoxifen or other antiestrogens may be added for slow-growing or mildly symptomatic tumors.^11,31,32 More aggressive desmoid tumors may be treated with chemotherapy. Vinblastin and methotrexate achieve some degree of response in 40% to 50% of patients.³³ For more rapidly growing desmoids, antisarcoma agents, such as doxorubicin and dacarbazine, may be administered.^34,35 Radiation therapy may also be effective, but it can result in substantial small bowel morbidity.

Surgical therapy. Surgical treatment of intra-abdominal desmoid tumors should be reserved for small, well-defined lesions with clear margins.¹¹ When intra-abdominal desmoids involve the small bowel mesentery, they should be treated according to their initial presentation and rate of growth. In patients with desmoid lesions that are refractory to all medical treatment and call for surgical treatment with extensive small bowel resection, small bowel transplantation may be feasible in selected cases.³⁶

Periampullary neoplasms In approximately 80% to 90% of persons with FAP, duodenal adenomas, periampullary adenomas, or both will develop.³⁷ Of these patients, 14% to 50% will eventually exhibit advanced polyposis, and as many as 6% will eventually have invasive cancer.^1,38–42 Although the risk of periampullary or duodenal cancer in FAP patients is relatively low, it is still several hundred times higher than that in the general population. Among FAP patients, those with APC mutations between codons 976 and 1067 appear to have the highest incidence of duodenal adenoma.

Surveillance should begin with side-viewing esophagogastroduodenoscopy (EGD) and biopsy of suspicious polyps either at the age of 20 years or at the time of prophylactic colectomy, whichever is earlier.¹¹ The purpose of screening is not to remove all disease but to watch for the development of high-grade dysplasia. Small, tubular adenomas without high-grade dysplasia may be biopsied and observed; adenomas that are larger than 1 cm or that exhibit high-grade dysplasia, villous changes, or ulceration should be removed. Surgical options include endoscopic removal and transduodenal excision, but both approaches have drawbacks: endoscopic ablation generally requires multiple settings,³⁸ and recurrence is high after either procedure.^38,43 Endoscopic ablation is a reasonable initial approach for most patients without invasive cancer and is an attractive alternative for patients who are unfit for duodenal resection. For patients with persistent or recurrent high-grade dysplasia in the papilla or duodenal adenomas and for patients with Spigelman stage IV disease, pancreas-preserving duodenectomy or pancreaticoduodenectomy is recommended.¹¹ The results reported for duodenal resection in patients with premalignant lesions are encouraging, with good local control and low morbidity.^38,44,45 Duodenectomy also greatly reduces the need for upper GI surveillance.

HNPCC, which accounts for 5% to 7% of CRCs, results from a mutation in one of the DNA mismatch repair (MMR) genes (MLH1, MSH2, MSH6, PMS1, PMS2, MLH3, and MSH3).^46,47 Two genes (MLH1, MSH2) may be responsible for as many as 90% of causative germline MMR mutations. However, only 50% to 70% of patients meeting clinical criteria for HNPCC have an identifiable germline MMR mutation, which suggests that one or more unidentified genes may be involved. A significant percentage of cases may be attributable to large germline deletions that are difficult to detect by means of direct sequencing. It appears that genomic deletions may account for as many as 7% of HNPCC cases defined on the basis of clinical criteria.⁴⁸

HNPCC is characterized by early-onset CRC, a predominance of lesions proximal to the splenic flexure (60% to 70% of cases), benign and malignant extracolonic tumors, and a predilection for synchronous and metachronous colorectal tumors.⁴ Microsatellite instability (MSI), reflecting a deficiency in DNA repair secondary to a mutation in the MMR genes, is noted in approximately 80% to 90% of HNPCC-related tumors.⁴ The lifetime risk of CRC in HNPCC patients is approximately 80%.^11,49

Establishing a clinical diagnosis of HNPCC is much more challenging than establishing a clinical diagnosis of FAP, in that it requires a careful and detailed family history. The Amsterdam II criteria [see Table 3] require that there be three relatives (of which one must be a first-degree relative of the other two) with an HNPCC-related cancer (of the colorectum, the endometrium, the small bowel, the ureter, or the renal pelvis), that two or more successive generations be involved, and that at least one relative have a CRC diagnosed before the age of 50.^50,51 Finally, FAP should be excluded. CRC occurs in 78% to 80% of MMR mutation—positive patients at a mean age of 46 years.^1,49,52 Endometrial cancer occurs in 43%, gastric cancer in 19%, urinary tract cancer in 18%, and ovarian cancer in 9%.⁵³

Adenomas in HNPCC patients show high-grade dysplasia and villous changes more frequently than adenomas in sporadic CRC patients.¹ Adenomas may also appear at an earlier age and are often larger than those found in the general population. Other pathologic features reported to be more common in HNPCC-related cancers include a mucinous or poorly differentiated histology, a solid or cribriform growth pattern, signet-ring cell tumors, and the presence of tumor-infiltrating and peritumoral lymphocytes. HNPCC-related CRCs have also been shown to have a lower rate of lymph node involvement.⁵⁴

Screening and Surveillance

CRC patients who belong to known HNPCC kindreds, who have a pedigree suggestive of HNPCC, or who meet the Bethesda criteria [see Table 4]⁵⁵ should be offered screening by MSI testing. MSI evaluation will yield positive results (i.e., an MSI-high tumor) in 80% to 90% of patients belonging to families that meet the Amsterdam criteria. Patients with MSI-high tumors should undergo testing for germline MMR mutations (tests for MSH2 and MLH1 are available commercially [see Table 2]). If tumor tissue is not available, initial germline testing may be considered. As in FAP, a mutation in an affected individual must first be established for testing in at-risk individuals to be informative.⁵

Recommended surveillance for HNPCC includes colonoscopy, initially every 1 to 2 years beginning at the age of 20 to 25, then annually after the age of 40.⁵⁶ Given the increasing evidence of an accelerated adenoma-carcinoma sequence in HNPCC, annual colonoscopy should be strongly considered.⁴ Female patients should undergo annual transvaginal ultrasonography and measurement of CA125 levels starting at 25 to 35 years of age, as well as annual endometrial aspiration.⁵⁶ Annual EGD is recommended for patients belonging to kindreds with a history of gastric cancer. Finally, ultrasonography and urine cytology every 1 to 2 years may be considered to screen for urinary tract malignancy.

Although the development of CRC in persons with HNPCC is not a certainty, the 80% lifetime risk,¹ the 45% rate of metachronous tumors, and the possibility of an accelerated adenoma-carcinoma sequence⁴ mandate consideration of prophylactic surgical options. Patients who have HNPCC as defined by their genotype or the Amsterdam criteria [see Table 3] and who have a colon cancer or more than one advanced adenoma should be offered either (1) prophylactic total abdominal colectomy with IRA or (2) segmental colectomy with yearly colonoscopy [see 5:34 Segmental Colon Resection].^11,57,58 (The first option, however, is open only to patients with normal rectal and anal sphincter function.) Although the risk of metachronous colon cancers may be higher after partial colectomy than after total colectomy with IRA, intensive colonoscopic surveillance and polypectomy may minimize the number of metachronous cancers in the remaining colon.^52,59 Careful surveillance is also necessary after total colectomy and IRA, given that the risk of metachronous rectal cancer after total colectomy is approximately 12% at 10 to 12 years.⁶⁰

HNPCC patients with an index rectal cancer that is amenable to a sphincter-preserving resection should be offered either (1) total proctocolectomy with IPAA [see 5:32 Procedures for Diverticular Disease].^11,58 The rationale for total proctocolectomy is based on the 17% to 45% rate of metachronous cancer in the remaining colon associated with an index rectal cancer in HNPCC patients.⁶¹ The decision between the two procedures depends in part on the patient's willingness to undergo intensive surveillance of the retained proximal colon, as well as on the level of bowel function.

Mutation-positive patients with a normal colorectum may also be offered prophylactic colectomy in selected cases.^56,62 This approach is supported by the similarity of lifetime cancer risk between patients with germline APC mutations and those with MMR mutations, as well as by the observation that total abdominal colectomy with IRA yields less functional disturbance than the prophylactic procedure recommended for FAP (total proctocolectomy with IPAA).^56,62 An alternative strategy in these patients is to carry out colonoscopic surveillance at 1- to 3-year intervals. This strategy has proved to be cost-effective⁶³ and to reduce both the rate of CRC development and overall mortality.^52,64,65 There is a risk that CRC may develop in the intervals between colonoscopies^64,66; however, when the surveillance interval is shorter than 2 years, tumors tend to be found in their early stages and to be curable when found.^52,64

A study using a decision-analysis model suggested that prophylactic total abdominal colectomy at the age of 25 might offer a survival benefit of 1.8 years when compared with colonoscopic surveillance. The benefit of prophylactic colectomy decreased when surgery was delayed until later in life and became negligible when it was performed at the time of cancer development.⁶⁵ However, surveillance provided a greater benefit with respect to quality of life (measured in quality-adjusted life years).⁶⁵ On the basis of this evidence, some surgeons recommend that prophylactic colectomy be performed only in highly selected situations (e.g., when colonoscopic surveillance is not technically possible or when a patient refuses to undergo regular surveillance). Thus, the decision between prophylactic surgery and surveillance for gene-positive unaffected patients is based on many factors, including the penetrance of disease in a family, the age of cancer onset in family members, functional and quality-of-life considerations, and the likelihood of patient compliance with surveillance.

Extracolonic Disease

Management of extracolonic cancers in HNPCC patients is not yet well defined. Female patients with a family history of uterine cancer should be offered prophylactic total abdominal hysterectomy (TAH) if their childbearing is complete or if they are undergoing abdominal surgery for other conditions.¹¹ This recommendation is based on the high (43%) rate of endometrial cancer in mutation-positive persons,⁵³ particularly those with hMSH2 mutations, and on the inefficacy of screening in some studies.⁶⁷ Oophorectomy should be added to TAH because of the high (9%) incidence of ovarian cancer in HNPCC patients⁵³ and the frequent coexistence of endometrial cancer with ovarian cancer.⁶⁸ The optimal timing for prophylactic TAH is unclear; however, endometrial cancer has been reported in HNPCC patients before the age of 35. At present, it seems reasonable to begin surveillance at the age of 25 and delay prophylactic surgery until childbearing is complete.¹¹

Like FAP and HNPCC, PJS follows an autosomal dominant pattern of inheritance with variable penetrance. It is caused in part by mutations in the gene LKB1/STK11, which maps to the telomeric region of chromosome 19p13.3. This gene, which codes for a multifunctional serine-threonine kinase, is thought to function as a tumor suppressor gene.^69–72 Germline mutations in LKB1/STK11 can be demonstrated in 18% to 63% of PJS patients, which suggests the existence of additional PJS loci.^72–75 Genetic testing for PJS can be accomplished through direct sequencing of the LKB1/STK11 gene [see Table 2]; however, such testing is not widely available. In families with an established mutation, genetic testing of at-risk individuals is informative, with a reported accuracy of 95%.⁷⁶

PJS is a hereditary polyposis syndrome characterized by hamartomas of the GI tract, as well as by mucocutaneous melanin pigmentation. Hamartomatous polyps may occur throughout the GI tract but are most frequently found in the small intestine (90%). Other common sites of hamartomas in PJS are the large intestine (50%) and the stomach; less common sites are the renal pelvis, the bile ducts, the urinary bladder, the lungs, and the nasopharynx.^1,77,78 Mucocutaneous pigmentation generally appears during infancy. The perioral and buccal areas are involved in 95% of cases; the periorbital and facial areas, the genital region, and the acral areas (including the hands and feet) may be involved as well.¹ The average age of diagnosis of PJS is 22 years in men and 26 years in women.

In as many as 86% of cases, the initial presentation of PJS is small bowel obstruction secondary to intussusception of hamartomas. Other presentations include acute or chronic GI bleeding, biliary and gastric outlet obstruction, and anal protrusion of polyps. The diagnosis of PJS is established by the presence of histologically confirmed hamartomas of the GI tract plus two of the following three criteria: (1) small bowel polyposis, (2) mucocutaneous melanotic pigmentation, and (3) a family history of PJS [see Table 1].⁷⁹

Patients with PJS are at significantly increased risk for both intestinal and extraintestinal malignancies. A meta-analysis found that in comparison with the general population, PJS patients were at a relative risk of 15.2 for the development of any malignancy.⁸⁰ The relative risks for the development of specific cancers were as follows: small bowel, 520; gastric, 213; pancreatic, 132; colorectal, 84; esophageal, 57; ovarian, 27; lung, 17; endometrial, 16; and breast, 15. The cumulative risk for the development of any cancer between the ages of 15 and 64 was 93%.⁸⁰ Other cancers associated with PJS are cholangiocarcinomas, testicular neoplasms, and duodenal tumors.¹

Although the relative risk for the development of CRC was high in this study,⁸⁰ the reported magnitude of risk in the individual studies included in the meta-analysis varied considerably. Previous studies also reported a wide range of CRC incidences in these patients.¹ Thus, the true incidence of CRC in PJS patients remains unclear.

The polyps seen in PJS are hamartomas characterized by hypertrophy or hyperplasia of the smooth muscle of the muscularis mucosa. Smooth muscle extends into the superficial epithelial layer of the bowel wall in a treelike fashion (a process referred to as arborization). Epithelial cells may become entrapped within the muscle layer, and this 'pseudoinvasion' can be mistaken for malignant transformation. Therefore, to diagnose a malignancy in a PJS polyp, cellular atypia or an elevated mitotic rate must be documented.⁸¹ Sporadic PJS polyps do occur, generally secondary to somatic LKB1/STK11 mutations in one or both alleles, and are histologically identical to their hereditary counterparts. These sporadic polyps appear not to be associated with an increased risk of GI cancer.⁸²

Histologically, areas of cutaneous pigmentation reveal an increased number of melanocytes at the dermal-epidermal junction, with elevated melanin levels in the basal cells. These lesions do not appear to have any malignant potential.

Screening and Surveillance

Clinical screening of asymptomatic persons is facilitated by the appearance of perioral hyperpigmentation during early childhood. Once the diagnosis of PJS is made, patients generally enter a surveillance program. Recommended surveillance for GI disease includes annual serum hemoglobin measurement and EGD every 2 to 3 years, beginning between the ages of 10 and 25.^79,83–85 Contrast radiography is employed to examine the remainder of the small bowel, beginning at the age of 10 and repeated every 2 to 3 years.^79,85 The frequency of surveillance examinations may be modified in individual circumstances. Colonoscopic surveillance is also important, commencing between puberty and the age of 25 and repeated every 2 to 3 years.^83,85 Sigmoidoscopy should not be employed for surveillance, because the rectum may be spared in some patients with more proximal disease. Organ-specific surveillance for other associated malignancies should also be initiated in accordance with current high-risk recommendations.

Cyclooxygenase-2 (COX-2) is known to be overexpressed in the hamartomatous tissue of PJS patients, and there is a correlation between expression of the COX-2 protein and expression of the LKB1/STK11 protein in PJS polyps and cancers.^86,87 These findings suggest that COX-2 may be a potential target for chemoprevention of PJS.

Surgical Therapy

Indications for surgical management of PJS include the presence of polyps larger than 1.5 cm that cannot be removed endoscopically, incomplete removal of polyps with adenomatous changes, the development of polyp-associated complications (e.g., obstruction, intussusception, and bleeding), and the management of malignant disease.⁸⁸

Endoscopic polypectomy is generally employed as initial therapy when it is technically feasible. For some polyps, however, operative polypectomy performed through an enterotomy is required. Segmental resection should be avoided. In the context of a laparotomy, intraoperative endoscopy (either peroral or via an enterotomy) allows direct visualization of the remainder of the small bowel and endoscopic clearance of any synchronous polyps. This procedure significantly reduces the need for subsequent laparotomy. The St. Mark's Hospital group in London found that none of 25 patients who underwent enteroscopy during laparotomy required subsequent laparotomy within a 4-year period, whereas 17% of historical control patients who did not undergo intraoperative enteroscopy required repeat laparotomy within a 1-year period.⁸⁹

Laparoscopy-assisted polypectomy and laparoscopic management of small bowel intussusception are additional surgical options.

Given the risk of CRC development in PJS patients, careful colonoscopic surveillance is clearly warranted. However, the role of prophylactic colectomy in patients who are at risk or are mutation positive is unclear. Because the true risk of CRC in these patients is unknown and genetic testing for PJS is not widely available, no recommendations can be made at present regarding the role of prophylactic colectomy in the PJS population.⁸⁸

Initial evidence suggested that mutations in the PTEN gene were responsible for JPS⁹⁰; however, subsequent evidence implicated SMAD4/DPC4 at 18q21.1 as a more common cause, accounting for as many as 50% of familial cases.^91–93 Mutations in BMPR1A at 10q22-q23 have also been reported to cause JPS but display variable penetrance [see Table 2].^94,95 Clonal genetic alterations are detected in stromal rather than epithelial cells, which suggests that the genetic changes in juvenile polyps originate in the nonepithelial component of the polyps.

Like PJS, JPS is characterized by the development of multiple hamartomas throughout the GI tract. Isolated juvenile polyps are common in children and are found in approximately 1% of persons younger than 21 years. Juvenile polyposis, however, is much less common. A family history of juvenile polyposis is present in 20% to 50% of patients.¹ Although JPS is an autosomal dominant disorder, its variable penetrance results in a less obvious pattern of inheritance than is seen with FAP or HNPCC.

JPS affects the two sexes equally and generally manifests itself during the first or second decade of life (mean age at diagnosis, 18.5 years).¹ Common presenting symptoms include chronic anemia, acute GI bleeding, prolapse of rectal polyps, protein-losing enteropathy, and intussusception with or without obstruction.¹

Extracolonic manifestations of JPS include gastroduodenal and small bowel polyps, malrotation of the midgut, and mesenteric lymphangiomas. Extraintestinal manifestations include clubbing, hypertrophic pulmonary osteoarthropathy, hydrocephalus and macrocephaly, alopecia, cleft lip and palate abnormalities, supernumerary teeth, porphyria, congenital cardiac and arteriovenous malformations, psoriasis, vitellointestinal duct abnormalities, renal structural abnormalities, and bifid uterus and vagina. JPS is also part of the phenotype for Ruvalcaba-Myhre-Smith syndrome and Gorlin syndrome. Cowden disease, which is characterized by hamartomatous polyposis and is associated with breast and thyroid cancer, may be a phenotypic variant of JPS.^1,95

The diagnostic criteria for JPS are as follows: (1) the presence of three or more juvenile polyps of the colon; (2) the presence of juvenile polyps throughout the entire GI tract; or (3) the presence of any number of polyps in a patient with known family history of JPS [see Table 1].⁸⁵ The clinical presentation of JPS can be divided into three main clinical variants: (1) JPS of infancy, which is a non—sex-linked recessive condition characterized by failure to thrive, susceptibility to infections, protein-losing enteropathy, bleeding, diarrhea, rectal prolapse, intussusception, and death by the age of 2 years in severe cases; (2) generalized JPS, which occurs in the first decade of life and is characterized by juvenile polyps throughout the GI tract; and (3) JPS of the colon, the most common presentation, which is characterized by colonic polyposis only.¹

Patients with JPS appear to be at increased risk for GI malignancies, especially CRC. One study estimated the risk of CRC to be 15% by age 35 and 68% by age 65.⁹⁶ In another study, GI malignancies (mostly CRC) were diagnosed in 36 (17%) of 218 JPS patients at a mean age of 33 years.⁹⁷ Associated gastric, pancreatic, and duodenal cancers have also been reported. CRCs are thought to arise from malignant transformation of dysplastic polyps.¹ Adenocarcinomas occur, on average, 15 years after diagnosis of JPS and generally are poorly differentiated or mucinous tumors with a poor prognosis.¹

Initial evaluation of the proband and the first-degree relatives, which ideally would be done in the middle to late teenage years, should include colonoscopy, EGD, and a small bowel series. If the initial evaluation yields negative results, a repeat evaluation should be performed in 3 years, then every 3 years thereafter as long as the results remain negative. If disease is encountered, random biopsies of polyps and intervening mucosa should be performed to detect adenomatous and dysplastic changes. Management depends on the presence of symptoms and on the extent and severity of polyposis. When polyposis is mild, endoscopic management may be feasible. Continued annual surveillance after endoscopic management is required; the surveillance interval may be lengthened to 3 years if subsequent evaluations reveal no disease.^1,85

When polyposis is severe or significant symptoms are apparent, prophylactic colectomy with IRA may be considered for suitable surgical candidates. Although rectal polyposis can generally be managed with rigid or flexible proctoscopy, IPAA may be considered if the polyposis is extensive. Continued annual surveillance of the rectal remnant (after IRA) or the ileal pouch (after IPAA) is required initially. Surveillance intervals may be increased to 3 years if subsequent evaluations find no evidence of disease.^1,85