Adenomas

A neoplastic polyp is an epithelial growth composed of abnormal glands of the large bowel. A neoplastic polyp has been termed an adenoma and is classified according to the amount of villous component. Those with 0 to 25% villous tissue are classified as tubular adenomas, 25 to 75% as tubulovillous adenomas, and 75 to 100% as villous adenomas. ¹ Tubular adenomas (▶ Fig. 19.1) account for 75% of all neoplastic polyps; villous adenomas (▶ Fig. 19.2), 10%; and tubulovillous adenomas (▶ Fig. 19.3), 15%. The villous growth pattern is most prominent in sessile large adenomas, particularly those located distally in the rectum. There remains considerable uncertainty as to the nature of villous growth, whether it is merely a manifestation of continued growth of tubular adenomas or whether it is a distinct phenotype that may reflect an acquired genetic change. In favor of the former is the rarity of small villous adenomas and large purely tubular adenomas. ¹

Dysplasia describes the histologic abnormality of an adenoma according to the degree of atypical cells, categorized as low-grade (mild), moderate, and high (severe). Thus, high-grade dysplasia designates a condition one step away from an invasive carcinoma. The frequency of high-grade dysplasia correlates with the size of the adenoma (▶ Fig. 19.4). The term carcinoma-in-situ, or “intramucosal carcinoma,” should be avoided, since it implies a biological potential for distant spread, which is unwarranted and could result in overtreatment. ¹

Neoplastic polyps are common. Since data on the clinical recording of adenomas may be biased due to selection of patients and diagnostic methods, most accurate epidemiologic data on adenomas are obtained from autopsy studies. However, data from screening colonoscopy support these data. In autopsy series, adenomas are present in 34 to 52% of males and 29 to 45% of females over 50 years of age. Most adenomas (87–89%) are less than 1 cm in size. ^{3 , 4} The number, but not the size, of adenomas, increases with age. ³ Carcinomas are found in 0 to 4%. ^{3 , 4 , 5 , 6} The National Polyp Study, a multicenter randomized clinical trial in the United States, included 3,371 adenomas in 1,867 patients detected by colonoscopy. ² This study gives valuable information regarding the natural history and characteristics of polyps: 66.5% of polyps were adenomas, 11.2% were hyperplastic, and 22.3% were classified as “other” (normal mucosa, inflammatory and juvenile polyps, lymphoid hamartomas, submucosal lipomas, carcinoids, and leiomyomas). The majority of the adenomas (69%) were in the left colon (▶ Table 19.1). The sizes of the adenomas were: 0.5 cm, 38%; 0.6 to 1 cm, 37%; and 1 cm, 25%. Colonoscopy studies suggest that the incidence of polyps is greater in males than females.

It is important to note that the size, the extent of villous component, and the increasing age are independent risk factors for high-grade dysplasia. The increased frequency of high-grade dysplasia in adenomas located distal to the splenic flexure is attributable mainly to increased size and villous component rather than to location per se. Multiplicity of adenomas affects the risk of high-grade dysplasia but is dependent on size and villous component and thus is not an independent factor. ² Invasive carcinomas are uncommon in adenomas < 1 cm, and the incidence increases with an increased size of the adenomas (▶ Table 19.2). ^{7 , 8}

Adenoma-Carcinoma Sequence

The Observation

The concept that carcinomas of the colon and rectum derived from benign adenoma was observed by Dukes ⁹ of St. Mark’s Hospital, London, in 1926. Jackman and Mayo ¹⁰ coined the term adenoma-carcinoma sequence in 1951. After decades of debates and challenges by those who believed that carcinoma of the colon and rectum derived de novo, ^{11 , 12} the adenoma-carcinoma sequence has finally become widely accepted and currently is the rationale of the approach to the secondary prevention of colorectal carcinoma by colonoscopic polypectomy. ^{1 , 13 , 14 , 15 , 16} Circumstantial evidence supporting the adenoma-carcinoma sequence abounds and explains the high concurrence rate of carcinoma and adenoma and the frequent findings of contiguous benign adenoma in the resected carcinoma. ¹⁷ Numerous studies (most of which are retrospective), based on tumor registry reports, hospital records, pathology reports, surgical specimens, and colonoscopy, show a coexistence of adenomas and adenocarcinomas of the colon and rectum ranging from 13 to 62%. ¹⁸ The cumulative incidence curve of adenomas based on data from the Norwegian Cancer Registry precedes the corresponding incidence curve of carcinomas by about 5 years (▶ Fig. 19.5). It should be kept in mind that adenomas are first diagnosed and reported to the cancer registry simultaneously with the diagnosis of colorectal carcinoma, indicating a longer time span between the two types of lesions than the curve indicates. It is manifested also in the natural history of both familial adenomatous polyposis (FAP) and hereditary nonpolyposis colon cancer (HNPCC) syndrome. The latter was originally thought to offer support to the de novo school of thought but several studies have since demonstrated coexisting and contiguous adenomas associated with HNPCC carcinoma with a frequency similar to that observed with sporadic carcinomas. ¹⁹ Due to the high prevalence of adenomas and the relatively far less frequent incidence of carcinomas, only a small proportion of adenomas give rise to carcinomas. ²⁰

Although the adenoma-carcinoma sequence concept has been favored by most authors as the main pathogenesis of colorectal carcinoma, the “de novo” origin of carcinoma developing from normal mucosa has received some attention in recent years as an alternative pathway. ¹⁹ In support of this de novo theory, authors ^{21 , 22 , 23} reported early colorectal carcinomas without evidence of adjacent adenomatous cells. In the series reported by Stolte and Bethke ²² of 155 such lesions, 59% of the lesions were polypoid and 34% were flat. However, proponents for the adenoma-carcinoma sequence may argue that these types of lesions are so aggressive that the infiltration destroys the adenomatous remnants. Muto et al ²⁴ thought that all genetic alterations may take place rapidly, one after another, without a chance for morphologic changes to be expressed as seen in the adenoma-carcinoma sequence. They said, “until a specific responsible gene for de novo carcinoma is detected, de novo carcinoma arising directly from normal mucosa is only an imaginary entity. Until then, the term ‘de novo’ carcinoma is better avoided and instead de novo-type carcinoma should be used.”

Molecular Genetics

Molecular genetic discoveries provide substantial support for the adenoma-carcinoma sequence concept. ²⁵ An adenoma represents an epithelial proliferation derived from a single cell (crypt). Its development occurs as a series of genetic mutations. The progression of colorectal epithelium from normal to adenoma to carcinoma can be simplified as in ▶ Fig. 19.6.

The initial step in colorectal carcinogenesis is the mutation in the adenomatous polyposis coli (APC) gene on chromosome 5q. The APC gene is inactivated, causing the affected cells to proliferate. These cells are thus primed for subsequent growth-enhancing mutation, which is more likely because of the increased rate of cell division.

Hypomethylation of DNA has been identified as the next factor involved in colorectal carcinogenesis. Loss of methylation of CpG dinucleotides occurs in cells that are already hyperproliferative because of the inactivation of the APC gene. These changes produce a growth of the affected cells resulting in adenoma formation. Hypomethylation of DNA may be directly linked to the K-ras (Kirsten rate sarcoma virus) activation that enhances the dysplasia so that the neoplasia can progress.

Because K-ras is an oncogene, mutation of one allele is enough to produce an effect. K-ras mutations can occur in the absence of APC gene mutations but, in this case, are usually limited to aberrant crypt foci (ACF) that do not progress to malignancy. In cells that have already suffered APC mutation (both alleles need two “hits”), K-ras mutation will drive progression. Small adenomas tend to advance to intermediate adenomas.

The transition from intermediate to advanced (or late) adenoma is associated with a distinct genetic alteration on the long arm of chromosome 18. This alteration is correlated with the mutation of a gene that maps to 18q21, named deleted in colon cancer (DCC). Specific DCC mutation has been detected in a number of colorectal carcinomas and carcinomas that have lost the capacity to differentiate into mucus-producing cells that have uniformly lost DCC expression.

The progress from advanced adenoma to carcinoma is frequently accompanied by loss of heterozygosity (i.e., mutation of one of two alleles) on chromosome 17p and mutation of the p53 gene that maps to 17p. These cumulative losses in tumor suppressor gene function accompanied by activation of dominant oncogenes drive the clonal expression of cells from the benign to the malignant site. ^{25 , 26} A fuller account of molecular genetics of colon and rectal adenocarcinoma is provided in Chapter 21.

Diagnosis of Large Bowel Adenomas

Clinically, there are two morphologic types of polyps, pedunculated and sessile. The pedunculated polyp has a stem lined with normal mucosa, called a stalk or a pedicle, and has the appearance of a mushroom (▶ Fig. 19.7). A sessile polyp grows flat on the mucosa (▶ Fig. 19.8). A pedunculated polyp rarely is > 4 cm in diameter, whereas a sessile polyp can encompass the entire circumference of the large bowel.

Adenomas of the large bowel are usually asymptomatic and are frequently discovered during routine radiologic studies or endoscopic examinations. Bleeding per rectum is the most common finding if the polyp is situated in the rectum or sigmoid colon. A large pedunculated polyp in the lower part of the rectum may prolapse through the anus. A large villous adenoma may manifest as watery diarrhea; in rare instances, it causes fluid and electrolyte imbalance. Intermittent abdominal pain from recurrent intussusception or spasm may occur with a large colonic polyp but is unusual. Mild anemia may follow chronic bleeding from an ulcerative polyp. With a small polyp, up to 8 mm, biopsy and electrocoagulation can be performed, preferably using a “hot” biopsy forceps for histopathologic examination. A large polyp should be completely snared or excised and sent for histopathologic examination. A biopsy of a large polyp may not represent the entire lesion and presents difficulty in the interpretation of an invasive carcinoma. Occasionally, biopsy may cause displacement of the gland into the submucosa and can be misinterpreted as an invasive carcinoma. ²⁷ This pseudoadenomatous invasion can also be caused by trauma from hard feces, repeated twisting of the stalk with subsequent ulceration of the surface. ²⁸

Management of Benign Adenomas

Colonoscopy has revolutionized the management of large bowel polyps. Most polyps throughout the entire colon and rectum can be excised through the colonoscope with minimal morbidity. At the present time, colonic resection or colotomy and polypectomy are reserved for cases in which colonoscopic polypectomy cannot be performed, such as lesions that are too large or too flat, or when the colonoscope cannot be passed to the site of the polyp.

Most pedunculated polyps can be snared in one piece since the pedicles are rarely > 2 cm in diameter. Sessile polyps < 2 cm usually can be snared in one piece. Large sessile polyps may be snared piecemeal and in more than one session or removed with endoscopic mucosal resection or endoscopic submucosal dissection (see Chapter 4). Excised polyps must be prepared properly and sectioned so that all the layers can be examined microscopically and the evidence of invasive carcinoma detected.

Adenomas in the rectum present a unique situation. These lesions can be palpated with finger, suction, or endoscope. If there is no induration, the chance that a lesion is benign is 90%. ^{29 , 30} There are a number of ways to remove a large adenoma in the rectum, including proctoscope or a colonoscope, per anal excision, transanal endoscopic microsurgery (TEMS), transanal minimally invasive surgery (TAMIS), and posterior proctotomy.

Patients with a neoplastic polyp have a higher risk of developing another polyp; so follow-up colonoscopy is advised. After the colon and rectum are cleared of polyps, follow-up colonoscopy every 3 to 5 years is recommended. A large sessile polyp, particularly villous type, is prone to recur, and a follow-up check of the polypectomy site should be done every 3 to 6 months the first year, every 6 to 12 months the second year, and every year thereafter to the fifth year. Then colonoscopic examination every 3 to 5 years is appropriate.

The Flat Polyp

In 1985, Muto et al ³¹ called attention to a separate type of polyp called a “flat” adenoma. This type of polyp is unique in that it is usually small and flat, often with a central depression, and is difficult to detect with colonoscopy or even with the resected colon and rectal specimens. Ninety percent of flat adenomas are < 1 cm and more than half are less than 5 mm. ³² The significance of flat adenomas is the high incidence of carcinomas, which occur in 6% of patients, even when the lesions are as small as 2 to 4 mm, and rapidly rise to 36% when the lesions are 9 to 10 mm. Approximately 10% of the adenomas in the Muto series were flat adenomas. They were most frequently located in the left colon and the rectum. Lynch et al ³³ found similar flat adenomas in patients who were members of the same kindred under study for HNPCC. Most of the lesions were in the right colon. The flat adenomas, originally thought to occur mostly among Japanese, have also been found in studies from Australia, Canada, and the United Kingdom. ³²

In a prospective study of 1,000 consecutive patients attending for colonoscopy, flat or depressed lesions were examined by Rembacken et al. ³⁴ Patients were not preselected and the indications were similar to other units in the United Kingdom. A flat adenoma was defined as mucosal elevations with a flat or slightly rounded surface and a height of less than half the diameter of the lesion. In practice, most flat adenomas were less than 2 mm in height and only very broad lesions were 5 mm high. During the examination, they used 0.2% indigo carmine dye, 3 to 6 mL, sprayed directly onto suspicious areas. Magnifying colonoscopy was also used.

The authors identified 321 adenomas; 119 (37%) were flat and 4 (1%) appeared depressed. Fifty-four percent of the flat or depressed lesions were situated between splenic flexure and rectum.

Seventy of the flat lesions (59%) were < 10 mm in size (mean, 5 mm) and 4% had early carcinoma (invasive into submucosa); 49 flat lesions (41%) were > 10 mm (mean, 21 mm), and 29% had early carcinoma. The mean size of the depressed lesions was 9 mm and three of four (75%) had early carcinoma, indicating their aggressiveness compared to other types of lesions.

Rembacken et al ³⁴ suggested, “Western colonoscopists refuse training in the recognition of flat, elevated and depressed lesion in order to detect colorectal neoplasms in their early stages.” The readers should note that in this study, all of the patients had indications for colonoscopic examinations and not as a screening examination for low-risk asymptomatic patients. In response to an editorial comment, ³⁵ Rembacken et al wrote, ³⁴ “The use of indigo carmine dye is paramount to the detection of flat and depressed lesions and only takes a few seconds. Without the dye, it is difficult to evaluate non-polypoid lesions because they generally appear to be erythematous patches, easily mistaken for scope trauma. The magnifying colonoscope does not help in the initial recognition of lesions but allows the endoscopists to assess the crypt pattern and predict the histology.” Recent molecular analysis of such flat adenomas suggests that they are etiologically distinct from other polypoid adenomas. ³⁶ The mutation rate and the K-ras gene are both significantly reduced (16% in flat adenomas compared to 50% in ordinary colorectal adenomas) and do not occur in the same codons. The management of flat adenomas is the same as for sessile adenomas.

Why Remove a Polyp?

It has generally been accepted that most colorectal carcinomas are derived from benign adenomas through the adenoma-carcinoma sequence. It is estimated that it takes about 5 years from a clean colon to the development of an adenoma and about 10 years from a clean colon to the development of invasive carcinoma. ¹³ Thus, removal of an adenoma is prophylactic against the development of colorectal carcinoma. Gilbertsen, ³⁷ in a retrospective study, showed that removal of rectal polyps in patients under surveillance with yearly rigid proctosigmoidoscopy results in a lower than expected incidence of rectal carcinoma. This result was confirmed by Selby et al ³⁸ in a case–control study using rigid proctosigmoidoscopy; screening examination produced a 70% reduction in the risk of death from rectal and distal sigmoid carcinoma. The National Polyp Study also showed that colonoscopic polypectomy results in a lower than expected incidence of colorectal carcinoma. ³⁹

Church studied diminutive (1–5 mm) and small (6–10 mm) adenomas of the colon and rectum, and found that although the risk of invasive carcinoma was low (0.1 and 0.2%, respectively), the risk of severe dysplasia was significant (4.4 and 15.6%, respectively). ⁴⁰ He advised a cold excision or a hot snare as appropriate (▶ Table 19.3).

Natural History of Untreated Large Bowel Adenomas

A retrospective review of patients from the precolonoscopic era by Stryker et al ⁴² analyzed 226 patients who had colonic polyps 210 mm in diameter and in whom periodic radiographic examination of the colon was elected over excision. Twenty-one invasive carcinomas were identified at the site of the index polyp at a mean follow-up of 108 months (range, 24–225 months). The risk of having a polyp ≥ 1 cm in size develop into an invasive carcinoma at 5, 10, and 20 years was 2.5, 8, and 24%, respectively.

Further study of this same group of patients by Otchy et al ⁴³ revealed that the cumulative probability of developing an invasive metachronous carcinoma at a site different from the index polyp was 2% at 5 years, 7% at 10 years, and 12% at 20 years. Over a median duration of polyp surveillance of 4.8 years (range, 1–27 years), 11 (5%) of the index polyps disappeared, 129 (57%) had no growth noted, and 86 demonstrated growth. Forty-two of the 86 polyps (49%) had at least a twofold increase in size. Seventy-one of the 86 polyps were removed, and 24 (34%) were carcinomatous. Fifteen of the 86 polyps that increased in size were not removed, and none of these patients developed a carcinoma. Forty-three of the 129 polyps that did not grow were eventually removed. Five of those polyps had carcinoma and one of these patients also developed a metachronous carcinoma at a later date. In addition, 2 of the 43 patients developed a colon carcinoma in areas distant from the site of the index polyp.

These data further support the recommendation for excision of all colonic polyps ≥ 10 mm in diameter and a periodic examination of the entire colon. Although this study has limitations inherent to any retrospective analysis, comparable prospective data are unlikely to be available in the future because of the widespread availability of colonoscopy and the compelling evidence to recommend the removal of neoplastic polyps.

What Happens to Smaller Adenomas?

Hofstad et al ⁴¹ prospectively studied the growth of colorectal polyps. Colonoscopy was performed in 58 subjects. Polyps ≥ 10 mm were removed; polyps < 5 mm and 5 to 9 mm were left behind for a follow-up study. Colonoscopy was followed up by one investigator once a year. On the third year, polyps were removed by snare or hot biopsy. The measurement of the polyps was performed by a measuring probe plus photography. On the third year, 7 of 58 patients had only hyperplastic polyps. Twenty-nine individuals had one adenoma, 17 individuals had two to three adenomas, and 5 individuals had four to five adenomas. Twenty-five percent of all the adenomas were unchanged in size, whereas 40% displayed growth and 35% showed regression or shrinking in size. Adenomatous polyps < 5 mm showed a tendency to growth, while the adenomas 5 to 9 mm showed a tendency to reduction in size. The hyperplastic polyps showed a similar pattern. There was a tendency to increase growth in the adenomatous polyps in the younger age groups reaching significance from initial examination to the third year and from the first to the second year of re-examination. Moreover, in the patients with four to five adenomas at the initial examination, the polyps showed larger growth than the polyps in patients with only one or two to three adenomas. There were no differences in polyp growth between the sexes. A similar prospective study by Bersentes et al ⁴⁴ on adenomas of the upper rectum or sigmoid colon, size 3 to 9 mm, showed no regression or consistent linear growth rates with a 2-year follow-up.

In the study by Hofstad et al, ⁴¹ 86% of the individuals had at least one new polyp during the 3 years and 75% had at least one new adenoma. The newly discovered polyps were significantly smaller than the average size at initial examination. They were also more frequent in the proximal part of the colon (71%) than the polyps discovered at initial examination (38%). There were more new adenomas among those with more than four to five adenomas at initial examination, than those with one adenoma, reaching significance from initial examination to the first year of examination and from initial examination to third year. There were more new adenomas among patients > 60 years of age than those < 60 years. No differences were found between the sexes.

Adenomas with Invasive Carcinoma

The term invasive carcinoma is applied only when the malignant cells have invaded the polyp, either sessile or pedunculated, partially or totally, through the muscularis mucosa into the submucosa. Severely dysplastic cells superficial to the muscularis mucosa do not metastasize and should be classified as atypia (rather than carcinoma in situ or superficial carcinoma). ¹³ For this type of lesion, complete excision is all that is necessary. Follow-up of these polyps is the same as for benign polyps.

A polyp with invasive carcinoma or a malignant polyp is an early carcinoma. For the TNM classification, it is a T1NxMx. Two classification systems for adenomas with invasive carcinoma have been proposed. In 1985, Haggitt et al ⁴⁵ proposed a classification for polyps with adenocarcinoma according to the depth of invasion as follows (▶ Fig. 19.9):

• 
  

  Level 0—Carcinoma in situ or intramucosal carcinoma. These are not invasive.

  
  
• 
  

  Level 1—Carcinoma invading through the muscularis mucosae into the submucosa but limited to the head of the polyp (i.e., above the junction between the adenoma and its stalk).

  
  
• 
  

  Level 2—Carcinoma invading the level of the neck of the adenoma (junction between adenoma and its stalk).

  
  
• 
  

  Level 3—Carcinoma invading any part of the stalk.

  
  
• 
  

  Level 4—Carcinoma invading into the submucosa of the bowel wall below the stalk of the polyp but above the muscularis propria. Although not specifically defined by Haggitt et al, sessile polyps with invasive carcinoma are considered to have level 4 invasion.

The risk of lymph node metastasis for pedunculated polyp (Haggitt level 1, 2, and 3) is low. ^{45 , 46 , 47 , 48}

In 1993, Kudo ⁴⁹ classified the submucosal invasion of the sessile lesions into three levels (▶ Fig. 19.10):

• 
  

  Sm1—invasion into the upper third of the submucosa.

  
  
• 
  

  Sm2—invasion into the middle third of the submucosa.

  
  
• 
  

  Sm3—invasion into the lower third of the submucosa.

The Sm system appears to be effective and practical and was recommended by a consensus workshop in Paris in 2002. ⁵⁰ In the series by Nascimbeni et al, ⁵¹ the pathologist could evaluate the depth of invasion into Sm1, Sm2, and Sm3 in 97% of the cases. In fact, the Haggitt level for the pedunculated lesion can be incorporated into the Sm system (▶ Fig. 19.10). However, the endoscopists must properly prepare the specimens and the pathologists must properly section them in order to examine the entire layers.

Risk factors for residual cancer or involvement of lymph nodes include: poor differentiation, lymphovascular invasion, depth of invasion in submucosa (e.g., Sm3), and a positive resection margin (< 2 mm from edge of tumor to resected margin). ^{51 , 52 , 53 , 54}

Patients having lesions with one or more risk factors should undergo an oncologic bowel resection. ^{52 , 55 , 56 , 57 , 58} Hagitt level 4 lesions will almost always have a positive margin and will usually require a resection. A malignant lesion that is removed piecemeal also requires further excision or resection. The incidence of residual cancer in the bowel or regional lymph nodes in specimens varies with respect to which risk factors were present but in reported studies has averaged 18% but as high as 50% in some series. ^{59 , 60}

For lesions with no risk factors, a complete snaring or a transanal excision is adequate. Close follow-up examination with endoscopy to detect a local recurrence may be performed every 3 to 6 months for the first year. This period can be extended to every 6 to 12 months in the second year and to every year for the next 2 years. Thereafter, endoscopy every 3 years is adequate.

Rectal polyps merit additional discussion. Those polyps that are early rectal cancers are managed as discussed in Chapter 22.

Transanal excision for a sessile polyp with invasive carcinoma, or a T1 carcinoma of the low rectum has a three- to fivefold higher risk of carcinoma recurrence compared with patients treated by radical resection ⁶¹ (▶ Table 19.4). Waiting to perform a radical resection after a local recurrence is a poor choice. In most series, the cancer-free survival for salvage resection in these patients is 50 to 56%. ^{62 , 63} On the other hand, an immediate radical resection after local excision (within 1 month) gives a better prognosis, 94% cancer-free survival at 10 years, and is comparable to primary resection in a case–control comparison. ⁶⁴ In short, local excision for a sessile polyp with invasive carcinoma (T1) of the lower third of rectum has high local recurrence. It appears that the early lesion at this site is a locally disseminated disease. To improve the outcome, the recurrence rate has to be improved: options include doing more radical resection in young and good health patients, finding a better adjuvant therapy, or finding better ways in selection of patients, such as molecular markers in the future.

Serrated Adenoma

This is the term coined by Longacre and Fenoglio-Preiser in 1990 ⁶⁹ to describe a new entity of mixed hyperplastic polyp/adenomatous polyp. In their study of 110 serrated adenomas, compared to 60 traditional adenomas and 40 hyperplastic polyps, they found that these lesions distributed throughout the colon and rectum, with a slight preponderance of large lesions (> 1 cm) occurring in the cecum and appendix.

There are two types of mixed epithelial polyps: one in which adenomatous and hyperplastic glands are mixed (▶ Fig. 19.11a), and one in which the adenoma has a serrated appearance on microscopic examination (▶ Fig. 19.11b). Microscopic examination of the lesions shows goblet cell immaturity, prominent architectural distortion, cytologically atypical nuclei, rare upper zone mitoses, and absence of a thickened collagen table. ^{66 , 69}

Grossly, the lesion is flat and smooth; it may look like a plaque or thickened mucosa on colonoscopic examination (▶ Fig. 19.12). This type of lesion can be easily missed on colonoscopy if the colon is overdistended (stretching it flat) or underdistended (causing wrinkle on mucosa to mask it). Unlike the classic hyperplastic polyps that are small and restricted to the rectum and rectosigmoid colon, serrated adenomas are larger and occur in both proximal and distal colon and rectum. ⁷⁰ Some of the individuals previously reported as having multiple hyperplastic polyps could instead have had multiple serrated adenomatous polyps. ⁶⁹

Based on the observation that 11% of serrated adenomas in the series of Longacre and Fenoglio-Preiser ⁶⁹ contained foci of intramucosal carcinoma, it was surmised that an individual lesion would carry a significant malignant potential. Nevertheless, the rarity of serrated adenoma (0.6% of colorectal polyps) would minimize their contribution to the overall burden of colorectal malignancy. ⁶⁵ Torlakovic and Snover ⁶⁶ reported six patients with serrated adenomatous polyposis. Each patient had at least 50 polyps, ranging from 0.3 to 4.5 cm in size, mostly sessile. Three patients had diffuse polyps, two patients had the polyps in the left colon, and one patient had them in the right colon. Four patients had carcinoma. Serrated adenomas are infrequently observed in endoscopic practice due to underdiagnosis and the potential for rapid evolution to carcinoma. The latter suggestion is supported by the demonstration of DNA microsatellite instability (MSI) in mixed polyps and serrated adenomas and by analogy with the aggressive adenomas in HNPCC. ⁶⁵

The known alterations include K-ras mutation, low and occasional high-level MSI, 1pLOH, and methylation of HPP1/TPEF (a putative antiadhesion molecule). Additional genetic alterations may be observed in neoplastic subclones occurring within or adjacent to hyperplastic polyps. These include loss of expression of MGMT or hMLH1. ⁶⁷

Sporadic MSI-low (MSI-L) and MSI-high (MSI-H) carcinomas may evolve through the serrated adenoma pathway. ⁶⁵ The serrated adenoma pathway is likely to show marked molecular heterogeneity, but patterns are beginning to emerge. The view that all, or even most, colorectal carcinomas are initiated by mutation of APC gene and evolve through the classical adenoma-carcinoma sequence may no longer be tenable. This understanding will surely transform our approach to the early detection and prevention of colorectal carcinoma. ⁶⁵

The molecular steps that determine growth of ACF into hyperplastic polyp are not known. Colorectal carcinoma is envisioned to arise from hyperplasticlike polyps (or sessile serrated polyps) in which the earliest events might be BRAF mutation synergizing with a methylated and silenced pro-apoptotic gene. Subsequent methylation of hMLH1 or MGMT then predisposes to mutation, dysplastic change, and finally to malignancy that is frequently characterized by MSI-H or MSI-L status. K-ras mutation may substitute for BRAF in methylator pathways culminating in MSI-L and some MSS colorectal carcinomas. ⁶⁸ Serrated adenomas are neoplastic polyps. The treatment is the same as in adenomatous polyps.

Juvenile Polyps and Juvenile Polyposis

Juvenile polyps characteristically occur in children, although they may present in adults at any age. This type of polyp is a hamartoma and is not premalignant. Macroscopically, they are pink, smooth, round, and usually pedunculated. The cut section shows a cheeselike appearance from dilated cystic spaces. Microscopic pictures show dilated glands filled with mucus and an abnormality of the lamina propria, which has a mesenchymal appearance (▶ Fig. 19.13). The muscularis mucosa does not participate in the structure of the polyp. Bleeding from the rectum is common. A moderate amount of bleeding can occur if the polyp is autoamputated, a phenomenon not seen in other types of polyps. Intussusception of the colon occasionally occurs if the polyp is large. Treatment is by excision or snaring through a colonoscope or a transanal excision.

Juvenile polyposis is an entity characteristically and biologically distinct from solitary juvenile polyp or other polyposis. The condition was first observed by McColl et al in 1964. ⁷¹ The term juvenile polyposis rather than juvenile polyposis coli is to be preferred as polyps are also found in the stomach and the small intestine. ⁷² There are two types of juvenile polyposis: in infancy and in other variable age of onset. ⁷³

Juvenile polyposis of infancy is a rare form without a family history. The infant presents with diarrhea, either bloody or mucinous, anemia, protein-losing enteropathy, and intussusception; rectal prolapse develops between 8 and 10 months of age and leads to significant morbidity. ^{73 , 74} The entire gastrointestinal (GI) tract is usually affected; the prognosis depends on the severity and extent of GI involvement. Death occurs before the age of 2 years in severe cases. ⁷² Surgery is indicated in cases of intussusception, or polypectomies in cases of rectal prolapse to reduce the leading point of the prolapse. Supportive care to replace fluid and electrolytes or total parenteral nutrition is indicated. ⁷⁴

The majority of patients with juvenile polyposis manifest in their first or second decade, but in 15% of patients, the diagnosis is delayed until they are adults. They usually present with rectal bleeding and anemia. Family history of juvenile polyposis is found in 20 to 50% of patients. Various extracolonic abnormalities, described in 11 to 20% of cases, have included digital clubbing, pulmonary arteriovenous fistula, macrocephaly, alopecia, bony swellings, cleft lip, cleft palate, supernumerary teeth, porphyria, arteriovenous malformation affecting the skin, psoriasis, congenital heart disease, malrotation of the gut, abnormalities involving the vitellointestinal duct, double renal pelvis and ureter, acute glomerulonephritis, undescended testes, and bifid uterus and vagina. ⁷²

Patients with juvenile polyposis usually have 50 to 200 colorectal polyps and a proportion have polyps in the stomach and small intestine. Some patients seem to have relatively few polyps, but these tend to be the parent of the prospectus. It is conceivable that the juvenile polyps are produced only within the first few decades and are subsequently lost through autoamputation. Thus, juvenile polyposis may be diagnosed when a relatively old and asymptomatic parent is screened colonoscopically and the smallest number of polyps found on this basis is 5. ⁷³

Jass et al ⁷³ proposed a working definition of juvenile polyposis:

• 
  

  More than five juvenile polyps of the colorectum.

  
  
• 
  

  Juvenile polyps throughout the GI tract.

  
  
• 
  

  Any number of juvenile polyps with a family history of juvenile polyposis.

On the other hand, Giardiello et al ⁷⁵ suggested that the patients with as few as three juvenile polyps should undergo screening for colorectal neoplasm.

Although there is no evidence that isolated juvenile polyp could be malignant, it is now well established that juvenile polyposis is a precancerous condition. ^{73 , 75 , 76 , 77 , 78} The risk of GI malignancy in affected members of juvenile polyposis kindred exceeds 50% in a series of kindred reported by Howe et al. ⁷⁶

In a classic paper on juvenile polyposis, Jass et al ⁷³ studied 87 patients with juvenile polyposis recorded in the St. Mark’s Polyposis Registry, including 1,032 polyps and 18 patients with colorectal carcinoma. They found that about 20% of juvenile polyps did not conform to the classical description. Grossly, they formed lobular masses (instead of spherical). These atypical juvenile polyps also revealed relatively less lamina propria and more epithelium than that found in the more typical variety and often adopted a villous or papillary configuration. Epithelial dysplasia occurred in both typical and atypical juvenile polyps but more frequently in the latter. Nearly 50% of the atypical juvenile polyps showed some degree of dysplasia similar to adenomas. The 18 patients with colorectal adenocarcinoma had a mean age of 34 years (range, 15–59 years). A high proportion of carcinomas were mucinous and/or poorly differentiated, which was in accord with case reports from other authors.

There is little direct information on the histogenesis of carcinoma in juvenile polyposis. Dysplasia has been shown to occur in two forms: (1) a focus of adenomatous change within a polyp and (2) an adenoma showing no residual juvenile features. ⁷⁶ On the mechanism of polyp-cancer sequence in juvenile polyposis, Kinzler and Vogelstein ⁷⁷ postulated,

“an abnormal stroma can affect the development of adjacent epithelial cells is not a new concept. Ulcerative colitis is an autoimmune disease that leads to inflammation and cystic epithelium in the mucosa of the colon. Initially, the imbedded epithelium shows no neoplastic changes, but foci of epithelial neoplasia and progression to cancer eventually develops in many cases. The regeneration that occurs to replace damaged epithelium may increase the probability of somatic mutations in this abnormal microenvironment. The increased risk of cancer in juvenile polyposis syndrome and ulcerative colitis patients, therefore, seems primarily the result of an altered terrain for epithelial cell growth and can be thought of as a landscaper defect.”

Juvenile polyposis is an autosomal dominant condition. ⁷⁶ The germline mutation is in the gene SMAD-4 (also known as DPC-4), located on chromosome 18q21.1. ^{75 , 79 , 80}

There is little information about prophylactic colectomy or proctocolectomy to prevent occurrence of carcinoma. The decision on performing the operation should be dictated by the number and the site of the polyps. Polyps of the colon and rectum that are too numerous for colonoscopy and polypectomies should have an abdominal colectomy with ileorectal anastomosis (IRA) or proctocolectomy with ileal pouch-anal anastomosis (IPAA) or an ileostomy. ^{72 , 74 , 81 , 82} In a series reported by Oncel et al, ⁸¹ 5 of 10 patients who underwent colectomy with IRA for juvenile polyposis required a subsequent proctectomy with a mean follow-up of 9 years (range, 6–34 years). This and other studies suggest that proctocolectomy with ileoanal pouch procedure may be a better option as an initial operation. ^{81 , 82}

The proband and relatives of the first degree should be screened, probably starting in the later teen years, by upper and lower GI endoscopy. If this initial screen is negative, a follow-up endoscopy should be performed every 3 years. ⁸³ For patients who have had a colectomy or an ileoanal pouch, surveillance should be performed periodically. ^{81 , 82} Howe et al ⁸⁴ recommended genetic testing as part of the workup. However, given the presumed genetic heterogeneity of this syndrome, failure to show a mutation in SMAD-4 does not support lengthening the surveillance interval to 10 years as they suggested. ⁸¹

Peutz–Jeghers Syndrome

PJS is a rare autosomal dominant disease characterized by GI hamartomatous polyposis and mucocutaneous pigmentation. It was originally described by Peutz in 1921 but was not clearly identified until attention was brought to it by Jeghers et al ⁸⁵ in 1949. The syndrome comprises melanin spots of buccal mucosa and lips; the face and digits may be involved to a variable extent, but mouth pigmentation is the sine qua non of this portion of the syndrome. The presence of polyps in the small bowel is a constant finding of this syndrome, but the stomach, colon, and rectum also may be involved. The characteristic Peutz–Jeghers polyp has an abnormal muscularis mucosa branching into the lamina propria, giving the appearance of a Christmas tree (▶ Fig. 19.14).

Cronkhite–Canada Syndrome

Cronkhite–Canada syndrome is characterized by generalized GI polyposis associated with alopecia, cutaneous pigmentation, and atrophy of fingernails and toenails (onychatrophia). It was first deducted in two patients and described by Cronkhite and Canada in 1955. ¹⁰³ The etiology is unknown. There is no familial inheritance pattern and no associated gene or mutation has been identified. ¹⁰⁴

Diarrhea is a prominent feature of this syndrome, accounting for 46 of 55 patients in the series of Daniel et al. ¹⁰⁵ The cause of diarrhea is unknown. Nardone et al ¹⁰⁶ reported a case of Cronkhite–Canada syndrome associated with achlorhydria and hypergastrinemia causing direct gastric wall invasion by gram-negative Campylobacter pylori. This may explain the diarrhea in those patients. Hair loss was noted in 49 of 55 patients. In most patients, hair loss took place simultaneously from the scalp, eyebrows, face, axillae, pubic areas, and extremities, but in some only loss of scalp hair was described. ¹⁰⁵ Nail changes were reported in 51 of 55 patients. In most of them, the nails showed varying degrees of dystrophy, such as thinning and splitting, and partial separation from the nail bed (onycholysis). Complete loss of all fingernails and toenails (onychomadesis), over a period of several weeks, was also noted in some patients. ¹⁰⁵

Hyperpigmentation was present in 45 of 55 patients, ranging from a few millimeters to 10 cm in diameter. The distribution of pigmentary skin changes could be anywhere, including extremities, face, palms, soles, neck, back, chest, scalp, and lips. ¹⁰⁵ Other manifestations include nausea, vomiting, weakness, weight loss, abdominal pain, numbness, and tingling of extremities. ¹⁰⁵ Electrolyte disturbances are a prominent feature and appear to reflect malabsorption and losses from the GI tract. Total serum protein is also found to be low in most patients due to excessive enteric protein loss. ¹⁰⁵

From radiologic, endoscopic, and autopsy data, the stomach and large intestine were involved in 53 of 55 cases. The actual frequency of small bowel involvement would be inaccurate because the small bowel X-rays and biopsies were not performed in every case, in the series of Daniel et al. ¹⁰⁵ From the autopsy data, the number of polyps was greatest in the duodenum, less in the jejunum and proximal ileum, and again increased in the terminal ileum. ¹⁰⁵ The polyps consist of cystic dilatation of the epithelial tubules similar to that of juvenile polyps, but the lesions are usually smaller and do not show marked excess of lamina propria. ^{83 , 105 , 106 , 107 , 108 , 109 , 110 , 111 , 112 , 113 , 114}

The true incidence of GI carcinoma in Cronkhite–Canada syndrome is unknown. In the review of literature by Daniel et al ¹⁰⁵ in 55 cases, they found six cases of carcinoma of the colon and/or rectum, including one case of carcinoma of the stomach. Some of these carcinomas were multiple. Watanabe et al ¹⁰⁸ reported a case of Cronkhite–Canada syndrome associated with triple gastric carcinoma. Histopathologic examination revealed that the polyp underwent malignant transformation without an adenoma component.

Cowden’s Disease

Cowden’s disease is an uncommon familial syndrome of combined ectodermal, endodermal, and mesodermal hamartomas. The disease was named for the propositus by Lloyd and Dennis in 1963. ¹¹³ Eighty percent of patients present with dermatologic manifestations, such as keratosis of extremities, the most common being a benign neoplasm of the hair shaft: a trichilemmoma. If a patient is diagnosed with more than one trichilemmoma, consideration should be given to the diagnosis of Cowden’s disease. The second most common area of involvement is the central nervous system. Cowden’s disease in concert with cerebella gangliocytomatosis is referred to as the Lhermitte–Duclos syndrome. Approximately 40% of affected individuals have macrocephaly as a component of the syndrome. Only 35% of patients who meet the diagnostic criteria for Cowden’s disease have GI polyposis. ⁸³

Polyps in patients with Cowden’s disease are small, typically < 5 mm in diameter. Microscopic features are consistent with hamartomas, characterized by disorganization and proliferation of the muscularis mucosa with minimally abnormal overlying mucosa. ¹¹⁴

Most patients with Cowden’s disease have been shown to subsume germline mutations in the PTEN gene located at 10q22. ¹¹⁵ PTEN is a tumor suppressor gene which has been shown to be involved with other forms of carcinoma such as familial thyroid carcinoma, inherited breast carcinoma, prostatic carcinoma, and malignant melanoma. ^{116 , 117 , 118 , 119 , 120 , 121} The majority of patients with Cowden’s disease will have some form of benign thyroid or breast disease. In addition, the projected lifetime risk of thyroid malignancy is 10% and of breast malignancy is approximately 30 to 50%. ^{116 , 117 , 118} There has been no reported increased risk of invasive GI malignancy to date. ⁸³

Screening and surveillance for breast malignancies should include a schedule of monthly breast self-examinations. Clinical examination should be undertaken annually, beginning in the late teen years or as clinically warranted by symptoms. Mammography should be implemented at the age of 25. Although no specific recommendations for thyroid surveillance have been published, annual screening by clinical examination should begin in the late teen years or as symptoms warrant. A thyroid ultrasound may be used in parallel every 1 to 2 years. ⁸³

GI polyposis should be addressed by endoscopic surveillance. Although no definitive increased risk of colorectal carcinoma has been documented, the syndrome is rare; thus, the true risk may be unrecognized. ⁸³