Polyps of the Large Intestine

Clinical Features

Adenomas are common lesions. They are almost always asymptomatic. However, some patients may develop either overt, or occult, rectal bleeding, and large polyps may lead to iron-deficiency anemia. The clinical importance of adenomas is almost entirely related to their well-established premalignant nature. In general, the prevalence of adenomas increases dramatically with age. By the fifth decade of life, approximately 12% of individuals have adenomas, of which approximately 25% are considered high-risk lesions (see later). After 50 years of age, the prevalence rate of adenomas continues to increase to approximately 50% of the population in high-risk Western countries such as the United States. The likelihood that adenomas will develop is strongly influenced by family history and by a variety of nutritional factors. ^,

Pathological Features

Adenomas are defined as in situ dysplastic clonal proliferations of epithelium ( Fig. 22.2 ). Microscopically, adenomas are categorized, architecturally, as either tubular, tubulovillous, or villous ( Fig. 22.3 ). However, precise histological criteria for each of these three categories vary widely. One reasonable rule of thumb is that villous adenomas should contain at least 75% villi, whereas pure tubular lesions should contain less than 25% villi. Thus tubulovillous lesions are those that contain between 25% and 75% villous epithelium. The degree of villous differentiation has been shown to increase with increasing size of the adenoma. Villous lesions are considered “advanced” for the purpose of clinical management (see later discussion).

The lamina propria of adenomas may contain a variable amount of lymphocytes, plasma cells, neutrophils, and eosinophils. Some adenomas, particularly those with high-grade dysplasia, may be ulcerated. Paneth cell or neuroendocrine cell metaplasia is a common finding and may be marked in some cases. Rarely, one may see squamous metaplasia (or squamous morules) in adenomas. Some adenomas, particularly those that are pedunculated, may contain dilated and ruptured crypts with mucin extravasation into the lamina propria. Often, these cases are associated with epithelial displacement into the submucosa. Desmoplasia is not normally seen in benign colonic adenomas. When present, it should raise a very strong suspicion of a submucosal invasive adenocarcinoma.

Very rarely, adenomas are associated with a minute proliferation of well-differentiated neuroendocrine cells either adjacent to, or underlying, dysplastic crypts in either the mucosa or submucosa. These lesions have been referred to as composite adenoma-microcarcinoid tumors. The neuroendocrine component consists of small clusters and nests of uniform cells with round nuclei and eosinophilic cytoplasm ( Fig. 22.4 ). The vast majority of studies indicate that composite adenoma-microcarcinoid have a benign clinical course, even when the neuroendocrine cells extend into the submucosa; however, care must be taken to exclude a high-grade neuroendocrine carcinoma arising from an adenoma.

Grading Dysplasia in Adenomas

Microscopically, by definition, all adenomas contain at least low-grade dysplasia. Dysplasia in adenomas is generally classified as either low grade or high grade based on a combination of cytological and architectural features. This classification system is favored because (1) a decrease in the degree of interobserver variability has been documented regarding interpretation of dysplasia by pathologists; (2) improved clinical pathological relevance is seen with regard to surveillance and treatment options; and (3) the term carcinoma in situ is often misinterpreted by clinicians as indicative of malignant behavior, which may lead to an unnecessary colonic resection.

Low-grade dysplasia is defined by the presence of architecturally noncomplex crypts containing nuclei that are pseudostratified, or partially stratified, such that the cell nuclei reach only the lower half of the cell cytoplasm. Mitotic activity may be brisk, but atypical mitoses, significant loss of polarity, and pleomorphism are minimal, if present at all. Apoptosis is usually readily identifiable throughout the adenomatous epithelium. The crypts are arranged in a parallel configuration without significant back-to-back configuration, cribriforming, or complex budding.

High-grade dysplasia is defined by marked pseudostratification, or stratification, of neoplastic nuclei that extend toward the luminal half of the cells and usually contain significant pleomorphism, increased mitotic activity, atypical mitoses, and marked loss of polarity. Architectural changes such as back-to-back gland configuration and cribriforming may also be observed. With progression of neoplasia, glands lose their orderly configuration and become more irregular and complex. In addition, neoplastic nuclei become more “open” in appearance and may contain prominent nucleoli ( Fig. 22.5 ). The N:C ratio of the cells increases, and loss of polarity becomes marked.

The pathological criteria and clinical importance of intramucosal adenocarcinoma in adenomatous polyps is a controversial topic. Intramucosal adenocarcinoma is defined as invasion into the lamina propria, including the muscularis mucosae, but not through the muscularis mucosae into the submucosa (see Fig. 22.5 ). Although this is a well-recognized diagnosis in other sites of the gastrointestinal (GI) tract, such as the stomach and esophagus, many studies have shown that invasion of the lamina propria of the colon has virtually no increased risk of lymph node metastasis provided that the polyp has been completely excised. This has led some authorities to conclude that the term intramucosal adenocarcinoma should not be used in the setting of colonic adenomatous polyps, given the lack of clinical significance. However, other experts argue that the term intramucosal adenocarcinoma should be used in any setting in which there is pathological evidence of invasion into the lamina propria, but that it is incumbent on the pathologist, gastroenterologist, and surgeon to understand the biological potential of this phenomenon, and that polypectomy is sufficient to treat these lesions. Ultimately, the goal is to avoid unnecessary surgery for patients with intramucosal adenocarcinoma. However, if one feels compelled to use the term adenoma with intramucosal adenocarcinoma, the diagnostic report should always include a statement regarding the absence of submucosally invasive cancer, the status of the resection margin, and the overall adequacy of resection and need for further therapy, which in most cases is best accomplished with endoscopic polypectomy.

Natural History and Treatment

Screening colonoscopy has played a major role in the steady decline in the incidence of colorectal carcinoma in the United States. The National Polyp Study, ^, a large study initiated in 1990, has demonstrated that removal of adenomas by endoscopic polypectomy significantly decreases both the incidence of colorectal adenocarcinoma and colorectal cancer–specific mortality. Patients with colon cancer identified on screening colonoscopy not only have lower-stage disease at presentation, but also have more favorable outcomes independent of their staging. The best predictor for the presence of malignancy within an adenoma at the time of excision is the polyp size. Among adenomas larger than 2 cm, there is a 10% to 20% risk of carcinoma in the polyp at the time of removal. Adenomas measuring between 1 and 2 cm have a 5% risk of harboring cancer, and those smaller than 1 cm have a much lower risk of adenocarcinoma (<1%).

The degree of dysplasia also represents an independent risk factor for malignancy in adenomas, regardless of polyp size. Most adenomas harbor only low-grade dysplasia. However, larger lesions have a greater likelihood of harboring high-grade dysplasia. ^{, ,} Nevertheless, high-grade dysplasia within a colorectal adenoma does not increase the risk of carcinoma elsewhere in the patient’s colon. Because of the time involved for an adenoma to acquire sufficient molecular changes for it to become invasive, most remain benign. In general, adenomas are slow-growing lesions. The lifetime prevalence rate of adenomas is approximately 50%, whereas the lifetime prevalence of colorectal adenocarcinoma is approximately 6%. Therefore it is evident that only a small minority of polyps ultimately develop adenocarcinoma. Actual longitudinal follow-up data have been difficult to obtain because adenomas are typically removed at the time of endoscopic identification. One cohort of 35 patients, all of whom had an adenoma smaller than 5 mm in maximum diameter, was followed up for 2 years. In 50% of the patients, an increase in size of the adenoma occurred; in the remainder, the lesion remained unchanged or even regressed slightly.

The appropriate treatment for all colorectal adenomas, regardless of their size, architectural type, or degree of dysplasia, is complete endoscopic removal with negative margins ( Box 22.1 ). ^,

Surveillance Guidelines

Colorectal adenocarcinoma prevention screening recommendations are based on the distribution of adenomas throughout the colon, the age of the patient, and the patient’s family history (see Box 22.1 ). Although most adenomas occur distal to the splenic flexure, up to 40% occur in the proximal colon. At least one-fourth of patients have adenomas that are present only proximal to the splenic flexure. For this reason, full colonoscopy is the preferred method for screening in the United States over all other modalities. Yearly fecal immunochemical testing (FIT) is suitable for those who decline colonoscopy. Second-tier tests include CT-colonography every 5 years, FIT-fecal-DNA test every 3 years, and flexible sigmoidoscopy every 5 to 10 years. The current consensus regarding screening guidelines supports the use of colonoscopy starting at 50 years of age, although there have been recent recommendations to reduce the age to 45. Persons with a family history of colorectal carcinoma in a first-degree relative diagnosed before 60 years of age should undergo colonoscopy every 5 years beginning at 40 years of age, or 10 years before the age at which the patient’s relative was diagnosed with an adenomatous polyp. ^, African Americans should undergo initial screening at 45 years of age. Colonoscopy should be repeated every 1 to 10 years, depending on the findings at the initial screening, and the patient’s family history. See Chapter 2 for details of screening and surveillance.

Key Reporting Issues for Pathologists

Aside from the diagnosis of an “adenoma,” pathologists should also report the architectural pattern (tubular, tubulovillous, or villous), particularly when villous; the presence or absence of high-grade dysplasia; and the status of the polypectomy margins whenever possible. Small, “diminutive” tubular adenomas are usually excised in total, and these are usually removed in such a manner that it is often not possible to evaluate the margins of the specimen. This is normally acceptable from a clinical point of view because the endoscopist is keenly aware of the method used to excise the polyp, and as a result, the possibility of a lack of ability of the pathologist to evaluate margins when removed as such. Similarly, margins cannot normally be evaluated in large adenomas that are removed by piecemeal polypectomy. In this situation, the endoscopist would normally know whether the polyp was removed in total (and as such, determination of the margins by the pathologist is not necessary) and also whether a certain fragment of polyp represented the stalk of the polyp where the deep margin resides. In this case, the pathologist may issue a comment regarding the status of the piece of tissue that was designated as the polyp stalk. For polyps that are removed in piecemeal, it would be expected to find multiple pieces of tissue with cautery effect at the edges of tissue fragments, either with or without dysplasia. These do not normally represent the final margins of the polyp, however. In situations in which the margin cannot be evaluated, it is usually helpful to indicate this in the diagnostic report. In contrast, if a polyp is removed in one piece, then the status of the resection margins should be evaluated and commented on in the pathology report. If possible, the number of each polyp type should be provided; however, this is often difficult when multiple polyps are placed in the same jar. The endoscopic size of the polyp is used for postpolypectomy surveillance guidelines; however, if the endoscopic size is markedly different from the histological size, it is helpful to comment on this finding as well.

A key issue is risk stratification of patients based on the likelihood of developing advanced neoplasia in the future. Patients are divided into those with low risk (defined as 1 to 2 tubular adenomas <10 mm in size), intermediate risk (3 to 4 tubular adenomas <10 mm), and high risk defined as an advanced adenoma (adenoma with villous features, high-grade dysplasia, or adenocarcinoma, size ≥10 mm, or five or more adenomas) (see Box 22.1 ). Advanced adenomas require more aggressive colonoscopic surveillance given the increased risk for further metachronous adenomas and adenocarcinoma. Unfortunately, the degree of interobserver variability is high for determination of the presence and degree of villous architecture and the presence or absence of high-grade dysplasia, even among expert GI pathologists. However, the vast majority of villous adenomas, or adenomas with high-grade dysplasia, are ≥10 mm in size. Thus these polyps are considered “advanced” regardless of the presence or absence of these high-risk features, therefore the lack of interobserver agreement for villosity and/or high-grade dysplasia would not affect surveillance. Importantly, the finding of advanced morphological features should not be considered an indication for colectomy because advanced lesions have no potential for metastasis unless they harbor an invasive cancer.

Clinical Features and Pathogenesis

Foci of displaced or “misplaced” epithelium or mucin extravasation within the submucosa is a rather common finding in adenomas, particularly in polyps that have a long or pedunculated stalk, and those from the left colon versus the right colon. This may lead to a misdiagnosis of invasive adenocarcinoma. This pathological reaction occurs in 2% to 4% of adenomatous polyps; it tends to occur more commonly in the left colon and particularly in the sigmoid colon, where the intraluminal pressures are higher and peristalsis is more vigorous. It occurs in patients of all ages and is slightly more common among males. Displaced epithelium in adenomas is believed to occur secondary to peristalsis-induced twisting and torsion of the polyp stalk, which leads to vascular compromise, inflammation, breakdown of the muscularis mucosae, and eventually herniation of adenomatous epithelium into the submucosa. Thus this reaction occurs much more commonly in polyps that have a long stalk, which are more susceptible to twisting and torsion. Rarely, adenomatous glands can herniate into a submucosal lymphoid aggregate. This results in formation of a “lymphoglandular complex.” This phenomenon tends to occur in regions of the mucosa in which the muscularis mucosae is incomplete or shows breaches in its integrity and sites in which blood vessels enter the mucosa from the submucosa. Such adenomas tend to be sessile in structure, in contrast with the more common pedunculated polyp that more often contains epithelial displacement resulting from tissue injury.

Pathological Features and Differential Diagnosis

The intramucosal portion of an adenoma that contains displaced epithelium may have features that range from low- to high-grade dysplasia, and it may be tubular, tubulovillous, or villous. More often than not, foci of displaced epithelium show a similar degree of dysplasia to that in the intramucosal portion of the polyp. The epithelium in the submucosa is usually well-lobulated and composed of well-circumscribed aggregates of crypts in a pattern reminiscent of the mucosal portion of the polyp. Displaced epithelium is usually surrounded by a distinct rim of lamina propria, is often associated with hemorrhage or hemosiderin deposition in the stroma, shows no evidence of desmoplasia, and lacks the cytological and architectural features of invasive carcinoma ( Fig. 22.6 ). Mucin pools, when present in association with displaced epithelium, are usually either acellular or lined by dysplastic epithelium of a grade similar to that in the surface mucosal portion of the polyp. Furthermore, mucin pools are typically smooth and regular, and they are usually associated with extravasated mucin associated with ruptured crypts in the surface of the polyp. In contrast, the presence of irregular, small, jagged mucin pools dissecting through the submucosal stroma or the presence of mucin pools associated with cytologically malignant cells floating in the pools of mucin are features suggestive of adenocarcinoma.

Table 22.1 outlines features of adenomas with displaced epithelium from invasive adenocarcinoma. Features suggestive of adenocarcinoma include (1) epithelium that contains a greater degree of cytological and architectural atypia than that of the intramucosal portion of the adenoma; (2) architectural complexity, such as the presence of irregular, tortuous, and jagged crypts; (3) back-to-back and cribriform crypt formation; (4) a nonlobular configuration of crypts; (5) single or small clusters of cells without a surrounding rim of lamina propria; (6) a desmoplastic reaction; and (7) the absence of hemorrhage or hemosiderin deposition in the stroma ( Fig. 22.7 ). Although displaced benign crypts usually demonstrate a communication of the intramucosal portion of the polyp on deeper sectioning, this feature may also occur in cases with adenocarcinoma, but it is more common in the former compared with the latter.

Treatment

Adenomas with epithelial displacement should be excised in total, with confirmation that the mucosal and deep cauterized margins of the polyp are negative, similar to adenomas without epithelial displacement. For management purposes, pools of mucin associated with a rim of dysplastic epithelium at the deep resection margin should be considered “positive” for adenomatous epithelium. The significance of the presence of acellular mucin pools at the cauterized deep resection margin is unclear. As such, this should probably warrant close endoscopic follow-up. Cases in which a definite distinction from invasive adenocarcinoma cannot be made with complete certainty should probably be managed individually according to other clinical health factors (see Malignant Polyps). If invasion cannot be ruled out with certainty, but there are no unfavorable histological features, then polypectomy would be considered adequate treatment regardless. In almost all cases in which uncertainty regarding the presence of a small focus of invasive carcinoma exists, complete polypectomy will usually be sufficient because “unfavorable” histological features are unlikely to be present. However, in cases in which invasion cannot be ruled out with certainty, but the foci of uncertainty involve a deep or lateral resection margin, then surgical resection should be considered if the clinical circumstances support surgery.

Familial Adenomatous Polyposis

FAP is an autosomal dominant condition caused by either inheritance of a mutated APC gene (located on chromosome 5q) or a new germline mutation in the same gene (up to one-third of cases). Adenomas then develop from loss of the second APC allele in colonic epithelial cells. APC is a tumor suppressor gene, and its absence allows for additional mutations in other genetic loci, such as in the KRAS and TP53 genes (see Chapter 23, Chapter 27 ). FAP patients develop large numbers of colorectal adenomas during late childhood, adolescence, and early adulthood. FAP has been traditionally defined by the presence of more than 100 adenomatous polyps in the colon, although many patients have several hundred or even thousands of polyps. Adenocarcinoma develops in most patients by their mid-30s, but it can occur as early as 17 years of age. FAP accounts for approximately 1% of all colon cancers.

The frequency of FAP is from 1 in 8000 to 1 in 14,000 in the general population, with equal sex representation worldwide. FAP patients may express a variety of extraintestinal phenotypes (see later discussion), but all include the presence of adenomatous polyps of the GI tract. They are predisposed to a high rate of adenocarcinoma. Upper GI tract adenomas and adenocarcinomas may also develop in FAP patients at a high frequency; they are particularly prominent in the first and second portions of the duodenum and in the periampullary region. Periampullary adenomas and adenocarcinomas have recently become the most common causes of morbidity and mortality in FAP patients who have undergone prophylactic colectomy. The prevalence of noncolonic adenomas in FAP ranges from 9% to 50% in the stomach and from 50% to 100% in the duodenum. Much less commonly, patients with FAP develop thyroid carcinoma, brain tumors (discussed later), pancreatic carcinoma, and hepatoblastoma.

Cyclooxygenase-2 (COX-2) inhibitors have received considerable attention lately as a method of decreasing polyp burden in FAP patients or reducing or eliminating rectal adenomas in patients who have had a subtotal colectomy. COX-2 inhibitors may also decrease the extent of duodenal polyposis in FAP patients. ^, Gardner’s syndrome, Turcot’s syndrome, and attenuated FAP are considered subtypes of FAP; these are discussed in the following sections ( Table 22.2 ).

For most patients, the treatment of choice is screening during adolescence followed by a postadolescent prophylactic colectomy.

Variants of Familial Adenomatous Polyposis

Desmoid fibromatosis affects 15% of patients with FAP and has been referred to as Gardner’s syndrome, although use of this term is no longer recommended by the World Health Organization (WHO). ^{, ,} Desmoid tumors in these patients occur most commonly in the small bowel mesentery and abdominal wall. Surgical incisions can also increase the incidence of desmoid fibromatosis. Osteomas (particularly in the mandible, skull, and long bones), epidermoid cysts, dental abnormalities, and congenital hypertrophy of the retinal pigmented epithelium (CHRPE) can also occur. Overall, CHRPE affects 70% to 80% of FAP patients.

Central nervous system tumors can occur in association with both FAP and Lynch syndrome. This is referred to as Turcot’s syndrome, although this term is also no longer recommended by the WHO as many patients have been found to have constitutional mismatch repair deficiency. ^{, ,} Although medulloblastomas are the most common primary central nervous system tumor in FAP patients, astrocytomas and ependymomas can also occur. In contrast, the most common primary central nervous system tumor in patients with Lynch syndrome and constitutional mismatch repair deficiency is glioblastoma.

Attenuated Familial Adenomatous Polyposis

The term attenuated FAP refers to a hereditary colon cancer syndrome with fewer than 100 colonic polyps (usually <30). ^, Many such patients, on closer endoscopic examination or with the assistance of methylene blue staining, are found to harbor more than 100 polyps, so these patients may, in fact, have conventional FAP. When patients are seen with a suspicious family history, or mild polyposis, or both, the diagnosis can be further investigated by assaying for mutations in the APC gene. In attenuated FAP, as opposed to classic FAP, APC mutations tend to occur at the most 5′ or 3′ aspect of the gene. ^, Once the mutations are discovered, these patients require close endoscopic screening for removal of polyps and prevention of adenocarcinoma development. In addition to being fewer in number, adenomas and adenocarcinomas develop at a later stage in life compared with classic FAP; the lifetime risk of colorectal carcinoma in attenuated FAP is approximately 80%. These adenomas, and adenocarcinomas, are otherwise of unremarkable morphology, although many of the adenomas in some patients with attenuated FAP are “flat” (see earlier discussion). Upper GI lesions and extraintestinal manifestations develop at a rate similar to that observed in classic FAP.

Mutyh -Associated Polyposis

There is an alternative genetic mechanism for the attenuated FAP phenotype, one that is inherited in an autosomal recessive fashion (unlike the mutations that cause FAP). ^, In 2002, Al-Tassan et al. studied a family whose members harbored multiple colorectal adenomas and carcinomas, but lacked APC mutations. They discovered that affected patients carried biallelic mutations in the mutY DNA glycosylase gene (MUTYH) located on the short arm of chromosome 1, which encodes a base excision repair enzyme responsible for preventing mutations after oxidative DNA damage. Mutations in MUTYH results in G→T conversions throughout the genome. Other groups have detected homozygous or compound heterozygous MUTYH mutations in 15% to 40% of patients who have between 10 and 100 adenomas. ^, The highest frequency of MUTYH -associated polyposis is seen in patients with more than 30 adenomatous polyps, but who have no family history of polyposis. Patients with MUTYH -associated polyposis can also have multiple serrated polyps, suggesting some overlap with serrated polyposis. Similar to FAP and attenuated FAP, duodenal adenomas and adenocarcinomas occur in these patients. Sebaceous tumors have also been described in affected patients.

Two specific missense mutations (p.Y179C and p.G396D) account for most MUTYH -associated polyposis in northern Europeans. Other mutations are more common in southern Europeans (p.E480del), Pakistanis (p.Y104∗), and Indians (p.E480∗). One large population-based study of 2239 cases by Farrington et al. found a 93-fold increased risk of colorectal cancer compared with wild-type controls. The lifetime risk of colorectal carcinoma in affected patients is 80%, and the risk of duodenal carcinoma is 4%. The risk of cancer in heterozygous carriers of an MUTYH mutation remains unknown but is probably slightly increased. ^{, ,}

Other Adenomatous Polyposis Syndromes

Recently, germline mutations in the exonuclease domain of POLE and POLD1 have been associated with a clinical phenotype that overlaps with FAP and attenuated FAP. ^, Similar to APC mutations, mutations in POLE and POLD1 are dominantly inherited. Patients present with oligo adenomatous polyposis and early-onset colorectal and endometrial cancer. The risk of colorectal carcinoma appears higher for POLD1 mutation carriers compared with POLE carriers; however, carriers of both mutations warrant colonoscopic surveillance. The cancers that develop in these patients are characterized by a hypermutated phenotype, and they are amenable to immunotherapy.

NTHL1 is a protein involved in base excision repair similar to MUTYH. Mutations in NTHL1 result in C→T transitions and predispose to adenomatous polyposis that presents in adulthood, generally by 50 years of age. ^, Endometrial carcinoma and sebaceous skin tumors are also seen, which results in overlap with both MUTYH -associated polyposis and Lynch syndrome. Other more rare polyposis syndromes include MSH3 -associated polyposis and AXIN2 -associated polyposis. MSH3 encodes the MMR gene that is involved in repairing di-, tri-, tetra-, and pentanucleotide repeats. Biallelic mutations in MSH3 result in multiple adenomas and colorectal carcinoma. Duodenal adenomas, gastric cancer, and early-onset astrocytomas have also been reported. AXIN2 regulates β-catenin degradation and functions in the same pathway as APC. Mutations in this gene result in colonic polyps, colorectal cancer, gastric polyps, and ectodermal dysplasia and is inherited in an autosomal dominant manner.

Clinical Features

Hyperplastic polyps are considered serrated polyps with normal architecture and normal hierarchical proliferative characteristics. They are small, innocuous lesions that may be found throughout the colon of adults, but they are especially common in the rectum. Up to 90% occur in the left colon. Their prevalence rate increases with age. Up to 35% of asymptomatic individuals older than 50 years of age demonstrate hyperplastic polyps. Specific lifestyle and dietary factors commonly associated with conventional adenomas are also associated with hyperplastic polyps. These include cigarette smoking, alcohol consumption, low folate intake, and obesity. ^, These lesions are typically asymptomatic and are often identified as an incidental finding during endoscopic examination of the colon. Endoscopically, hyperplastic polyps are small, sessile, smooth bumps or nodules that appear pale and often flatten on insufflation of air.

Pathological Features

Microvesicular hyperplastic polyps are usually small or diminutive sessile polyps. Endoscopically, they can often have a stellate pit pattern, as described by Kudo et al. Histologically, they are composed of serrated epithelium with funnel-shaped, evenly spaced crypts and with proliferative zones confined to the crypt bases ( Fig. 22.9 ). Nuclear atypia is minimal, but some polyps show a mild degree of nuclear stratification at the base of the polyp, which should not be confused with a conventional adenoma in poorly oriented tissue sections. Dystrophic goblet cells and cells with round vesicular nuclei and prominent nucleoli are rare; if present, they are usually confined to the lower portions of the crypts. Most notably, these lesions show distinct evidence of surface maturation and an absence of cells with eosinophilic cytoplasm characteristic of SSLs and TSAs (see later discussion). The cells lining the crypts consist mainly of goblet cells and cells with abundant fine apical vacuoles containing microvesicular mucin (see Fig. 22.9 ). The nuclei in the cells in the surface epithelium are basally located and quite small in size. Importantly, the luminal epithelial serrations are usually limited to the upper one-half to two-thirds of the crypt in contrast with SSL (see later). When cut in cross section, the serrated crypts of these polyps have a uniform stellate appearance.

Some polyps also show thickening, and even extension, of the muscularis mucosae into the lamina propria. Occasionally, these prolapse-type changes can be quite striking and result in distortion of the serrated crypts, mimicking an SSL. In addition, larger-sized microvesicular hyperplastic polyps (>0.5 cm) may be slightly distorted and even show mild, often symmetric crypt dilation. In this instance, care should be taken to distinguish these changes from an SSL (see latter discussion). Occasional reports of inverted hyperplastic polyps exist in the literature. However, it is now apparent that an inverted growth phenomenon, wherein serrated crypts are present within the submucosa, occurs almost exclusively in SSLs. ^, It is similarly unclear how these lesions differ, if at all, from SSL with epithelial displacement.

Goblet cell hyperplastic polyps are the second most common type of hyperplastic polyp. They are typically small (<0.5 cm) sessile lesions and are commonly found in the left colon, but they can certainly be present in the right colon as well. Goblet cell hyperplastic polyps are often overlooked because their morphological alterations are usually quite subtle. These polyps show elongated crypts that are rich in goblet cells without microvesicular mucin ( Fig. 22.10 ). They are normally sessile lesions but demonstrate a far lesser degree of luminal serration than microvesicular hyperplastic polyps. The luminal section is usually limited only to the surface and uppermost portions of the crypts. The crypts may be mildly elongated as well. Nuclear atypia, stratification, and mitoses are not seen in this type of hyperplastic polyp. In these polyps, the crypts are taller and wider than normal. Occasional branching or tortuosity may occur, but this is rare.

Natural History and Treatment

Until recently, small hyperplastic polyps were not believed to require definitive treatment, although they are typically removed in the process of endoscopy with biopsy. However, the biological characteristics and natural history of hyperplastic polyps, particularly the microvesicular type, are now considered controversial. Large lesions, particularly those located in the right colon, should be removed in total. ^, Small proximal hyperplastic polyps may carry the same risk of synchronous advanced neoplasia as similarly sized SSLs, although the risk of metachronous neoplasia is unclear. Further studies are needed to determine the natural history and malignant potential of hyperplastic polyps and each of the different subtypes. There is a strong correlation between the occurrence of hyperplastic polyps and conventional adenomas in similar populations worldwide. Although some studies suggest that distally located hyperplastic polyps predict an increased risk of proximal adenomas or even advanced adenomas, this is controversial. Thus at present, the finding of a distal hyperplastic polyp at sigmoidoscopy is not considered to be an absolute indication for full colonoscopy. Furthermore, surveillance for detection of metachronous hyperplastic polyps, or adenomas, after removal of a hyperplastic polyp is not currently recommended.

Clinical Features

In contrast with hyperplastic polyps, SSLs are usually larger in size (>5 mm) and occur more often in the proximal colon (75%). ^, The prevalence of SSLs varies according to endoscopist experience and has been found in up to 20% of individuals undergoing screening colonoscopy by skilled endoscopists. SSLs are associated with female sex, smoking, high body mass index, and alcohol intake. ^, SSLs can be difficult to identify at endoscopy. However, awareness and specific training can improve detection. They often appear as ill-defined, sessile, pale lesions with an irregular shape and a “cloudlike” surface ( Fig. 22.11 ). ^, These polyps are often covered by adherent mucin, and a rim of bubbles or fecal debris can sometimes collect at the periphery of these lesions. They have a stellate configuration (Kudo pit pattern II) and appear as superficially elevated lesions (Paris classification 0-IIa). ^{, ,} Incomplete resection is common in SSLs given the lack of a sharp demarcation between the SSL and adjacent normal colonic mucosa.

Pathological Features

Morphologically, SSLs show distinctive features, characterized more specifically by the presence of asymmetric crypt dilation, crypt irregularity (horizontally shaped crypts), prominent and often exaggerated lower (and upper) crypt serration, mitoses in the upper levels of the crypts, vesicular nuclei in the upper crypts, reduced amounts of lamina propria between crypts, hypermucinous epithelium, and, occasionally, an inverted (epithelial displacement) growth pattern ( Fig. 22.12 ; see Table 22.3 ). Rows of cells with eosinophilic (“pink”) cytoplasm are not uncommon. These may represent “senescent” cells or possibly early dysplastic cells, but this topic is controversial. The basal portions of the crypts are usually branched and may appear flask- or boot-shaped (similar to an inverted T ), indicative of horizontal rather than vertical growth. An expert panel of pathologists suggested that the presence of one or more unequivocally architecturally distorted crypts of this kind situated at the base of the lesion is enough evidence to establish a diagnosis of an SSL rather than a hyperplastic polyp in the appropriate clinical and endoscopic setting. This definition has been adopted in the current WHO classification. Mature goblet cells, or mucinous cells, show an irregular distribution by being uncharacteristically located at the bases of the crypts. Some polyps also contain perineurial-like proliferations in the lamina propria.

SSLs may show a variable degree of nuclear atypia. At the low end of the spectrum, lesions show little or no stratification, a low mitotic rate, and clear evidence of surface maturation. SSLs may occasionally show an increased degree of nuclear stratification, consisting of cells with open vesicular chromatin and small nucleoli as well as increased eosinophilia of the cytoplasm. These features approach those of a traditional serrated adenoma, but the cytological changes are often focal and present only in the surface epithelium, not along the entire length of the crypts. This focal eosinophilic change should not be considered indicative of a TSA or an SSL with dysplasia.

It is important to emphasize that the size, location, number, and endoscopic appearance are not part of the pathological diagnosis of a hyperplastic polyp or an SSL. Using these parameters can hamper future efforts to refine the risk profile of patients with serrated polyps, especially in scenarios such as small proximal hyperplastic polyps or large distal hyperplastic polyps. Only a proper morphological classification of serrated polyps will allow for refinement of the risk of metachronous neoplasia in these patients. Results from such analyses can then be used to modify the diagnostic criteria and include other features for the diagnosis if necessary. From a practical perspective, it is extremely rare to encounter a microvesicular hyperplastic polyp >9 mm, and such polyps should be considered advanced serrated polyps for the purpose of subsequent surveillance intervals. The main dilemma occurs when encountering small or diminutive proximal serrated polyps that morphologically are best classified as a microvesicular hyperplastic polyp (see Fig. 22.8 ). Furthermore, goblet cell hyperplastic polyps are not infrequently seen in the proximal colon. Despite the lack of diagnostic features, there has been a tendency to consider all proximal serrated polyps as SSLs because of the presumption that hyperplastic polyps do not exist in the proximal colon. Such a diagnostic strategy will lead to an increased frequency of colonoscopy for these patients and an increased burden on surveillance programs. In rare cases of polyps with marginal histological features, size and location can be used to tip the diagnosis one way or another.

Natural History

The natural history and risk of progression to malignancy of SSLs are poorly understood but are under intense investigation. The mean age of patients with SSLs with dysplasia is similar to SSLs that harbor invasive carcinoma, suggesting that once dysplasia occurs in an SSL, progression to invasive adenocarcinoma occurs fairly rapidly. It is commonly believed that loss of DNA repair ability and subsequent microsatellite instability are mechanisms that drive rapid neoplastic progression. Although SSLs that harbor dysplasia progress rapidly to carcinoma, the acquisition of dysplasia in SSLs takes years to occur. ^{, ,} In one study of 55 patients with SSL who were monitored for a mean of 7 years, colorectal cancer or high-grade dysplasia developed in 15% of patients. In other studies, patients with SSL were found to be at increased risk for synchronous advanced neoplasia and interval adenomas during surveillance. One rational proposal for the management of SSL is based on the fact that most traditional hyperplastic polyps (goblet cell or microvesicular type) are unlikely to progress to carcinoma, and that SSLs without dysplasia are probably slow, but progressive, lesions. Table 22.4 provides a summary of the recommendations from the U.S. Multi-Society Task Force on Colorectal Cancer, British Society of Gastroenterology, and Cancer Council Australia.

Treatment

For SSLs without dysplasia, complete endoscopic removal is recommended. ^, However, if this is not possible, then repeat endoscopy with biopsies should be performed within 1 year to evaluate for signs of dysplasia, and this should be followed by continued surveillance at shorter intervals until the lesion is completely removed. Surgical excision for large SSLs without dysplasia, particularly those that are recurrent, is also a reasonable alternative, although this is becoming less common with the increased use of advanced endoscopic techniques. In contrast, complete excision is considered mandatory for SSLs with dysplasia based on the likelihood that these lesions have undergone hypermethylation and acquired microsatellite instability and therefore are prone to carcinoma progression. Complete excision should be accomplished by endoscopy or surgical resection. If removed by piecemeal polypectomy, patients usually undergo repeat endoscopy within 6 months to 1 year to ensure that the lesion has been removed in its entirety and that there has been no progression.

Pathological Features

Most TSAs are composed of epithelial cells with abundant and intensely eosinophilic cytoplasm and bland oval-shaped palisaded nuclei, usually without significant mitotic activity ( Fig. 22.14 ; see Table 22.3 ). ^{, ,} The second and probably most typical feature of TSA is the presence of distinctive slitlike serrations, similar to the narrow slits present in normal small-intestinal epithelium. ^{, , , ,} The association of both the traditional serrated adenoma cytology and slitlike serrations makes the diagnosis of traditional serrated adenoma straightforward. A third characteristic feature is the presence of ectopic crypt foci. Ectopic crypt foci are small buds of epithelial cells resembling the bases of normal crypts that are not anchored to the muscularis mucosae, but instead are situated perpendicular to the villous projections of the polyp. Ectopic crypt foci are not specific for TSAs. They occur in other polyps, such as conventional adenomas, and as such may simply represent small zones of proliferation in polyps with a villiform growth pattern. ^, Interestingly, flat TSA often lack ectopic crypt foci. Some TSAs with typical slitlike serration and at least focal ectopic crypt foci may show a predominance of goblet cells (mucin-rich variant) rather than the characteristic eosinophilic cells with abundant cytoplasm. ^,

Architecturally, TSAs are often protuberant and villous, but they can grow as flat, superficially elevated lesions as well. Filiform traditional serrated adenomas show elongated, finger-like villous projections and often show inflammation, ulceration, and dilated lymphatics within the lamina propria. Up to 50% of TSAs reveal an adjacent area of nondysplastic serrated precursor, either a hyperplastic polyp or SSL. Distal TSAs may show an adjacent “shoulder” area of flat growth at the edge of the protuberant polyp. In a recent study, small (<10 mm) polyps resembling these shoulder areas were shown to be early forms of TSA that can be recognized by subtle superficial serrations and cytoplasmic eosinophilia. ^,

Despite being a polyp with malignant potential, most TSAs do not reveal the classic cytological characteristics of “dysplasia” because they lack significant nuclear hyperchromasia and mitotic activity outside of the areas of ectopic crypt foci. It is controversial among experts whether the “baseline atypia” of TSAs represents a very low grade of dysplasia (because molecular aberrations are common in these lesions) versus truly non-dysplastic “senescent epithelium” (based on the absence of significant cytological atypia). Regardless of the term used to describe the cytological features of TSAs without traditional high-grade dysplasia, TSAs are nonetheless considered premalignant polyps because most, if not all, will eventually progress through various stages of overt (conventional) dysplasia to cancer if left untreated. Overt (conventional) dysplasia does develop in TSAs with progression. The dysplasia in TSAs can be either intestinal type, resembling dysplasia seen in conventional adenomas, or serrated type, characterized by tightly packed serrated glands composed of epithelial cells with prominent cytoplasmic eosinophilia, enlarged nuclei, and prominent nucleoli (see Fig. 22.14 ). ^, Both high-grade intestinal and serrated dysplasia should be reported when present because this indicates a more advanced stage of the lesion. However, it is also unclear whether “advanced” TSAs, as such, portend any increased risk of further synchronous or metachronous neoplasia or risk of cancer after excision compared with conventional TSAs without high-grade dysplasia. Furthermore, it is similarly unclear whether the risk of progression is affected by the type (intestinal vs. serrated) or extent of dysplasia in the polyp.

Natural History and Treatment

The natural history of TSA and the risk of progression to malignancy are poorly understood given its infrequent occurrence as an isolated polyp. ^, Most TSAs occur synchronously with other polyps. An additional complication is the occasional lack of use of consistent terminology, as mentioned earlier. For instance, the terms TSA with associated SSL versus SSL with TSA-like dysplasia when the lesion occurs on a background of SSL or a hyperplastic polyp. Thus evaluation or progression has been challenging. Nevertheless, progression to high-grade dysplasia was reported in 37% of TSAs in one study. In this study, 11% contained intramucosal adenocarcinoma. In another study of filiform TSAs by Yantiss et al., 22% showed high-grade dysplasia, and another 6% showed invasive adenocarcinoma. Some studies have suggested that the rate of malignant transformation in TSAs is similar to that of conventional adenomas. ^, It is likely that the risk of progression is related to the size and location of the lesion. Large TSAs in the proximal colon may progress at a more rapid rate than those in the left colon.

Prior studies suggesting a high risk of concurrent colorectal carcinoma and metachronous adenomas in patients with a baseline TSA were limited by a small number of patients. ^, More recently, in a relatively large study of Korean patients with TSAs, the risk of metachronous serrated polyps, conventional adenomas, and high-risk adenomas was higher than a control population with only baseline conventional adenomas. A recent population-based study from Denmark also showed that patients with TSA were nearly five times more likely to develop colorectal carcinoma on follow-up compared with those without TSA. The treatment of TSAs is complete endoscopic removal. The surveillance interval is similar to that for patients with conventional adenomas (i.e., every 3 to 5 years).