There are numerous terms applied to the anal canal (see terminology at the end of the chapter), but the only important one is that relevant to tumors. This includes the dentate line; the transitional mucosa that extends proximally for a variable distance, which varies from millimeters to a few centimeters, but cannot be seen with the naked eye; and the squamous mucosa distal to the dentate line.

Critical terms that are used for various anatomic components that are helpful in understanding the anatomy and histology of this region are as follows (Fig. 21-1):

There is much confusion over the normal anatomy and histology of the anal canal for a variety of reasons. This is primarily because of the persistent use of terminology that has become obsolete but which is still perpetuated. We attempt to walk you safely through this minefield. The one landmark to which all definitions (and problems) need to be related is the dentate line, also known as the pectinate line, which is the single, reproducible, constant, but wavy line that runs circumferentially around the anal mucosa and which is easily visible in the opened, resected specimen (Fig. 21-1) or on retroflexion of the endoscope at sigmoidoscopy; it marks the embryologic junction between the cloaca above and the proctodeum below. It is at the lowest part of the anal transitional zone (ATZ), which is lined largely by transitional or squamous mucosa that extends upward for a variable distance before becoming colorectal mucosa (Fig. 21-2). The more commonly used name for the line is dentate, not pectinate, and that is the name used throughout the rest of this chapter. Immediately below the dentate line is the nonhair pecten (Fig. 21-3) (or anal
margin—another term we could do without), which is covered by moist, nonkeratinizing squamous mucosa, similar to that in the mouth and vagina, and which continues down to merge imperceptibly with the hair-bearing perianal skin (Fig. 21-3).

The surgical anal canal is defined as that part of the distal intestinal tract enclosed by the internal sphincter muscle, namely, from the levator ani muscles of the pelvic floor above to the skin of the anal verge below. This definition, attributed to Symington in 1888,² certainly made life easy for practicing surgeons, who can see the levators from either direction. It is probably already apparent that the concept of the surgical anal canal has at least two major drawbacks:

1. It includes the rectum, which is a separately recognized part of the gut with clear definitions of its location.

2. In classifying tumors of this region, we ignore the concept of the surgical anal canal and its all-inclusive mucosae and skin. Thus, using the definition of surgical anal canal literally, we should call many rectal adenocarcinomas adenocarcinomas of the anal canal, which we obviously do not, because we use the definition of anatomic rectum to classify them. We include squamous mucosa of the ATZ and pecten in our concept of anal carcinomas, because they arise in the most distal part of the gut, which we call the anus. Nevertheless the AJCC also defines the anal canal as extending from the apex of the anal sphincter complex to the palpable intersphincteric groove at the distal edge of the internal sphincter muscle. This may be useful in patients, but is tough to translate to histology in an opened abdomino-perineal resection.

The obvious solution is to utilize the concept of the histologic anal canal, which extends from the termination of rectal mucosa to the perianal skin. It includes the ATZ and pecten. The histologic anal canal is largely surrounded by the external sphincter muscle. Everything that is lined by the rectal epithelium is assumed to be rectal, and all adenocarcinomas arising from this mucosa are, by definition, rectal. Using this definition of the histologic anal canal, carcinomas arising in the mucosa of the ATZ, from the dentate line up to the junction with rectal mucosa, should be included in carcinomas of the anal canal as the transition zone is the proximal part of the anal canal. Those arising from the squamous pecten immediately below the dentate line are also included in carcinomas of the anal canal. Finally, carcinomas arising in the hair-bearing perianal skin are called carcinomas of the perianal skin (logic wins in the end).

Distally, the pecten merges imperceptibly with the perianal skin. In the late 19th century, Hilton thought that he could see this junction, as marked by the white line that bears his name.³ However, there is no histologic counterpart to this line, and in many but not all patients, it is indiscernible; the term has therefore largely fallen out of use. This junction is, however, marked grossly by a change over a few centimeters during which the skin becomes more pigmented and bears hairs, gradually becoming the skin of the anal verge and the perianal region; the same changes are reflected histologically, but keratinization also occurs in the perianal skin.

With this insight, let us take a more detailed look at the remaining anatomic structures and microscopic appearance of the anal canal.

The anal canal develops from the keratinizing proctodeum distally and the nonkeratinizing cloaca proximally, their junction being the site of the dentate line. We include all of these in the histological anal canal. Immediately above the dentate line, 6 to 10 longitudinally oriented anal columns (of Morgagni) may be visible merging with the rectal mucosa above, particularly in young patients (Fig. 21-1). They may disappear with age. These columns are joined at their bases by small semilunar valves, which form papillae, and these structures define the dentate line. The dentate line may be difficult to identify when valves and papillae are not readily identified; it is then best defined by the opening of the lowest visible sinus. Immediately behind the columns, small pockets, the anal sinuses (or crypts), may be found. Some of the openings of the anal ducts and their glands (Fig. 21-4) are into these sinuses, but occasionally they are found further up the transition zone. The proximal extent of the ATZ is the rectum, and the junction between the two mucosae can usually be appreciated grossly because of the pink color to the rectum and the more gray color of the ATZ.

Distal to the dentate line, the squamous-lined pecten is smooth, whitish, and shiny in life, but in resected specimens that have been pinned out, it is often transversely wrinkled, and the wrinkling may extend up into the transition zone. The junction between pecten and perianal skin often is not

sharp but is marked grossly by a change over a few centimeters during which the skin becomes more pigmented and bears hairs, gradually becoming the skin of the anal verge and the perianal region.

This is far easier to see in diagrams than it is to demonstrate anatomically (Figs. 21-1A and 21-4A). It consists of three parts. The first is the internal anal sphincter, which is a continuation and modification of the circular muscle of the rectum. This is largely surrounded by the external sphincter muscle, the second part of the anal musculature, which has subcutaneous, superficial, and deep parts. The subcutaneous part is immediately related to the perianal skin and the most external parts of the anal canal. The superficial part sweeps around to reinforce the bulk of the internal sphincter on all sides, and the deep part runs into the third part of the anal musculature, the levator ani, which is attached to the lateral pelvic walls. Part of this is the sling of muscle called puborectalis, which causes angulation of the rectum, partially preventing incontinence. This sling of muscle is held in place posteriorly by the anococcygeal ligament and anteriorly is inserted into the perineal body.

The blood supply of the upper part of the anal canal comes from the inferior mesenteric artery through its termination in the superior rectal artery (Fig. 21-5). The latter bifurcates on the posterior surface of the rectum, each branch passing laterally (the lateral rectal arteries), piercing the muscle at several sites to enter the submucosa down to the level of the columns of Morgagni. The middle sacral artery arises from the posterior wall of the aorta immediately above its bifurcation and supplies a branch to the posterior rectal wall. The distal anal vasculature comes primarily from the internal pudendal arteries, which divide into the inferior rectal artery and several other branches that supply the sphincter muscles and then the mucosa. These anastomose with branches from the lateral rectal arteries and the gluteal and perineal arteries.

The venous system is of greater clinical significance in view of its frequent dilatation to form hemorrhoids. Two major submucosal venous plexuses are present, the internal and external, which communicate with each other across the dentate line (Fig. 21-6).⁴ The internal hemorrhoidal plexus is located immediately above the dentate line. On penetrating the muscularis propria, there is a third perimuscular rectal plexus of veins. Most of the veins from this plexus drain up through the superior rectal vein into the portal circulation via the inferior mesenteric vein, but there is also a well-defined anastomosis with the external hemorrhoidal plexus, which is located distal to the dentate line. The greatest concentration of veins is immediately above the dentate line in the columns of Morgagni, and it is these veins and/or those of the external plexus that enlarge and prolapse and may ultimately undergo all of the complications of hemorrhoids. For this reason, excision of internal hemorrhoids invariably includes part of the dentate line and the accompanying transitional epithelium. The external hemorrhoidal plexus is also submucosal and drains through the inferior and middle hemorrhoidal veins back into the internal pudendal veins, and thence into the internal iliac veins and vena cava. In addition, the communications between the internal and external hemorrhoidal plexus form an anastomosis between the portal and systemic venous systems, so that hemorrhoidal bleeding may be a reflection of portal hypertension.

Lymphatic drainage from the pecten is primarily to the inguinal and from there to the iliac nodes. Above the dentate line, lymphatics go to both to the internal iliac nodes and the inferior mesenteric nodes, the latter via the lateral (perirectal) and superior rectal artery groups of nodes. Tumors of the anal canal therefore always require careful observation of these groups of nodes. Fortunately, the inguinal nodes are superficial and therefore accessible to fine needle aspiration, biopsy, or excision.

From proximal to distal, the epithelial types found in the distal rectum and anal canal are the following.

Columnar zone. It is the most distal part of the histologic rectum but included in some of the archaic definitions of the anal canal (see previous discussion). It consists of typical colonic crypts; however, the last centimeter or so often appears atrophic, with thin mucosa and a reduced number of crypts that may also be bifid (Fig. 21-2). Further, occasional crypts may be found in isolation within the ATZ; although by light microscopy these often appear to be buried beneath the transitional epithelium (Fig. 21-2C), by scanning electron microscopy, they open onto the surface, forming typical but isolated crypt units virtually identical to their more proximal counterparts.⁴ Occasionally, Paneth cells may be found in the crypts embedded within the transition zone. Rarely, the


crypts appear to be more like pyloric glands and may even have fundic gland mucosa complete with parietal cells.⁵ These metaplasias are probably a reaction to injury, although the presence of parietal cells suggests that they may be heterotopic mucosa.

Anal transition zone. This zone extends from the crypt-bearing distal rectal mucosa above to the uninterrupted pectinate squamous mucosa below. The ATZ usually reaches 0.5 to 1 cm above the dentate line. It is sometimes very short and on occasion may not be present at all, although that is unusual.⁶ On the other hand, it may extend as much as 20 mm above the dentate line to 6 mm below it. The typical ATZ epithelium bears a superficial resemblance to urothelium (hence “transitional”). It consists of four to nine layers of cuboidal cells with their nuclei oriented perpendicular to the basement membrane (Fig. 21-2). However, the superficial cells can be flattened, cuboidal, polygonal, resemble umbrella cells of the bladder, or columnar. In ultrastructural studies, it was found to share features of squamous epithelium and urothelium. The length of the ATZ lined by transitional epithelium varies greatly from one person to another, and in the ATZ it may be mixed with rectal crypts proximally and mature squamous epithelium, especially distally. The amount and distribution of each epithelium in the ATZ is totally unpredictable. Really rarely, the transition zone may be lined by simple columnar mucosa resembling gastric surface epithelium but with larger ATZ.⁷ If a histologic section is taken through an apparent junction between rectal mucosa and the ATZ, the transitional-type epithelial sometime extends proximally much further than anticipated macroscopically, and the rectal mucosa may extend more distally. Sometimes transitional-type mucosa may also extend distally for a few millimeters beneath the dentate line.

Squamous zone distal to the dentate line. This consists of nonkeratinizing squamous mucosa with short, stubby stromal papillae but no skin appendages. Distally the mucosa contains increased numbers of melanocytes, and at some point it forms a granular layer with subsequent keratinization (Fig. 21-3).

Perianal skin. This is keratinized squamous epithelium typical of skin with appendages, primarily hair follicles, sebaceous glands, and apocrine glands. The epidermis has well-developed papillae and prominent melanocytes (Fig. 21-4).

Anal ducts and glands. These are tubular structures that extend from the ATZ into the stroma, including the internal sphincter muscle at the level of the dentate line. Their lining is as unpredictable as is the ATX epithelium. The ducts may be lined by transitional epithelium, pseudostratified epithelium resembling Fallopian tube epithelium, but without cilia, and goblet cells may be scattered around as well. The glands are simple tubular mucous glands that empty into the ducts (Fig. 21-5). We have seen a few cases in which the glands contain typical prostatic glandular epithelium. The ducts open into the anal valves and sinuses at the level of the dentate line and often above. At the lower end of the ATZ a further series of simple perianal glands are occasionally found. These glands secrete a mixture of sulfomucins and sialomucins⁸ that forms a thin layer over the ATZ.⁴

The upward extension of squamous epithelium from the dentate line has distinct implications if suction biopsy is being carried out for possible Hirschsprung’s disease. It is well known that the most distal part of the large bowel immediately above the dentate line is normally relatively deficient in ganglion cells and may also have thickened nerve trunks, thereby mimicking the changes found in Hirschsprung’s disease. For this reason, it is always advocated that biopsies be taken at least 2 cm above the dentate line. When biopsies contain both changes in the nerves, suggesting Hirschsprung’s disease but also transitional zone epithelium, the only sensible conclusion is that the specimen is too far distal for diagnostic purposes and that the biopsy should be repeated.

Hemorrhoids tend to be a disease of Western society and have been traditionally associated with a low-residue diet. It has been noted that hemorrhoids are relatively uncommon in the Third World countries and are virtually unknown in animals, suggesting that straining to pass hard stool plays a role. In addition, the underlying elastic and connective tissue may lose its abilities to recoil with aging. Prolapse may be potentiated by hyperactivity of the internal sphincter muscle. Although they are traditionally thought of as dilated varicose veins of the hemorrhoidal plexus, bleeding from them is characteristically bright red, suggesting that they may be a variation of arteriovenous fistula reminiscent of erectile tissue.¹⁷, ¹⁸ In older patients and particularly in women, hemorrhoids may be associated with anal prolapse.

Small hemorrhoids cause virtually no problem. The most common manifestation is bleeding, manifested as streaks of blood on the outside of the stool. When hemorrhoids become larger, they may prolapse and fail to return following defecation, thereby drawing attention to themselves. Prolapsed hemorrhoids may produce a mucoid discharge, but pain is usually a feature only when they are acutely prolapsed, inflamed, thrombosed, or infarcted.¹⁹ Internal hemorrhoids can usually be distinguished from external hemorrhoids by the reddish rectal mucosa that covers the upper portion and the ATZ mucosa distally (Fig. 21-8).

A variety of surgical and nonsurgical techniques are used to treat hemorrhoids, including injection with sclerosants, banding, cryosurgery, electrocoagulation, and laser therapy.

There is a surprising variability in the microscopic appearance of hemorrhoids. They are often less prominent immediately following resection because blood is lost from the numerous channels present, which collapse and are less impressive histologically than they would have been if distended. The dominant change is the presence of dilated vessels, mainly veins, of variable wall thickness in the submucosa. Other hemorrhoids consist primarily of numerous small vessels, along with fewer larger vessels (Fig. 21-8). Thrombosis and recanalization may be seen, and superficial ulceration may also be present. When colonic crypts are present, the lamina propria between them may be partially obliterated by fibromuscular tissue derived from but perpendicular to the muscularis mucosae. The fibromuscular proliferation is a reaction to local prolapse. In the adjacent tissue these may be marked neuronal hyperplasia, although the significance of this is unclear.²⁰ The overlying epithelium may consist primarily of transitional or rectal mucosa in internal hemorrhoids or nonkeratinizing squamous pectinate mucosa and perianal skin in external ones. In addition, the epithelium may be acanthotic and keratinized, presumably in reaction to trauma. This may make the surface look white, that is, leukoplakia. One annoying surprise is the rare finding of squamous epithelial dysplasia.²¹ This will be dealt with later in the section of squamous carcinoma and its precursors.

In general, hemorrhoids are sent to the pathologist because hospital regulations require that all excised tissue be sent. Some pathologists therefore document that tissue has been received but do not examine it microscopically. We routinely examine hemorrhoids to detect unsuspected lesions such as dysplasia or even invasive squamous carcinoma, although the yield is minute, and such examination may not be cost effective. On the rare occasions when we have found invasive squamous carcinoma, the surgeon admitted
that the appearance of the hemorrhoids was “not quite right” but was still surprised at the diagnosis.

Even if hemorrhoids are examined microscopically, there are virtually no data regarding the number and size of submucosal veins in the normal tissue for comparison with the abnormal tissue. Most samples with the typical dilated vessels are signed out simply as hemorrhoids. If thrombi or ulcers are present, they do not change the diagnosis. We do not mention if they are internal or external hemorrhoids, since that is a clinical decision, despite the fact that the mucosa covering them is different.

Minor degrees of acute and chronic inflammation are frequently present in the submucosa and around anal ducts. Some fibroepithelial polyps may be secondary to resolved fissures. Although the anatomy of fistulae is important surgically, and their anatomic origin and course are clinically important, specimens received for histologic examination usually consist of acute and chronic inflammatory cells, granulation tissue, sometimes abscesses, and fibrosis. The latter may also be present in the internal sphincter muscle away from the acute lesion, but not in controls, raising the question of a primary underlying lesion in the muscle.²³ Mucosa or perianal skin may be present.
The major clinical question when the tissue from a fissure or fistula is seen is whether, in a patient without known underlying Crohn’s disease, the fissure or fistula may be the anal or perianal component of Crohn’s disease (see the chapter on Crohn’s disease). In these tissues, occasional foreign body giant cells or even rare small granulomas are found, resulting from a reaction to contents of the bowel or sometimes to substances that have been applied topically. These should not be interpreted as Crohn’s disease. If more substantial granulomas are found, then Crohn’s disease or even infections become possibilities, and the clinicians should be told. Even more suggestive of Crohn’s disease are numerous lymphoid aggregates similar to those in the transmural inflammation in other parts of the bowel, especially if accompanied by granulomas (Fig. 21-10). Like Crohn’s disease, tuberculosis affects the anus and perianal skin. As with all anal specimens, the overlying mucosa should be examined for the presence of condylomas and dysplasia.²¹

The anus is subject to a variety of infections, some of which occur simultaneously with rectal involvement. The most common anal infections are discussed below. Refer to more detailed discussion on GI infections in Chapter 19.

Viral infections—herpes and papillomavirus. Anal and perianal herpes virus infection occurs, but it is usually part of more extensive rectal involvement. The lesions involving the pecten and perianal skin are identical to those in other squamous surfaces, with the typical nuclear inclusions in squamous cells. Human papilloma virus infection is the most important anal and perianal infection because of its association with anal carcinomas. This is discussed subsequently.

Syphilis. Primary chancres are usually suspected clinically, and it is only if the presentation is atypical, such as a fissure, that the lesion is likely to be biopsied. The diffuse plasma cell infiltrate, particularly in association with endarteritis, should alert the pathologist to the diagnosis. If there is a high level of clinical suspicion, this diagnosis should be considered. Appropriate silver or immuno stain should be carried out. However zoon can produce a similar dense plasmacytic infiltrate.

Condylomata lata, the proliferative lesion, can also occur in the anal region as a manifestation of secondary syphilis. The lesions have no specific features but can mimic a variety of other dermatological conditions or look like inflamed skin tags. The inflammatory infiltrate can be a mixture of all types of chronic inflammatory cells including immunoblasts that may lead to a consideration of lymphoma (see lesion in gastritis Chapter 13), and plasma cells may or may not predominate. While the endothelium may be prominent it is usually no more so than in many other inflammatory lesions. The overlying epithelium is usually hyperplastic, spongiotic, and has focal neutrophils (Fig. 21-9). The diagnosis usually take a high degree of clinical suspicion as all tissue examined from inflammatory anal lesions other than overt abscesses are
potentially good starters for syphilis, but it is obviously impractical to screen them all. Unless really “on the ball,” something other than the morphological appearances usually has to trigger the search for the myriad organisms present. On immunostains organisms predominate in the basal zone of the squamous epithelium, usually in the spongiotic intercellular space (Fig. 21-9D). They are less frequent in the lamina propria, and uncommon in the deep inflammatory infiltrate, but when present do tend to be around endothelium.

Lymphogranuloma venereum. This disease caused by the bacterium, Chlamydia trachomatis, is another disease in which the histologic findings are nonspecific, consisting primarily of granulation tissue with lymphoid aggregates. Conversely in cases with rectal lymphoid hyperplasia (rectal tonsil) a possibility of Chlamydia infection should be considered in the differential diagnosis. In the later stages, the resultant scar may cause severe strictures.²⁴

Granuloma inguinale. This disease frequently presents with perianal ulcers. The biopsy findings display a heavy plasma cell and neutrophil infiltrate. Pseudoepitheliomatous hyperplasia may be present. The characteristic Leishman-Donovan bodies may be found in macrophages and are best identified using either Warthin-Starry or Giemsa stain. The causative organism is the bacterium Calymmatobacterium granulomatis.

Squamous epithelial hyperplasia, also known as acanthosis, is analogous to similar lesions occurring elsewhere, particularly in the oral mucosa. It may be found in the epithelium covering hemorrhoids as a reactive change, usually in the ATZ, possibly due to prolapse, but it may also be an upgrowth of the pecten over the transition zone following ulcers (Fig. 21-13). Thick isolated plaques of squamous hyperplasia occur in the midst of chronic inflammation. Some of these are so proliferative that they look invasive at their bases, making the low-power distinction from an unusually well-differentiated squamous carcinoma nearly impossible. However, the basal layer lacks the atypia associated with squamous carcinomas, supporting the notion that these are reactive and pseudoepitheliomatous hyperplasia. Unfortunately, there is virtually no literature on these lesions, so we have no clue as to their behavior and etiology. A few cases may have been reported as keratoacanthomas, but they lack the central keratin-filled cup.

On occasions, reactive acanthosis is keratinized, so that it looks white, much like comparable lesions in the perianal skin, where the squamous hyperplasias commonly is covered by a thick layer of hyperkeratosis, which results in a clinically visible thick, whitish plaque, sometimes well circumscribed, that is referred to as leukoplakia. However, even dysplasia can be hyperkeratotic and look like any other white plaque or leukoplakia.

These squamous proliferations are venereally transmitted human papillomavirus (HPV) infections. Many are warty or papillary growths that are soft and have a predilection for moist surfaces, such as the genitalia and perineum, as well as the anal canal and perianal skin. Others are flat or plaque-like lesions, and are more likely to be found in the anal mucosa. They are mostly small lesions. The largest lesions are discussed below as giant condylomas, now known as verrucous carcinomas. Multiple lesions are common (Fig. 21-14).
HPV types 6 and 11 are by far the most common types in simple condyloma, but oncogenic types, especially 16 and 18 can also be detected but less frequently.²⁹, ³⁰, ³¹, ³², ³³, ³⁴, ³⁵, ³⁶, ³⁷, ³⁸, ³⁹, ⁴⁰

Condylomas are characterized by thick acanthotic epithelium with papillomatosis, that is, thin columns (papillae) of fibrovascular stroma extending high into the thick epithelium (Fig. 21-15). The rete ridges are elongated, sometimes branched and broad. The basal layer is only one to three cells thick; its nuclei are round or slightly elongated, but polarity is maintained. Maturation is normal. Some superficial epithelial cells have koilocytic features, the histologic hallmark of HPV infection. Koilocytosis is discussed in much more detail in textbooks and papers dealing with uterine cervical squamous proliferations. Koilocytosis is characterized by the presence of large squamous cells that contain one or two nuclei, many of which are crinkled resembling raisins (sometimes referred to colloquially as “raisinoid nuclei”), and these nuclei are surrounded by a halo of clear cytoplasm. There is little nuclear atypia. Immunoreactivity to HPV is commonly present. Condylomas that arise in the anus sometimes extend proximally to cover the distal most rectal mucosa. In fact, we have seen a few cases that appeared to endoscopists as if they were rectal polyps. A major concern associated with condylomata is the development of dysplasias and squamous carcinomas within them. There is also a strong association with genital condylomas especially in the uterine cervix. It is suggested that in a woman with anal condylomas, cervical smears be performed if they have not been done recently, as anal, vulvar, vaginal, and cervical lesions can all coexist. However, HPV typing has yet to emerge in the management algorithm of these lesions as it tends to be reflected in the grade of dysplasia. When this is clearly low grade or high grade this is easy, but this is not always easy.