The most common type of cancer arising in the anus is squamous or epidermoid carcinoma (Table 1) 356. This is, however, a rare tumour, and is 20 to 30 times less common than colorectal cancer. Of the other tumours that occasionally arise in the anus, only malignant melanoma is seen with any frequency; this is eight times less common than epidermoid carcinoma. Other tumours are so rare that there is only anecdotal experience of their treatment.

Anal epidermoid cancer is sensitive to both chemotherapy and irradiation. This has led to reappraisal of the surgical treatment for this condition, with a trend away from abdominoperineal resection to a much more conservative surgical approach that allows preservation of the anal sphincter mechanism.

Much of the early information on anal cancer was derived from small, retrospective studies that used a variety of pathological terms and staging systems. Other studies combined their results with other types of anal cancer such as melanoma, adenocarcinoma, etc. Data were therefore conflicting. In addition, there has never been any agreement as to the anatomical boundaries which distinguish the anal canal from its margin; epidermoid cancers arising at these two sites have, therefore, not been accurately differentiated. As epidermoid cancer arising in the canal differs fundamentally in its clinical presentation, pathological features, treatment, and prognosis from that at the margin, distinction between the two conditions is mandatory.

Anatomists regard the anal canal as that part of the alimentary tract distal to the rectal ampulla. However, this definition does not correlate with the clinical and pathological characteristics of anal cancer in that region. Although there is general agreement that the proximal end of the anal canal correlates with the anorectal ring, its distal limit is more poorly defined; both the dentate (pectinate) line and the anal verge have been used (Fig. 1) 1188. Similarly, there is no clear definition of the lateral border of the anal margin, although one authority has recommended that it lies within a 5-cm radius of the verge itself.

The definition of the distal limit of the anal canal as the dentate line or the anal verge is of paramount importance since the anatomical boundary used to differentiate between the anal canal and its margin determines the relative incidence of tumours at these two sites. When the anal verge is used as the boundary, less than 15 per cent of cancers are ascribed to the margin, whereas those authorities using the more proximal dentate line claim that 30 per cent of tumours occur at this site.

The dentate line itself is a cause of confusion. It corresponds to the site of the anal valves and represents the junction of the postallantoic gut with the proctodeum. It is described as the ‘anal transitional zone’ because of its transitional cell appearance with characteristics of both the cuboidal rectal mucosa above and the modified squamous epithelium of the pecten below. The transitional zone does not, however, correspond to the macroscopic limits of the dentate line. Transitional epithelium, with islands of squamous cells within it, extends up to 2 cm above the line, explaining how tumours of this epithelial type occur in the more proximal anal canal.

Epidermoid carcinoma of the margin accounts for only about 15 to 30 per cent of all squamous anal cancers. Eighty-five per cent are well differentiated and produce keratin. Basaloid features are rare and presumably result from the development of an invasive squamous component within a pre-existing basal cell carcinoma.

Epidermoid cancers of the canal are usually more poorly differentiated, and only about 30 per cent produce keratin. Basaloid (cloacogenic or transitional) features, found in about 40 per cent of tumours, imply derivation from the anal transitional zone. When basaloid features are present there is usually a squamous component as well: classification as squamous or basaloid is based on the predominant cell type, although this is somewhat academic and subject to differences in interpretation.

The most common presenting features of anal cancer are pain, bleeding, change in bowel habit, and pruritus ani. Less common presentations include inguinal lymphadenopathy from secondary spread to that site, and as an unexpected histological finding in a haemorrhoidectomy specimen.

About 60 per cent of patients with epidermoid anal cancer at the anal margin have associated perianal conditions, including condylomata, chronic fistula, leucoplakia, or the effects of previous irradiation. Such features are found in less than 10 per cent of patients with tumours arising in the anal canal.

An area of particular interest lies in the relationship of anal cancer to sexually transmitted disease. Earlier studies suggested an association with lymphogranuloma venereum and syphilis, although this was not subsequently confirmed. However, more recent epidemiological data have provided evidence in support of the hypothesis that a sexually transmitted agent is involved in the pathogenesis of anal squamous cancer, especially in the homosexual population. Human papilloma virus type 16 DNA has been detected in more than 50 per cent of patients with anal cancer, indicating that this may be the transmissible agent involved.

Epidermoid anal cancer is diagnosed on the basis of examination of biopsy specimens. The differential diagnosis includes conditions such as rectal adenocarcinoma, Bowen's disease, Paget's disease, condyloma acuminata, leucoplakia, Crohn's disease, and certain intrinsic skin disorders such as lichen sclerosus et atrophicus. At the time of diagnosis it is necessary to determine accurately the stage of the disease. Difficulties encountered in staging on a clinical basis have been overcome by the increasing use of local ultrasound to assess muscle invasion. Computerized tomography or MRI imaging may also help determine the presence of local invasion or metastatic disease.

Local excision is sufficient treatment in the majority of patients with epidermoid cancer of the anal margin, giving a 5-year survival rate in excess of 80 per cent, although large tumours may require skin grafting. The prognosis may be determined by the depth of invasion.

Abdominoperineal resection has not been widely adopted for patients with tumours at the anal margin. When it has been used it has been usually reserved for deeper infiltrative tumours or for patients with persistent recurrence; the survival rate of about 50 per cent is therefore somewhat less than that obtained after simple local excision. A potential advantage of abdominoperineal resection is that inferior mesenteric nodal metastases will be resected, although such spread is rare in cancers arising at the margin.

There are few data on primary irradiation or combined chemotherapy and irradiation of epidermoid cancer of the anal margin. Although responses have been observed there seems little need for this approach as simple local excision will suffice in the vast majority of patients.

If recurrence occurs after local excision of anal margin cancer, further simple excision may be attempted. If this is not possible, radiotherapy, using either external beam irradiation or an implant, may be considered. Occasional patients require abdomino-perineal resection although good local tumour control may be difficult to achieve if there is persistent recurrence.

The good survival following local excision of epidermoid cancer of the anal margin results from the relatively benign characteristics of the tumour. These relatively well differentiated, keratinizing squamous tumours therefore behave rather like other primary skin cancers arising elsewhere on the body.

Epidermoid cancer of the anal canal has a worse prognosis and requires more complex treatment than its relatively benign counterpart at the margin.

A wide variety of treatments have been adopted for patients with tumours in the anal canal, ranging from local excision to preoperative neoadjuvant chemotherapy and radiotherapy followed by abdominoperineal resection. It is its potential for chemo- and radiosensitivity that has caused much interest in the past few years.

Simple local excision alone is only rarely applicable to patients with epidermoid cancer of the anal canal. When performed the prognosis has been relatively good, with a 5-year survival of up to 65 per cent: this presumably reflects the early stage of disease for which the operation was performed. It is probably only suitable for those superficially invasive tumours that are less than 2 cm in diameter, and is not generally recommended.

Abdominoperineal resection has been the standard treatment for epidermoid cancer of the anal canal, with a 5-year survival rate of 50 per cent. A wide perineal phase of the operation is recommended; if inguinal lymph nodes are involved, inguinal lymphadenectomy is required as a secondary procedure. Survival depends on tumour size, histological grade, depth of invasion, and overall staging. Earlier suggestions that predominantly basaloid tumours have a better prognosis than the more usual squamous type have not been substantiated.

Primary irradiation has been used in the treatment of epidermoid anal cancer for over 50 years. In the past the various radiotherapeutic techniques were inadequate and produced disappointing results in terms of both tumour control and local morbidity. However, more modern megavoltage X-ray therapy has produced encouraging results, with 5-year survival rates in excess of 50 per cent, and with the added advantage of allowing anal sphincter preservation. Interstitial techniques have also been used with success in some centres.

Primary irradiation therapy may sometimes be associated with severe local morbidity and patients may require a colostomy because of anal necrosis or stenosis.

The original, somewhat empirical, observation by Nigro and his colleagues in 1974 of the effects of preoperative treatment of epidermoid anal cancer with combination chemotherapy and irradiation attracted much attention. Their use of 5-fluorouracil and mitomycin C was somewhat arbitrary, although both had demonstrated cytotoxic activity against a wide range of gastrointestinal tumours and also exhibited properties of radiosensitization.

The chemoirradiation regimen now currently used varies between institutions. Most give mitomycin C on day 1 of the treatment followed by infusion of 5-fluorouracil over a 5- to 7-day period. Irradiation starts either concurrently with chemotherapy or 3 days after the mitomycin C has been given. The average radiation dose is about 30 to 50 Gy, given over 3 to 7 weeks, a lower dose than is usually recommended for patients treated by radiation alone. These regimes are associated with side-effects such as radiation-induced proctitis and dermatitis, leucopenia, thrombocytopenia, stomatitis, and diarrhoea.

In most centres patients are re-examined about 4 to 6 weeks after completion of radiotherapy: About 70 per cent will show a complete response to their treatment and in only 10 per cent is there no effect. Those patients who fail to respond require abdominoperineal resection. The major question is whether after a complete response simple excision of residual scar tissue will suffice, or whether these patients should be treated by abdomino-perineal resection. Current data provide no real answer, but do indicate that neoadjuvant chemoirradiation may reduce the need for abdominoperineal resection. Whether it improves survival is less clear. These questions are currently being addressed by randomized controlled studies.

Malignant melanoma of the anus is rare and has a poor prognosis. Whereas one can usually expect a 50 to 70 per cent 5-year survival for cutaneous melanoma occurring elsewhere in the body the average 5-year cure rate reported in the literature for malignant melanoma in the anal region is only 6 per cent.

Malignant melanoma of the anal region usually presents with bleeding or a mass. It often appears as a slightly pigmented lesion, although amelanotic melanoma is recognized at this site. Unfortunately, many patients have disease affecting the inguinal lymph nodes at the time of presentation.

Because of its rarity, virulence, and poor prognosis the treatment of this tumour has varied from simple local excision to radical abdominoperineal resection with bilateral prophylactic groin dissection. Adjuvant chemotherapy and radiation treatment seem to provide little benefit.

The majority of long-term survivors have been treated by abdominoperineal resection. There seems to be little benefit of prophylactic inguinal node dissection, although staging by nodal biopsy at the time of abdominoperineal resection may be clinically useful.

Adenocarcinoma of the anal canal may result from downgrowth of a primary rectal tumour or from carcinoma developing de novo in an anal gland or fistula. Such tumours are rare, the prognosis is poor, and treatment is by abdominoperineal resection.

Basal cell carcinoma is another rare anal condition. Experience is anecdotal but full thickness local excision appears to assure a cure.

Bowen's disease is an intraepithelial in-situ squamous cancer of the skin and is seen more commonly on the trunk, hands, and face than in perianal and genital areas. It is associated with synchronous and metachronous cancers at other sites.

Clinically, it appears as a slowly growing, minimally elevated, erythematous plaque-like lesion. The histological picture is that of an in-situ squamous cancer; above an intact basement membrane the cells demonstrate loss of polarity, acanthosis, and an inflammatory infiltrate. The characteristic Bowenoid cells have large hyperchromatic nuclei, with vacuoles providing a haloed effect.

The treatment of Bowen's disease is wide local excision, with skin grafting if necessary: special attention must be paid to the margin of excision. Long-term follow-up is necessary because of the possibility of recurrent disease in an area predisposed to the development of this condition.

Extramammary Paget's disease is an extremely rare intraepithelial neoplasm. Unlike its counterpart in the breast it is not always associated with frank underlying malignancy. The clinical appearance of Paget's disease is that of a pale grey, crusting, scaly lesion. The diagnosis is confirmed by demonstrating the characteristic Paget cells with their large pale cytoplasm that stains positively with aldehyde-fuscin. Any associated invasive cancer arises from a skin appendage such as a sweat gland rather than from rectal mucosa.

If no frank cancer is present then simple excision is suitable treatment for Paget's disease. Negative margins of resection must be obtained or recurrence is likely. If there is underlying malignancy then a deeper, wider excision is needed, but the outlook is generally poor.

Fenger C. The anal transitional zone. Location and extent. Acta Pathol Microbiol Immunol Scand (A) 1979; 87: 379–86.

Greenall MJ, Quan SHQ, DeCosse JJ. Epidermoid cancer of the anus. Br J Surg 1985; Suppl. S97 – S103.

Greenall MJ, Quan SHQ, Stearns MW, Urmacher C, DeCosse JJ. Epidermoid cancer of the anal margin: pathologic features, treatment and clinical results. Am J Surg 1985; 149: 95–101.

Nigro ND, Vaitkevicius VK, Considine B. Combined therapy for cancer of the anal canal: a preliminary report. Dis Colon Rect 1974; 17: 354–56.

Palmer JG, et al. Anal cancer and human papillomavirus. Dis Colon Rect 1989; 32: 1016–22.