Anal Canal Adenocarcinoma in a Patient with Longstanding Crohn’s Disease Arising From Rectal Mucosa that Migrated From a Previously Treated Rectovaginal Fistula

Background

Crohn’s disease (CD) has been associated with perianal fistulas and abscesses, as well as an increased risk of malignancy [1–4]. The frequency of malignancy associated with CD has been reported to be 0.6–3.1% [5–7]. Although tumor location varies and is not always predictable, cancer arising from a fistula is rare [8–11]. Several hypotheses have been proposed to explain carcinoma development in fistulas, but the pathogenic mechanisms by which carcinomas arise in patients with CD remain unclear [12–14]. Indeed, the risk of carcinoma as a complication of CD may be cumulative. Most of these tumors are diagnosed at an advanced stage, with late diagnosis and delayed treatment associated with poor patient prognosis [15]. This report describes an early symptomatic anal canal adenocarcinoma in a patient with a 33-year history of CD and a history of rectovaginal fistula, which had been cured by surgical treatment 26 years earlier. The pathological and immunohistological findings in this patient suggest the origin and progression of CD-associated cancer.

Case Report

A 50-year-old woman with a 33-year history of CD, mostly concentrated in the ileum and colon, presented with perianal pain of several months’ duration. She had undergone surgical correction of a rectovaginal fistula 26 years earlier. Postoperatively, she did not report any symptoms related to this fistula, which remained quiescent for 26 years until a few months before presentation to us. She developed significant perianal pain in the 12 o’clock position. Infliximab therapy was initiated at age 46 years, and uncontrolled bleeding from multiple ileocecal ulcers required surgical resection at the age of 49 years.

She was admitted to our department for severe perianal pain. Computed tomography (CT) at presentation showed an 18-mm isodense collection of perianal fluid on the anterior rectal wall with unclear margins (Figure 1A). CT performed 10 months earlier during routine surveillance showed perianal fluid collection around the same area, but it was 7 mm in diameter at that time and asymptomatic. Endoscopic examination showed no lesions in the rectum or anus. Magnetic resonance imaging (MRI) of the pelvis showed a multiple conglomerated cystic mass at the same location, with no invasion of the vaginal wall (Figure 1B). Colonoscopy revealed an ulcerative lesion in the anal canal 10 mm in diameter and located from the 12 o’clock to the 2 o’clock position, corresponding to the collection of perianal fluid identified by CT and MRI (Figure 2).

A digital rectal examination, performed under local anesthesia of the anorectum, showed a small amount of jelly-like fluid at the site of ulcerative lesion, which was located on the level from 2 cm from the anus as CT and colonoscopy showed, along with compression of the posterior vaginal wall, but no fistula connected with the distal posterior vagina. The lesion was biopsied because mucinous adenocarcinoma was suspected. Pathologic examination showed disappearance of the crypts and a lack of mucous membranes. The pathological diagnosis was mucinous adenocarcinoma. Based on radiological findings, the stage was T1N0M0. Abdominoperineal resection was performed, with negative resection margins (Figure 3). Stenosis of the small bowel was observed during surgery, necessitating small bowel resection with a Kono-S anastomosis.

The final pathological diagnosis consisted of two types of adenocarcinoma: well differentiated tubular adenocarcinoma in the rectal mucosa and mucinous adenocarcinoma in the submucosa, measuring 1.5 cm and 2.0 cm in size, respectively (Figure 4). The tubular adenocarcinoma had invaded to the level of the submucosa (Figure 5A) and the mucinous adenocarcinoma had invaded to the level of the muscularis propria (Figure 5B).

There was no evidence of lymph node metastasis, but invasion of the lymph channels was observed. In addition, the tumor was surrounded by thick fibrous tissue with no evidence of infiltration by inflammatory cells (Figure 5C, 5D). The tumor cells expressed CDX2 and cytokeratin20 but were negative for cytokeratin (CK) 7 (Figure 6). Immunostaining showed the tumor cells were positive for MUC2 and MUC5AC but negative for MUC1 and MUC6 (Figure 7). These findings indicate that the tumor originated from the rectal mucosa, not from the anal glands. The final diagnosis was anal canal adenocarcinoma arising from the rectal mucosa.

Postoperatively, the patient developed catheter-related sepsis, which resolved after a course of treatment with antibiotics. One month after surgery, she started IFX therapy, which was scheduled once every 8 weeks. Based on the Japanese Society for Cancer of the Colon and Rectum’s Guidelines 2014 for the Treatment of Colorectal Cancer, adjuvant chemotherapy was not administered. She continues to be followed with no signs or symptoms of recurrent carcinoma 18 months after surgery.

Discussion

CD increases the risk of cancer. The incidence of malignant transformation has been reported to be 0.6–3.1% [5,6] in Europe and the United States.

A diagnosis of anorectal cancer arising from a former fistula requires not only pathological findings, but knowledge of the clinical course of the patient. Five clinical features have been reported necessary for a diagnosis of anorectal cancer arising from a fistula: 1) an anal fistula with continuous inflammation for more than 10 years, 2) pain or induration of the anal fistula, 3) mucinous secretions, 4) no evidence of any other primary cancer, and 5) the primary opening being an anal crypt or the anal canal [16–19]. However, this patient did not meet the first diagnostic criterion.

Her rectovaginal fistula had been successfully repaired 26 years earlier, with no evidence of exacerbation until she experienced severe pain several months prior to presentation. Pathological and clinical findings did not demonstrate the presence of a rectovaginal fistula or chronic inflammation. However, the tumor was located at the site of the former rectovaginal fistula. Pathological examination also showed that the thick fibers surrounding the tumor were likely part of the former rectovaginal fistula. Immunohistological methods are useful in determining the origin of anorectal adenocarcinomas [20,21]. For instance, adenocarcinomas arising from anal glands are positive for CK7 but negative for CK20, whereas adenocarcinomas arising from the rectal mucosa are negative for CK7 but positive for CK20. This patient’s tumor was positive for CK7 but negative for CK20, suggesting that its origin was the rectal mucosa, not the anal gland.

Pathological examination also showed that the mucinous adenocarcinoma component of the tumor was larger than the tubular adenocarcinoma component. In addition, the former showed deeper invasion. Radiological examination 10 months earlier showed that a small asymptomatic mass was present in the same location, but this lesion was not apparent on endoscopic examination (i.e., an aphthous ulcer).

These findings suggested that rectal epithelial cells had migrated from the primary opening of the rectovaginal fistula, and that the rectal epithelium lining the former rectovaginal fistula was the origin of the anorectal adenocarcinoma in the absence of chronic inflammation (Figure 8). However, we cannot completely exclude the possibility that the rectal mucosa in the anal canal became malignant and that tumor cells in the mucosal epithelium invaded the submucosa along the fistula and ultimately became mucinous adenocarcinoma.

Although several studies have suggested that carcinogenesis may be due to long-term treatment with IFX, an anti-tumor necrosis factor α (TNFα) antibody that has been shown effective for inducing and maintaining CD remission, the role of IFX in cancer progression remains unclear [22–24]. The current patient had been treated with IFX for 4 years. IFX therapy may suppress TNFα and thus enhance tumor growth. IFX therapy may also delay cancer diagnosis by controlling inflammation and ameliorating anal symptoms such as pain. Therefore, particular attention should be paid to malignant transformation in patients with CD on long-term IFX therapy, especially when they complain of increased symptoms (i.e. increased pain or fluid discharge) from a perianal lesion.

Conclusions

This report describes an anal canal adenocarcinoma diagnosed early in a patient with longstanding CD and long-term IFX treatment. Pathological and clinical findings suggest that adenocarcinoma might arise in patients with CD, even in the absence of longstanding perianal disease, and that it might be associated with adenomatous transformation of the migrated epithelial lining of a former fistula tract.