|Home | About | Journals | Submit | Contact Us | Français|
Colonic mucosal involvement with endometriosis can result in gastrointestinal bleeding1. The mucosal abnormalities associated with colonic endometriosis may be difficult or impossible to identify at colonoscopy2.
A woman age 37 was referred for investigation of bright red rectal bleeding of seven months' duration and usually coinciding with menstruation. Pelvic endometriosis had been diagnosed 5 years earlier and had been treated with the oral contraceptive pill and laparoscopic laser ablation. Cleft palate and hare lip, uterus didelphus and septate vagina had all been surgically corrected. On examination, tender masses were palpable in both iliac fossae. Abdominal MRI scans suggested endometriotic cysts arising from both ovaries. Colonoscopy revealed mild mucosal inflammation at 35 cm from the anal verge, and a polypoid lesion 3 cm in diameter was seen within the ascending colon 65 cm from the anal verge. On histological examination, biopsies from the ‘inflamed’ area of distal colon showed ‘inflammatory changes’, while two separate biopsies from the polypoid lesion in the ascending colon showed fragments with mixed features of an inflammatory polyp and adenomatous glands: areas of low-grade dysplasia were surrounded by inflamed fibrotic stroma with focal surface ulceration and granulation tissue (Figure 1). The clinical diagnosis was distal colonic endometriosis and a proximal adenomatous polyp. The caecal adenomatous lesion could not be removed colonoscopically because of its size and location. At laparotomy there was a diffuse mass involving the posterior caecal wall, adjacent retroperitoneal muscles, right ovary and right fallopian tube. After joint intra-abdominal examination by colorectal and gynaecological surgeons, it was concluded that these findings were not typical of endometriosis. Typical endometriotic appearances were seen in the left ovary and on the serosa of the adjacent sigmoid colon. Frozen section histological examination did not identify the nature of the mass, which was resected en bloc with ileocolic anastomosis. On histological examination the polypoid lesion removed from the ascending colon contained adenomatous glands with low grade dysplasia admixed with numerous foci of endometriosis resulting in marked expansion of the polypoid fronds (Figure 2 a and b). Immunohistochemical staining clearly differentiated between colonic glands (cytokeratin 20 positive, cytokeratin 7 negative) and endometrial glands (cytokeratin 7 positive, cytokeratin 20 negative, Figure 2 b and c). Sections from the retroperitoneal mass showed endometriosis. The patient has had no further colonic bleeding during 25 months' follow-up.
Colonic endometriosis can mimic colorectal cancer by producing an invasive abdominal mass associated with abdominal pain, bleeding1 or an ulcer2,3, but ‘adenomatous’ change within colonic mucosa does not seem to have been reported previously. The present case suggests that colonic mucosal endometriosis can produce reactive changes in colonic glands, with cellular atypia and glandular hyperplasia mimicking dysplasia, thus making the lesions harder to discriminate from a neoplastic adenoma. It is possible that the ‘adenomatous’ lesion in this case resulted from invasion of a pre-existing neoplastic adenoma by endometriosis, but this is unlikely in view of the patient's young age and the absence of a history of colonic adenomas in the patient or her family. Although the colonic lesion in the present case was more fronded than a typical neoplastic adenoma, it would be difficult to use this appearance at colonoscopy to differentiate between the two conditions, since surface irregularity is also associated with malignant transformation within neoplastic adenomas.
Although endometriosis can mimic all stages of colorectal neoplasia, discrimination between colonic glands and endometrial glands is possible with cytokeratin immunostaining.