Portal vein thrombosis

Malignant transformation is an infrequent complication of endometriosis. The ovary is the primary site in 76% of cases, and extragonadal sites are identified in 24%. Endometrioid carcinoma is the most common histologic type; sarcoma is very rare. We report a case of low-grade endometrial stromal sarcoma of the rectosigmoid colon presenting with epigastric pain due to portal vein thrombosis. This tumor arose from extragonadal endometriosis in a 61 -- year-old woman and was treated by surgical resection. The main differential diagnosis of this unusual colonic neoplasm includes primary mesenchymal tumors, such as gastrointestinal stromal tumors.

(Arch Pathol Lab Med. 2001;125:1088-1090)

Endometriosis is one of the most common benign gynecologic conditions and has been estimated to affect approximately 10% to 25% of women presenting with gynecologic symptoms in the United States. Intestinal involvement occurs in 3% to 37% of cases and involves areas where the peritoneum is irregularly folded, such as the antimesocolic border of the sigmoid colon. It also affects those parts of the bowel that lie in proximity to the genital organs. The sigmoid colon and rectum are involved in 15% to 72% of the cases with intestinal involvement.1

Malignant transformation is a rare but well-documented complication of endometriosis, occurring in 0.7% to 1% of cases. The ovary is the primary site in 76% of the cases, whereas extragonadal sites represent 24%.2,3 The colorecturn is involved in only 5% of cases.2 Theoretically, any histologic pattern can arise in endometriosis, but endometrioid adenocarcinoma is the most frequent. Endometrial stromal sarcoma is extremely rare, particularly in the rectosigmoid colon.

In this article, we present an endometrial stromal sarcoma of the rectosigmoid colon arising in extragonadal endometriosis. To our knowledge, this is the first case revealed by portal vein thrombosis.

REPORT OF A CASE

A 61-year-old woman, para 1, presented with epigastric pain in December 1997. She had no history of, or symptoms to suggest, the presence of endometriosis and had been receiving hormone replacement therapy for 13 years.

Abdominal ultrasound revealed a portal vein thrombosis, and computed tomographic scan of the pelvis revealed a malignant tumor in the rectosigmoid colon. Her serum carcinoembryonic antigen and CA 125 levels were within normal limits. Colonoscopic examination showed an unusual polypoid tumor with stenosis of the lumen at the rectosigmoid junction, but endoscopic biopsy showed only nonspecific inflammation.

Because a metastatic workup was negative, the patient underwent laparotomy with rectosigmoid colon resection and low anterior reanastomosis. At the time of surgery there appeared to be tumor nodules (1-2 cm in diameter) on the posterior wall of the right board ligament of the uterus. Frozen section examination of these granulations showed endometrial glands and surrounding stroma, in keeping with endometriosis (Figure 1). In addition, a postoperative hysterography and curetting biopsy revealed no mass lesion, making a uterine primary unlikely.

The patient did not receive any adjuvant treatment. She was alive without evidence of disease for 30 months of follow-up.

MATERIALS AND METHODS

Routine hematoxylin-eosin sections were prepared from formalin-fixed, paraffin-embedded tumor tissue. Immunohistochemical studies were performed by indirect staining methods using antibodies against cytokeratin (clone KL1, 1:50, Immunotech, Marseille, France), vimentin (3B4, 1:100, Dako, Glostrup, Denmark), epithelial membrane antigen (E29, 1:75, Dako), smooth muscle actin (1A4, 1:50, Dako), desmin (D33, 1:20, Dako), 5100 protein (polyclonal, 1:200, Dako), neurofilaments (2F11, 1: 75, Dako), CD34 (QBEnd 10, 1:50, Dako), CD117 (polyclonal, 1: 100, Santa Cruz Biotechnology, Santa Cruz, Calif), estrogen receptor (1D5, 1:75, Dako), and progesterone receptor (PR88, 1:40, BioGenex, San Ramon, Calif).

PATHOLOGIC FINDINGS

The specimen consisted of a 20-cm resection of rectosigmoid colon with attached mesenteric fat. A 2.7-cm tumor was found at 3 cm of rectal margin and was responsible for stenosis of the lumen. The tumor grossly involved all layers of the rectal wall and was associated with overlying polypoid bluish mucosa. Ten lymph nodes were identified in the attached mesentery.

Microscopic examination revealed a typical tonguelike growth of tumor nodules that invaded all layers of the rectal wall (Figure 2). These nodules were composed of densely packed, plump spindle cells in short fascicles, interspersed with prominent small arterioles (Figure 3). The tumor cells resembled those of normal endometrial stroma of proliferative phase, with scanty ill-defined cytoplasm and round or ovoid nuclei with dispersed chromatin. The cells exhibited little nuclear pleomorphism and few mitotic figures were identified, the mitotic count being less than 1 in 10 high-power fields. There was 1 area of epithelial-like configuration. Prominent lymphatic, vascular, and perineural invasion was seen. The histologic features were typical of low-grade endometrial stromal sarcoma.

There was no lymph node metastasis, and all surgical margins were free of disease.

Tumor cells were strongly positive for vimentin (Figure 4), estrogen receptor, and progesterone receptor, but were negative for cytokeratin, epithelial membrane antigen, S100 protein, neurofilament, CD34, and CD117. A few tumor cells in an area of epithelial-like configuration were positive for desmin and smooth muscle actin.

COMMENT

Malignant transformation of endometriosis has been well documented since it was first reported by Sampson 4 in 1925, who recommended 3 criteria for a definitive diagnosis of malignancy arising in endometriosis: (1) close proximity of benign endometriosis to the malignant tumor, (2) no other primary site identified, and (3) tumor histology compatible with an endometrial primary. In 1953, Scoff5 suggested that a more stringent qualification should be applied, requiring that microscopic benign endometriosis was contiguous with malignant tissue. For the extraovarian case, adequate evidence for such an association is coexistence of the tumor and endometriotic tissue, even without demonstrable continuity, if the 2 processes appear in an uncommon site or at an unusual age, and the malignant tumor is of a histologic type that has been well established to arise from endometrial-type tissue.6 These less strict criteria are justified, because transitional areas between endometriosis and cancer can be destroyed by the growth of the tumor and therefore are shown only in 5% to 10% of cases.6 Our case is consistent with these criteria.

In their review of preneoplastic and neoplastic changes in gastrointestinal endometriosis, Yantiss et al7 found fewer than 50 cases of neoplasms arising in endometriosis of the gastrointestinal tract, reported as small series or single cases. The classic presentation is that of increasing abdominal pain, bloating, or rectal bleeding in a postmenopausal patient with or without a history of endometriosis. Presenting symptoms due to vascular thrombosis, as in our case, were reported only once in the literature.8 In that case, a 52-year-old woman with sarcomatous change in chronic pelvic endometriosis presented with a swollen right leg due to an extensive right iliofemoral vein thrombosis. Epigastric pain due to portal vein thrombosis, resuiting probably from extensive venous invasion, was the first symptom in our patient.

The most common histologic type of cancer arising in endometriosis is endometrioid adenocarcinoma. Endometrial stromal sarcoma is rare in gastrointestinal and other sites of malignant transformation of endometriosis.3 Palladiano et al (cited in Yantiss et al7) reported 5 endometrial stromal sarcomas involving the rectovaginal septum, rectum, and colon, and Mostoufizadeh and Scully6 reported 2 additional cases of sarcoma arising in endometriosis involving the gastrointestinal tract. Finally, Baiocchi et al (cited in Yantiss et a17) reported 2 cases occurring in the colon.

The differential diagnosis of endometrial stromal sarcoma arising in the gastrointestinal tract includes mesenchymal neoplasms, particularly if the underlying benign endometriosis is obscured.7 Most of the mesenchymal neoplasms (fibromatosis, schwannoma, and leiomyoma) can be immediately excluded from the differential diagnosis based on their histologic features. However, given their variable gross and histologic appearances, gastrointestinal stromal tumors can be confused with endometrial stromal sarcoma. In contrast to endometrial stromal sarcoma, these tumors tend to be well-circumscribed with broad, pushing borders and rare vascular invasion. The cells are arranged in short fascicles with a vaguely organoid arrangement reminiscent of smooth muscle neoplasms. Nuclear atypia and pleomorphism may be marked. Endometrial stromal sarcomas have characteristically invasive tongues of tumor at the periphery and are usually composed of short regular fascicles or sheets of monomorphic plump spindle cells. The presence of prominent arterioles and extensive vascular invasion should argue against the diagnosis of stromal tumor. Finally, immunohistochemical stains are useful in distinguishing between these entities, as stromal tumors are well known to stain diffusely for CD117 and CD34,9 and endometrial stromal sarcoma for estrogen and progesterone receptors.

The association between exogenous hormone therapy and the development of malignancy in endometriosis is well known and should alert both pathologists and clinicians to the possibility of malignant transformation of endometriosis in these patients.10 Our patient had received hormone replacement therapy for 13 years.

The management of extrauterine endometrial stromal sarcoma is difficult, but primary surgical treatment with complete resection of all disease should be performed when feasible.10 The prognosis correlates well with stage. A 100% 5-year survival rate has been noted for malignant transformation in extragonadal endometriosis, confined to the site of origin.3 However, the 5-year survival rate was only 12% in cases of disseminated intraperitoneal disease.2

In summary, this case of endometrial stromal sarcoma of the rectosigmoid colon arising from extragonadal endometriosis is of interest because the atypical clinical presentation was the epigastric pain probably due to a portal vein thrombosis. Although rare, malignant transformation of extragonadal endometriosis should be included in the differential diagnosis of unusual colorectal tumors affecting postmenopausal patients, particularly if they have received hormone therapy.

References

1. Fenoglio-Preiser CM, Pascal RR, Perzin KH. Tumors of the Intestines. Washington, DC: Armed Forces Institute of Pathology; 1990:413-425. Atlas of Tumor Pathology, 2nd series, fascicle 27.

2. Irvin W, Pelkey T, Rice L, Anderson W. Endometrial stromal sarcoma of the vulva arising in extraovarian endometriosis: a case report and literature review. Gynecol Oncol. 1998;71:313-316.

3. Heaps JM, Nieberg RK, Berek JS. Malignant neoplasms arising in endometriosis. Obstet Gynecol. 1990;75:1023-1028.

4. Sampson JA. Endometrial carcinoma of ovary arising in endometrial tissue in that organ. Arch Surg. 1925;10:1-72.

5. Scott RB. Malignant changes in endometriosis. Obstet Gynecol. 1953;2: 283-289.

6. Mostoufizadeh M, Scully RE. Malignant tumors arising in endometriosis. Clin Obstet Gynecol. 1980;23:951-963.

7. Yantiss RK, Clement PH, Young R. Neoplastic and preneoplastic changes in gastrointestinal endometriosis: a study of 17 cases. Am J Surg Pathol. 2000;24: 513-524.

8. Strinfellow JM, Hawnaur JM. CT and MRI appearances of sarcomatous change in chronic pelvic endometriosis.Br J Radiol. 1998;71:90-93.

9. Sarlomo-Rikala M, Kovatich AJ, Barusevicius A, Miettinen M. CDI 17: a sensitive marker for gastrointestinal stromal tumors that is more specific than CD34. Mod Pathol. 1998;11:728-734.

10. McCluggage WG, Bailie C, Weir P. Endometrial stromal sarcoma arising in pelvic endometriosis in a patient receiving unopposed oestrogen therapy. BrI Obstet Gynaecol. 1996;103:1252-1254.

Najat Mourra, MD; Emmanuel Tiret, MD; Yann Parc, MD; Paul de Saint-Maur, MD; Rolland Parc, MD; Jean-Francois Flejou, MD

Accepted for publication February 20, 2001.

From the Departments of Pathology (Drs Mourra, Saint-Maur, and Flejou) and Surgery (Drs Tiret, Y. Parc, and R. Parc), Hopital Saint-Antoine, Paris, France.

Reprints: Najat Mourra, Department of Pathology, Hdpital Saint-Antoine, AP-HP, 184 rue faubourg St-Antoine, 75012 Paris, France (e-mail: najat.mourra@free.fr).

Copyright College of American Pathologists Aug 2001
Provided by ProQuest Information and Learning Company. All rights Reserved