Diagnostic challenge in distinguishing intestinal endometriosis-associated carcinoma from primary colorectal adenocarcinoma: a case report

A 43-year-old nulliparous woman presented with melena. She had undergone a left salpingo-oophorectomy for an ovarian endometriotic cyst at 36 years of age. Colonoscopy demonstrated a semicircumferential rectosigmoid tumor, and imaging revealed a locally advanced rectal mass with suspected uterine invasion and multiple lymph node metastases ( Fig. 1 ). She was clinically diagnosed with advanced rectal adenocarcinoma. Initial clinical findings at presentation; (A) colonoscopy revealed a semicircumferential tumor in the rectosigmoid colon; (B) T2-weighted magnetic resonance imaging demonstrated a mass lesion involving the upper rectum; (C) fluorodeoxyglucose positron emission tomography/computed tomography (FDG-PET/CT) showed increased uptake in the primary tumor and multiple enlarged lymph nodes, including pararectal, internal iliac, obturator, common iliac, and para-aortic lymph nodes; these findings initially suggested advanced colorectal adenocarcinoma. The patient underwent a laparoscopic colectomy involving the sigmoid colon and rectum, a total hysterectomy, a right salpingo-oophorectomy, and a colostomy creation. Histopathological examination demonstrated a moderately differentiated adenocarcinoma invading the myometrium with lymphovascular invasion and multiple lymph node metastases. Endometriotic lesions were present in the muscularis propria and subserosa of the rectum and sigmoid colon and on the uterine serosa ( Fig. 2 ). However, no definitive histological transition between benign endometriosis and carcinoma was identified. The overall clinicopathological findings were therefore considered most consistent with primary colorectal adenocarcinoma. Histopathological findings of the original surgical specimen; (A) gross examination of the resected specimen demonstrated a tumor involving the rectosigmoid colon; (B) low-power view showing moderately differentiated adenocarcinoma with irregular glandular and cribriform structures; (C) high-power view of the adenocarcinoma; (D) endometriotic lesions identified in the muscularis propria of the intestine; no definitive histological transition between benign endometriosis and carcinoma was observed; hematoxylin and eosin staining; scale bars: 400 μm (B), 100 μm (C), and 400 μm (M). Following surgery, she received systemic chemotherapy for colorectal cancer. Five years after the initial diagnosis, progressive pelvic and distant recurrence developed, and she was referred to our department because of vaginal bleeding. A biopsy of a vaginal stump lesion revealed adenocarcinoma ( Fig. 3 ). Immunohistochemical analysis showed positivity for estrogen receptor (ER), paired box gene 8 (PAX8), and cytokeratin 7 (CK7), and negativity for cytokeratin 20 (CK20) and caudal-type homeobox transcription factor 2 (CDX2), suggesting a Müllerian rather than a colorectal origin. Recurrent vaginal stump lesion; (A) colposcopic image of a recurrent lesion at the vaginal stump; (B) histopathological examination of the biopsy specimen revealed adenocarcinoma; immunohistochemical findings of the recurrent lesion prompted reconsideration of the original diagnosis and retrospective immunohistochemical reassessment of the original tumor; hematoxylin and eosin staining; scale bar: 100 μm. Retrospective immunohistochemical evaluation of the original surgical specimen demonstrated the same immunophenotype ( Fig. 4 ). Although a direct transition between endometriosis and carcinoma could not be demonstrated, the tumor distribution, Müllerian immunophenotype, and intestinal endometriosis in the original specimen supported a final diagnosis of malignant transformation of intestinal endometriosis. Immunohistochemical findings of the original surgical specimen; representative immunohistochemical staining of the original tumor showing positivity for ER, PAX8, and CK7, and negativity for CK20 and CDX2; this immunophenotype was inconsistent with primary colorectal adenocarcinoma and supported a Müllerian origin of the tumor; scale bars: 100 μm.

The patient was deceased at the time of manuscript preparation, and her relatives could not be traced; therefore, consent for publication could not be obtained. All clinical information and images have been sufficiently anonymized to protect patient confidentiality. The circumstances regarding patient consent were discussed with the Journal of Surgical Case Reports Editorial Office prior to submission and were confirmed with the Senior Editor.

Hisashi Kataoka is the guarantor of this work.

Malignant transformation of endometriosis was originally defined using criteria proposed by Sampson and subsequently modified by Scott, including the coexistence of benign and malignant endometrial tissue, histological compatibility with endometrial origin, exclusion of another primary tumor, and the demonstration of a transition from endometriosis to carcinoma [ 4 , 5 ]. However, these criteria may not always be fulfilled because extensive tumor growth can obscure or destroy the precursor endometriotic lesion. The initial diagnosis of colorectal adenocarcinoma in our patient was clinically reasonable. She presented with melena, the colonoscopy showed a rectosigmoid mass, imaging demonstrated locally advanced disease with lymph node metastases, and histopathology revealed a moderately differentiated adenocarcinoma with uterine invasion. Although intestinal endometriosis was identified, no transition between benign endometriosis and carcinoma was demonstrated. The coexistence of endometriosis alone was therefore insufficient to establish malignant transformation. Endometriosis-associated intestinal carcinoma generally arises from the outer bowel wall and extends inward, whereas primary colorectal adenocarcinoma originates from the mucosa and progresses outward [ 3 ]. Once the tumor involves the full thickness of the bowel wall, however, this distinction may become difficult on morphology alone. In the present case, the recurrence provided the critical diagnostic clue. ER, PAX8, and CK7 positivity together with CK20 and CDX2 negativity favored a Müllerian origin [ 6 ]. Demonstration of the same immunophenotype in the original tumor led to a reconsideration of the initial diagnosis. Intestinal endometriosis-associated carcinoma should therefore be considered in women with colorectal adenocarcinoma and a history or pathological evidence of endometriosis, particularly when the clinical course or immunophenotype is atypical. Importantly, the presence of endometriosis alone does not establish malignant transformation. Integration of clinical history, tumor distribution, morphology, and immunohistochemistry is essential for determining tumor origin in diagnostically challenging cases.

Malignant transformation occurs in ~1% of women with endometriosis, predominantly in the ovaries [ 1 , 2 ]. Extraovarian malignancy is uncommon, and malignant transformation of intestinal endometriosis is particularly rare [ 2 , 3 ]. Because these tumors may present with gastrointestinal bleeding, bowel obstruction, or a colorectal mass, they can closely mimic primary colorectal adenocarcinoma. Histological distinction may become especially difficult when tumor invasion involves the full thickness of the bowel wall. We report a case initially diagnosed and treated as primary colorectal adenocarcinoma in which recurrence prompted immunohistochemical reassessment, ultimately supporting the malignant transformation of intestinal endometriosis.