Carcinoma, endometrial

Postmenopausal bleeding is the most important presenting symptom of endometrial carcinoma. There are multiple benign causes of postmenopausal bleeding including oestrogen therapy, endometrial hypertrophy, atrophic vaginitis, endometrial or cervical polyps. Only 1020% of patients presenting with this symptom actually have a malignancy. A purulent or serous discharge is a less common presenting complaint. Less than 5% of patients with carcinoma of the endometrium are asymptomatic at the time of presentation. These cases are usually detected by abnormal cervical cytology or histological examination of a uterus removed for a benign cause. Endometrial carcinoma in premenopausal women usually presents as abnormal uterine bleeding, ranging from oligomenorrhoea to menometrorrhagia.

The diagnosis of endometrial carcinoma is made by histological examination of endometrial tissue. Endometrial biopsy with endocervical curettage is a safe and effective procedure in the evaluation of patients with postmenopausal bleeding. This procedure can be performed in the outpatient setting and has largely replaced fractional curettage under anaesthesia as the first step in diagnosing endometrial carcinoma. Curettage under anaesthesia or hysteroscopy can be performed in appropriate instances. Atypical or malignant endometrial cells are usually seen on cervical cytology. The pap smear, however, is inadequate as the sole test to exclude endometrial cancer. Up to 13% of postmenopausal patients with cancer of the endometrium may present with normal cervical cytology. An elevated serum (carcinoembryonic antigen) CA-125 usually indicates extrauterine disease.

Approximately 90% of uterine malignancies are true epithelial carcinomas. The remaining 10% of nonepithelial uterine cancers are comprised of sarcomas, mixed tumours and secondary malignancies. Nearly 90% of the epithelial cancers are typical adenocarcinomas. Somewhere between 10 and 20% of adenocarcinomas demonstrate some degree of squamous differentiation (adenosquamous variety). Other less common epithelial varieties include clear cell, mucinous, secretory and papillary serous carcinoma. The clear cell and papillary serous types demonstrate highly malignant behaviour, and thus, have a much worse prognosis. Typical endometrial carcinomas are assigned one of three grades based on the cellular architecture. Grade 1 cancers are predominately glandular and are well differentiated. Grade 3 lesions are poorly differentiated adenocarcinomas, with predominantly solid elements. A high degree of nuclear atypia can also increase tumour grade.

Endometrial carcinoma typically arises from the endometrial muc the lymphatic drainage system results in marked variability in the pattern of nodal metastases. Haematogenous dissemination tends to occur late in the disease. The most common sites of distant metastases are lung, liver, brain and bone. Depth of myometrial invasion appears to be a very important independent and confounding prognostic variable.

Staging

In 1988, the International Federation of Gynecology and Obstetrics (FIGO) recommended a surgical staging system for endometrial cancer. Stage I disease is confined to the uterine corpus. Stage II denotes spread into the uterine cervix. A number of factors define stage III disease, including vaginal extension, spread into the adnexa or parametrium, positive peritoneal cytology, or nodal metastases to the iliac or para-aortic chains. Bowel or bladder mucosal involvement, as well as distant metastases define stage IV disease. In recognition that tumour grade is a major determinant of prognosis, each surgical stage is additionally stratified by assigning a grading system. Tumours are graded from one to three, with lower scores representing greater tumour differentiation.

Approximately 70% of patients present with stage I disease. The five-year survival is seen in 76% of these patients. The five-year survival rate decreases to 59% for stage II and 29% for stage III. Tumour grade also has an effect on five year survival. Retrospective analysis of patients with stage Ia and Ib disease showed a consistent drop in the five-year survival rate with increasing tumour grade.

Imaging

A chest radiograph is obtained in almost every patient diagnosed with endometrial carcinoma. In addition to detecting lung metastases, chest radiography is useful in identifying comorbid pulmonary conditions. Other conventional radiographic studies are not routinely performed. Skeletal scintigraphy and bone radiography can be performed when skeletal metastases are suspected. Intravenous pyelography and lymphangiography have largely been replaced by cross-sectional modalities. Endoscopy is the preferred method for detection of suspected rectal or colonic involvement over barium enema.

On ultrasound, stage I endometrial carcinoma typically appears as widening of the echogenic endometrial stripe. The postmenopausal endometrium atrophies, generally measuring less than 3 mm. A double layer thickness measurement of greater than 5 mm is almost always considered abnormal (Fig.1). The exception is in women on estrogen therapy. In these individuals, an endometrial stripe thickness of 7 mm or less is considered to be within normal limits. Endovaginal ultrasound may be a useful tool for screening women with postmenopausal bleeding. Several studies have shown that demonstration of a normal endometrium virtually excludes the presence of endometrial carcinoma. This is important because 8090% of patients with postmenopausal bleeding do not have endometrial carcinoma. Ultrasound has been utilized to evaluate the depth of myometrial invasion. The inner myometrium is hypoechoic relative to the outer myometrium. Focal thinning of this hypoechoic ring is suggestive of superficial invasion. Obliteration of the hypoechoic inner layer suggests deep (> 50%) myometrial invasion. The accuracy of ultrasound in the evaluation of myometrial invasion is 6869%. Ultrasound has been advocated in the staging of known endometrial carcinoma. Relatively poor contrast resolution, however, may hamper the ability of ultrasound to differentiate between tumour, normal myometrium and benign uterine diseases. In addition, large body habitus or extreme uterine flexion can hinder study quality. Polypoid lesions, pyometria, adenomyosis, leiomyomas and myometrial atrophy all have the potential to mimic myometrial invasion, resulting in false positives. False negatives can occur with superficial tumour spread or microinvasion. The value of ultrasound in the detection of lymph node metastases or cervical invasion has not been established.

Carcinoma of the endometrium appears as diffuse or foc demonstrate small volume or microscopic spread to the cervix, parametrium, lymph nodes, bladder or rectum. False positive errors generally occur with reactive adenopathy or parametritis. There is insufficient data on the value of spiral CT in staging endometrial carcinoma.

Endometrial carcinoma can have a variable appearance on MRI. Typically, the signal characteristics are similar to the normal endometrium on nonenhanced MR studies. Tumours are generally hyperintense with respect to myometrium on T2-weighted images and isointense on T1-weighted images. The presence of blood products within these lesions can make them heterogeneous. Larger tumours are often seen as a polypoid mass expanding the endometrial cavity. Secondary signs of small tumours include increased thickness or lobulation of the endometrium. Endometrial carcinoma enhances to a variable degree on dynamic post-gadolinium images. This enhancement is typically greater than or less than that of the myometrium and normal endometrium. Contrast enhancement improves the detection of small tumours and differentiation between lesions and fluid or necrosis. Myometrial invasion is best determined on post-gadolinium and T2-weighted images (Fig.4). An intact junctional zone on T2-weighted images implies that there is no myometrial invasion (stage Ia) (Fig.5). Stage Ib lesions are present when there is disruption of the junctional zone and an intact outer two thirds of the myometrium. Invasion of more than 50% of the myometrium on dynamic gadolinium images indicates stage Ic disease. False positive and negative results can occur in patients with coexisting benign pathology, elderly patients without normal zonal anatomy, or patients with bulky tumours that thin, rather than invade, the myometrium. The reported accuracy of contrast-enhanced MR in estimating the depth of myometrial invasion varies from 8594%. Cervical involvement (stage II) is best demonstrated on sagittal contrast-enhanced and T2-weighted images (Fig.6). Enhancing or high signal tumour is readily identified against the background of normal cervical tissue (Fig.7). Postcurrettage haemorrhage can simulate cervical involvement if only unenhanced sequences are utilized. The accuracy of contrast-enhanced MR in the evaluation of cervical extension ranges between 91% and 95%.

Extrauterine disease

Local extrauterine disease (stage III) is seen as transmyometrial extension of tumour, adnexal masses, vaginal metastases or pelvic lymphadenopathy. Disruption or irregularity of the uterine serosal surface adjacent to tumour is indicative of transmyometrial spread. T1-weighted images are helpful in this regard. Increased nodal size (greater than 1 cm) is required for the detection of metastatic lymphadenopathy. Detection rates of nodal metastases are similar to those of CT. Interruption of normal tissue planes and focal increased T2 signal intensity within the rectal or bladder wall is indicative of tumour invasion (stage IVa). MRI has been shown to be useful as an adjunct to clinical staging in biopsy-proven cases. The overall accuracy of gadolinium-enhanced MRI in the staging of endometrial carcinoma is between 84% and 94%. The staging accuracy drops to as low as 75% for unenhanced MRI.

Recurrence

The pattern of tumour recurrence usually depends on the type of therapy. Patients treated with radiation and surgery typically present with distant metastases without local pelvic recurrence. Patients treated with surgery alone, on the other hand, usually present with parametrial, pelvic side wall, or vaginal apex treatment failures. The vast majority of recurrences occur within 3 years of treatment initiation. It is not uncommon, however, for early stage, low-grade tumours to recur late, often more than 5 years after the initiation of treatment therapy. When pelvic failures do occur, factors improving 5-year survival include late recurrence, size less than 2 cm, and vaginal location. Post treatment surveillance should be performed with greatest frequency during the first 3 years, as between 75 and 95% of recurrences occur during this time. Ultrasonography, CT, or MRI can be performed to confirm suspected pelvic recurrences. Contrast-enhanced abdominal and pelvic CT should be obtained periodically in patients with advanced disease or known nodal metastases.

HH