The value of ultrasound in preoperative assessment of the myometrial and cervical invasion in endometrial carcinoma.

This study was undertaken to determine the accuracy of ultrasonography in depicting the depth of myometrial invasion and cervical involvement in women with adenocarcinoma of the endometrium. During the past 2 years, pelvic ultrasound using abdominal and vaginal probes was carried out as part of routine preoperative work-up in 69 consecutive patients with carcinoma of the endometrium. Fifty-eight patients were primarily operated on; the remaining 11 had preoperative intracavitary irradiation followed by surgery at the referral hospital. All patients underwent total abdominal hysterectomy with bilateral salpingo-oophorectomy and staging laparotomy. A Hitachi EUB 450 ultrasound machine with a high-resolution 6.5-MHz transvaginal transducer was used to assess the intrauterine extension of the tumor. Myometrial extension was measured from the endometrial-myometrial interface to the deepest edge of the tumor extension into the myometrium and was categorized as 0%, < 50%, or > 50% myometrial invasion. Cervical involvement was determined based on irregular echogenicity of the tumor in the cervix of the uterus. Sixty-three of 69 patients had no cervical involvement on ultrasound, whereas the cervix was negative in 60 subjects on pathological examination; i.e., there were 3 false-negative and no false-positive results. The correlation was highly significant (P < 0.001). Of these, the endocervix was involved in 5 cases (stage IIa) and there were 4 patients with stage IIb disease. All 3 false-negative cases were stage IIa, with microscopical involvement in 2 patients. In the third case histology demonstrated disease extending low to the inner cervical os and was considered involvement of the cervix. The inner half of myometrium was involved in 28 patients and the outer half in 31, and there was no myometrial invasion in 10 cases. The corresponding values for ultrasound were 29, 31, and 9, respectively; i.e., there was only 1 false-positive and no false-negative results. The correlation is highly significant (P < 0.001). Sonography revealed irregular interface between the endometrium and myometrium, giving the appearance of measurable myometrial involvement in the only woman with a false-positive result. Our results indicate that transvaginal sonography correctly predicts endometrial and cervical involvement in endometrial carcinoma. In evaluating these factors, ultrasound using transvaginal probes is probably more accurate than intraoperative gross evaluation or frozen section analysis. Preoperative knowledge of these variables allows the gynecologists an appropriate consultation in terms of extended surgery and to select those patients who might benefit from referral to gynecologic oncology centers.