Transvaginal sonography vs. hysteroscopy in the preoperative staging of endometrial carcinoma.

The aim of this study was to evaluate the diagnostic accuracy of transvaginal sonography vs. hysteroscopy in the preoperative staging of endometrial carcinoma, in order to establish the most appropriate surgical therapy. Transvaginal sonography and hysteroscopy were used preoperatively in 67 women with histologically proven endometrial carcinoma. Deep myometrial invasion (> 50%) was present at postoperative pathology in 26/67 (39%) women and spread to the cervix occurred in 11/67 (16%) women. Transvaginal sonographic examination was initially directed at assessing myometrial invasion, which was correctly predicted in 52/67 (78%) women. Transvaginal sonography demonstrated a sensitivity of 88% (23/26) and a specificity of 71% (29/41) for deep invasion, with a positive predictive value (PPV) of 66% (23/35) and a negative predictive value (NPV) of 91% (29/32). The accuracy of transvaginal sonography in detecting cervical involvement was 82% (55/67), and that of hysteroscopy was 72% (48/67): transvaginal sonography was slightly less sensitive (54% vs. 64%), but more specific (87% vs. 73%) than hysteroscopy. When cervical invasion was present, the PPVs of transvaginal sonography and hysteroscopy were 46% (6/13) and 32% (7/22), respectively, while the NPV was 91% for both techniques (49/54; 41/45). Our data show that the accuracy of transvaginal sonography was comparable with that of hysteroscopy in detecting cervical involvement. Therefore, in the majority of cases, when both techniques showed that disease was limited, the appropriately limited type of surgery would be performed. Conversely, detection of a myometrial invasion of > 50% or an extension to the cervix would lead to an unnecessarily extensive operation if this was the sole criterion used for making the decision.