Introduction/Background Preoperative assessment of myometrial infiltration in endometrial cancer may help to identify cases that may benefit from extended surgery. The aim of this study is to analyse the diagnostic accuracy of two-dimensional transvaginal ultrasound (2D-TVU) for detecting deep myometrial invasion (MI) and cervical stromal involvement (CSI) of endometrial cancer.

Methodology All patients with confirmed endometrial cancer underwent a 2D-TVS before surgery. MI and CSI were assessed by subjective impression and recorded prospectively. At the time of the operation the specimens were sent for frozen section and the decision of performing lymph node dissection (LND) was taken intraoperatively. Final histology after hysterectomy served as gold standard.

Results Twenty-two patients were included. The mean age was 64.3 years old (43–84). The histology was endometrioid in 19 patients (86.4%), serous in 2 (9.1%) and Mixed Müllerian Malignant Tumour in 1 (4.5%). Superficial MI (<50%) on ultrasound was reported in 15 patients (68.2%) and deep MI (>50%) in 6 (27.3%). Just in one case the cervix was thought to be affected (4.5%). All patients had hysterectomy and bilateral salpingo-oophorectomy, 10 patients underwent bilateral pelvic LND and 4 para-aortic LND, based on MI on frozen section and histological type. Final histology showed pT1A in 14 patients (63.6%), pT1B in 6 (27.3%) and pT2 in 2 (9%). There was one case of no correlation between frozen section and final histology and the patient had a re-operation for LND. Two patients had pelvic and para-aortic lymph node involvement. Accuracy for diagnosing deep MI with 2D-TVU was 86.4%, and for CSI was 95.5%.

Conclusion Ultrasound evaluation by subjective impression seems to be a good method for assessing MI and CSI in endometrial cancer. The use of preoperative ultrasound in combination with frozen section improve diagnostic performance.

Disclosure Nothing to disclose.