433 Myometrial infiltration assessment in low-risk endometrial cancer by 3D transvaginal ultrasound and diffusion-weighted magnetic resonance imaging

Introduction/Background* In patients with early-stage, grade 1-2, endometrioid endometrial cancer, preoperative assessment of myometrial invasion is essential to define the need of pelvic and paraaortic lymph node dissection. Our aim was to evaluate the role of three-dimensional transvaginal ultrasound (3D-TVUS) and diffusion-weighted magnetic resonance imaging (DW-RMI) for the assessment of myometrial infiltration in patients with well-differentiated (G1) or moderately differentiated (G2) endometrioid endometrial carcinoma (EC). Methodology We performed a retrospective observational study. Myometrial infiltration was assessed by 3D-TVUS and DW-MRI in 152 women with G1 or G2 endometroid EC who underwent surgical treatment in a tertiary referral center between 2012 and 2019. Sensitivity, specificity, predictive values and accuracy for the two techniques and for a combination of both were computed. Definitive histopathological data in the surgical specimen regarding myometrial infiltration was used as ‘Gold Standard’. Result(s)* One hundred and fifty-two patients were included, 120 (79%) patients presented myometrial infiltration 50%). 3D-TVUS and DW-MRI showed an agreement of 78.9% with a kappa index of 0.44 for the detection of deep myometrial infiltration. Sensitivity, specificity and accuracy of 3D-TVUS for the detection of deep myometrial infiltration were 71.0%, 80.5% and 78.5% respectively. Evaluation of myometrial infiltration with DW-MRI had a sensitivity, specificity and accuracy of 76.2%, 84.4% and 82.9% respectively. Association of both techniques increased sensitivity and specificity up to 84.6% and 93.2% and provided a low false negative rate (2.3%). Conclusion* The combination of 3D-TVUS and DW-RMI offers a high sensitivity and specificity to identify deep myometrial infiltration in patients with endometrioid G1 or G2 EC, thus these patients might benefit from performance of both techniques in preoperative evaluation.