A pre-operative radiological score to predict lymph node metastasis in advanced epithelial ovarian cancer

International audienceBackground: Lymphadenectomy is part of cytoreductive surgery for patients with advanced epithelial ovarian cancer (AEOC) in case of abnormal lymph nodes before and during surgery. The aim of this study was to develop and validate a pre-operative radiological score to predict pelvic and/or para-aortic lymph node metastasis (LNM) in patients with AEOC undergoing cytoreductive surgery. Methods: We conducted a multicentre retrospective study. The construction sample was composed of 53 patients operated within two referral centers. The validation sample was composed of 39 patients operated in a third referral center. The score was built with a logistic regression model with internal validation by bootstrap. Results: Two variables were associated with the prediction of pelvic and/or para-aortic LNM at computerized tomography (CT) and/or positron emission tomography (PET/CT): “para-aortic lymph node involvement” (adjusted diagnostic odds ratio) (aDOR) = 8.77 95CI [1.42–54.09], p = 0.02) and “colon involvement” (aDOR = 7.97 95CI [1.28–49.58], p = 0.03). Bootstrap procedure showed that the model was stable. The 2-points LNM pre-operative radiological score was derived from these 2 radiological variables and a high-risk group was identified for a score ≥ 1: the probability of pelvic and/or para-aortic LNM was 76%, the specificity was 85.7% 95CI [67.3–96.0] and the positive likelihood ratio was 3.6 95CI [1.4–9.7]. In the validation sample, a score ≥ 1 had a specificity of 78.3% and a LR+ of 1.2. Conclusion: LNM pre-operative radiological score could help the surgeon's decision to perform pelvic and para-aortic lymphadenectomy in patients with AEOC undergoing cytoreductive surgery. Trial registration: The research protocol was approved by the Ethics Committee for Research in Obstetrics and Gynecology (CEROG 2016-GYN 1003)