Article Text

Download PDFPDF

168 Preoperative evaluation of lipid markers of malignant epithelial ovarian tumors
  1. Mariia Iurova1,
  2. Stanislav Pavlovich1,
  3. Nataliia Starodubtseva2,
  4. Vitaliy Chagovets2,
  5. Grigorii Khabas2 and
  6. Vladimir Frankevich2
  1. 1Federal State Autonomous Educational Institution of Higher Education I.M. Sechenov First Moscow State Medical University of the Ministry of Health of the Russian Federation (Sechenov University); Federal State Budget Institution ‘national Medical Research Center for Obstetrics, Gynecology and Perinatology Named after Academician V.I. Kulakov’ Ministry of Health of Russia
  2. 2Federal State Budget Institution ‘National Medical Research Center for Obstetrics, Gynecology and Perinatology named after Academician V.I. Kulakov’ Ministry of Health of Russia, Moscow, Russian Federation


Introduction/Background The venous blood is repleted with abundant tumor-promoting factors and lipids, that play an essential role in ovarian high-grade serous carcinoma (HGSC). A comprehensive picture of mediators impacting HGSC progression is, however, not available.

Research question to determine the value of the serum lipid profile in HGSC for diagnosis.

Methodology This study was approved by the Institute Research Medical Ethics Committee. Analysis of blood serum lipids of healthy volunteers (n = 13, control group) and patients with verified HGSOC (I-IV stages, n = 28, main group): I-II stages (n=5), III-IV stages (n=23)) has been performed. Patients with HGSOC managed in the Department of Innovative Oncology and Gynecology (National Medical Research Center for Obstetrics, Gynecology and Perinatology named after Academician V.I. Kulakov) were comparable in age, body mass index, grade and FIGO stages. Lipids were analysed by high performance mass spectrometry liquid chromatography (HPLC-MS). The Orthogonal Projections to Latent Structures Discriminant Analysis (OPLS-DA) multifactor analysis method and non-parametric t-test, have been applied for statistical data processing. Random forest model was used to evaluate predictive performance of potential biomarkers based on leave-one-out cross-validation in terms of area under the receiver operating characteristic (ROC). The predictive accuracy of the predictive lipids was performed using the logistic regression modeling with AUC value.

Results In main group the levels of 128 of 345 studied lipids differed significantly compared to the control group (p≤0.05), the parameters of the OPLS-DA model were: R2 = 0.87, Q2 = 0.80; AUC=0.99. ROC curve sensitivity = 96% and specificity =1%, the AUC value of these metabolite combinations for predicting HGOC recurrence was 1. Lipid profile changes significantly differed between the groups: control group vs I-II stages (p≤0.05), control group vs III-IV stages (p≤0.05).

11 patients who developed the disease relapse or progression had significant preoperative increase of oxidized lysophosphatidylcholine (OxLPC) and phosphatidylethanolamine (PE) in contrast to 17 patients who showed no evidence of recurrence after at least 14 months of follow up.

Conclusion Lipid profile changes in HGSC may have considerable prognostic value for the disease after treatment. The signatures defined by our work may provide a basis for the development of prognostic tools and may predict the clinical course of HGSC patients.

This work was supported by RSF grant № 20-65-46014.

Disclosures Nothing to discloser.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.