ORIGINAL RESEARCH article
Front. Med.
Sec. Obstetrics and Gynecology
Volume 12 - 2025 | doi: 10.3389/fmed.2025.1562558
This article is part of the Research TopicAdvances in New Biomarkers for the Diagnosis and Therapy of Gynaecological TumoursView all 5 articles
Proteomic Alterations in Ovarian Cancer: Predicting Residual Disease Status Using Artificial Intelligence and SHAP-Based Biomarker Interpretation
Provisionally accepted
- 1Faculty of Medicine, İnönü University, Malatya, Türkiye
- 2İnönü University, Malatya, Malatya, Türkiye
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
High-grade serous ovarian cancer (HGSOC) is the most aggressive and prevalent subtype of ovarian cancer, and treatment outcomes are significantly influenced by residual disease status following neoadjuvant chemotherapy (NACT). This study aimed to develop and validate machine learning models for predicting residual disease status via pre-NACT proteomic data and to identify potential biomarkers through explainable artificial intelligence (XAI) methods. Proteomic data from 20 HGSOC patients treated with NACT were analysed and categorized into two groups on the basis of surgical outcomes: no residual disease (R0, n=14) and suboptimal residual disease (R1, n=6). Using the BORUTA variable selection method, 18 significant proteins were identified from an initial set of 97 differentially expressed proteins. Three machine learning models-random forest, support vector machines, and bootstrap aggregation with classification and regression trees-were developed and compared. The random forest model demonstrated superior performance, with an AUC of 0.955, an accuracy of 0.830, a sensitivity of 0.904, a specificity of 0.763, and an F1 score of 0.839. SHAP analysis identified five key proteins (P48637, O43491, O95302, Q96CX2, and P49189) as the most significant predictors of residual disease status. These proteins, including glutathione synthetase and the peptide-prolyl cis-trans isomerase FKBP9, were found to be significantly associated with chemotherapy resistance mechanisms. This study demonstrates the potential of integrating proteomic data with machine learning approaches for predicting treatment outcomes in HGSOC, offering insights for personalized therapeutic strategies and improved patient care. The identified protein signatures may serve as valuable biomarkers for predicting NACT response and guiding treatment decisions in HGSOC patients.
Keywords: High-grade serous ovarian cancer (HGSOC)1, Neoadjuvant chemotherapy (NACT)2, Machine Learning3, proteomic biomarkers4, SHAP analysis5
Received: 17 Jan 2025; Accepted: 03 Jul 2025.
Copyright: © 2025 Yasar and Melekoglu. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence:
Seyma Yasar, Faculty of Medicine, İnönü University, Malatya, Türkiye
Rauf Melekoglu, Faculty of Medicine, İnönü University, Malatya, Türkiye
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.