Back to search

National and Kapodistrian University of Athens

Τεχνητή νοημοσύνη και γυναικολογικό υπερηχογράφημα

Abstract

dc:description

BACKROUND: Preoperative discrimination between benign, malignant, and borderline ovarian tumors remains a major clinical challenge, as it directly influences therapeutic planning and patient prognosis. The aim of this study was to develop and evaluate an artificial intelligence–based system for automated classification of ovarian masses using ultrasound images. MATERIALS AND METHODS: This retrospective diagnostic accuracy study was conducted in two sequential phases. In phase I, a binary classification model (benign vs malignant) based on convolutional neural networks was developed. In phase II, the system was extended to a three-class classification framework (benign, malignant, borderline ovarian tumors), incorporating synthetic data augmentation using generative adversarial networks to address class imbalance. All images underwent anonymization, normalization, and data augmentation prior to model training. RESULTS: Model performance was assessed using accuracy, sensitivity, specificity, receiver operating characteristic curves, and area under the curve (AUC) across multiple decision thresholds, with validation performed through k-fold cross-validation. The system demonstrated high diagnostic performance, achieving AUC values up to 0.96 for benign lesions, 0.94 for malignant lesions, and 0.91 for borderline tumors. Transitioning from binary to multiclass classification combined with synthetic data generation resulted in measurable and consistent performance improvement, particularly for borderline ovarian tumors. CONCLUSIONS: The proposed artificial intelligence model demonstrates high classification accuracy and has the potential to function as a clinical decision support tool for the evaluation of ovarian masses using ultrasound images. Integration of such technologies into clinical practice may enhance diagnostic accuracy, reduce operator dependence, and contribute to personalized patient management.

Degree

thesis:*
Grantor dc:publisher
National and Kapodistrian University of Athens
Year dc:date
2026

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Giourga, Maria
  • Γιούργα, Μαρία

Subjects

dc:subject × 21

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/61575

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Giourga, Maria; Γιούργα, Μαρία. Τεχνητή νοημοσύνη και γυναικολογικό υπερηχογράφημα. National and Kapodistrian University of Athens, 2026. http://hdl.handle.net/10442/hedi/61575