Analysis of Epithelial Ovarian Cancer Subtypes Classification on Histopathological Images Using Deep Learning Techniques

Epithelial ovarian cancer is an aggressive form of gynecologic malignancy that significantly contributes to morbidity and mortality rates among women globally. Effective diagnosis and classification of its subtypes such as High Grade Serous Carcinoma (HGSC), Clear Cell Ovarian Carcinoma (CC), Endometrioid Carcinoma (EC), Low Grade Serous Carcinoma (LGSC), and Mucinous Carcinoma (MC) are crucial for accurate treatment planning and prognosis. This study employs advanced deep learning techniques to automate the classification and analysis of these subtypes using histopathology images. The preprocessing steps involve grayscale conversion, edge detection, histogram equalization, and Contrast Limited Adaptive Histogram Equalization (CLAHE) to improve image quality and maintain consistency during feature extraction. To achieve effective classification, multiple deep learning models are evaluated, including a modified Convolutional Neural Network (CNN), VGG16, ResNet, and DenseNet. These architectures are further integrated with techniques for precise tumor region localization. A thorough accuracy assessment is conducted to determine the most effective model for subtype classification. This research aims to provide a scalable and reliable framework for diagnostic support, enhancing patient care by facilitating personalized treatment strategies through innovative deep learning approaches.