Investigator
Alexandra Hospital
A Deep Learning Approach for Classifying Benign, Malignant, and Borderline Ovarian Tumors Using Convolutional Neural Networks and Generative Adversarial Networks
Background/Objectives: Accurate preoperative characterization of ovarian masses is essential for appropriate clinical management, particularly for borderline ovarian tumors (BOTs), which are less common and often difficult to distinguish from benign or malignant lesions on ultrasound. Although expert subjective ultrasound assessment achieves high diagnostic accuracy, limited availability of highly trained sonologists restricts its widespread application. Artificial intelligence-based approaches offer a potential solution; however, the low prevalence of BOTs restricts the development of robust deep learning models due to severe class imbalance. This study aimed to develop a Convolutional Neural Network (CNN)-based classifier enhanced with Generative Adversarial Networks (GANs) to improve the discrimination of ovarian masses as benign, malignant, or BOT using ultrasound images. Methods: A total of 3816 ultrasound images from 636 ovarian masses were retrospectively analyzed, including 390 benign lesions, 202 malignant tumors, and 44 BOTs. To address class imbalance, a Deep Convolutional GAN (DCGAN) was used to generate 2000 synthetic BOT images for data augmentation. A three-class ensemble CNN model integrating VGG16, ResNet50, and InceptionNetV3 architectures was developed. Performance was assessed on an independent test set and compared with a baseline model trained without DCGAN augmentation. Results: The incorporation of DCGAN-generated BOT images significantly enhanced classification performance. The BOT F1-score increased from 68.4% to 86.5%, while overall accuracy improved from 84.7% to 91.5%. For BOT identification, the final model achieved a sensitivity of 88.2% and specificity of 85.1%. Class-specific AUCs were 0.96 for benign lesions, 0.94 for malignant tumors, and 0.91 for BOTs. Conclusions: DCGAN-based augmentation effectively expands limited ultrasound datasets and improves CNN performance, particularly for BOT detection. This approach demonstrates potential as a decision support tool for preoperative assessment of ovarian masses.
Molecular Profiling and Treatment Outcomes in Uterine Serous Carcinoma: Prognostic Role of Estrogen Receptor Expression
Background: Uterine serous carcinoma (USC) represents a rare but aggressive subtype of endometrial cancer, accounting for a disproportionate number of disease-related deaths. Although molecular classification has improved risk stratification, prognostic heterogeneity highlights the need for new prognostic markers. Methods: We retrospectively analyzed 83 patients with USC treated at our institution between 1 January 2015 and 31 December 2023. Clinicopathological characteristics, treatment strategies, molecular biomarkers accessed by immunohistology (TP53, ER, PR, HER2, and MMR status), and survival outcomes were collected. Patients were first staged by FIGO 2009 and retrospectively reclassified by FIGO 2023. Disease-free survival (DFS), progression-free survival (PFS), and overall survival (OS) were assessed using Kaplan–Meier and Cox regression analyses. Results: The majority of patients were presented with advanced disease (FIGO stage IIIC-IV). TP53 mutations were found in 88% of cases, HER2 amplification in 18%, and ER expression in 57.8%. ER-positive patients showed significantly improved DFS in the adjuvant setting compared with ER-negative patients, whereas no significant associations were observed for first-line PFS or OS in multivariable analyses. HER2 amplification was not associated with inferior survival in our cohort. The advanced stage remained an independent predictor of worse OS. Conclusions: USC is a biologically heterogeneous disease, and its treatment should be guided by its molecular profile. ER expression identifies a subset of patients with improved DFS, suggesting potential prognostic relevance in this high-risk histology.
