George Daskalakis

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.

Neoadjuvant Chemotherapy in Advanced Stage Endometrial Cancer: A Systematic Review and Meta-Analysis

Background and Objectives: Endometrial cancer is the most common gynecological malignancy in developed countries and is becoming increasingly prevalent. Early diagnosis and treatment may lead to lower rates of morbidity and mortality. The aim of the present meta-analysis is to investigate whether neoadjuvant chemotherapy (NACT) can enhance resectability, reduce tumor burden, and ultimately improve survival rates compared to primary surgery in patients with advanced endometrial cancer. Materials and Methods: All studies that examined the impact of NACT on survival outcomes of patients with advanced endometrial cancer were eligible for inclusion, including randomized and non-randomized interventional studies. Studies were identified by searching MEDLINE (1945–2024), Scopus (1941–2024), Google Scholar (2004–2024) and ClinicalTrials.gov (2000–2024). Data was selected and extracted by two reviewers based on the PRISMA guidelines. Results: Five retrospective studies with a cumulative total of 8658 patients were included. No statistically significant difference in overall survival was observed between patients who received NACT and those who underwent primary surgery (HR 0.91, 95% CI 0.79–1.04). NACT was associated with some perioperative advantages, though these did not translate into a survival benefit. Conclusions: The currently available evidence, which is limited to retrospective studies with significant heterogeneity, suggests that NACT does not confer a survival advantage over primary debulking surgery in advanced endometrial cancer. These findings should be considered hypothesis-generating, underscoring the need for prospective trials. NACT may still be a reasonable option for selected subgroups, such as frail patients, those with extensive peritoneal disease, or cases in which complete cytoreduction is unlikely with upfront surgery.

The Impact of Positive Peritoneal Cytology on the Survival Rates of Early-Stage-Disease Endometrial Cancer Patients: Systematic Review and Meta-Analysis

Background and Objectives: The impact of positive peritoneal cytology has been a matter of controversy in early-stage endometrial cancer for several years. The latest staging systems do not take into consideration its presence; however, emerging evidence about its potential harmful effect on patient survival outcomes suggests otherwise. In the present systematic review and meta-analysis, we sought to accumulate current evidence. Materials and Methods: Medline, Scopus, the Cochrane Central Register of Controlled Trials CENTRAL, Google Scholar and Clinicaltrials.gov databases were searched for relevant articles. Effect sizes were calculated in Rstudio using the meta function. A sensitivity analysis was carried out to evaluate the possibility of small-study effects and p-hacking. Trial sequential analysis was used to evaluate the adequacy of the sample size. The methodological quality of the included studies was assessed using the Newcastle–Ottawa scale. Results: Fifteen articles were finally included in the present systematic review that involved 19,255 women with early-stage endometrial cancer. The Newcastle–Ottawa scale indicated that the majority of included studies had a moderate risk of bias in their selection of participants, a moderate risk of bias in terms of the comparability of groups (positive peritoneal cytology vs. negative peritoneal cytology) and a low risk of bias concerning the assessment of the outcome. The results of the meta-analysis indicated that women with early-stage endometrial cancer and positive peritoneal cytology had significantly lower 5-year recurrence-free survival (RFS) (hazards ratio (HR) 0.26, 95% CI 0.09, 0.71). As a result of the decreased recurrence-free survival, patients with positive peritoneal cytology also exhibited reduced 5-year overall survival outcomes (HR 0.50, 95% CI 0.27, 0.92). The overall survival of the included patients was considerably higher among those that did not have positive peritoneal cytology (HR 12.76, 95% CI 2.78, 58.51). Conclusions: Positive peritoneal cytology seems to be a negative prognostic indicator of survival outcomes of patients with endometrial cancer. Considering the absence of data related to the molecular profile of patients, further research is needed to evaluate if this factor should be reinstituted in future staging systems.

The p16/ki-67 assay is a safe, effective and rapid approach to triage women with mild cervical lesions

Objective The aim of this study was to evaluate the diagnostic accuracy and efficiency of p16/ki-67 dual stain in the identification of CIN2+ lesions, in Greek women with ASCUS or LSIL cytology. Methods A total of 200 women, 20 to 60 years old, were enrolled in the study. All samples were cytologically evaluated and performed for p16/ki-67 and high-risk HPV (HR-HPV) test. All patients were referred to colposcopy for biopsy and histological evaluation. Three cervical cancer (CC) screening strategies were designed and the total direct medical costs of the procedures during our clinical trial were evaluated, from a healthcare perspective. Results HPV 16 as expected was the most common HR-HPV type followed by HPV 31 and HPV 51. The risk for CIN2+ was significantly higher in HPV 16/18 positive cases. p16/ki-67 demonstrated a high sensitivity for CIN2+ identification in both ASCUS and LSIL groups (90.4% and 95%, respectively). HR-HPV test with sensitivity 52.3% and 65.5%, as well as colposcopy with sensitivity 14.3% and 36% respectively in ASCUS and LSIL group, showed inferior results compared to p16/ki-67. The specificity of p16/ki-67 for ASCUS and LSIL was 97.2% and 95.2% respectively, inferior only to colposcopy: 100% and 100%, lacking however statistical significance. HR-HPV test instead, presented the lowest specificity: 76.4% and 71.4% respectively in comparison to the other two methods. From a healthcare perspective, the costs and benefits of the tests implementation for the annual screening and triaging, in three CC screening strategies, were also calculated and discussed. Conclusions The results of the study indicate that p16/ki-67 is a safe and rapid assay that could be used to detect CIN2+ among women with mild cervical lesions, presenting both high sensitivity and specificity and could minimize the psychological and economic burden of HPV screening.