Evaluation of survey and perioperative results using artificial intelligence in stage II-IV ovarian cancer

Introduction: Stage II-IV ovarian cancer is associated with the highest mortality rates among gynecological cancers. Predicting prognostic factors is critical for making accurate surgical and treatment planning decisions. This study evaluates the effectiveness of artificial intelligence and machine learning algorithms in predicting perioperative outcomes such as overall survival, hospital stay duration, and ICU stay duration in ovarian cancer patients. Materials and Methods: This retrospective study included 174 patients who underwent surgery for ovarian cancer between 2007 and 2024. Clinical and pathological data were collected, and several machine learning algorithms, including Gradient Boosting Classifier, Logistic Regression, Dummy Regressor, and Bayesian Ridge, were applied to analyze survival time, hospital stay duration, and ICU stay duration. Results: The Gradient Boosting Classifier and Logistic Regression algorithms showed the best performance in predicting survival/mortality outcomes with an 84% accuracy. However, Dummy Regressor and Bayesian Ridge models demonstrated low performance in predicting hospital and ICU stay durations. Optimal cytoreduction was associated with significantly longer survival, whereas suboptimal surgeries (R2) had a negative impact on prognosis. Conclusion: Artificial intelligence and machine learning algorithms can be effective tools for evaluating prognostic factors in stage II-IV ovarian cancer. However, further optimization is required for predicting complex outcomes such as hospital stay duration. Future prospective studies with larger datasets could facilitate the integration of these technologies into clinical practice. Key Words: Ovarian Cancer, Artificial Intelligence, Machine Learning, Cytoreduction