Artificial Intelligence Systems as Prognostic and Predictive Tools in Ovarian Cancer

Area of Science:

• Oncology
• Medical Informatics
• Artificial Intelligence

Background:

• Personalized medicine requires accurate prognostic and predictive information for heterogeneous diseases like epithelial ovarian cancer.
• Conventional algorithms for epithelial ovarian cancer are often too complex for routine clinical application.
• There is a need for advanced tools to assist in clinical decision-making for ovarian cancer.

Purpose of the Study:

• To investigate the potential of an artificial intelligence (AI) model for providing prognostic and predictive information in epithelial ovarian cancer.
• To compare the performance of an AI model against conventional statistical approaches.
• To assess AI's utility in predicting patient survival and surgical outcomes.

Main Methods:

• A database of 668 epithelial ovarian cancer cases over 10 years was compiled with routinely collected clinical data and survival information.
• An AI model was developed to compare various algorithms and classifiers.
• Conventional statistical methods, including logistic regression, were used as a benchmark.

Main Results:

• An artificial neural network (ANN) algorithm accurately predicted overall survival with 93% accuracy and an AUC of 0.74.
• The ANN model outperformed logistic regression in survival prediction.
• ANN also predicted surgical outcomes (cytoreduction status) with 77% accuracy and an AUC of 0.73.

Conclusions:

• AI systems demonstrate potential in delivering prognostic and predictive data for epithelial ovarian cancer patients.
• The findings suggest a future role for AI in personalized cancer care.
• Further research with larger datasets is recommended to enhance AI model performance.