Computational pathology in ovarian cancer

Area of Science:

• Pathology
• Computational Imaging
• Artificial Intelligence in Oncology

Background:

• Histopathology is crucial for ovarian cancer diagnosis and grading, relying on visual assessment of cellular and architectural features.
• Computational imaging and AI can extract thousands of quantitative features from digital pathology slides, including those imperceptible to the human eye.
• Morphologic and spatial patterns from computational pathology are anticipated to be more informative biomarkers than current -omics data for complex tumor biology.

Purpose of the Study:

• To highlight the potential of computational pathology and AI in ovarian cancer research.
• To encourage the application and development of computational pathology tools for ovarian cancer.
• To bridge the gap in computational pathology adoption within the ovarian cancer field.

Main Methods:

• Utilizing computational imaging to analyze digital pathology slides.
• Applying artificial intelligence and machine learning to interpret image data.
• Quantifying thousands of visual and subvisual features, including nuclear characteristics, entropy, eccentricity, and fractal dimensions.

Main Results:

• Computational pathology can extract quantitative biomarkers beyond human visual assessment.
• AI and machine learning can explore and quantify the spatial organization of tissues and cells.
• Subvisual alterations in the pre-cancer microenvironment may aid early detection and prevention research.

Conclusions:

• Computational pathology offers a new paradigm for understanding ovarian cancer biology by integrating spatial and molecular data.
• There is a significant need to advance computational pathology applications in ovarian cancer for improved prevention, early detection, and treatment.
• Ovarian cancer research teams are encouraged to adopt and develop computational pathology tools to drive progress in the field.