Computational pathology: A survey review and the way forward

Area of Science:

• Computational Pathology (CPath) as an interdisciplinary field.
• Integration of computational approaches for medical histopathology image analysis.

Background:

• Digital pathology enables data flow for deep learning and computer vision in cancer diagnostics.
• A significant gap exists in adopting CPath algorithms into clinical practice despite extensive research.

Purpose of the Study:

• To provide a comprehensive review of over 800 papers in Computational Pathology.
• To address challenges from problem design to application and implementation of CPath tools.
• To map the current landscape and future directions of CPath.

Main Methods:

• Cataloging papers into model-cards by examining key works and challenges.
• Analyzing CPath developments across data-centric, model-centric, and application-centric perspectives.
• Reviewing challenges in the full cycle of CPath development and integration.

Main Results:

• Identification of key challenges in CPath development and clinical integration.
• A structured overview of the current state of Computational Pathology research.
• Cataloging of over 800 papers with associated model cards for community reference.

Conclusions:

• Highlighting remaining challenges in technical development and clinical integration of CPath.
• Providing directions for future advancements in Computational Pathology.
• Facilitating community understanding and adoption of CPath tools for cancer diagnosis and treatment.