The tumour histopathology "glossary" for AI developers

Area of Science:

• Computational pathology
• Digital pathology
• Artificial intelligence in oncology

Background:

• Artificial intelligence (AI) and deep learning (DL) show promise in analyzing histopathology images for cancer research.
• Translating AI/DL methods into clinical practice requires computational researchers to understand histopathology.
• A knowledge gap exists between computational methods and histopathological interpretation.

Purpose of the Study:

• To bridge the knowledge gap between computational researchers and histopathology.
• To provide essential histopathology concepts for AI developers in cancer research.
• To accelerate the development of AI algorithms for cancer research.

Main Methods:

• Introduction to key cell types in histopathology (epithelial, stromal, immune cells).
• Explanation of malignancy, precursor lesions, and the tumor microenvironment (TME).
• Overview of foundational histopathology techniques: Hematoxylin and Eosin (HE) staining, immunohistochemistry, and multiplexed antibody staining.

Main Results:

• Essential histopathology concepts are defined and illustrated.
• Foundational staining techniques are introduced.
• Key features of cancer pathology relevant to AI analysis are discussed.

Conclusions:

• Understanding histopathology is crucial for effective AI development in cancer research.
• This work equips AI developers with foundational knowledge to enhance AI-driven cancer diagnostics.
• Bridging the expertise gap will accelerate AI applications in oncology.