Annotating for Artificial Intelligence Applications in Digital Pathology: A Practical Guide for Pathologists and...

Area of Science:

• Pathology
• Artificial Intelligence
• Machine Learning

Background:

• Training machine learning models for AI in pathology requires extensive expert annotation.
• There is a lack of standardized guidance for this annotation process.
• Annotation methodologies vary based on study objectives, yet common difficulties persist.

Purpose of the Study:

• To describe experiences and provide a practical guide for annotation strategies in computational pathology AI development.
• To address common difficulties encountered during the annotation process.
• To assist pathologists, researchers, and AI developers in pathology annotation.

Main Methods:

• Summarizing key aspects of annotation methodology.
• Issuing guiding principles for team interaction and ground-truth quality assessment.
• Discussing different annotation types, software, and hardware options.

Main Results:

• A practical guide for annotation strategies in computational pathology.
• Identification of common difficulties and solutions.
• Principles for effective team collaboration and quality control.

Conclusions:

• Effective annotation is crucial for successful AI development in pathology.
• This guide provides a framework to overcome common annotation challenges.
• It aims to enhance the quality and efficiency of pathology data annotation for AI.