Best Practices and Checklist for Reviewing Artificial Intelligence-Based Medical Imaging Papers: Classification.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Machine Learning

Background:

• Artificial Intelligence (AI) is rapidly advancing medical imaging analysis, leading to numerous classification studies.
• Current review criteria for AI studies are often subjective, hindering reproducible science.
• The Society for Imaging Informatics in Medicine (SIIM) identified a need for standardized review guidelines.

Purpose of the Study:

• To establish best practices and guidelines for reviewing AI-based medical imaging classification studies.
• To ensure AI research papers are cohesive, reproducible, accurate, and self-contained.
• To provide a focused approach for evaluating AI image classification tasks.

Main Methods:

• The study outlines essential sections and information required for AI image classification manuscripts.
• It discusses key elements including dataset curation, preprocessing, reference standards, partitioning, model architecture, and training.
• The approach is presented from a machine learning practitioner's perspective.

Main Results:

• The work provides a structured framework for assessing the quality and reproducibility of AI medical imaging studies.
• It details necessary information for publication consideration, focusing specifically on image classification.
• This serves as a focused checklist compared to broader review guidelines.

Conclusions:

• Implementing these guidelines will enhance the review process for AI medical imaging papers.
• The proposed standards aim to improve the quality and reliability of published AI research.
• This initiative facilitates a consistent approach to presenting AI research in medical imaging.