AJNR Study-Specific Guidelines for AI in Medical Imaging: Bridging Gaps in Reference Standard and Clinical Evaluation

Area of Science:

• Neuroradiology
• Artificial Intelligence
• Medical Imaging

Background:

• AI applications in neuroradiology are rapidly increasing.
• Heterogeneous study designs and reporting hinder reproducibility and clinical translation of AI tools.
• Existing guidelines may not fully address the nuances of AI imaging research.

Purpose of the Study:

• To introduce six new study-type-specific reporting checklists for AI imaging research submitted to AJNR.
• To provide authors with standardized frameworks for reporting AI studies.
• To enhance the quality, reproducibility, and clinical relevance of AI research in neuroradiology.

Main Methods:

• Development of six distinct checklists tailored to specific AI study types: Classification/Prediction, Detection/Segmentation, Image Transformation, Clinical AI Tool Evaluation, Workflow Optimization, and Technical Developments.
• Informed by collective experience from manuscript reviews and real-world AI deployments.
• Summarization of key elements within each checklist in a technical note format.

Main Results:

• Introduction of six comprehensive reporting checklists for AI neuroradiology studies.
• Checklists address common pitfalls and gaps identified in AI research.
• The checklists cover diverse AI applications within neuroradiology.

Conclusions:

• The proposed checklists will support authors in submitting high-quality AI imaging research.
• Standardized reporting is crucial for bridging the gap between AI development and clinical radiology practice.
• This initiative complements existing guidelines, emphasizing rigorous standards and clinical relevance.