Checklist for Artificial Intelligence in Medical Imaging (CLAIM)
There is the necessity of ensuring that scientific results are reproducible and that publications provide sufficient detail about the work to allow for validation of its rigor, quality, and generalizability. This allows other researchers to potentially reproduce the results.
So RSNA has come up with checklists and standards that have become widely utilized in the scientific community. This CLAIM is to the intended audiences such as
- Researchers and Scientists in the field of medical imaging and artificial intelligence: The article might provide insights, guidelines, or findings that are relevant to ongoing research in the domain of AI-driven medical imaging.
- Medical Professionals who utilize imaging technologies: This includes radiologists and clinicians who might be using or evaluating AI technologies in their diagnostic processes.
- AI Developers and Engineers in the health tech domain: They may find insights into best practices and challenges related to the deployment of AI in medical imaging.
- Students and Academics specializing in AI, data science, or medical imaging: For educational purposes, understanding developments, and challenges in the field.
- Policy Makers and Regulatory Authorities: Those who are involved in setting guidelines and regulations for AI applications in healthcare might utilize such articles to comprehend the scientific and ethical considerations.
Basic checklist items an author has to follow while Writing and Reporting AI in Medical Imaging Research:
Clarity and Reproducibility
- Clearly describe the methodology and techniques used.
- Ensure that the work is presented in a way that allows for reproducibility by other researchers.
Adherence to Reporting Standard
Follow relevant reporting standards and guidelines (e.g., STARD, STROBE, and CONSORT) to ensure comprehensive and standardized reporting.
Data Handling
- Clearly describe data sources, collection methods, and preprocessing steps.
- Address data privacy and ethical considerations.
Model Development and Validation
- Provide detailed information on model development, including architecture, training, and validation.
- Clearly report model performance metrics and validation procedures.
Results Presentation
- Ensure that results are clearly presented and supported with appropriate statistical analyses.
- Include visual aids (like graphs and charts) for better clarity and comprehension.
Discussion and Implications
- Clearly discuss the findings, their significance, and potential implications.
- Address limitations and propose future work.
Ethical Considerations
- Address ethical considerations related to AI model development and deployment.
- Ensure that patient data is handled ethically and in compliance with relevant regulations.
References and Citations
- Properly cite all sources, data, and methodologies referenced in the article.
- Ensure that all claims and statements are supported with appropriate references.
Peer Review and Feedbac
Consider obtaining peer review and feedback to enhance the quality and rigor of the article.
Accessibility and Inclusivity
- Ensure that the article is written in an accessible manner for a wide audience.
- Consider the broader impacts and inclusivity aspects of the work.
Transparency and Conflict of Interest
- Be transparent about any potential conflicts of interest.
- Clearly state funding sources and affiliations.
Remember, specific journals or publications might have their own guidelines and checklists that authors should adhere to. Always refer to these when preparing a manuscript for submission to a particular venue.
Checklist and standards used for reproducing the scientific results and ensure the intended information is exactly transferred without loss of rigor and quality.
