T2. Not all reporting guidelines are suitable for AI research in medical imaging #AI specific challenges require tailored guidelines, such as #CLAIM to ensure accurate reporting #RadAIchat doi.org/10.1148/ryai.2…

T2. Really great table that summarizes various guidelines Mike Klontzas #RadAIchat

T2. Creating a guideline tailored to medical imaging AI was the motivation behind the creation of the first version of CLAIM in 2020 #RadAIchat

T2. CLAIM focuses on all aspects of imaging AI taking into account details such as the image acquisition (point 13) #RadAIchat

T3. #CLAIM focuses on #AI research in medical imaging, while TRIPODStatement +AI deals with #AI powered prediction models Both certainly aim to improve reporting, but CLAIM includes specific aspects of medical imaging #RadAIchat

T3. 🚨In comparison to other guidelines CLAIM has been developed by 72 panelists across a variety of disciplines dealing with medical imaging AI from radiologists to pathologists and computer scientists and many others ✅ #RadAIchat

T3. 🚨Compared to TRIPOD+AI and other guidelines, CLAIM has abandoned the term “validation” since it has been deemed ambiguous. Authors are encouraged to use “internal testing” and “external testing” #RadAIchat

T3. #CLAIM focuses on #AI research in all types of medical imaging #radiology #pathology #opthomalogy #RadAIchat

T3. 🚨CLAIM does not have a separate point for disclosures and conflicts of interest which can be found in TRIPOD+AI #RadAIchat

T4. #CLAIM2024 requires details about image acquisition #RadAIchat

T4. #CLAIM is tailored to medical imaging research and addresses imaging-specific features such as image data acquisition, image pre-processing and model performance on medical images #RadAIchat

T4. CLAIM also focuses on data curation, reference standard, model training, evaluation and performance #AImodels #RadAIchat

T4. 📖 #CLAIM2024 necessitates the description of image de-identification methods #RadAIchat

T5. Panelists have voted not to include a separate section on imaging biomarkers/radiomics since this is a different topic, already covered in other guidelines such as CLEAR insightsimaging.springeropen.com/articles/10.11… #RadAIchat

T5. CLAIM could certainly help researchers to ensure high quality also in radiomics research I’d recommend, however, the CheckList for EvaluAtion of Radiomics research (CLEAR), which is specifically tailored to radiomics myesr EuSoMII #RadAIchat doi.org/10.1186/s13244…

T5. 🆘 Issues such as biomarker extraction, feature selection or region segmentation that are common in imaging biomarker research do not have a separate section in 📖CLAIM #RadAIchat

T5. The update panel members agreed that the checklist should not be extended to research on imaging biomarkers #RadAIchat

T5. The METhodological RadiomICs Score (METRICS), a quality assessment tool for radiomics research based on the CLEAR guideline, is now also available EuSoMII #RadAIchat doi.org/10.1186/s13244…

T5. CLEAR comes with an explanations, elaborations and examples paper to facilitate its use #RadAIchat doi.org/10.1186/s41747…

T5. If you still wonder if checklists like #CLAIM2024 should be used when preparing manuscripts, you can have a look at a recent commentary doi.org/10.4274/dir.20… #RadAIchat