Week in Review: MRI and ultraprocessed food | Data transparency for AI | MICI Q3 results

Erik Ridley Headshot

Dear AuntMinnie Member,

Although it’s well known that ultraprocessed food increases risk for cardiovascular events, the association between ultraprocessed consumption and subclinical disease has remained unclear. 

With the use of MRI, however, researchers have now tied higher levels of consumption with greater atherosclerotic burden. You can get all of the details by reading our most popular story from last week. 

A lack of transparency on training data for AI algorithms can raise concerns about performance and generalizability. In a new study, researchers weren't able to find information on training or validation sets for half of the 24 AI- and machine learning-based medical devices cleared for Alzheimer’s disease and related dementia. And very few included external validation data.

Radiology administrators have mixed feelings about how medical imaging will fare in the coming months, according to The MarkeTech Group's Medical Imaging Confidence Index (MICI) third quarter 2025 report. In five areas of practice, administrators had an overall "neutral" confidence score

Other highly-read stories on AuntMinnie last week included coverage of a brain PET tracer being delivered intranasally and a report on how integrating radiologist eye gaze data into AI algorithms may make these models more of a collaborative partner. And visual reads of amyloid PET scans were deemed effective in real-world settings.

See below for the full list of our top stories from last week. 

  1. MRI reveals effects of ultraprocessed food on the heart

  2. Data transparency is lacking for FDA-cleared Alzheimer’s AI technology

  3. MICI Q3: Radiology administrators mixed on the state of medical imaging

  4. Brain PET tracer delivered intranasally for first time in humans

  5. Can eye gaze data from radiologists make AI more human?

  6. Visual reads of amyloid PET scans effective in real-world settings

  7. AI reduces false-negative rates on screening DBT

  8. ASRT survey shows vacancy rates still high among imaging departments

  9. AI spots breast lesions on targeted ultrasound

  10. fMRI confirms yet again that exercise really does improve mood

  11. Doximity: Radiology makes top 5 specialties for compensation growth

  12. AI helps clinicians identify misplaced endotracheal tubes

  13. 3 predictions on how AI could affect radiology

  14. CT may best conventional measurements for assessing BMD in diabetics

Erik Ridley
Editor in Chief
AuntMinnie.com

Page 1 of 13
Next Page