What the study found
The study found that cognitive and communication impairment, along with respiratory compromise, were independently associated with poor cardiac MR image quality. The authors also reported that the image-quality labels assigned by a large language model (LLM) agreed substantially with interpreters' assessments.
Why the authors say this matters
The abstract states that poor cardiac MR image quality can prompt repeat examinations and hinder clinical decision-making. The study suggests that identifying pre-imaging clinical factors associated with image quality may be relevant for understanding which patients are more likely to have poor-quality scans.
What the researchers tested
The researchers conducted a retrospective study of 1,006 adults undergoing clinical cardiac MR examinations on 1.5 T and 3 T scanners. They used a HIPAA-compliant LLM to assign image-quality labels from radiology reports, extracted pre-imaging clinical and patient variables from electronic health records, and tested associations with image quality using univariate analysis and multivariable logistic regression.
What worked and what didn't
The binarized image-quality labels showed substantial agreement with interpreters' assessments in both the primary dataset and the repeat-adjusted sensitivity dataset. Cognitive and communication impairment and respiratory issues were independently associated with poor image quality, and these associations remained significant in the repeat-adjusted analysis. Other clinical variables were not independently associated after adjustment.
What to keep in mind
This was a retrospective study at a single clinical setting, based on adult cardiac MR examinations. The abstract does not describe limitations beyond the note that the sensitivity analysis used a repeat-imaging-adjusted image-quality label.
Key points
- Cognitive and communication impairment was independently associated with poor cardiac MR image quality.
- Respiratory compromise was independently associated with poor cardiac MR image quality.
- An LLM-based labeling approach showed substantial agreement with interpreters' assessments.
- Other clinical variables were not independently associated after adjustment.
- The study analyzed 1,006 adults undergoing clinical cardiac MR examinations.
Disclosure
- Research title:
- Cognitive and respiratory issues were linked to poorer cardiac MR image quality
- Authors:
- Hong Yu, Masha Bondarenko, Ali Nowroozi, Y T. Lee, Adrian Serapio, Punita Kaveti, Jae Ho Sohn
- Institutions:
- Affiliated Hospital of Zunyi Medical College, University of California System, University of California System, University of California System, University of California System, University of California System, University of California System, University of California, San Francisco, University of California, San Francisco, University of California, San Francisco, University of California, San Francisco, University of California, San Francisco, University of California, San Francisco, University of California, San Francisco, Zunyi Medical University
- Publication date:
- 2026-04-19
- DOI:
- 10.1002/jmri.70336
- OpenAlex record:
- View
Get the weekly research newsletter
Stay current with scholarly research without reading academic papers — one filtered digest, every Friday.