Comparison between multimodal foundation models and radiologists for the diagnosis of challenging neuroradiology cases with text and images
Radiologists show a specific ability to benefit from the integration of textual and visual information, whereas multimodal models mostly rely on the clinical context to suggest diagnoses.
Authors
- Riyad Hanafi (ORCID: https://orcid.org/0000-0003-2282-5081)
- Yann Gaillandre
- Grégory Kuchcinski (ORCID: https://orcid.org/0000-0001-8206-198X)
- Emma de Sousa
- Matthieu Masy
- Najib Chalhoub (ORCID: https://orcid.org/0000-0001-7774-2967)
- Bastien Le Guellec (ORCID: https://orcid.org/0000-0002-3618-4791)
- Quentin Vannod‐Michel
- Aghilès Hamroun (ORCID: https://orcid.org/0000-0003-4988-206X)
- Cyril Bruge
- Victor Chaton
Institutions
- Inserm (FR)
- Université de Lille (FR)
- Institut Pasteur de Lille (FR)
- Centre Hospitalier Universitaire de Lille (FR)
- Hôpital Roger Salengro (FR)
- Centre Hospitalier de Lens (FR)
- Hôpital Saint-Philibert (FR)
- Centre Hospitalier de Valenciennes (FR)
Publication Details
- Journal
- Diagnostic and Interventional Imaging
- Published
- 2025-05-09
- DOI
- https://doi.org/10.1016/j.diii.2025.04.006
- Citations
- 17
- Primary Topic
- Radiology practices and education
- Type
- article
- Field-Weighted Citation Impact
- 23.12