Vision Foundation Models in Radiology: A Scoping Review of Data, Methodology, Evaluation and Clinical Translation
arXiv:2607.07219v1 Announce Type: cross Abstract: Vision foundation models (VFMs) are increasingly being developed for radiological imaging, yet their definition, development and evaluation remain heterogeneous. We conducted a PRISMAScR scoping review of peer-reviewed studies published between January 2017 and March 2026 describing foundation models trained exclusively on radiological imaging data. Sixty-seven studies were included and mapped across three pillars: data scale and heterogeneity, architectural and pretraining scalability, and downstream transferability and generalization. Dataset
The proliferation and increasing specialization of large AI models necessitates systematic reviews to consolidate understanding and identify challenges in application areas like radiology.
This scoping review provides critical insights into the current state, challenges, and future directions of Vision Foundation Models in a highly sensitive and regulated field, indicating where investment and research are concentrating.
The systematic mapping of data, methodology, and evaluation in radiology VFMs highlights the need for standardized practices, potentially accelerating the development and clinical translation of reliable AI diagnostic tools.
- · AI developers in medical imaging
- · Healthcare providers
- · Patients
- · Medical AI research institutions
- · Companies with proprietary, non-interoperable AI solutions
- · Traditional radiology training programs slow to adapt to AI integration
Improved and more accessible AI-driven diagnostic tools in radiology due to clearer development pathways.
Increased competition and consolidation among medical AI companies as best practices and standards emerge.
Potential for AI to reshape medical education and specialist training, focusing more on AI oversight and integration rather than purely manual interpretation.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG