Search papers, labs, and topics across Lattice.
This scoping review analyzes 67 peer-reviewed studies on Vision Foundation Models (VFMs) in radiology, focusing on data, methodology, evaluation, and clinical translation. The findings reveal significant variability in data scale, model architecture, and evaluation practices, highlighting a predominance of transformer-based architectures and self-supervised pretraining techniques. While VFMs demonstrate potential for transferability across various imaging modalities, challenges such as limited data representativeness and inconsistent evaluation hinder their clinical application.
Radiology's Vision Foundation Models show promise, but their clinical impact is stymied by inconsistent evaluation practices and data limitations.
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. Datasets primarily covered brain MRI, thoracoabdominal CT, and chest X-ray, ranging from fewer than 100,000 samples to multi-million-image cohorts. Transformer-based architectures and self-supervised pretraining predominated, particularly masked image modeling, contrastive learning and multi-stage approaches. Evaluation focused mainly on segmentation and classification, whereas cross-center, cross-scanner, anatomical and modality-shift validation was inconsistently reported. Alignment with FUTURE-AI principles was uneven. Overall, radiology-specific VFMs show promising transferability, but clinical translation remains constrained by limited data representativeness, heterogeneous benchmarks, incomplete reporting and insufficient deployment-oriented evaluation.
