A4-MR: Abnormality-aware Anatomical Alignment for MR
We introduce A4-MR, a 3D vision-language foundation model trained on approximately 50,000 MRI studies comprising 491,000 series, roughly half of which are paired with radiology reports.
ProofPaper ↗
Key points
- Recent advances in large-scale foundation models have enabled strong generalization across a wide range of clinical and research tasks in magnetic resonance imaging (MRI).
- A4-MR is designed to: (1) learn anatomy localization through large-scale pretraining with noisy labels; (2) align image and text representations at the anatomical level; and (3) learn distinct representations of healthy and abnormal anatomy.
- Using a frozen encoder, A4-MR also surpasses an end-to-end task-specific MedNeXt model for knee segmentation by +1.4 Dice and performs comparably to the best task-specific model for abdominal organ segmentation on AMOS, within 0.4 Dice.
- Together, these results demonstrate that anatomy-level multimodal alignment and abnormality-aware representation learning can enable a single MRI foundation model to support localization, segmentation, and pathology-related tasks.
Sources (1)
- [1]A4-MR: Abnormality-aware Anatomical Alignment for MRarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 9, 06:10 AM
We introduce A4-MR, a 3D vision-language foundation model trained on approximately 50,000 MRI studies comprising 491,000 series, roughly half of which are paired with radiology reports.
Recent advances in large-scale foundation models have enabled strong generalization across a wide range of clinical and research tasks in magnetic resonance imaging (MRI).
Extractive summary: sentences quoted from the sources.