Lamps: Learning Anatomy from Multiple Perspectives via Self-supervision in Chest Radiographs
This work addresses the challenge of developing foundation models for medical imaging that effectively capture human anatomy, which is incremental as it builds on existing self-supervised learning methods by incorporating anatomical perspectives.
The paper tackled the problem of learning anatomical features in medical imaging by proposing Lamps, a self-supervised learning method that uses consistency, coherence, and hierarchy of human anatomy as supervision signals on chest radiographs, achieving superior robustness and transferability across 10 datasets compared to 10 baseline models.
Foundation models have been successful in natural language processing and computer vision because they are capable of capturing the underlying structures (foundation) of natural languages. However, in medical imaging, the key foundation lies in human anatomy, as these images directly represent the internal structures of the body, reflecting the consistency, coherence, and hierarchy of human anatomy. Yet, existing self-supervised learning (SSL) methods often overlook these perspectives, limiting their ability to effectively learn anatomical features. To overcome the limitation, we built Lamps (learning anatomy from multiple perspectives via self-supervision) pre-trained on large-scale chest radiographs by harmoniously utilizing the consistency, coherence, and hierarchy of human anatomy as the supervision signal. Extensive experiments across 10 datasets evaluated through fine-tuning and emergent property analysis demonstrate Lamps' superior robustness, transferability, and clinical potential when compared to 10 baseline models. By learning from multiple perspectives, Lamps presents a unique opportunity for foundation models to develop meaningful, robust representations that are aligned with the structure of human anatomy.