Robust and Efficient Medical Imaging with Self-Supervision
This work addresses the challenge of data-efficient generalization for medical imaging AI, which is crucial for deploying robust systems across diverse clinical settings without extensive site-specific data, representing an incremental but important step forward in the field.
The paper tackles the problem of sub-optimal out-of-distribution performance and data inefficiency in medical imaging AI by introducing REMEDIS, a unified representation learning strategy that combines large-scale supervised transfer learning with self-supervised learning, resulting in up to 11.5% relative improvement in diagnostic accuracy and matching strong supervised baselines using only 1% to 33% of retraining data.
Recent progress in Medical Artificial Intelligence (AI) has delivered systems that can reach clinical expert level performance. However, such systems tend to demonstrate sub-optimal "out-of-distribution" performance when evaluated in clinical settings different from the training environment. A common mitigation strategy is to develop separate systems for each clinical setting using site-specific data [1]. However, this quickly becomes impractical as medical data is time-consuming to acquire and expensive to annotate [2]. Thus, the problem of "data-efficient generalization" presents an ongoing difficulty for Medical AI development. Although progress in representation learning shows promise, their benefits have not been rigorously studied, specifically for out-of-distribution settings. To meet these challenges, we present REMEDIS, a unified representation learning strategy to improve robustness and data-efficiency of medical imaging AI. REMEDIS uses a generic combination of large-scale supervised transfer learning with self-supervised learning and requires little task-specific customization. We study a diverse range of medical imaging tasks and simulate three realistic application scenarios using retrospective data. REMEDIS exhibits significantly improved in-distribution performance with up to 11.5% relative improvement in diagnostic accuracy over a strong supervised baseline. More importantly, our strategy leads to strong data-efficient generalization of medical imaging AI, matching strong supervised baselines using between 1% to 33% of retraining data across tasks. These results suggest that REMEDIS can significantly accelerate the life-cycle of medical imaging AI development thereby presenting an important step forward for medical imaging AI to deliver broad impact.