Continual atlas-based segmentation of prostate MRI
This work addresses the challenge of continual learning in medical imaging for prostate cancer diagnosis, which is incremental as it adapts an existing atlas-based approach to a new context.
The paper tackled the problem of continual learning for medical image segmentation, specifically prostate MRI, by proposing Atlas Replay, which uses atlas-based segmentation with privacy-preserving prototypes to maintain consistency across changing training distributions. The results showed that Atlas Replay is robust and generalizes well to unseen domains while maintaining knowledge, outperforming end-to-end methods across seven publicly available datasets.
Continual learning (CL) methods designed for natural image classification often fail to reach basic quality standards for medical image segmentation. Atlas-based segmentation, a well-established approach in medical imaging, incorporates domain knowledge on the region of interest, leading to semantically coherent predictions. This is especially promising for CL, as it allows us to leverage structural information and strike an optimal balance between model rigidity and plasticity over time. When combined with privacy-preserving prototypes, this process offers the advantages of rehearsal-based CL without compromising patient privacy. We propose Atlas Replay, an atlas-based segmentation approach that uses prototypes to generate high-quality segmentation masks through image registration that maintain consistency even as the training distribution changes. We explore how our proposed method performs compared to state-of-the-art CL methods in terms of knowledge transferability across seven publicly available prostate segmentation datasets. Prostate segmentation plays a vital role in diagnosing prostate cancer, however, it poses challenges due to substantial anatomical variations, benign structural differences in older age groups, and fluctuating acquisition parameters. Our results show that Atlas Replay is both robust and generalizes well to yet-unseen domains while being able to maintain knowledge, unlike end-to-end segmentation methods. Our code base is available under https://github.com/MECLabTUDA/Atlas-Replay.