Li Yuheng, Wynne Jacob F, Wu Yizhou, Qiu Richard L J, Tian Sibo, Wang Tonghe, Patel Pretesh R, Yu David S, Yang Xiaofeng
Department of Radiation Oncology and Winship Cancer Institute, Emory University, Atlanta, GA 30322, USA; Department of Biomedical Engineering, Emory University and Georgia Institute of Technology Atlanta, GA 30308, USA.
Department of Radiation Oncology and Winship Cancer Institute, Emory University, Atlanta, GA 30322, USA.
Radiother Oncol. 2025 Mar;204:110711. doi: 10.1016/j.radonc.2025.110711. Epub 2025 Jan 9.
This study aims to develop a robust, large-scale deep learning model for medical image segmentation, leveraging self-supervised learning to overcome the limitations of supervised learning and data variability in clinical settings.
We curated a substantial multi-center CT dataset for self-supervised pre-training using masked image modeling with sparse submanifold convolution. We designed a series of Sparse Submanifold U-Nets (SS-UNets) of varying sizes and performed self-supervised pre-training. We fine-tuned the SS-UNets on the TotalSegmentator dataset. The evaluation encompassed robustness tests on four unseen datasets and transferability assessments on three additional datasets.
Our SS-UNets exhibited superior performance in comparison to state-of-the-art self-supervised methods, demonstrating higher Dice Similarity Coefficient (DSC) and Surface Dice Coefficient (SDC) metrics. SS-UNet-B achieved 84.3 % DSC and 88.0 % SDC in TotalSegmentator. We further demonstrated the scalability of our networks, with segmentation performance increasing with model size, demonstrated from 58 million to 1.4 billion parameters:4.6 % DSC and 3.2 % SDC improvement in TotalSegmentator from SS-UNet-B to SS-UNet-H.
We demonstrate the efficacy of self-supervised learning for medical image segmentation in the CT, MRI and PET domains. Our approach significantly reduces reliance on extensively labeled data, mitigates risks of overfitting, and enhances model generalizability. Future applications may allow accurate segmentation of organs and lesions across several imaging domains, potentially streamlining cancer detection and radiotherapy treatment planning.
本研究旨在开发一种强大的大规模深度学习模型用于医学图像分割，利用自监督学习克服监督学习的局限性以及临床环境中的数据变异性。
我们使用带有稀疏子流形卷积的掩蔽图像建模来策划一个用于自监督预训练的大量多中心CT数据集。我们设计了一系列不同大小的稀疏子流形U-Net（SS-UNet）并进行自监督预训练。我们在TotalSegmentator数据集上对SS-UNet进行微调。评估包括对四个未见数据集的稳健性测试以及对另外三个数据集的可迁移性评估。
与最先进的自监督方法相比，我们的SS-UNet表现出卓越的性能，展示出更高的骰子相似系数（DSC）和表面骰子系数（SDC）指标。SS-UNet-B在TotalSegmentator中实现了84.3%的DSC和88.0%的SDC。我们进一步证明了我们网络的可扩展性，随着模型大小从5800万参数增加到14亿参数，分割性能提高：在TotalSegmentator中，从SS-UNet-B到SS-UNet-H，DSC提高了4.6%，SDC提高了3.2%。
我们证明了自监督学习在CT、MRI和PET领域医学图像分割中的有效性。我们的方法显著减少了对大量标注数据的依赖，降低了过拟合风险，并增强了模型的通用性。未来的应用可能允许在多个成像领域对器官和病变进行准确分割，潜在地简化癌症检测和放射治疗计划。
