DeepAdapter:整合自监督学习与无监督域适应的早产儿视网膜病变筛查算法
DeepAdapter 是一种整合自监督学习与无监督域适应的深度学习算法,先用 50 万张无标签婴儿彩色眼底照片进行自监督预训练,再用少量标注数据结合域适应模块,在外部测试集和公共跨种族数据集上准确率显著优于监督学习,并开发了网页筛查系统。
要点
- DeepAdapter 在外部测试集准确率 0.828,优于监督学习 0.739(p<0.001)。
- 在公共跨种族数据集准确率 0.839,优于监督学习 0.813(p=0.014)。
- CORAL 距离从监督学习的 0.297 降至 3.550×10-6,域偏移得到缓解。
核心结论
- DeepAdapter 有效缓解域偏移,显著提升 ROP 筛查模型的泛化能力。
研究设计与样本
- 两阶段开发:先用 50 万张无标签婴儿彩色眼底照片进行自监督学习,再用标注数据训练并加入无监督域适应模块。
- 共纳入 500 000 张无重叠无标签彩色眼底照片和 39 456 张有标签彩色眼底照片。
- 外部验证使用一个独立临床中心和两个公共多民族数据集。
主要发现
- 与监督方法相比,DeepAdapter 将 CORAL 距离从 0.297 降至 3.550×10-6。
- 内部测试中两种算法 ROP 预测准确率均为 0.989。
- 外部测试准确率:DeepAdapter 0.828(95%CI 0.823–0.833)vs 监督学习 0.739(95%CI 0.733–0.745),p<0.001。
- 公共跨种族数据集准确率:DeepAdapter 0.839(95%CI 0.818–0.859)vs 监督学习 0.813(95%CI 0.791–0.835),p=0.014。
临床怎么用
- 可集成质控与 ROP 诊断的网页系统,用于大规模多中心临床筛查。
- 为其他医学专科开发泛化模型提供参考。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
AIMS: To develop and validate DeepAdapter, a novel deep learning algorithm that integrates self-supervised learning (SSL) and unsupervised domain adaptation (UDA) to enhance model generalisability for retinopathy of prematurity (ROP) screening. METHODS: DeepAdapter was developed in two stages. First, SSL was applied to 500 000 unlabelled infantile colour fundus photographs (CFPs) retrospectively collected from four Chinese clinical centres to learn general fundus representations. Second, a much smaller labelled dataset of Normal and ROP CFPs was constructed, and a UDA module was incorporated to mitigate domain shift, defined as distribution differences between training and deployment datasets that may impair model performance. Correlation Alignment (CORAL) distance was employed to quantify these domain shifts. External validation was performed using data from an independent clinical centre and two public multiethnic datasets. RESULTS: A total of 500 000 non-overlapping unlabelled CFPs and 39 456 labelled CFPs were included. Compared with the supervised method, DeepAdapter decreased CORAL distance from 0.297 to 3.550×10-6. Both algorithms achieved high internal accuracy (0.989) in ROP prediction, but DeepAdapter significantly outperformed the supervised method on external testing (accuracy: 0.828 (95% CI 0.823 to 0.833) vs 0.739 (95% CI 0.733 to 0.745), p <0.001) and on the public cross-ethnicity dataset (accuracy: 0.839 (95% CI 0.818 to 0.859) vs 0.813 (95% CI 0.791 to 0.835), p=0.014). A web-based system integrating quality control and ROP diagnosis was developed for large-scale, multicentre clinical screening. CONCLUSIONS: DeepAdapter effectively mitigated domain shift and significantly improved model generalisability for ROP screening, providing a valuable reference for developing generalisable models in other medical specialties.
引用
- Zhao J, Li L, Lin Z, et al. DeepAdapter: a generalisable algorithm integrating self-supervised learning and unsupervised domain adaptation for robust retinopathy of prematurity screening. Br J Ophthalmol. 2026. PMID: 42285743. doi:10.1136/bjo-2025-328559
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。