模型到数据方法用于OCT视网膜内液分割的深度学习
这项单中心横断面研究纳入2018年8月至2019年2月在新英格兰眼科中心接受OCT的活动性渗出性年龄相关性黄斑变性患者,采用模型到数据方法训练深度学习模型识别OCT B扫描上的视网膜内液。使用128名参与者的400张OCT B扫描训练模型,学习曲线Dice系数>80%,模型与人工分级在Dice系数和交并比评分上无统计学显著差异(P>.05)。
要点
- 模型到数据方法在眼科深度学习中是可行的,无需数据传输即可验证算法。
- 模型训练使用128名参与者的400张OCT B扫描,学习曲线Dice系数>80%。
- 模型与人工分级IRF的Dice系数和交并比评分无统计学显著差异(P>.05)。
核心结论
- 模型到数据深度学习方法应用于眼科,避免了数据共享、安全和隐私等传统障碍。
- 该概念验证研究提示此类范式应在更大规模多中心深度学习研究中进一步检验。
研究设计与样本
- 单中心横断面研究,纳入2018年8月1日至2019年2月28日在新英格兰眼科中心接受OCT的活动性渗出性年龄相关性黄斑变性患者。
- 数据主要分析时间为2019年3月1日至6月20日。
- 采用模型到数据方法训练深度学习模型识别OCT B扫描上的视网膜内液。
主要发现
- 模型使用128名参与者(69名女性[54%],59名男性[46%];平均[SD]年龄77.5[9.1]岁)的400张OCT B扫描进行训练,学习曲线Dice系数>80%。
- 模型与人工分级视网膜内液口袋相比,Dice系数或交并比评分无统计学显著差异(P>.05)。
临床怎么用
- 模型到数据方法可避免数据传输,有助于解决眼科深度学习中的数据隐私和安全问题。
- 目前结果的临床相关性有限,暂不能直接用于临床决策。
局限性
- 单中心研究,样本量有限。
- 结果的临床相关性在现阶段有限。
- 仅为概念验证研究,需更大规模多中心研究进一步验证。
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
IMPORTANCE: Amid an explosion of interest in deep learning in medicine, including within ophthalmology, concerns regarding data privacy, security, and sharing are of increasing importance. A model-to-data approach, in which the model itself is transferred rather than data, can circumvent many of these challenges but has not been previously demonstrated in ophthalmology. OBJECTIVE: To determine whether a model-to-data deep learning approach (ie, validation of the algorithm without any data transfer) can be applied in ophthalmology. DESIGN, SETTING, AND PARTICIPANTS: This single-center cross-sectional study included patients with active exudative age-related macular degeneration undergoing optical coherence tomography (OCT) at the New England Eye Center from August 1, 2018, to February 28, 2019. Data were primarily analyzed from March 1 to June 20, 2019. MAIN OUTCOMES AND MEASURES: Training of the deep learning model, using a model-to-data approach, in recognizing intraretinal fluid (IRF) on OCT B-scans. RESULTS: The model was trained (learning curve Dice coefficient, >80%) using 400 OCT B-scans from 128 participants (69 female [54%] and 59 male [46%]; mean [SD] age, 77.5 [9.1] years). In comparing the model with manual human grading of IRF pockets, no statistically significant difference in Dice coefficients or intersection over union scores was found (P > .05). CONCLUSIONS AND RELEVANCE: A model-to-data approach to deep learning applied in ophthalmology avoided many of the traditional hurdles in large-scale deep learning, including data sharing, security, and privacy concerns. Although the clinical relevance of these results is limited at this time, this proof-of-concept study suggests that such a paradigm should be further examined in larger-scale, multicenter deep learning studies.
引用
- Mehta N, Lee CS, Mendonça LSM, et al. Model-to-Data Approach for Deep Learning in Optical Coherence Tomography Intraretinal Fluid Segmentation. JAMA Ophthalmol. 2020. PMID: 32761143. doi:10.1001/jamaophthalmol.2020.2769
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。