Byers Eye Institute, Department of Ophthalmology, Stanford University, 2370 Watson Court, Palo Alto, CA, 94303, United States.
Center for Biomedical Informatics Research, School of Medicine, Stanford University, 1265 Welch Road, Stanford, CA, 94305, United States.
Int J Med Inform. 2021 Jun;150:104464. doi: 10.1016/j.ijmedinf.2021.104464. Epub 2021 Apr 16.
To develop and evaluate novel word embeddings (WEs) specific to ophthalmology, using text corpora from published literature and electronic health records (EHR).
We trained ophthalmology-specific WEs using 121,740 PubMed abstracts and 89,282 EHR notes using word2vec continuous bag-of-words architecture. PubMed and EHR WEs were compared to general domain GloVe WEs and general biomedical domain BioWordVec embeddings using a novel ophthalmology-domain-specific 200-question analogy test and prediction of prognosis in 5547 low vision patients using EHR notes as inputs to a deep learning model.
We found that many words representing important ophthalmic concepts in the EHR were missing from the general domain GloVe vocabulary, but covered in the ophthalmology abstract corpus. On ophthalmology analogy testing, PubMed WEs scored 95.0 %, outperforming EHR (86.0 %) and GloVe (91.0 %) but less than BioWordVec (99.5 %). On predicting low vision prognosis, PubMed and EHR WEs resulted in similar AUROC (0.830; 0.826), outperforming GloVe (0.778) and BioWordVec (0.784).
We found that using ophthalmology domain-specific WEs improved performance in ophthalmology-related clinical prediction compared to general WEs. Deep learning models using clinical notes as inputs can predict the prognosis of visually impaired patients. This work provides a framework to improve predictive models using domain-specific WEs.
使用来自已发表文献和电子健康记录 (EHR) 的文本语料库，开发和评估特定于眼科的新型词嵌入 (WE)。
我们使用 word2vec 连续词袋架构，通过 121740 篇 PubMed 摘要和 89282 篇 EHR 笔记训练眼科特定的 WE。通过一项新颖的眼科特定领域的 200 个问题类比测试和使用 EHR 笔记作为输入的深度学习模型对 5547 名低视力患者预后的预测，将 PubMed 和 EHR WE 与一般领域 GloVe WE 和一般生物医学领域 BioWordVec 嵌入进行比较。
我们发现，EHR 中代表重要眼科概念的许多词在一般领域 GloVe 词汇中缺失，但在眼科摘要语料库中有所涵盖。在眼科类比测试中，PubMed WE 得分为 95.0%，优于 EHR（86.0%）和 GloVe（91.0%），但低于 BioWordVec（99.5%）。在预测低视力预后方面，PubMed 和 EHR WE 的 AUROC 相似（0.830；0.826），优于 GloVe（0.778）和 BioWordVec（0.784）。
我们发现，与一般 WE 相比，使用眼科领域特定的 WE 可提高与眼科相关的临床预测性能。使用临床笔记作为输入的深度学习模型可以预测视力受损患者的预后。这项工作为使用领域特定的 WE 改进预测模型提供了框架。
