Forecasting adverse surgical events using self-supervised transfer learning for physiological signals.
Forecasting adverse surgical events using self-supervised transfer learning for physiological signals.
复制标题
DOI：
10.1038/s41746-021-00536-y
复制
发表时间：
2021-12-08
影响因子：
15.2
通讯作者：
Lee SI
中科院分区：
文献类型：
--
作者：
Chen H;Lundberg SM;Erion G;Kim JH;Lee SI
全世界每年进行数以亿计的外科手术，产生一种普遍类型的电子健康记录（EHR）数据，包括时间序列生理信号。在这里，我们提出了一种可转移的嵌入方法（即将时间序列信号转换为预测机器学习模型的输入特征的方法），称为PHASE（生理信号嵌入），使我们能够更准确地预测基于生理信号的不良手术结果。我们根据来自两个手术室（OR）数据集和重症监护病房（ICU）数据集的5万多例手术的每分钟电子病历数据对PHASE进行评估。PHASE在预测六种不同的结果（低氧血症、低碳酸血症、低血压、高血压、苯肾上腺素和肾上腺素）方面优于其他最先进的方法，如在原始数据上训练的长短期记忆网络和在手工特征上训练的梯度增强树。在迁移学习设置中，我们在一个数据集中训练嵌入模型，然后在未见数据中嵌入信号并预测不良事件，PHASE与传统方法相比，以更低的计算成本实现了更高的预测精度。最后，考虑到理解模型在临床应用中的重要性，我们证明PHASE是可解释的，并使用局部特征归因方法验证了我们的预测模型。
Hundreds of millions of surgical procedures take place annually across the world, which generate a prevalent type of electronic health record (EHR) data comprising time series physiological signals. Here, we present a transferable embedding method (i.e., a method to transform time series signals into input features for predictive machine learning models) named PHASE (PHysiologicAl Signal Embeddings) that enables us to more accurately forecast adverse surgical outcomes based on physiological signals. We evaluate PHASE on minute-by-minute EHR data of more than 50,000 surgeries from two operating room (OR) datasets and patient stays in an intensive care unit (ICU) dataset. PHASE outperforms other state-of-the-art approaches, such as long-short term memory networks trained on raw data and gradient boosted trees trained on handcrafted features, in predicting six distinct outcomes: hypoxemia, hypocapnia, hypotension, hypertension, phenylephrine, and epinephrine. In a transfer learning setting where we train embedding models in one dataset then embed signals and predict adverse events in unseen data, PHASE achieves significantly higher prediction accuracy at lower computational cost compared to conventional approaches. Finally, given the importance of understanding models in clinical applications we demonstrate that PHASE is explainable and validate our predictive models using local feature attribution methods.
登录
查看更多内容
影响因子：
4.4
作者：
Kohli MD;Summers RM;Geis JR
通讯作者：
Geis JR
影响因子：
2.6
作者：
Kable, AK;Gibberd, RW;Spigelman, AD
通讯作者：
Spigelman, AD
影响因子：
9.8
作者：
Johnson AE;Pollard TJ;Shen L;Lehman LW;Feng M;Ghassemi M;Moody B;Szolovits P;Celi LA;Mark RG
通讯作者：
Mark RG
影响因子：
8.8
作者：
Basali, A;Mascha, EJ;Schubert, A
通讯作者：
Schubert, A
影响因子：
5.9
作者：
Gupta, Priyanka;Malhotra, Pankaj;Shroff, Gautam
通讯作者：
Shroff, Gautam