一种用于提高急性呼吸窘迫综合征患者结局预测中模型泛化能力的预训练-微调方法。
A pretrain-finetune approach for improving model generalizability in outcome prediction of acute respiratory distress syndrome patients.
机构信息
Instrument Science and Electrical Engineering, Jilin University, Changchun, China; Biomedical Diagnostics Lab, Department of Electrical Engineering, Eindhoven University of Technology, the Netherlands.
Biomedical Diagnostics Lab, Department of Electrical Engineering, Eindhoven University of Technology, the Netherlands; State Key Laboratory of Digital Medical Engineering, School of Instrument Science and Engineering, Southeast University, Nanjing, China.
出版信息
Int J Med Inform. 2024 Jun;186:105397. doi: 10.1016/j.ijmedinf.2024.105397. Epub 2024 Mar 2.
BACKGROUND
Early prediction of acute respiratory distress syndrome (ARDS) of critically ill patients in intensive care units (ICUs) has been intensively studied in the past years. Yet a prediction model trained on data from one hospital might not be well generalized to other hospitals. It is therefore essential to develop an accurate and generalizable ARDS prediction model adaptive to different hospital or medical centers.
METHODS
We analyzed electronic medical records of 200,859 and 50,920 hospitalized patients within 24 h after being diagnosed with ARDS from the Philips eICU Institute (eICU-CRD) and the Medical Information Mart for Intensive Care (MIMIC-IV) dataset, respectively. Patients were sorted into three groups, including rapid death, long stay, and recovery, based on their condition or outcome between 24 and 72 h after ARDS diagnosis. To improve prediction performance and generalizability, a "pretrain-finetune" approach was applied, where we pretrained models on the eICU-CRD dataset and performed model finetuning using only a part (35%) of the MIMIC-IV dataset, and then tested the finetuned models on the remaining data from the MIMIC-IV dataset. Well-known machine-learning algorithms, including logistic regression, random forest, extreme gradient boosting, and multilayer perceptron neural networks, were employed to predict ARDS outcomes. Prediction performance was evaluated using the area under the receiver-operating characteristic curve (AUC).
RESULTS
Results show that, in general, multilayer perceptron neural networks outperformed the other models. The use of pretrain-finetune yielded improved performance in predicting ARDS outcomes achieving a micro-AUC of 0.870 for the MIMIC-IV dataset, an improvement of 0.046 over the pretrain model.
CONCLUSIONS
The proposed pretrain-finetune approach can effectively improve model generalizability from one to another dataset in ARDS prediction.
背景
在过去的几年中，人们一直在深入研究如何在重症监护病房（ICU）中对重症患者的急性呼吸窘迫综合征（ARDS）进行早期预测。然而，在一个医院训练的预测模型可能无法很好地推广到其他医院。因此，开发一种适用于不同医院或医疗中心的准确且可推广的 ARDS 预测模型至关重要。
方法
我们分析了来自 Philips eICU 研究所（eICU-CRD）和 Medical Information Mart for Intensive Care（MIMIC-IV）数据集的 200859 名和 50920 名在被诊断为 ARDS 后 24 小时内住院的患者的电子病历。根据患者在 ARDS 诊断后 24 至 72 小时内的病情或结局，将患者分为快速死亡、住院时间长和康复三个组。为了提高预测性能和可推广性，我们采用了“预训练-微调”方法，即在 eICU-CRD 数据集上进行模型预训练，然后仅使用 MIMIC-IV 数据集的 35%部分进行模型微调，最后在 MIMIC-IV 数据集的其余数据上测试微调后的模型。我们使用了逻辑回归、随机森林、极端梯度提升和多层感知机神经网络等知名机器学习算法来预测 ARDS 结局。使用接收者操作特征曲线下的面积（AUC）来评估预测性能。
结果
结果表明，总体而言，多层感知机神经网络的表现优于其他模型。使用预训练-微调方法可以提高 ARDS 结局预测的性能，在 MIMIC-IV 数据集上的微 AUC 为 0.870，比预训练模型提高了 0.046。
结论
所提出的预训练-微调方法可以有效地提高 ARDS 预测中从一个数据集到另一个数据集的模型可推广性。
Zotero 插件