A pretrain-finetune approach for improving model generalizability in outcome prediction of acute respiratory distress syndrome patients

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.

Using Weakly Supervised Deep Learning to Classify and Segment Single-Molecule Break-Junction Conductance Traces

In order to design molecular electronic devices with high performance and stability, it is crucial to understand their structure-to-property relationships. Single-molecule break junction measurements yield a large number of conductance-distance traces, which are inherently highly stochastic. Here we propose a weakly supervised deep learning algorithm to classify and segment these conductance traces, a method that is mainly based on transfer learning with the pretrain-finetune technique. By exploiting the powerful feature extraction capabilities of the pretrained VGG-16 network, our convolutional neural network model not only achieves high accuracy in the classification of the conductance traces, but also segments precisely the conductance plateau from an entire trace with very few manually labeled traces. Thus, we can produce a more reliable estimation of the junction conductance and quantify the junction stability. These findings show that our model has achieved a better accuracy-to-manpower efficiency balance, opening up the possibility of using weakly supervised deep learning approaches in the studies of single-molecule junctions.