OEDL: an optimized ensemble deep learning method for the prediction of acute ischemic stroke prognoses using union features

Abstract

Background: Early stroke prognosis assessments are critical for decision-making regarding therapeutic intervention. We introduced the concepts of data combination, method integration, and algorithm parallelization, aiming to build an integrated deep learning model based on a combination of clinical and radiomics features and analyze its application value in prognosis prediction.

Methods: The research steps in this study include data source and feature extraction, data processing and feature fusion, model building and optimization, model training, and so on. Using data from 441 stroke patients, clinical and radiomics features were extracted, and feature selection was performed. Clinical, radiomics, and combined features were included to construct predictive models. We applied the concept of deep integration to the joint analysis of multiple deep learning methods, used a metaheuristic algorithm to improve the parameter search efficiency, and finally, developed an acute ischemic stroke (AIS) prognosis prediction method, namely, the optimized ensemble of deep learning (OEDL) method.

Results: Among the clinical features, 17 features passed the correlation check. Among the radiomics features, 19 features were selected. In the comparison of the prediction performance of each method, the OEDL method based on the concept of ensemble optimization had the best classification performance. In the comparison to the predictive performance of each feature, the inclusion of the combined features resulted in better classification performance than that of the clinical and radiomics features. In the comparison to the prediction performance of each balanced method, SMOTEENN, which is based on a hybrid sampling method, achieved the best classification performance than that of the unbalanced, oversampled, and undersampled methods. The OEDL method with combined features and mixed sampling achieved the best classification performance, with 97.89, 95.74, 94.75, 94.03, and 94.35% for Macro-AUC, ACC, Macro-R, Macro-P, and Macro-F1, respectively, and achieved advanced performance in comparison with that of methods in previous studies.

Conclusion: The OEDL approach proposed herein could effectively achieve improved stroke prognosis prediction performance, the effect of using combined data modeling was significantly better than that of single clinical or radiomics feature models, and the proposed method had a better intervention guidance value. Our approach is beneficial for optimizing the early clinical intervention process and providing the necessary clinical decision support for personalized treatment.

Keywords: MRI; deep learning; ensemble learning; ischemic stroke; metaheuristic algorithms; radiomics.

Figure 1

Whole pipeline of the proposed method. The data source and feature extraction, data processing and feature fusion, model construction, model optimization, and other processes are included. Each step is represented by a dotted box.

Figure 2

2D mask labeling process for patients.

Figure 3

Feature fusion process.

Figure 4

SMOTEENN balancing algorithm.

Figure 5

Construction process of the deep integration learning method.

Figure 6

Schematic diagram of each base learner. (A) DNN, (B) LSTM-RNN, and (C) DBN.

Figure 7

BBOA schematic diagram. To solve the parameter optimization problem faced by deep networks, we used the Big Bang optimization algorithm (BBOA). (A) The concept of a cosmic explosion; (B) the BBOA pipeline (the PUOA algorithm procedure).

Figure 8

Correlation test of characteristics. (A) Correlation test of clinical characteristics and (B) correlation test of the iconographic features.

Figure 9

Rose plot of feature weights. The 19 extracted features are represented by A to P, and the feature weights are shown.

Figure 10

Visual interpretation of the importance of selected features. (A) Feature density scatterplot: each column represents a sample, and each row represents a feature; the features are sorted by their average absolute SHAP values; red represents the positive direction, and blue represents the negative direction. (B) Feature distribution heatmap: each point represents a sample, the samples are sorted by their SHAP values, and the absolute SHAP value of a feature represents its contribution to the model. (C) Feature decision diagram: this figure represents the accumulation of all samples and features as well as model's decision-making process.

Figure 11

Scatter plot display of the classification results of OEDL. (A) Clinical, (B) Radiomics, and (C) Joint.