Two-Step Machine Learning to Diagnose and Predict Involvement of Lungs in COVID-19 and Pneumonia using CT Radiomics

Abstract Objective We aimed to develop a two-step machine learning (ML) based model to diagnose and predict involvement of lungs in COVID-19 and non COVID-19 pneumonia patients using CT chest radiomic features. Methods Three hundred CT scans (3-classes: 100 COVID-19, 100 pneumonia, and 100 healthy subjects) were enrolled in this study. Diagnostic task included 3-class classification. For severity prediction, two radiologists scored involvement of lungs in COVID-19 and pneumonia scans based on percentage of involvement in all 5 lobes. Datasets were classified into mild (0-25%), moderate (26-50%), and severe (>50%). Whole lungs were segmented utilizing deep learning-based segmentation method. Altogether, 107 features including shape, first-order histogram, second and high order texture features were extracted. For both tasks, datasets were randomly divided into 90% training sets (70% and 30% for training and validation, respectively) and 10% test sets. Pearson correlation coefficient (PCC≥90%) was performed to exclude highly correlated features. Subsequently, different feature selection algorithms (Correlation attribute evaluation, Information gain attribute, Wrapper Subset selection algorithm, Relief method, and Correlation-based feature selection) were assessed. The most pertinent features were finally selected using voting method based on the evaluation of all algorithms. Several ML-based supervised algorithms were utilized, namely Naïve Bays, Support Vector Machine, Bagging, Random Forest, K-nearest neighbors, Decision Tree and Ensemble Meta voting. The synthetic minority oversampling technique (SMOTE) was used to balance the three classes in training sets. The optimal model was first selected based on precision, recall and area-under-curve (AUC) by randomizing the training/validation sets 20 times, followed by testing using the test set. To ensure the repeatability of the results, the entire process was repeated 50 times. Results Nine pertinent features (2 shape, 1 first-order, and 6 second-order features) were obtained after feature selection for both phases. In diagnostic task, the performance of 3-class classification using Random Forest was 0.909±0.026, 0.907±0.056, 0.902±0.044, 0.939±0.031, and 0.982±0.010 for precision, recall, F1-score, accuracy, and AUC, respectively. The severity prediction task using Random Forest achieved 0.868±0.123 precision, 0.865±0.121 recall, 0.853±0.139 F1-score, 0.934±0.024 accuracy, and 0.969±0.022 AUC. Conclusion The two-phase ML-based model accurately classified COVID-19 and pneumonia patients using CT radiomics, and adequately predicted severity of lungs involvement. This 2-steps model showed great potential in assessing COVID-19 CT images towards improved management of patients.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.