A rapid screening classifier for diagnosing COVID-19

Abstract

Rationale: Coronavirus disease 2019 (COVID-19) has caused a global pandemic. A classifier combining chest X-ray (CXR) with clinical features may serve as a rapid screening approach. Methods: The study included 512 patients with COVID-19 and 106 with influenza A/B pneumonia. A deep neural network (DNN) was applied, and deep features derived from CXR and clinical findings formed fused features for diagnosis prediction. Results: The clinical features of COVID-19 and influenza showed different patterns. Patients with COVID-19 experienced less fever, more diarrhea, and more salient hypercoagulability. Classifiers constructed using the clinical features or CXR had an area under the receiver operating curve (AUC) of 0.909 and 0.919, respectively. The diagnostic efficacy of the classifier combining the clinical features and CXR was dramatically improved and the AUC was 0.952 with 91.5% sensitivity and 81.2% specificity. Moreover, combined classifier was functional in both severe and non-serve COVID-19, with an AUC of 0.971 with 96.9% sensitivity in non-severe cases, which was on par with the computed tomography (CT)-based classifier, but had relatively inferior efficacy in severe cases compared to CT. In extension, we performed a reader study involving three experienced pulmonary physicians, artificial intelligence (AI) system demonstrated superiority in turn-around time and diagnostic accuracy compared with experienced pulmonary physicians. Conclusions: The classifier constructed using clinical and CXR features is efficient, economical, and radiation safe for distinguishing COVID-19 from influenza A/B pneumonia, serving as an ideal rapid screening tool during the COVID-19 pandemic.

Keywords: COVID-19; chest X-ray; clinical feature; deep learning.

Figure 1

Development of the classifier for differentiating coronavirus disease 2019 (COVID-19) from influenza A/B and structure of the deep neural network (DNN). (A) A total of 525 patients with COVID-19 and 107 patients with influenza A/B were enrolled and separated into a training set of 290 cases and a test set of 328 cases after exclusion. A DNN was applied for feature extraction, selection, classification. The proposed fusion network, clinical network, chest x-ray (CXR) network and computed tomography (CT) network were established for final diagnosis. (B)The combined network system has two input streams: image data and clinical data. The two kinds of data are processed by two streams of deep neural layers, which are ultimately concatenated. When processing CXR image or clinical data only, the other one data stream is removed.

Figure 2

Diagnostic performance of the proposed classifiers in the whole cohort, non-severe subset, and severe subset. (A) Diagnostic performance in the whole cohort. According to the receiver operating characteristic (ROC) curves of our proposed method in the whole cohort (A1), combining the chest x-ray (CXR) and clinical data (green) improves the performance compared to both individually (blue and orange). A2-A5: Confusion metrics for clinical only (A2), CXR only (A3), combined (A4), and computed tomography (CT) (A5). Both the CXR and clinical data can diagnose coronavirus disease 2019 (COVID-19) and influenza. While the accuracy for diagnosing influenza using clinical features is relatively low and that for COVID-19 using CXR is lower, combining the clinical features and CXR improves both. (B) Diagnostic performance in the non-severe subset. As shown in the ROC curves for the non-severe subset (B1), clinical data (blue) perform better than chest x-ray (CXR) (orange). B2-B5: confusion metrics for clinical only (B2), CXR only (B3), combined (B4), and CT (B5) in non-severe patients. Combining the CXR and clinical data improves the diagnostic accuracy of COVID-19; although the diagnostic accuracy for influenza is slightly lower than with the clinical features only, the overall area under the curve is improved in the combined method. (C) Diagnostic performance in the severe subset. As presented in the ROC curves for the severe subset (C1), the diagnostic accuracy of CT outperformed the clinical feature or CXR. The area under the curve of the combined method is no better than for CXR only (p = 0.46). C2-C5: The confusion metrics for clinical only (C2), CXR only (C3), combined (C4), and CT (C5). AUC: area under the receiver operating curve.

Figure 3

Cluster heatmap and principal component analysis (PCA) of deep features in coronavirus disease 2019 (COVID-19) and influenza. (A) Cluster heatmap of deep features in COVID-19 and influenza. Heatmap visualized most predominant 500 deep features among 5120 features and showed clear differences between individuals with COVID-19 and influenza A/B. (B) PCA of deep features in COVID-19 and influenza. The deep features separate COVID-19 from influenza A/B along the principal component.

Figure 4

Comparison between pulmonary physicians and artificial intelligence (AI) system. The blue line is the receiver operating characteristic (ROC) curve of proposed AI system using fused clinical and chest x-ray (CXR) data, while the yellow one is the performance for CXR only. The round points are readers' results using only CXR and the star points are performances of pulmonary physicians using clinical data together with images.