Severity and Consolidation Quantification of COVID-19 From CT Images Using Deep Learning Based on Hybrid Weak Labels

Ensemble Deep Learning Derived from Transfer Learning for Classification of COVID-19 Patients on Hybrid Deep-Learning-Based Lung Segmentation: A Data Augmentation and Balancing Framework

BACKGROUND AND MOTIVATION

Lung computed tomography (CT) techniques are high-resolution and are well adopted in the intensive care unit (ICU) for COVID-19 disease control classification. Most artificial intelligence (AI) systems do not undergo generalization and are typically overfitted. Such trained AI systems are not practical for clinical settings and therefore do not give accurate results when executed on unseen data sets. We hypothesize that ensemble deep learning (EDL) is superior to deep transfer learning (TL) in both non-augmented and augmented frameworks.

METHODOLOGY

The system consists of a cascade of quality control, ResNet-UNet-based hybrid deep learning for lung segmentation, and seven models using TL-based classification followed by five types of EDL's. To prove our hypothesis, five different kinds of data combinations (DC) were designed using a combination of two multicenter cohorts-Croatia (80 COVID) and Italy (72 COVID and 30 controls)-leading to 12,000 CT slices. As part of generalization, the system was tested on unseen data and statistically tested for reliability/stability.

RESULTS

Using the K5 (80:20) cross-validation protocol on the balanced and augmented dataset, the five DC datasets improved TL mean accuracy by 3.32%, 6.56%, 12.96%, 47.1%, and 2.78%, respectively. The five EDL systems showed improvements in accuracy of 2.12%, 5.78%, 6.72%, 32.05%, and 2.40%, thus validating our hypothesis. All statistical tests proved positive for reliability and stability.

CONCLUSION

EDL showed superior performance to TL systems for both (a) unbalanced and unaugmented and (b) balanced and augmented datasets for both (i) seen and (ii) unseen paradigms, validating both our hypotheses.

Computational Intelligence-Based Disease Severity Identification: A Review of Multidisciplinary Domains

Disease severity identification using computational intelligence-based approaches is gaining popularity nowadays. Artificial intelligence and deep-learning-assisted approaches are proving to be significant in the rapid and accurate diagnosis of several diseases. In addition to disease identification, these approaches have the potential to identify the severity of a disease. The problem of disease severity identification can be considered multi-class classification, where the class labels are the severity levels of the disease. Plenty of computational intelligence-based solutions have been presented by researchers for severity identification. This paper presents a comprehensive review of recent approaches for identifying disease severity levels using computational intelligence-based approaches. We followed the PRISMA guidelines and compiled several works related to the severity identification of multidisciplinary diseases of the last decade from well-known publishers, such as MDPI, Springer, IEEE, Elsevier, etc. This article is devoted toward the severity identification of two main diseases, viz. Parkinson's Disease and Diabetic Retinopathy. However, severity identification of a few other diseases, such as COVID-19, autonomic nervous system dysfunction, tuberculosis, sepsis, sleep apnea, psychosis, traumatic brain injury, breast cancer, knee osteoarthritis, and Alzheimer's disease, was also briefly covered. Each work has been carefully examined against its methodology, dataset used, and the type of disease on several performance metrics, accuracy, specificity, etc. In addition to this, we also presented a few public repositories that can be utilized to conduct research on disease severity identification. We hope that this review not only acts as a compendium but also provides insights to the researchers working on disease severity identification using computational intelligence-based approaches.

Review of Semantic Segmentation of Medical Images Using Modified Architectures of UNET

In biomedical image analysis, information about the location and appearance of tumors and lesions is indispensable to aid doctors in treating and identifying the severity of diseases. Therefore, it is essential to segment the tumors and lesions. MRI, CT, PET, ultrasound, and X-ray are the different imaging systems to obtain this information. The well-known semantic segmentation technique is used in medical image analysis to identify and label regions of images. The semantic segmentation aims to divide the images into regions with comparable characteristics, including intensity, homogeneity, and texture. UNET is the deep learning network that segments the critical features. However, UNETs basic architecture cannot accurately segment complex MRI images. This review introduces the modified and improved models of UNET suitable for increasing segmentation accuracy.

Application of Machine Learning and Deep Learning Techniques for COVID-19 Screening Using Radiological Imaging: A Comprehensive Review

Lung, being one of the most important organs in human body, is often affected by various SARS diseases, among which COVID-19 has been found to be the most fatal disease in recent times. In fact, SARS-COVID 19 led to pandemic that spreads fast among the community causing respiratory problems. Under such situation, radiological imaging-based screening [mostly chest X-ray and computer tomography (CT) modalities] has been performed for rapid screening of the disease as it is a non-invasive approach. Due to scarcity of physician/chest specialist/expert doctors, technology-enabled disease screening techniques have been developed by several researchers with the help of artificial intelligence and machine learning (AI/ML). It can be remarkably observed that the researchers have introduced several AI/ML/DL (deep learning) algorithms for computer-assisted detection of COVID-19 using chest X-ray and CT images. In this paper, a comprehensive review has been conducted to summarize the works related to applications of AI/ML/DL for diagnostic prediction of COVID-19, mainly using X-ray and CT images. Following the PRISMA guidelines, total 265 articles have been selected out of 1715 published articles till the third quarter of 2021. Furthermore, this review summarizes and compares varieties of ML/DL techniques, various datasets, and their results using X-ray and CT imaging. A detailed discussion has been made on the novelty of the published works, along with advantages and limitations.

Recommender System for the Efficient Treatment of COVID-19 Using a Convolutional Neural Network Model and Image Similarity

Background: Hospitals face a significant problem meeting patients' medical needs during epidemics, especially when the number of patients increases rapidly, as seen during the recent COVID-19 pandemic. This study designs a treatment recommender system (RS) for the efficient management of human capital and resources such as doctors, medicines, and resources in hospitals. We hypothesize that a deep learning framework, when combined with search paradigms in an image framework, can make the RS very efficient. Methodology: This study uses a Convolutional neural network (CNN) model for the feature extraction of the images and discovers the most similar patients. The input queries patients from the hospital database with similar chest X-ray images. It uses a similarity metric for the similarity computation of the images. Results: This methodology recommends the doctors, medicines, and resources associated with similar patients to a COVID-19 patients being admitted to the hospital. The performance of the proposed RS is verified with five different feature extraction CNN models and four similarity measures. The proposed RS with a ResNet-50 CNN feature extraction model and Maxwell-Boltzmann similarity is found to be a proper framework for treatment recommendation with a mean average precision of more than 0.90 for threshold similarities in the range of 0.7 to 0.9 and an average highest cosine similarity of more than 0.95. Conclusions: Overall, an RS with a CNN model and image similarity is proven as an efficient tool for the proper management of resources during the peak period of pandemics and can be adopted in clinical settings.

Application of Deep Learning Techniques in Diagnosis of Covid-19 (Coronavirus): A Systematic Review

Covid-19 is now one of the most incredibly intense and severe illnesses of the twentieth century. Covid-19 has already endangered the lives of millions of people worldwide due to its acute pulmonary effects. Image-based diagnostic techniques like X-ray, CT, and ultrasound are commonly employed to get a quick and reliable clinical condition. Covid-19 identification out of such clinical scans is exceedingly time-consuming, labor-intensive, and susceptible to silly intervention. As a result, radiography imaging approaches using Deep Learning (DL) are consistently employed to achieve great results. Various artificial intelligence-based systems have been developed for the early prediction of coronavirus using radiography pictures. Specific DL methods such as CNN and RNN noticeably extract extremely critical characteristics, primarily in diagnostic imaging. Recent coronavirus studies have used these techniques to utilize radiography image scans significantly. The disease, as well as the present pandemic, was studied using public and private data. A total of 64 pre-trained and custom DL models concerning imaging modality as taxonomies are selected from the studied articles. The constraints relevant to DL-based techniques are the sample selection, network architecture, training with minimal annotated database, and security issues. This includes evaluating causal agents, pathophysiology, immunological reactions, and epidemiological illness. DL-based Covid-19 detection systems are the key focus of this review article. Covid-19 work is intended to be accelerated as a result of this study.

Magnetic Resonance Imaging under Image Enhancement Algorithm to Analyze the Clinical Value of Placement of Drainage Tube on Incision Healing after Hepatobiliary Surgery

This study was aimed at exploring the application value of magnetic resonance imaging (MRI) based on image enhancement algorithm in analyzing the placement of drainage tubes in the healing of incisions after hepatobiliary surgery. A total of 70 patients with liver cancer undergoing laparoscopic hepatobiliary surgery were selected, including 34 males and 36 females. According to the detection method of postoperative recovery, they were divided into a group A (conventional MRI detection) and a group B (MRI detection based on Retinex algorithm). Patients were divided into two groups according to whether subcutaneous drainage tubes were placed: group C (no subcutaneous drainage tubes were placed) and group D (subcutaneous drainage tubes were placed), with 35 patients in each group. The results showed that there was no significant difference between group A and group B in tumor residual or recurrence. The detection rate of tumor capsule in group B was significantly higher than that in group A (P < 0.05). The sensitivity, specificity, and accuracy of group A for the detection of recurrent lesions were 63.40%, 86.90%, and 78.60%, respectively; those in group B were 82.70%, 98.50%, and 93.20%, respectively. Therefore, the difference between the two groups was statistically significant (P < 0.05). The incidence of poor wound healing and infection in group C were significantly lower than those in group D (P < 0.05). Therefore, the effect of MRI detection based on image enhancement algorithm was more conducive to the evaluation of postoperative recovery due to the traditional MRI detection. In addition, the drainage tube was helpful to the postoperative wound healing and showed high clinical value.

Diagnostic performance of CT lung severity score and quantitative chest CT for stratification of COVID-19 patients

Purpose

Lung severity score (LSS) and quantitative chest CT (QCCT) analysis could have a relevant impact to stratify patients affected by COVID-19 pneumonia at the hospital admission. The study aims to assess LSS and QCCT performances in severity stratification of COVID-19 patients.

Materials and methods

From April 19, 2020, until May 3, 2020, patients with chest CT suggestive for interstitial pneumonia and tested positive for COVID-19 were retrospectively enrolled and stratified for hospital admission as Group 1, 2 and 3 (home isolation, low intensive care and intensive care, respectively). For LSS, lungs were divided in 20 regions and visually assessed by two radiologists who scored for each region from non-lung involvement as 0, < 50% assigned as 1 and > 50% as 2. QCCT was performed with a dedicated software that extracts pulmonary involvement expressed in liters and percentage. LSS and QCCT were analyzed with ROC curve analysis to predict the performance of both methods. P values < 0.05 were considered statistically significant.

Results

Final population enrolled included 136 patients (87 males, mean age 66 ± 16), 19 patients in Group 1, 86 in Group 2 and 31 in Group 3. Significant differences for LSS were observed in almost all comparisons, especially in Group 1 vs 3 (AUC 0.850, P < 0,0001) and Group 1 + 2 vs 3 (AUC 0.783, P < 0,0001). QCCT showed significant results in almost all comparisons, especially between Group 1 vs 3 (AUC 0.869, P < 0,0001). LSS and QCCT comparison between Group 1 and Group 2 did not show significant differences.

Conclusions

LSS and QCCT could represent promising tools to stratify COVID-19 patient severity at the admission.

Detection of COVID-19 in smartphone-based breathing recordings: A pre-screening deep learning tool

This study was sought to investigate the feasibility of using smartphone-based breathing sounds within a deep learning framework to discriminate between COVID-19, including asymptomatic, and healthy subjects. A total of 480 breathing sounds (240 shallow and 240 deep) were obtained from a publicly available database named Coswara. These sounds were recorded by 120 COVID-19 and 120 healthy subjects via a smartphone microphone through a website application. A deep learning framework was proposed herein that relies on hand-crafted features extracted from the original recordings and from the mel-frequency cepstral coefficients (MFCC) as well as deep-activated features learned by a combination of convolutional neural network and bi-directional long short-term memory units (CNN-BiLSTM). The statistical analysis of patient profiles has shown a significant difference (p-value: 0.041) for ischemic heart disease between COVID-19 and healthy subjects. The Analysis of the normal distribution of the combined MFCC values showed that COVID-19 subjects tended to have a distribution that is skewed more towards the right side of the zero mean (shallow: 0.59±1.74, deep: 0.65±4.35, p-value: <0.001). In addition, the proposed deep learning approach had an overall discrimination accuracy of 94.58% and 92.08% using shallow and deep recordings, respectively. Furthermore, it detected COVID-19 subjects successfully with a maximum sensitivity of 94.21%, specificity of 94.96%, and area under the receiver operating characteristic (AUROC) curves of 0.90. Among the 120 COVID-19 participants, asymptomatic subjects (18 subjects) were successfully detected with 100.00% accuracy using shallow recordings and 88.89% using deep recordings. This study paves the way towards utilizing smartphone-based breathing sounds for the purpose of COVID-19 detection. The observations found in this study were promising to suggest deep learning and smartphone-based breathing sounds as an effective pre-screening tool for COVID-19 alongside the current reverse-transcription polymerase chain reaction (RT-PCR) assay. It can be considered as an early, rapid, easily distributed, time-efficient, and almost no-cost diagnosis technique complying with social distancing restrictions during COVID-19 pandemic.

Value of Ultrasonic Image Features in Diagnosis of Perinatal Outcomes of Severe Preeclampsia on account of Deep Learning Algorithm

This study is aimed at discussing the value of ultrasonic image features in diagnosis of perinatal outcomes of severe preeclampsia on account of deep learning algorithm. 140 pregnant women singleton with severe preeclampsia were selected as the observation group. At the same time, 140 normal singleton pregnant women were selected as the control group. The hemodynamic indexes were detected by color Doppler ultrasound. The CNN algorithm was used to classify ultrasound images of two groups of pregnant women. The differential scanning calorimetry (DSC), mean pixel accuracy (MPA), and mean intersection of union (MIOU) values of CNN algorithm were 0.9410, 0.9228, and 0.8968, respectively. Accuracy, precision, recall, and F1-score were 93.44%, 95.13%, 95.09%, and 94.87%, respectively. The differences were statistically significant (P < 0.05). Compared with the normal control group, the umbilical artery (UA), uterine artery-systolic/diastolic (UTA-S/D), uterine artery (UTA), and digital video (DV) of pregnant women in the observation group were remarkably increased; the minimum alveolar effective concentration (MCA) of the observation group was obviously lower than the MCA of the control group, and the differences between groups were statistically valid (P < 0.05). Logistic regression analysis showed that UA-S/D, UA-resistance index (UA-RI), UTA-S/D, UTA-pulsatility index (UTA-PI), DV-peak velocity index for veins (DV-PVIV), and MCA-S/D were independent risk factors for the outcome of perinatal children with severe preeclampsia. In the perinatal management of severe epilepsy, the combination of the above blood flow indexes to select the appropriate delivery time had positive significance to improve the pregnancy outcome and reduce the perinatal mortality.

Lung detection and severity prediction of pneumonia patients based on COVID-19 DET-PRE network

BACKGROUND

The sudden outbreak of COVID-19 pneumonia has brought a heavy disaster to individuals globally. Facing this new virus, the clinicians have no automatic tools to assess the severity of pneumonia patients.

METHODS

In the current work, a COVID-19 DET-PRE network with two pipelines was proposed. Firstly, the lungs in X-rays were detected and segmented through the improved YOLOv3 Dense network to remove redundant features. Then, the VGG16 classifier was pre-trained on the source domain, and the severity of the disease was predicted on the target domain by means of transfer learning.

RESULTS

The experiment results demonstrated that the COVID-19 DET-PRE network can effectively detect the lungs from X-rays and accurately predict the severity of the disease. The mean average precisions (mAPs) of lung detection in patients with mild and severe illness were 0.976 and 0.983 respectively. Moreover, the accuracy of severity prediction of COVID-19 pneumonia can reach 86.1%.

CONCLUSIONS

The proposed neural network has high accuracy, which is suitable for the clinical diagnosis of COVID-19 pneumonia.

SD-UNet: A Novel Segmentation Framework for CT Images of Lung Infections

Due to the outbreak of lung infections caused by the coronavirus disease (COVID-19), humans have to face an unprecedented and devastating global health crisis. Since chest computed tomography (CT) images of COVID-19 patients contain abundant pathological features closely related to this disease, rapid detection and diagnosis based on CT images is of great significance for the treatment of patients and blocking the spread of the disease. In particular, the segmentation of the COVID-19 CT lung-infected area can quantify and evaluate the severity of the disease. However, due to the blurred boundaries and low contrast between the infected and the non-infected areas in COVID-19 CT images, the manual segmentation of the COVID-19 lesion is laborious and places high demands on the operator. Quick and accurate segmentation of COVID-19 lesions from CT images based on deep learning has drawn increasing attention. To effectively improve the segmentation effect of COVID-19 lung infection, a modified UNet network that combines the squeeze-and-attention (SA) and dense atrous spatial pyramid pooling (Dense ASPP) modules) (SD-UNet) is proposed, fusing global context and multi-scale information. Specifically, the SA module is introduced to strengthen the attention of pixel grouping and fully exploit the global context information, allowing the network to better mine the differences and connections between pixels. The Dense ASPP module is utilized to capture multi-scale information of COVID-19 lesions. Moreover, to eliminate the interference of background noise outside the lungs and highlight the texture features of the lung lesion area, we extract in advance the lung area from the CT images in the pre-processing stage. Finally, we evaluate our method using the binary-class and multi-class COVID-19 lung infection segmentation datasets. The experimental results show that the metrics of Sensitivity, Dice Similarity Coefficient, Accuracy, Specificity, and Jaccard Similarity are 0.8988 (0.6169), 0.8696 (0.5936), 0.9906 (0.9821), 0.9932 (0.9907), and 0.7702 (0.4788), respectively, for the binary-class (multi-class) segmentation task in the proposed SD-UNet. The result of the COVID-19 lung infection area segmented by SD-UNet is closer to the ground truth compared to several existing models such as CE-Net, DeepLab v3+, UNet++, and other models, which further proves that a more accurate segmentation effect can be achieved by our method. It has the potential to assist doctors in making more accurate and rapid diagnosis and quantitative assessment of COVID-19.

The Differences and Changes of Semi-Quantitative and Quantitative CT Features of Coronavirus Disease 2019 Pneumonia in Patients With or Without Smoking History

Objective: To assess CT features of COVID-19 patients with different smoking status using quantitative and semi-quantitative technologies and to investigate changes of CT features in different disease states between the two groups.

Methods: 30 COVID-19 patients with current smoking status (29 men, 1 woman) admitted in our database were enrolled as smoking group and 56 COVID-19 patients without smoking history (24 men, 32 women) admitted during the same period were enrolled as a control group. Twenty-seven smoking cases and 55 control cases reached recovery standard and were discharged. Initial and follow-up CT during hospitalization and follow-up CT after discharge were acquired. Thirty quantitative features, including the ratio of infection volume and visual-assessed interstitial changes score including total score, score of ground glass opacity, consolidation, septal thickening, reticulation and honeycombing sign, were analyzed.

Results: Initial CT images of the smoking group showed higher scores of septal thickening [4.5 (0–5) vs. 0 (0–4), p = 0.001] and reticulation [0 (0–5.25) vs 0 (0–0), p = 0.001] as well as higher total score [7 (5–12.25) vs. 6 (5–7), p = 0.008] with statistical significance than in the control group. The score of reticulation was higher in the smoking group than in the control group when discharged [0.89 (0–0) vs. 0.09 (0–0), p = 0.02]. The score of septal thickening tended to be higher in the smoking group than the control group [4 (0–4) vs. 0 (0–4), p = 0.007] after being discharged. Quantitative CT features including infection ratio of whole lung and left lung as well as infection ratio within HU (−750, −300) and within HU (−300, 49) were higher in the control group of initial CT with statistical differences. The infection ratio of whole lung and left lung, infection ratio within HU (−750), and within HU (−750, −300) were higher in the control group with statistical differences when discharged. This trend turned adverse after discharge and the values of quantitative features were generally higher in the smoking group than in the control group without statistical differences.

Conclusions: Patients with a history of smoking presented more severe interstitial manifestations and more residual lesion after being discharged. More support should be given for COVID-19 patients with a smoking history during hospitalization and after discharge.

COVID-19 Screening in Chest X-Ray Images Using Lung Region Priors

Early screening of COVID-19 is essential for pandemic control, and thus to relieve stress on the health care system. Lung segmentation from chest X-ray (CXR) is a promising method for early diagnoses of pulmonary diseases. Recently, deep learning has achieved great success in supervised lung segmentation. However, how to effectively utilize the lung region in screening COVID-19 still remains a challenge due to domain shift and lack of manual pixel-level annotations. We hereby propose a multi-appearance COVID-19 screening framework by using lung region priors derived from CXR images. Firstly, we propose a multi-scale adversarial domain adaptation network (MS-AdaNet) to boost the cross-domain lung segmentation task as the prior knowledge to the classification network. Then, we construct a multi-appearance network (MA-Net), which is composed of three sub-networks to realize multi-appearance feature extraction and fusion using lung region priors. At last, we can obtain prediction results from normal, viral pneumonia, and COVID-19 using the proposed MA-Net. We extend the proposed MS-AdaNet for lung segmentation task on three different public CXR datasets. The results suggest that the MS-AdaNet outperforms contrastive methods in cross-domain lung segmentation. Moreover, experiments reveal that the proposed MA-Net achieves accuracy of 98.83% and F1-score of 98.71% on COVID-19 screening. The results indicate that the proposed MA-Net can obtain significant performance on COVID-19 screening.

A prediction of hematoma expansion in hemorrhagic patients using a novel dual-modal machine learning strategy

Objective.Hematoma expansion is closely associated with adverse functional outcomes in patients with intracerebral hemorrhage (ICH). Prediction of hematoma expansion would therefore be of great clinical significance. We therefore attempted to predict hematoma expansion using a dual-modal machine learning (ML) strategy which combines information from non-contrast computed tomography (NCCT) images and multiple clinical variables.Approach.We retrospectively identified 140 ICH patients (57 with hematoma expansion) with 5616 NCCT images of hematoma (2635 with hematoma expansion) and 10 clinical variables. The dual-modal ML strategy consists of two steps. The first step is to derive a mono-modal predictor from a deep convolutional neural network using solely NCCT images. The second step is to achieve a dual-modal predictor by combining the mono-modal predictor with 10 clinical variables to predict hematoma growth using a multi-layer perception network.Main results. For the mono-modal predictor, the best performance was merely 69.5% in accuracy with solely the NCCT images, whereas the dual-modal predictor could boost the accuracy greatly to be 86.5% by combining clinical variables.Significance.To our knowledge, this is the best performance from using ML to predict hematoma expansion. It could be potentially useful as a screening tool for high-risk patients with ICH, though further clinical tests would be necessary to show its performance on a larger cohort of patients.

CT Quantification of COVID-19 Pneumonia at Admission Can Predict Progression to Critical Illness: A Retrospective Multicenter Cohort Study

Objective: Early identification of coronavirus disease 2019 (COVID-19) patients with worse outcomes may benefit clinical management of patients. We aimed to quantify pneumonia findings on CT at admission to predict progression to critical illness in COVID-19 patients.

Methods: This retrospective study included laboratory-confirmed adult patients with COVID-19. All patients underwent a thin-section chest computed tomography (CT) scans showing evidence of pneumonia. CT images with severe moving artifacts were excluded from analysis. Patients' clinical and laboratory data were collected from medical records. Three quantitative CT features of pneumonia lesions were automatically calculated using a care.ai Intelligent Multi-disciplinary Imaging Diagnosis Platform Intelligent Evaluation System of Chest CT for COVID-19, denoting the percentage of pneumonia volume (PPV), ground-glass opacity volume (PGV), and consolidation volume (PCV). According to Chinese COVID-19 guidelines (trial version 7), patients were divided into noncritical and critical groups. Critical illness was defined as a composite of admission to the intensive care unit, respiratory failure requiring mechanical ventilation, shock, or death. The performance of PPV, PGV, and PCV in discrimination of critical illness was assessed. The correlations between PPV and laboratory variables were assessed by Pearson correlation analysis.

Results: A total of 140 patients were included, with mean age of 58.6 years, and 85 (60.7%) were male. Thirty-two (22.9%) patients were critical. Using a cutoff value of 22.6%, the PPV had the highest performance in predicting critical illness, with an area under the curve of 0.868, sensitivity of 81.3%, and specificity of 80.6%. The PPV had moderately positive correlation with neutrophil (%) (r = 0.535, p < 0.001), erythrocyte sedimentation rate (r = 0.567, p < 0.001), d-Dimer (r = 0.444, p < 0.001), high-sensitivity C-reactive protein (r = 0.495, p < 0.001), aspartate aminotransferase (r = 0.410, p < 0.001), lactate dehydrogenase (r = 0.644, p < 0.001), and urea nitrogen (r = 0.439, p < 0.001), whereas the PPV had moderately negative correlation with lymphocyte (%) (r = −0.535, p < 0.001).

Conclusions: Pneumonia volume quantified on initial CT can non-invasively predict the progression to critical illness in advance, which serve as a prognostic marker of COVID-19.

A multi-center study of COVID-19 patient prognosis using deep learning-based CT image analysis and electronic health records

PURPOSE

As of August 30th, there were in total 25.1 million confirmed cases and 845 thousand deaths caused by coronavirus disease of 2019 (COVID-19) worldwide. With overwhelming demands on medical resources, patient stratification based on their risks is essential. In this multi-center study, we built prognosis models to predict severity outcomes, combining patients' electronic health records (EHR), which included vital signs and laboratory data, with deep learning- and CT-based severity prediction.

METHOD

We first developed a CT segmentation network using datasets from multiple institutions worldwide. Two biomarkers were extracted from the CT images: total opacity ratio (TOR) and consolidation ratio (CR). After obtaining TOR and CR, further prognosis analysis was conducted on datasets from INSTITUTE-1, INSTITUTE-2 and INSTITUTE-3. For each data cohort, generalized linear model (GLM) was applied for prognosis prediction.

RESULTS

For the deep learning model, the correlation coefficient of the network prediction and manual segmentation was 0.755, 0.919, and 0.824 for the three cohorts, respectively. The AUC (95 % CI) of the final prognosis models was 0.85(0.77,0.92), 0.93(0.87,0.98), and 0.86(0.75,0.94) for INSTITUTE-1, INSTITUTE-2 and INSTITUTE-3 cohorts, respectively. Either TOR or CR exist in all three final prognosis models. Age, white blood cell (WBC), and platelet (PLT) were chosen predictors in two cohorts. Oxygen saturation (SpO2) was a chosen predictor in one cohort.

CONCLUSION

The developed deep learning method can segment lung infection regions. Prognosis results indicated that age, SpO2, CT biomarkers, PLT, and WBC were the most important prognostic predictors of COVID-19 in our prognosis model.