Prognostic value of cellular population data in patients with COVID-19

Background and aims Beckman Coulter hematology analysers identify leukocytes by their volume (V), conductivity (C) and scatter (S) of a laser beam at different angles. Each leukocyte sub-population [neutrophils (NE), lymphocytes (LY), monocytes (MO)] is characterized by the mean (MN) and the standard deviation (SD) of 7 measurements called "cellular population data” (@CPD), corresponding to morphological analysis of the leukocytes. As severe forms of infections to SARS-CoV-2 are characterized by a functional activation of mononuclear cells, leading to a cytokine storm, we evaluated whether CPD variations are correlated to the inflammation state, oxygen requirement and lung damage and whether CPD analysis could be useful for a triage of patients with COVID-19 in the Emergency Department (ED) and could help to identify patients with a high risk of worsening. Materials and method The CPD of 825 consecutive patients with proven COVID-19 presenting to the ED were recorded and compared to classical biochemical parameters, the need for hospitalization in the ward or ICU, the need for oxygen, or lung injury on CT-scan. Results 40 of the 42 CPD were significantly modified in COVID-19 patients in comparison to 245 controls. @MN-V-MO and @SD-V-MO were highly correlated with C-reactive protein, procalcitonin, ferritin and D-dimers. SD-UMALS-LY > 21.45 and > 23.92 identified, respectively, patients with critical lung injuries (>75%) and requiring tracheal intubation. @SD-V-MO > 25.03 and @SD-V-NE > 19.4 identified patients required immediate ICU admission, whereas a @MN-V-MO < 183 suggested that the patient could be immediately discharged. Using logistic regression, the combination of 8 CPD with platelet and basophil counts and the existence of diabetes or obesity could identify patients requiring ICU after a first stay in conventional wards (area under the curve = 0.843). Conclusion CPD analysis constitutes an easy and inexpensive tool for triage and prognosis of COVID-19 patients in the ED.

Prognostic value of cellular population data in patients with COVID-19.

Background and aims: Beckman Coulter hematology analysers identify leukocytes by their volume (V), conductivity (C) and scatter (S) of a laser beam at different angles. Each leukocyte sub-population [neutrophils (NE), lymphocytes (LY), monocytes (MO)] is characterized by the mean (MN) and the standard deviation (SD) of 7 measurements called "cellular population data" (@CPD), corresponding to morphological analysis of the leukocytes. As severe forms of infections to SARS-CoV-2 are characterized by a functional activation of mononuclear cells, leading to a cytokine storm, we evaluated whether CPD variations are correlated to the inflammation state, oxygen requirement and lung damage and whether CPD analysis could be useful for a triage of patients with COVID-19 in the Emergency Department (ED) and could help to identify patients with a high risk of worsening. Materials and method: The CPD of 825 consecutive patients with proven COVID-19 presenting to the ED were recorded and compared to classical biochemical parameters, the need for hospitalization in the ward or ICU, the need for oxygen, or lung injury on CT-scan. Results: 40 of the 42 CPD were significantly modified in COVID-19 patients in comparison to 245 controls. @MN-V-MO and @SD-V-MO were highly correlated with C-reactive protein, procalcitonin, ferritin and D-dimers. SD-UMALS-LY > 21.45 and > 23.92 identified, respectively, patients with critical lung injuries (>75%) and requiring tracheal intubation. @SD-V-MO > 25.03 and @SD-V-NE > 19.4 identified patients required immediate ICU admission, whereas a @MN-V-MO < 183 suggested that the patient could be immediately discharged. Using logistic regression, the combination of 8 CPD with platelet and basophil counts and the existence of diabetes or obesity could identify patients requiring ICU after a first stay in conventional wards (area under the curve = 0.843). Conclusion: CPD analysis constitutes an easy and inexpensive tool for triage and prognosis of COVID-19 patients in the ED.

Stratification of COVID-19 Severity Using SeptiCyte RAPID, a Novel Host Immune Response Test.

SeptiCyte® RAPID is a gene expression assay measuring the relative expression levels of host response genes PLA2G7 and PLAC8, indicative of a dysregulated immune response during sepsis. As severe forms of COVID-19 may be considered viral sepsis, we evaluated SeptiCyte RAPID in a series of 94 patients admitted to Foch Hospital (Suresnes, France) with proven SARS-CoV-2 infection. EDTA blood was collected in the emergency department (ED) in 67 cases, in the intensive care unit (ICU) in 23 cases and in conventional units in 4 cases. SeptiScore (0-15 scale) increased with COVID-19 severity. Patients in ICU had the highest SeptiScores, producing values comparable to 8 patients with culture-confirmed bacterial sepsis. Receiver operating characteristic (ROC) curve analysis had an area under the curve (AUC) of 0.81 for discriminating patients requiring ICU admission from patients who were immediately discharged or from patients requiring hospitalization in conventional units. SeptiScores increased with the extent of the lung injury. For 68 patients, a chest computed tomography (CT) scan was performed within 24 h of COVID-19 diagnosis. SeptiScore >7 suggested lung injury ≥50% (AUC = 0.86). SeptiCyte RAPID was compared to other biomarkers for discriminating Critical + Severe COVID-19 in ICU, versus Moderate + Mild COVID-19 not in ICU. The mean AUC for SeptiCyte RAPID was superior to that of any individual biomarker or combination thereof. In contrast to C-reactive protein (CRP), correlation of SeptiScore with lung injury was not impacted by treatment with anti-inflammatory agents. SeptiCyte RAPID can be a useful tool to identify patients with severe forms of COVID-19 in ED, as well as during follow-up.

Automated Quantification of CT Patterns Associated with COVID-19 from Chest CT.

PURPOSE: To present a method that automatically segments and quantifies abnormal CT patterns commonly present in coronavirus disease 2019 (COVID-19), namely ground glass opacities and consolidations. MATERIALS AND METHODS: In this retrospective study, the proposed method takes as input a non-contrasted chest CT and segments the lesions, lungs, and lobes in three dimensions, based on a dataset of 9749 chest CT volumes. The method outputs two combined measures of the severity of lung and lobe involvement, quantifying both the extent of COVID-19 abnormalities and presence of high opacities, based on deep learning and deep reinforcement learning. The first measure of (PO, PHO) is global, while the second of (LSS, LHOS) is lobe-wise. Evaluation of the algorithm is reported on CTs of 200 participants (100 COVID-19 confirmed patients and 100 healthy controls) from institutions from Canada, Europe and the United States collected between 2002-Present (April 2020). Ground truth is established by manual annotations of lesions, lungs, and lobes. Correlation and regression analyses were performed to compare the prediction to the ground truth. RESULTS: Pearson correlation coefficient between method prediction and ground truth for COVID-19 cases was calculated as 0.92 for PO (P < .001), 0.97 for PHO (P < .001), 0.91 for LSS (P < .001), 0.90 for LHOS (P < .001). 98 of 100 healthy controls had a predicted PO of less than 1%, 2 had between 1-2%. Automated processing time to compute the severity scores was 10 seconds per case compared to 30 minutes required for manual annotations. CONCLUSION: A new method segments regions of CT abnormalities associated with COVID-19 and computes (PO, PHO), as well as (LSS, LHOS) severity scores.

Automated AI-Driven CT Quantification of Lung Disease Predicts Adverse Outcomes in Patients Hospitalized for COVID-19 Pneumonia.

The purpose of our work was to assess the independent and incremental value of AI-derived quantitative determination of lung lesions extent on initial CT scan for the prediction of clinical deterioration or death in patients hospitalized with COVID-19 pneumonia. 323 consecutive patients (mean age 65 ± 15 years, 192 men), with laboratory-confirmed COVID-19 and an abnormal chest CT scan, were admitted to the hospital between March and December 2020. The extent of consolidation and all lung opacities were quantified on an initial CT scan using a 3D automatic AI-based software. The outcome was known for all these patients. 85 (26.3%) patients died or experienced clinical deterioration, defined as intensive care unit admission. In multivariate regression based on clinical, biological and CT parameters, the extent of all opacities, and extent of consolidation were independent predictors of adverse outcomes, as were diabetes, heart disease, C-reactive protein, and neutrophils/lymphocytes ratio. The association of CT-derived measures with clinical and biological parameters significantly improved the risk prediction (p = 0.049). Automated quantification of lung disease at CT in COVID-19 pneumonia is useful to predict clinical deterioration or in-hospital death. Its combination with clinical and biological data improves risk prediction.

Machine learning automatically detects COVID-19 using chest CTs in a large multicenter cohort.

OBJECTIVES: To investigate machine learning classifiers and interpretable models using chest CT for detection of COVID-19 and differentiation from other pneumonias, interstitial lung disease (ILD) and normal CTs. METHODS: Our retrospective multi-institutional study obtained 2446 chest CTs from 16 institutions (including 1161 COVID-19 patients). Training/validation/testing cohorts included 1011/50/100 COVID-19, 388/16/33 ILD, 189/16/33 other pneumonias, and 559/17/34 normal (no pathologies) CTs. A metric-based approach for the classification of COVID-19 used interpretable features, relying on logistic regression and random forests. A deep learning-based classifier differentiated COVID-19 via 3D features extracted directly from CT attenuation and probability distribution of airspace opacities. RESULTS: Most discriminative features of COVID-19 are the percentage of airspace opacity and peripheral and basal predominant opacities, concordant with the typical characterization of COVID-19 in the literature. Unsupervised hierarchical clustering compares feature distribution across COVID-19 and control cohorts. The metrics-based classifier achieved AUC = 0.83, sensitivity = 0.74, and specificity = 0.79 versus respectively 0.93, 0.90, and 0.83 for the DL-based classifier. Most of ambiguity comes from non-COVID-19 pneumonia with manifestations that overlap with COVID-19, as well as mild COVID-19 cases. Non-COVID-19 classification performance is 91% for ILD, 64% for other pneumonias, and 94% for no pathologies, which demonstrates the robustness of our method against different compositions of control groups. CONCLUSIONS: Our new method accurately discriminates COVID-19 from other types of pneumonia, ILD, and CTs with no pathologies, using quantitative imaging features derived from chest CT, while balancing interpretability of results and classification performance and, therefore, may be useful to facilitate diagnosis of COVID-19. KEY POINTS: • Unsupervised clustering reveals the key tomographic features including percent airspace opacity and peripheral and basal opacities most typical of COVID-19 relative to control groups. • COVID-19-positive CTs were compared with COVID-19-negative chest CTs (including a balanced distribution of non-COVID-19 pneumonia, ILD, and no pathologies). Classification accuracies for COVID-19, pneumonia, ILD, and CT scans with no pathologies are respectively 90%, 64%, 91%, and 94%. • Our deep learning (DL)-based classification method demonstrates an AUC of 0.93 (sensitivity 90%, specificity 83%). Machine learning methods applied to quantitative chest CT metrics can therefore improve diagnostic accuracy in suspected COVID-19, particularly in resource-constrained environments.

Quantification of Tomographic Patterns associated with COVID-19 from Chest CT.

PURPOSE: To present a method that automatically segments and quantifies abnormal CT patterns commonly present in coronavirus disease 2019 (COVID-19), namely ground glass opacities and consolidations. MATERIALS AND METHODS: In this retrospective study, the proposed method takes as input a non-contrasted chest CT and segments the lesions, lungs, and lobes in three dimensions, based on a dataset of 9749 chest CT volumes. The method outputs two combined measures of the severity of lung and lobe involvement, quantifying both the extent of COVID-19 abnormalities and presence of high opacities, based on deep learning and deep reinforcement learning. The first measure of (PO, PHO) is global, while the second of (LSS, LHOS) is lobewise. Evaluation of the algorithm is reported on CTs of 200 participants (100 COVID-19 confirmed patients and 100 healthy controls) from institutions from Canada, Europe and the United States collected between 2002-Present (April, 2020). Ground truth is established by manual annotations of lesions, lungs, and lobes. Correlation and regression analyses were performed to compare the prediction to the ground truth. RESULTS: Pearson correlation coefficient between method prediction and ground truth for COVID-19 cases was calculated as 0.92 for PO (P < .001), 0.97 for PHO(P < .001), 0.91 for LSS (P < .001), 0.90 for LHOS (P < .001). 98 of 100 healthy controls had a predicted PO of less than 1%, 2 had between 1-2%. Automated processing time to compute the severity scores was 10 seconds per case compared to 30 minutes required for manual annotations. CONCLUSION: A new method segments regions of CT abnormalities associated with COVID-19 and computes (PO, PHO), as well as (LSS, LHOS) severity scores.

COVID-19 pneumonia: high diagnostic accuracy of chest CT in patients with intermediate clinical probability.

OBJECTIVES: To assess inter-reader agreements and diagnostic accuracy of chest CT to identify COVID-19 pneumonia in patients with intermediate clinical probability during an acute disease outbreak. METHODS: From March 20 to April 8, 319 patients (mean age 62.3 years old) consecutive patients with an intermediate clinical probability of COVID-19 pneumonia underwent a chest CT scan. Two independent chest radiologists blinded to clinical information and RT-PCR results retrospectively reviewed and classified images on a 1-5 confidence level scale for COVID-19 pneumonia. Agreements between radiologists were assessed with kappa statistics. Diagnostic accuracy of chest CT compared with RT-PCR assay and patient outcomes was measured using receiver operating characteristics (ROC). Positive predictive value (PPV) and negative predictive value (NPV) for COVID-19 pneumonia were calculated. RESULTS: Inter-observer agreement for highly probable (kappa: 0.83 [p < .001]) and highly probable or probable (kappa: 0.82 [p < .001]) diagnosis of COVID-19 pneumonia was very good. RT-PCR tests performed in 307 patients were positive in 174 and negative in 133. The areas under the curve (AUC) were 0.94 and 0.92 respectively. With a disease prevalence of 61.2%, PPV were 95.9% and 94.3%, and NPV 84.4% and 77.1%. CONCLUSION: During acute COVID-19 outbreak, chest CT scan may be used for triage of patients with intermediate clinical probability with very good inter-observer agreements and diagnostic accuracy. KEY POINTS: • Concordances between two chest radiologists to diagnose or exclude a COVID-19 pneumonia in 319 consecutive patients with intermediate clinical probability were very good (kappa: 0.82; p < .001). • When compared with RT-PCR results and patient outcomes, the diagnostic accuracy of CT to identify COVID-19 pneumonia was high for both radiologists (AUC: 0.94 and 0.92). • With a disease prevalence of 61.2% in the studied population, the positive predictive values of CT for diagnosing COVID-19 pneumonia were 95.9% and 94.3% with negative predictive values of 84.4% and 77.1%.