ABSTRACT
Multiple clinical phenotypes have been proposed for COVID-19, but few have stemmed from data-driven methods. We aimed to identify distinct phenotypes in patients admitted with COVID-19 using cluster analysis, and compare their respective characteristics and clinical outcomes. We analyzed the data from 547 patients hospitalized with COVID-19 in a Canadian academic hospital from January 1, 2020, to January 30, 2021. We compared four clustering algorithms: K-means, PAM (partition around medoids), divisive and agglomerative hierarchical clustering. We used imaging data and 34 clinical variables collected within the first 24 hours of admission to train our algorithm. We then conducted survival analysis to compare clinical outcomes across phenotypes and trained a classification and regression tree (CART) to facilitate phenotype interpretation and phenotype assignment. We identified three clinical phenotypes, with 61 patients (17%) in Cluster 1, 221 patients (40%) in Cluster 2 and 235 (43%) in Cluster 3. Cluster 2 and Cluster 3 were both characterized by a low-risk respiratory and inflammatory profile, but differed in terms of demographics. Compared with Cluster 3, Cluster 2 comprised older patients with more comorbidities. Cluster 1 represented the group with the most severe clinical presentation, as inferred by the highest rate of hypoxemia and the highest radiological burden. Mortality, mechanical ventilation and ICU admission risk were all significantly different across phenotypes. We conducted a phenotypic analysis of adult inpatients with COVID-19 and identified three distinct phenotypes associated with different clinical outcomes. Further research is needed to determine how to properly incorporate those phenotypes in the management of patients with COVID-19.
ABSTRACT
BackgroundDecision scores and ethically mindful algorithms are being established to adjudicate mechanical ventilation in the context of potential resources shortage due to the current onslaught of COVID-19 cases. There is a need for a reproducible and objective method to provide quantitative information for those scores. PurposeTowards this goal, we present a retrospective study testing the ability of a deep learning algorithm at extracting features from chest x-rays (CXR) to track and predict radiological evolution. Materials and MethodsWe trained a repurposed deep learning algorithm on the CheXnet open dataset (224,316 chest X-ray images of 65,240 unique patients) to extract features that mapped to radiological labels. We collected CXRs of COVID-19-positive patients from two open-source datasets (last accessed on April 9, 2020)(Italian Society for Medical and Interventional Radiology and MILA). Data collected form 60 pairs of sequential CXRs from 40 COVID patients (mean age {+/-} standard deviation: 56 {+/-} 13 years; 23 men, 10 women, seven not reported) and were categorized in three categories: "Worse", "Stable", or "Improved" on the basis of radiological evolution ascertained from images and reports. Receiver operating characteristic analyses, Mann-Whitney tests were performed. ResultsOn patients from the CheXnet dataset, the area under ROC curves ranged from 0.71 to 0.93 for seven imaging features and one diagnosis. Deep learning features between "Worse" and "Improved" outcome categories were significantly different for three radiological signs and one diagnostic ("Consolidation", "Lung Lesion", "Pleural effusion" and "Pneumonia"; all P < 0.05). Features from the first CXR of each pair could correctly predict the outcome category between "Worse" and "Improved" cases with 82.7% accuracy. ConclusionCXR deep learning features show promise for classifying the disease trajectory. Once validated in studies incorporating clinical data and with larger sample sizes, this information may be considered to inform triage decisions.
