Accepted for/Published in: Journal of Medical Internet Research
Date Submitted: Aug 11, 2020
Date Accepted: Feb 1, 2021
Date Submitted to PubMed: Feb 3, 2021
The Classifiers Established with Clinical Laboratory Indicators to Distinguish COVID-19 from Community-Acquired Pneumonia: Retrospective Cohort Study
ABSTRACT
Background:
The initial symptoms of the patients with COVID-19 are very much alike with those of the patients with community-acquired pneumonia (CAPN), and it is difficult to distinguish COVID-19 from CAPN by clinical symptoms and imaging examination.
Objective:
The objective of our study was to construct an effective model for early identification of COVID-19 from CAPN.
Methods:
The clinical laboratory indicators (CLIs) of 61 COVID-19 patients and 60 CAPN patients were analyzed retrospectively. Random combinations of various CLIs (CLI_combinations) were utilized to establish COVID19_vs_CAPN classifiers with machine learning algorithms including Random Forest Classifier (RFC), Logistic Regression (LR) and Gradient Boosting Classifier (GBC). The performance of the classifiers was assessed using the area under the receiver operating characteristic curve (AUC) and recall rate in COVID-19 prediction with the test data.
Results:
The classifiers constructed with three algorithms from 43 CLI_combinations showed high performance (recall rate > 0.9 and AUC > 0.85) in COVID-19 prediction for the test_set. In the high performance classifiers, the CLIs including PCT (procalcitonin), MCHC (mean corpuscular hemoglobin concentration), UA (urine acid), albumin, AGR (ratio of albumin to globulin), NEUTC (neutrophil count), RBC (red blood cell count), monocyte coun, BASOC (basophil count) and WBC (white blood count) showed a high usage rate, they also had high feature_importance except BASOC. The CLI_combination of [PCT, AGR, UA, WBC, NEUTC, BASOC, RBC, MCHC] was the representative one of nine optimal CLI_combinations capable of constructing perfect classifiers (AUC = 1.0) with RFC or GBC, and replacing any CLI in these CLI_combinations would lead to a significant degradation in the performance of classifiers built with them.
Conclusions:
The classifiers constructed with only a few specific CLIs could perfectly distinguish COVID-19 from CAPN, which will help clinicians with early isolation and centralized management of COVID-19 patients.
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.