Volume 32, Issue 2 pp. 435-443
RESEARCH ARTICLE

Can laboratory parameters be an alternative to CT and RT-PCR in the diagnosis of COVID-19? A machine learning approach

Mehmet Kalaycı

Mehmet Kalaycı

Department of Medical Biochemistry, Elazig Fethi Sekin City Hospital, Elazığ, Turkey

Search for more papers by this author
Hakan Ayyıldız

Hakan Ayyıldız

Department of Medical Biochemistry, Elazig Fethi Sekin City Hospital, Elazığ, Turkey

Search for more papers by this author
Seda Arslan Tuncer

Corresponding Author

Seda Arslan Tuncer

Department of Software Engineering, Firat University, Elazığ, Turkey

Correspondence

Seda Arslan Tuncer, Department of Software Engineering, Firat University, Elazığ, Turkey.

Email: [email protected]

Search for more papers by this author
Pinar Gundogan Bozdag

Pinar Gundogan Bozdag

Department of Radiology, Elazig Fethi Sekin City Hospital, Elazığ, Turkey

Search for more papers by this author
Gulden Eser Karlidag

Gulden Eser Karlidag

Department of Clinic of Infectious Diseases and Clinical Microbiology, University of Health Sciences, Elazig Fethi Sekin City Hospital, Elazığ, Turkey

Search for more papers by this author
First published: 22 January 2022
Citations: 1

Abstract

In this study, a machine learning-based decision support system that uses routine laboratory parameters has been proposed in order to increase the diagnostic success in COVID-19. The main goal of the proposed method was to reduce the number of misdiagnoses in the RT-PCR and CT scans and to reduce the cost of testing. In this study, we retrospectively reviewed the files of patients who presented to the coronavirus outpatient. The demographic, thoracic CT, and laboratory data of the individuals without any symptoms of the disease, who had negative RT-PCR test and who had positive RT-PCR test were analyzed. CT images were classified using hybrid CNN methods to show the superiority of the decision support system using laboratory parameters. Detection of COVID-19 from CT images achieved an accuracy of 97.56% with the AlexNet-SVM hybrid method, while COVID-19 was classified with an accuracy of 97.86% with the proposed method using laboratory parameters.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT

The dataset used for analysis during the current study are available from the corresponding author on reasonable request.

The full text of this article hosted at iucr.org is unavailable due to technical difficulties.