Development and validation of chest CT-based imaging biomarkers for early stage COVID-19 screening

Area of Science:

• Radiology
• Medical Imaging
• Infectious Diseases

Background:

• Coronavirus Disease 2019 (COVID-19) is a global pandemic requiring effective early screening for control and treatment.
• Chest computed tomography (CT) is a crucial tool for diagnosing respiratory illnesses.

Purpose of the Study:

• To develop and validate novel chest CT-based imaging biomarkers for the accurate screening of COVID-19 patients.
• To assess the performance of these biomarkers in differentiating COVID-19 from other conditions like community-acquired pneumonia (CAP) and healthy controls.

Main Methods:

• Development and validation of imaging biomarkers using chest CT scans from 419 patients across two independent hospitals.
• Identification and analysis of vasculature-like signals in CT images.
• Application of unsupervised feature learning to discover clinical-relevant imaging biomarkers.
• Double-blind validation of the discovered biomarkers in an independent hospital cohort.

Main Results:

• COVID-19 patients exhibited a significantly higher abundance of vasculature-like signals compared to healthy individuals and CAP patients.
• The developed imaging biomarkers demonstrated high diagnostic performance: sensitivity 0.941, specificity 0.920, AUC 0.971, accuracy 0.931, and F1 score 0.929.
• The biomarkers were successfully validated in an independent hospital setting.

Conclusions:

• Chest CT-based imaging biomarkers derived from vasculature-like signals show significant potential for accurate and sensitive COVID-19 screening.
• These findings offer a promising new approach to aid in the early detection and management of COVID-19 patients.
• The validated biomarkers could serve as a valuable tool in clinical settings for rapid COVID-19 patient identification.