Chest CT Images for COVID-19: Radiologists and Computer-Based Detection

Area of Science:

• Radiology
• Infectious Diseases
• Pulmonology

Background:

• Chest CT is a diagnostic criterion for COVID-19 in China.
• The ability of CT to differentiate COVID-19 from other viral pneumonias is not well-established.

Purpose of the Study:

• To determine if radiologists or computer-based systems can differentiate COVID-19 from other viral pneumonias using chest CT.
• To evaluate the diagnostic utility of CT imaging in COVID-19.

Main Methods:

• Retrospective case-control study of 52 COVID-19 and 80 non-COVID-19 pneumonia patients.
• Chest CT images evaluated by radiologists and a computer-based analysis system.
• Comparison of lesion detection and infected lung volume between groups.

Main Results:

• Key CT findings for COVID-19 included peripheral ground-glass opacities and air space infiltrates.
• No significant differences in CT features or infected lung volumes were found between COVID-19 and other viral pneumonia groups.
• Demographics, comorbidities, and clinical manifestations were similar between groups.

Conclusions:

• Initial chest CT findings do not reliably differentiate COVID-19 from other viral pneumonias.
• CT findings should not replace microbiological diagnosis for COVID-19.
• Re-evaluation of isolated patients without laboratory confirmation is suggested.