Chest CT texture-based radiomics analysis in differentiating COVID-19 from other interstitial pneumonia

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Distinguishing Coronavirus Disease-19 (COVID-19) pneumonia from other interstitial pneumonias on chest CT is clinically important.
• Radiomics, a method of extracting quantitative features from medical images, offers potential for objective image analysis.

Purpose of the Study:

• To evaluate the utility of texture-based radiomics analysis in differentiating COVID-19 pneumonia from pneumonia of other etiologies using chest CT.
• To assess the diagnostic performance of CT texture analysis (CTTA) in identifying COVID-19 pneumonia.

Main Methods:

• Retrospective analysis of 120 patients with suspected COVID-19 and CT findings of interstitial pneumonia.
• Sixty patients were COVID-19 positive and 60 were negative by RT-PCR.
• CT texture analysis (CTTA) was performed on filtered and unfiltered images to extract textural features.

Main Results:

• Unfiltered CTTA showed significant differences in mean intensity, mean of positive pixels (MPP), and kurtosis between COVID-19 positive and negative patients.
• Filtered CTTA, particularly fine texture scales, revealed lower standard deviation (SD) and MPP in COVID-19 positive patients.
• A combination of texture features achieved 60% sensitivity and 80% specificity in differentiating COVID-19 positive from negative patients.

Conclusions:

• Preliminary findings suggest that CT texture analysis (CTTA) holds potential for distinguishing COVID-19 pneumonia from other interstitial pneumonias on chest CT.
• Radiomics-based CTTA may serve as a valuable tool in the diagnostic workup of COVID-19 pneumonia.