Quantification of pulmonary opacities using artificial intelligence in chest CT scans during SARS-CoV-2 pandemic: validation and prognostic assessment

Egyptian Journal of Radiology and Nuclear Medicine

Table 1 Variables of the study

Radiological variables	AI-based analysis variables	Clinical variables	Secondary variables
Codified from the radiology report	Obtained from AI-Rad Companion	Obtained from the medical record	Obtained from the medical record
Presence of ground glass opacity in the left lung (yes, no)	Percentage of low opacity in the left lung (yes, no)	PCR test result (positive, negative)	Age (in years)
Presence of ground glass opacity in the right lung (yes, no)	Percentage of low opacity in the right lung (yes, no)	Hospital admission (yes, no)	Sex (male, female)
Presence of ground glass opacity in any lung (yes, no)	Percentage of low opacity in any lung (yes, no)	Intensive Care Unit admission (yes, no)	Date of CT scan
Presence of consolidation in the left lung (yes, no)	Percentage of high opacity in the left lung (yes, no)	Death (yes, no)
Presence of consolidation in the right lung (yes, no)	Percentage of high opacity in the right lung (yes, no)	Hospital stay (in days)
Presence of consolidation in any lung (yes, no)	Presence of high opacity in any lung (yes, no)	Lung diseases: COVID-19 pneumonia (yes or no), non-COVID-19 pneumonia (yes or no), pulmonary edema (yes or no)

The term “any lung” refers to both lungs combined. For example, if the presence of ground glass opacity in the left lung is positive, it is also positive in “any lung.” On the other hand, the percentage of ground glass opacity in “any lung” is calculated by the AI using the combined volume of both lungs.