Predicted versus CT-derived total lung volume in a general population: The ImaLife study

Area of Science:

• Pulmonary Medicine
• Radiology
• Medical Imaging

Background:

• Global Lung Function Initiative (GLI) model predictions are standard for assessing lung volumes in pulmonary disease.
• The accuracy of GLI-predicted lung volumes against computed tomography (CT)-derived total lung volume (TLV) is not well established.

Purpose of the Study:

• To compare the Global Lung Function Initiative 2021 (GLI-2021) model's predicted total lung capacity (TLC) with CT-derived total lung volume (TLV) in a healthy adult cohort.

Main Methods:

• 151 females and 139 males (age 45-65) from the Imaging in Lifelines (ImaLife) cohort underwent low-dose inspiratory chest CT.
• Total lung volume (TLV) was measured using automated CT analysis.
• Bland-Altman analysis compared CT-derived TLV with GLI-2021 predicted TLC, with repeat analysis in never-smokers.

Main Results:

• CT-derived TLV averaged 4.7±0.9 L in women and 6.2±1.2 L in men.
• GLI-2021 predicted TLC overestimated TLV, showing a systematic bias of 1.0 L (women) and 1.6 L (men).
• High variability was observed, with limits of agreement ranging from 3.2 L (women) to 4.2 L (men), consistent in never-smokers.

Conclusions:

• Predicted TLC from the GLI-2021 model substantially overestimates CT-derived TLV in healthy adults, exhibiting low precision and accuracy.
• For clinical scenarios demanding accurate or precise lung volume quantification, direct measurement of lung volume is recommended over GLI predictions.