Visual assessment versus Quantitative densitometry by computed tomography for detection of mild emphysema

Emphysema may be used for subclassification of chronic obstructive pulmonary disease (COPD) and is also a risk factor for lung cancer. Emphysema can be directly visualized by low dose computed tomography (CT). As visual assessment of emphysema is time-consuming, our aim was to detect emphysema by quantitative densitometry (% Low Attenuation Value below -950 Hounsfield Units, LAV950) using the software CTPulmo3D (Syngo.via, Siemens Healthineers). Participants from the pilot study of The Swedish CArdioPulmonary Bioimage Study (SCAPIS) with visually detected emphysema (n=100, 56 males and 44 females) and 100 matched controls (gender, age, height and weight) without emphysema were included in this study. In more than 80% of the participants the emphysema was visually graded as mild (1% to ≤25% of the lung parenchyma). Inter- and intraobserver variability for visual assessment of emphysema was reliable with Krippendorff alpha values ranging from 0.80 to 0.85. Patients with emphysema were more obstructive (FEV1/FVC 0.72 versus 0.78) and had a higher tobacco burden (27 versus 20 pack-years) compared to the matched controls. The mean difference in LAV950 between matched pairs of emphysema and controls was -0.1% (CI -1.3 to 1.2%), limits of agreement from -12.3 to 12.1%. When defining emphysema as LAV950 being ≥ 10%, the sensitivity of finding emphysema was 27%. In conclusion, visual assessment of emphysema is reliable. Quantitative measurements of LAV950 by the software used in this study were not able to distinguish patients with or without visually detected mild emphysema. Other quantitative methods need to be evaluated for use in large-scale studies.