Characterization of emphysema in low dose computed tomography images using maximum likelihood estimation of the scale exponent

Low attenuation clusters (LACs) are associated with the presence of emphysema in low dose computer tomography (LDCT) scans. It is believed that the LACs´ frequency-size distribution can be appropriately fit with power law and be well described with the scale exponent. In this work, we investigated whether the LACs´ frequency-size distribution of a group of 80 LDCT scans taken from a lung cancer-screening program followed a power-law behavior. Contrary to the current thinking, only a subset of the data (60%) actually allowed a power-law fit. This suggests that the claims of power-law behavior in LAC´s frequency-size distribution is a delicate issue and should be regarded with care.