Emphysema remains an enigma. Visual classification of emphysema on CT imaging has remained unchanged for over 30 years and essentially constitutes the identification of three disease patterns: centrilobular,1 2 paraseptal3 or panlobular4 emphysema. This classification originates from anatomical descriptions dating from the 1950s specifying the location of airspace destruction within the acinus or lobule and the proximity of emphysema to the visceral pleura. The extrapolation of histopathological scale features of damage5 6 to the clinical CT scale was a key milestone in the early years of lung CT interpretation.

Yet, when considering emphysema and its role in disease pathogenesis and progression, one is led to wonder what has been lost by only considering emphysema in terms of three anatomical patterns. Centrilobular emphysema can itself comprise a range of imaging phenotypes, from frank destruction of an entire secondary pulmonary lobule to subtle reductions in peribronchiolar lung density that may be easily missed at first glance. While visual evaluation has focused on measuring emphysema extent, could we be inadequately capturing emphysema severity?

The advent of computer analysis of lung CT imaging dramatically improved the ability with which emphysema extent could be quantified using density masks7 or parametric response maps.8 Computer tools typically evaluate the entirety of the lungs in discretised small voxel volume units. While valuable information pertaining to the co-ordinates and morphology of emphysema and …