The nonlinear partial volume effect and computed tomography densitometry of foam and lung

A quantitative study was performed to assess the magnitude of the nonlinear partial volume effect (NLPVE) in computed tomography (CT) densitometry of polyethene foam and lung. This effect arises in materials having density variations on the scale of the sampling area of an individual CT-detector element. It causes a systematic underestimation of the density determined with CT. Foam samples and a resected lung of a goat were imaged with high resolution (20 lp/mm) using a mammography system, and the observed optical density variation in the images was converted into a distribution of pathlengths that x rays penetrate within the solid component of the cellular material. The obtained pathlength distribution was used to calculate the transmission, as seen by a single detector in computed tomography. Comparison with the transmission through homogeneous material of the same thickness gave an estimate of the NLPVE. For the foams studied, the CT-determined density was found to be too low by approximately 0.3%-0.5% due to this effect. Although these density errors are small, in calibrations of a CT scanner they may be of significance. For lung the underestimation of the density was less than 0.1%. These experimentally derived, NLPVE related CT-density errors are 32%-84% of those calculated from a simple model of a cellular solid.