CT image-based quantification of sub-pixel diameter microparticle accumulations in tissues using á priori biological information

With the increasing use of microspheres and nanoparticles for diagnostic and therapeutic purposes, the need to quantify the spatial distribution and concentration of those particles in a minimally invasive manner, such as by imaging, is required. In the case of CT-imaging, labelling of those particles with elements that have high contrast, and when possible that is specific for that element, is an obvious approach, but this still begs the question as to what extent particles that are smaller than the detector pixel can be quantified over relatively large volumes of tissue. This study is an exploration of three approaches to quantify the spatial distribution and/or size of those microscopic particles by use of; (i) a model of the impact of high contrast opaque particle on the detected x-ray attenuation, (ii) quasi-monochromatic energy CT and (iii) the statistics of random clustering of particles resulting in clusters that are larger than detector pixels and using that information to extrapolate to sub-resolution information about individual particles. To explore the role of particle size relative to detector pixel size we recorded x-ray attenuation in detector pixels smaller than the particle and then retrospectively increased the effective detector pixel size by summing the x-ray signal in contiguous pixels around the particle location.