Quantitative imaging with the micro-pet small-animal pet tomograph

Quantitative imaging of complex biological processes is a critical technology of the post-sequencing era. In particular, positron emission tomography (PET), using small-animal models, has emerged as a powerful technique to explore physiology in a flexible, noninvasive, and potentially highly, quantitative way. With the recent advent of commercial high-resolution, small-animal imagers, such as the micro-PET scanners from Siemens (formerly Concorde Microsystems), functional imaging of rodent models using PET has found increasing acceptance. However, a broad class of PET research, particularly neuroimaging, requires quantitative accuracy which, for the new small-animal systems, has generally been slow to reach the standards of state-of-the-art clinical research cameras. An essential first step in a quantitative PET study is the generation of a faithful representation of the radioactivity distribution in the subject as a function of time, which can be subsequently interpreted in terms of biological processes using methods such as tracer kinetic modeling. Since the accuracy of the input images is critical to the effectiveness of such models, the development of methods to improve image quantification is an important endeavor. These issues in the physics of imaging comprise the focus of this manuscript. Many factors impact PET image quantification including system setup and calibration, prereconstruction corrections for physical effects (e.g., deadtime, randoms, scatter, and attenuation), the type of image reconstruction algorithm, and postreconstruction methods that delineate anatomical regions and correct for spatial-resolution effects (i.e., partial volume effects). While most of these quantitative issues are applicable to all small-animal PET systems, they will be described in the specific context of the popular micro-PET R4 rodent tomograph in order to provide concrete recommendations.