The influence of cold wall thickness and background activity on recovery coefficients in PET and SPECT
Purpose: The partial volume effect caused by limited spatial resolution consists of signal spill-out from and, in the case of background, spill-in into objects. This can be described by the recovery coefficient (RC), which can be measured in phantoms. With background activity present the RC of an object inside a phantom is influenced by its confining cold wall. This is relevant for the use of RC-values in quantitative imaging, for standard uptake values, and dosimetry. Methods: A theoretical equation for RC as a function of object-to-background concentration ratio (OBR) is formulated and combined with a correction for wall thickness. Novel analytical expressions for spheres and large objects of arbitrary size are derived for the RC of the mean (RCmean). For spheres this expression was compared to previously published experimental data exploring thick (1.5 mm) and thin (0.18 mm) walls. Results: For 10, 37 and 60 mm spheres, 1 mm wall thickness and isotropic spatially invariant Gaussian point spread function with a full width half maximum in the range 2-20 mm, the minimum-maximum values for the change in RCmean times OBR are 0.03-0.19, 0.05-0.06, 0.03-0.04, respectively. Experimental and theoretical differences in RCmean between spheres with thick and thin walls were in good agreement for 10, 15, 38 and 58 mm spheres, but not for the 20 and 30 mm spheres. Conclusion: When using phantom measurements for determining RC-values it is important to consider the cold wall thickness, otherwise significant errors can arise in the presence of background activity.