Background: High-resolution small-animal SPECT imaging plays a pivotal role in preclinical research by enabling quantitative visualization of radiopharmaceutical distribution in animal models.However, challenges, including physical degrading factors, such as photon attenuation, scattering, and limited spatial resolution hinder SPECT's quantitative accuracy.Advanced reconstruction algorithms are essential to mitigate these effects and enhance image fidelity.Methods: We developed a quantitative iterative reconstruction algorithm for the developed silicon photomultipliers (SiPMs)-based HiReSPECT II preclinical SPECT system, incorporating resolution recovery, attenuation correction, and scatter compensation.Resolution recovery was modelled using a distance-dependent Gaussian collimator-detector response (CDR).Attenuation correction was implemented via binary attenuation maps with photon path-length calculation using Siddon's ray-tracing method.Scatter correction was performed on list-mode data by estimating and subtracting scattered events using the Dual-Energy Window (DEW) technique.The reconstructed images were calibrated to absolute radiotracer concentration units.Quantitative imaging performance was validated using uniformity and NEMA NU 4-2008 image quality phantoms, followed by in vivo mouse imaging studies.Results: The proposed algorithm significantly enhanced image quality and quantitative accuracy.Resolution recovery improved sharpness, while scatter and attenuation corrections minimized artifacts and ensured uniformity.Key metrics demonstrated substantial improvements in contrast, uniformity and resolution.Quantitative accuracy also significantly improved.Conclusion: Incorporating resolution recovery, scatter correction, and attenuation correction into preclinical SPECT image reconstruction is critical for achieving high quantitative accuracy.The integration of these techniques in systems, such as HiReSPECT II, establishes a new benchmark for preclinical imaging, facilitating radiotracer development and advancing translational research.
Dareyni et al. (2026) studied this question.