Implementation of Cascade Gamma and Positron Range Corrections for I-124 Small Animal PET

• Published in: IEEE transactions on nuclear science 2014-02, Vol.61 (1), p.142-153
• Main Authors: Harzmann, S., Braun, F., Zakhnini, A., Weber, W. A., Pietrzyk, U., Mix, M.
• Format: Article
• Language: English
• Subjects: Algorithms; Animals; Cascade gamma coincidences; Cascades; Computer simulation; Image reconstruction; Image resolution; Logic gates; Monte Carlo simulations; non-pure positron emitter; Phantoms; Positron emission; Positron emission tomography; positron emission tomography (PET); positron range; Positrons; Spatial resolution; Tomography
• ISSN: 0018-9499; 1558-1578
• DOI: 10.1109/TNS.2013.2293914

Small animal Positron Emission Tomography (PET) should provide accurate quantification of regional radiotracer concentrations and high spatial resolution. This is challenging for non-pure positron emitters with high positron endpoint energies, such as I-124: On the one hand the cascade gammas emitted from this isotope can produce coincidence events with the 511 keV annihilation photons leading to quantification errors. On the other hand the long range of the high energy positron degrades spatial resolution. This paper presents the implementation of a comprehensive correction technique for both of these effects. The established corrections include a modified sinogram-based tail-fitting approach to correct for scatter, random and cascade gamma coincidences and a compensation for resolution degradation effects during the image reconstruction. Resolution losses were compensated for by an iterative algorithm which incorporates a convolution kernel derived from line source measurements for the microPET Focus 120 system. The entire processing chain for these corrections was implemented, whereas previous work has only addressed parts of this process. Monte Carlo simulations with GATE and measurements of mice with the microPET Focus 120 show that the proposed method reduces absolute quantification errors on average to 2.6% compared to 15.6% for the ordinary Maximum Likelihood Expectation Maximization algorithm. Furthermore resolution was improved in the order of 11-29% depending on the number of convolution iterations. In summary, a comprehensive, fast and robust algorithm for the correction of small animal PET studies with I-124 was developed which improves quantitative accuracy and spatial resolution.