Thesis (M.Sc.) - Cairo University - Faculty of Science - Department of Biophysics

This work is aiming to quantify the effect of incorporation of PSF on image quality in different size lesions and show its effect on PVE and contrast. A cylindrical phantom equipped with different volumes hollow spheres was used. The lesion detectability and the SUV accuracy were evaluated at the same activity in different sphere volumes (i.e., different activity concentration ratios). Two reconstruction algorithms were used (TrueX and OSEM), one of them was modified with point spread function (PSF). The effect of recovery coefficient (RC) model on SUV values of different inner size diameters filled with different activity concentration ratios was evaluated. The model was applied on small cohort of patients and look uUp table (LUT) for different lesions with different sizes was determined. The image quality parameters such as contrast and noise, of the spheres used in this work, obtained for the two reconstruction algorithms. The results showed that PSF algorithm improves lesion detectability, contrast and slightly increase the noise. Contrast and coefficient of variation (CV) were generally superior compared to those obtained by the algorithms without the PSF correction. Also the results showed that this phantom study has successfully provided a practical "look up table" that are typically noted in human PET imaging. With knowledge of the size of the lesion, these RC curves will be utilized for a more accurate measurement of SUV, which could help in the quantification of the true activity concentration in lesions. We had applied these algorithms successfully in a small cohort of patients. The implications of increased accuracy in SUV determination using partial volume correction are important both in clinical and various research settings

Issued also as CD