Cold and hot region detectability and image quality enhancement in TC99m spect by reduction of scattered gamma photons prior to detectio

• Main Authors: Sayed, Inayatullah Shah, Harun, Siti Zawani
• Format: Conference or Workshop Item
• Language: English
• Published: 2011
• Subjects: QC Physics
• Online Access: http://irep.iium.edu.my/8457/1/Cold_and_hot_region_%28Article_Journal%29.pdf; http://irep.iium.edu.my/8457/; http://seafomp.org/?page_id=265
• Institution: Universiti Islam Antarabangsa Malaysia

Introduction: Single Photon Emission Computed Tomography (SPECT) inherits a serious problem of scattered gamma photons emanating from the object to be scanned. It is well recognised that, presence of scattered gamma photons in SPECT projection data, degrade the image quality and limit the accuracy in quantification of radioactivity distribution. Therefore, it is essential to remove the scattered gamma photons from SPECT projection data for better image quality and quantitative analysis. In this context, some techniques are available, which are window subtraction, mathematical (Monte Carlo simulation) and convolution technique. In this work a new technique has been introduced, that is a material filter technique, which absorbs some fraction of scattered gamma photons before they can reach to the surface of gamma camera detector. Methods: A novel cylindrical PET/SPECT phantom provides similar cross sectional views of cold and hot regions insert was scanned. Which was filled with water and 20.8 mCi of Tc99m radioactivity was distributed in the phantom. Philip ADAC Forte dual head gamma camera was used with LEHR collimator. Symmetric energy window (20%) centred at 140 keV was adjusted. Matrix size 128x128 was selected for data acquisition. Filter (absorber sheet) was mounted (each separately) on the outer surface of the collimator - copper (thickness 0.127 mm) and aluminum filter (thickness 0.2 mm). Image data were acquired without and with filter (absorber sheet). Ninety (90) views were taken over 360o with anti clockwise orientation. Filtered back projection image reconstruction technique was used by selecting Butterworth digital filter with cut-off frequency 0.35 and 0.45 cycles/cm, and order 5. Chang’s attenuation correction technique with linear attenuation coefficient (LAC) values 0.13, 0.14 and 0.15 cm-1 was also applied. Results: Images were visually inspected in order to observe the cold and hot region detectability. Image quality was measured by calculating the contrast, SNR, and standard deviation in the count density of uniform, cold and hot region images. Results show that improvement in uniformity of radioactivity distribution is achieved by applying the aluminum filter and using Butterworth filter cut-off frequency 0.35cycles/cm, order 5 with LAC 0.13 cm-1. Overall, cold and hot regions image quality and detectability with copper filter in conjunction with Butterworth filter cut-off frequency 0.35cycles/cm, and order 5 with LAC 0.13 cm-1 was improved. Discussion: Images reconstructed from the data collected with copper and aluminum filters showed improvement in the uniformity of the radioactivity distribution in the uniform region compared to images without filter - corrected with Chang’s attenuation correction technique with LAC value 0.13 cm-1 for both 0.35 and 0.45 cycles/cm cut-off frequencies of Butterworth filter. Hot region images of three pairs (22.4 mm, 17.9 mm, and 14.3 mm diameter) appeared clearer with both types of filters as compared to without filter. Furthermore, images obtained from without filtered data showed almost all sizes of hot regions were attached to each other either with the same pair or to the next pair. Relatively, improvement in the gap between the hot region images and sharper edges with copper and aluminum filters was observed, that reflects the removal of some fraction of scattered gamma photons by absorber filters. Visual analysis of cold region images reconstructed from copper filter data showed better detection of regions where lower part of ‘V’ shape of cold region images relative to without filter and aluminum filter images was clear. Furthermore, images produced using Butterworth filter with cut-off frequency 0.35cycles/cm and order 5, the cold regions appeared more circular as compared to images by selecting cut-off frequency 0.45cycles/cm and order 5. However, in this case contrast, SNR and standard deviation values were equivocal (without and with filters). Conclusions: It is concluded that, cold and hot region detectability, uniformity in the image of a uniform section of phantom, SNR, standard deviation and cold and hot region image contrast is marginally improved with physical filters.