The use of aluminium filter to improve the image quality in Tc-99m brain SPECT imaging : a phantom study

Abstract

Scattered gamma photons are the result of Compton scattering interaction of radiation within the matter. They have low energy as compared to unscattered photons and their contribution is adverse in terms of degradation of spatial resolution and image quality in single photon emission imaging. The objective to apply a physical filter is to reduce the amount of scattered gamma photons before reaching the scintillation detection system. The physical filter material aluminium, Al sheet (0.10mm and 0.20mm thickness) was chosen to perform this study. The type and thickness of physical filter was selected on the basis of percentage attenuation calculations of different gamma ray energies by various thicknesses and materials. The parameters that were investigated in this study are spectra of Tc-99m, spatial resolution, uniformity and system volume sensitivity. For the specific organ study, the image acquisitions from the Hoffman 3-D brain phantom and the RSD Striatal brain phantom were investigated. Data were acquired using Infinia® GE Healthcare dual-head gamma camera without and with physical filter with LEHR collimator installed. Spatial resolution study was done by scanning a Tc-99m line source (0.8mm inner diameter) at various source-to-collimator distances in air and in scattering medium. For spectra, uniformity and sensitivity, a cylindrical source tank filled with water added with Tc-99m was scanned. Both Hoffman 3-D and RSD Striatal brain phantoms with water with added Tc-99m were scanned. The Hoffman 3-D brain phantom was prepared with two conditions; normal brain and cold defect brain. Both brain phantoms images were analysed qualitative (visual analysis) and quantitatively (grey-to-white ratio, thalamus size and contrast ratio). All SPECT images were reconstructed with filtered-back projection (FBP) method by applying Butterworth filter of order 5 with different cut-off frequencies 0.30, 0.40 and 0.50 cycles/cm. The images were corrected by Chang’s attenuation correction method using linear attenuation coefficient, LAC 0.11cm-1 and 0.12cm-1 for without physical filter and 0.12cm-1, 0.13cm-1 and 0.14cm-1 for with physical filter. The spatial resolution with the physical filter improved as the source-to-collimator distance increases. There was a significant reduction in count rate from Compton and photopeak regions of Tc-99m spectra with physical filter compared to without physical filter. There was a substantial reduction in the counts from the photopeak region where the decrease in the scatter radiation is probably higher. System volume sensitivity was reduced. Despite reduction in system volume sensitivity, the system uniformity improved with physical filter. Enhancement in image quality of Hoffman 3-D brain phantom and the RSD Striatal brain phantom by applying 0.10 mm Al physical filter was achieved. Therefore, this technique is considered suitable for further investigation for clinical trials to validate the applicability of physical filter technique in clinical SPECT imaging.