A method for attenuation correction of brain phantom using SPECT projection data

The image quality of SPECT images varies depending on the collection conditions, image processing conditions, and equipment. The attenuation correction of the Hoffman 3D brain phantom in the SPECT device is performed by the Chang method assuming a uniform absorber, but this method cannot perform accurate attenuation correction because the contour of the reconstructed image is drawn by the threshold value. This time, we devised a method (μ-map of Hoffman phantom by SPECT Projection data: μHSP) to create an attenuation map of Hoffman 3D brain phantom from SPECT projection data assuming clinical examination, and examined its validity. The % Contrast of the Hoffman 3D brain phantom image encapsulating the ^99mTc preparation and the ¹²³I preparation was 8% lower on average by performing attenuation correction with μHSP regardless of the pixel size. The % Contrast of the ¹²³I preparation was 5% lower on average than that of the ^99mTc preparation. The Hoffman 3D brain phantom image was corrected by μHSP, and the TGR (thalamus to gray matter ratio) value was about 40% higher for both formulations, and the x-axis direction profile curve for each pixel size was 4 to 5 times higher than the relative count value. It was concluded that both formulations are capable of producing attenuation correction maps with sufficiently high accuracy. (S.S.)