[Effect of the skull on brain perfusion SPECT imaging]

Because attenuation gradually decreases reconstructed counts from the periphery to the center portion of the object, the deep region far from the detector is not clearly observed if attenuation is not compensated. In brain perfusion SPECT, diagnosis is sometimes made using filtered back projection images without attenuation compensation (FBP). Brain perfusion SPECT has the unique characteristic that the radiopharmaceutical accumulates only within the brain and is never taken up by the surrounding skull. This study investigated the effect of skull on brain perfusion SPECT reconstructed with FBP. We theoretically derived the relation between the counts of brain and the linear attenuation coefficient of skull. It was found that the difference in reconstructed counts between the deep gray matter and peripheral gray matter decreased due to the existence of the skull. This result indicated that the deep gray matter was inclined to be visible if the FBP images were displayed according to relative counts normalized to the maximum count of each image. In order to confirm this, we made a numerical phantom with realistic human brain and skull contours on the basis of MR images from a normal volunteer. The linear attenuation coefficient of brain was assumed to be 0.15 cm-1, while that of skull was assumed to be 0.26 cm-1 (denoted as BONE+) or 0 cm-1 (BONE-). In accordance with the theoretical results, the deep gray matter of BONE+ images was more clearly observed than that of BONE- images, if these were displayed using the relative counts of each images. The physical phantom experiments also supported the theoretical and numerical phantom studies.