A clinical protocol for attenuation and scatter correction of thallium-201 and Sestamibi myocardial perfusion SPECT images
Photon attenuation and scatter in myocardial perfusion SPECT are major sources of diagnostic uncertainty because they affect both the qualitative and quantitive characteristics of patient images. Recent studies have demonstrated improved performance in the diagnostic accuracy of Sestamibi myocardial perfusion SPECT images when using attenuation corrected images . One technique for correcting for attenuation uses a short (65 cm) focal length fan beam collimator and a line Methods for correcting images for scatter using adjacent energy windows have also been proposed. Though methods for correcting for these problems exist, implementation in a clinical setting has been difficult.
We have developed a clinical protocol for performing attenuation and scatter correction. The system uses a fan-beam collimator and a line source of activity on a PRISM 3000, triple head camera system for acquiring transmission data. For acquiring emission data, parallel hole collimators are placed on the two other camera heads to determine: (1) the photo-peak image, (2) the scatter contribution to the photo-peak image, and (3) a large scatter window for estimating the body outline. The body outline is used as a spatially varying gamma prior for correcting the transmission image for the effects of object truncation. Presented herein is also a method of automatically selecting a spatially varying Gibbs prior to control the noise in the iterative reconstruction. The method of Ordered Subsets-Expectation Maximization is also used to accelerate the reconstruction of the images .
We observed in phantom studies that correcting images for attenuation and scatter will reduce the effects of these artifacts but may emphasize other effects, i.e. detector response, patient motion, etc. Improvements in uniformity of myocardial images were observed for Sestamibi phantom studies using a breast attenuator which partially obscured the heart. However, less significant improvements were observed using 201Tl where the heart was completely obscured by the breast attenuator.
This protocol demonstrated that clinical implementation of attenuation and scatter correction is possible without significant adjustment of clinical routine. However, because of the residual artifacts present in the corrected images, it is necessary to obtain simultaneously data for reconstructing images using traditional FBP protocols. in conjunction with correction protocols.