Applications of Iterative Reconstruction Methods in SPECT

We have applied iterative reconstruction methods to compensate for the major image degrading effects of photon attenuation and geometric response in SPECT imaging. The compensation methods implement, in the projection and backprojection operations of the iterative reconstruction algorithms, an accurate model of photon attenuation in the patient’s body and the geometric response of the collimated-detector system. To evaluate the corrective iterative reconstruction methods, data from a computer-generated phantom which simulated T1-201 distribution in the thoracic region was used. Also, the techniques were assessed using data from a cardiac SPECT study with T1-201. Our studies indicate that compensations for attenuation and detector response in SPECT are possible using iterative reconstruction techniques. The compensations are especially important when the attenuation coefficient distribution in the body region, such as the thorax, is non-uniform. The attenuation compensation scheme using the iterative maximum likelihood-EM (ML-EM) algorithm can provide reconstructed images with low noise amplification, and accurate quantitative information without distortions and artifacts. Also, compensation for detector response gives additional improvement in spatial resolution.