Improving wrist SPECT with fanbeam collimators and an MLEM algorithm

The human wrist, composed of 8 small carpal bones is one of the most complicated joints in the human body. A planar X-ray is often the first test ordered to determine the cause of the injured wrist but it is not always conclusive. The metabolic information provided by single photon emission computed tomography (SPECT) has potential to clarify the situation of radiographically occult wrist imaging. However, a drawback of SPECT has been its poor spatial resolution, which has made it difficult to distinguish pathology within and between the small carpal bones of the wrist. We have evaluated the possible improvements in spatial resolution which can be obtained in wrist SPECT imaging through the use of fanbeam collimation and resolution recovery through the incorporation of the system response function into the MLEM (MLEM-RR) and MAP-EM (MAP-RR) iterative reconstruction algorithm. Spatial resolution and image noise measured from a wrist phantom were compared for these reconstruction techniques as well as for FBP and MLEM. Resolution improved from 8.4 mm FWHM for FBP to 6.3 mm FWHM for MLEM-RR, however noise also increased. Wrist SPECT was also performed in 5 volunteers asymptomatic for wrist injury. Better quality images were observed when incorporating the system response functions.