Evaluation Attenuation Correction using MRI in brain SPECT images

Background: SPECT imaging is very useful in clinical applications and this is a consequence of its functional information which represents body physiology and metabolism. Quantitative parameters extracted from SPECT images would increase reliability of reports. Attenuation and scatter correction have a significant role in quantitation of nuclear medicine images. In this study uniform attenuation correction- which its validity in brain SPECT images has shown before- is compared with nonuniform attenuation correction using MR images.Method: Ten normal subjects were selected for performing SPECT and MR imaging in this research. SPECT images were reconstructed rudimentarily -without attenuation correction- using OSEM algorithm. After extraction of extracranial tissues from T1_MRI the resulted image was registered with SPECT image. Nonuniform attenuation map was reached by segmentation of T1_MR images into four classes of tissue type, including skull, soft tissue, nasal sinus and air cavity and Uniform attenuation map was considered as a class of tissue type -soft tissue-. Eventually with attributing related attenuation coefficients and applying the registration transformation matrix, the two attenuation maps were used in reconstruction algorithm. Brain lobes maps as an output of Free Surfer software was concerned to calculate the activity concentration (ACC) of each lobe of the two hemisphere for the two series of SPECT images.Results: The comparison of ACC quantities demonstrates a strong correlation between Uniform attenuation corrected and Nonuniform attenuation corrected SPECT images. Furthermore, the results represent existence of underestimation for UAC in comparison with NUAC. This underestimation has a spatial dependency. The lowest percentage was concluded for the Insula areas, while the maximum percentage was related to the frontal area.

Keywords:

MRI attenuation correction , quantitative brain SPECT