CT-based attenuation and scatter correction compared with uniform attenuation correction in brain perfusion SPECT imaging for dementia

This study investigated if the appearance and diagnostic accuracy of HMPAO brain perfusion SPECT images could be improved by using CT-based attenuation and scatter correction compared with the uniform attenuation correction method.

A cohort of subjects who were clinically categorized as Alzheimer's Disease ($n=38$), Dementia with Lewy Bodies ($n=29$) or healthy normal controls ($n=30$), underwent SPECT imaging with Tc-99m HMPAO and a separate CT scan. The SPECT images were processed using: (a) correction map derived from the subject's CT scan or (b) the Chang uniform approximation for correction or (c) no attenuation correction. Images were visually inspected. The ratios between key regions of interest known to be affected or spared in each condition were calculated for each correction method, and the differences between these ratios were evaluated.

The images produced using the different corrections were noted to be visually different. However, ROI analysis found similar statistically significant differences between control and dementia groups and between AD and DLB groups regardless of the correction map used.

We did not identify an improvement in diagnostic accuracy in images which were corrected using CT-based attenuation and scatter correction, compared with those corrected using a uniform correction map. (paper)