Impact of Measured (CT) Versus Calculated (Chang) Attenuation Correction (AC) on Visual and Quantitative Analysis of SPECT Brain Scans.

Objectives When performing a SPECT brain scan, the count distribution might be impacted by the data acquisition and processing techniques employed. A critical factor in image production and quantitative analysis is how attenuation correction (AC) is applied. The standard processing technique uses a Chang AC technique that was created using a standardized atlas. With the new SPECT/CT scanners available, AC using a low dose CT is now available. The objective of this study was to evaluate the differences between scans that have been processed using CT and Change AC.

Methods TC-99m ECD SPECT brain scans were performed on a dual detector SPECT/16 slice CT. A low dose CT AC scan was performed on all subjects. The scans were then processed identically using both Chang and CT AC. The differences between the Chang and CT were examined using visual comparison, subtraction, and quantitative analysis. In visual comparison, the regions of interest are examined for their visual distribution. In subtraction techniques, the scans were aligned and subtracted from each other, demonstrating the differences between the two scans. Quantitative analysis was performed using NeuroSTAT, which can normalize the images to global counts, Thalamus, Cerebellum, and Pons.

Results In all 10 of the scans analyzed, it was found that the Chang AC resulted in greater relative activity in the Basal Ganglia and Thalamus in all subjects. Scans that used the CT AC had deep brain structures that demonstrated reduced relative activity compared to the Chang AC scans. The cerebellum demonstrated greater relative activity in the CT AC scans than in the Change AC scans. These findings were also confirmed visually and using the subtraction technique. Using quantification, the differences in the scans can be visualized depending on the region of interest used to normalize. The Change AC technique over-corrected the Thalamus, making it to appear it had greater relative activity compared to CT AC. When normalizing to that “overcorrected” region, the rest of the brain appears visually to be decreased by comparison, also demonstrated in the quantitative analysis. A similar finding was found when comparing to the “under-corrected” cerebellum. When normalized to the whole brain counts, the difference between Chang and CT AC was minimal.

Conclusions Visual and quantitative differences in the SPECT brain studies were demonstrated depending on the use of a Chang or measured attenuation (CT) and may be especially important if quantification is used. If quantification is used, normalization to whole-brain counts appears to be desired as minimal differences were found between Chang and CT AC. When comparing results to a normal database, it would be desirable for the database to be processed with similar techniques.

• Download figure
• Open in new tab
• Download powerpoint