Skip to main content
. 2011 Jun 15;11(1):56–66. doi: 10.1102/1470-7330.2011.0012

Table 3.

QSPECT processing workflow in Syngo MI Applications

Step Activity Description
1 Data selection Selection of raw SPECT acquisition and CTAC series
2 Scatter estimation Extraction of the photopeak acquisition series (energy window A) and generation the lower scatter series (from energy window B) using the DEW technique
3 Series reconstruction Non-corrected filtered backprojection reconstruction, for the purpose of generating the SPECT matrix for CTAC resampling
4 Series registration Resampling of CTAC to the SPECT matrix
5 Attenuation map Conversion of resampled CTAC into an attenuation map using 208 keV narrow beam geometry settings (no filtering)
6 Series reconstructiona Iterative reconstruction (OSEM Flash 3D, 4 iterations, 8 subsets) with AC and SC (no zoom, no filtering)
7 Series arithmetic Multiplication (rescaling) of the raw SPECT acquisition series by 0.01 to avoid pixel saturation (>65,535 counts/pixel) during the next step. The Use Scale Factor option was enabled in this activityc
8 Series arithmetic Sum of the rescaled raw SPECT acquisition series frames (96 × 4 energy windows = 384 frames) into one static image. The displayed count sum of this static image times 100 is the wide-spectrum counts (CWo)
9 Series arithmeticb Multiplication of the reconstructed SPECT series by a coefficient (K) converting counts to dead-time corrected quantitative uptake data. The resulting series is saved as QSPECT

aFor the calibration studies, the photopeak and scatter series were scaled down before reconstruction (by a factor up to 100) in order to avoid voxel saturation (>65,535 counts/voxel) in the reconstructed SPECT dataset.

bThis step was not performed for the calibration studies.

cThis allows fractions of counts (e.g. 2 decimal digits) to be preserved in subsequent step.