a) Center of rotation (COR).
An alignment error between the mechanical center of the rotational gantry and the center of the electronic matrix can produce reconstruction artifacts in the image. The effect is evident when the error is greater than 2 pixels in a 64x64 matrix. Smaller errors, however, can generate a loss in spatial resolution and image contrast. To a certain degree, COR errors can be electronically compensated but if a progressive deterioration is detected the camera should be serviced. COR checking should be performed frequently, usually on a weekly basis or depending on each manufacturer's recommendation and stability of the system.
b) Detector uniformity.
In SPECT, it is assumed that photon detection (sensitivity) is constant throughout the collimated surface of the detector when exposed to a flood source. Errors in flood field uniformity are produced with significant variations in detection efficiency, which can be due to physical or electronic factors. Uniformity is much more critical for SPECT as compared to planar imaging because more severe artifacts can be introduced. Most systems use previously acquired "correction maps" kept in memory to compensate for non-uniformities before reconstruction. Extrinsic uniformity (with collimator in place) should be daily checked for at least 3,000K counts in a 64x64 matrix using a refillable 99mTc phantom or a 57Co flat source. Larger matrices will require larger number of counts. Correction or sensitivity maps are to be acquired periodically for a total of 30,000-40,000K counts. Non-uniformity should be kept lower than 3% for SPECT.
c) Attenuation correction.
Although still expensive and not universally available, attenuation correction represents a rapidly evolving standard for myocardial SPECT. Accurate attenuation correction is mainly dependent upon high-quality transmission images. Therefore, QC of attenuation correction performance should include the following verifications: uniformity, variability and temporal drift of the reference transmission scan; consistency of hardware performance; pre-scanning methods to ensure adequate transmission scan counts; and algorithms that assist the technologist and physician in assessing the sufficiency of the data. Implementation of all these QC techniques, however, has not been incorporated in the current releases of all available attenuation correction protocols. Furthermore, QC of transmission data and attenuation-corrected reconstructed images should be performed for each patient and interpreted in comparison with non-corrected conventional tomograms.
d) Other verifications.
Energy resolution, spatial resolution, linearity, pixel size, tomographic resolution and tomographic uniformity should also be checked periodically in accordance to manufacturer's recommendations or established protocols. Detailed description of these procedures are beyond the scope of this presentation and can be found elsewhere.
e) Quality control of clinical studies.
After acquisition, raw images should be reviewed in cine mode to check for appropriate radiopharmaceutical biodistribution, total count density, data integrity, "upward creep" of the heart, attenuation, subdiaphragmatic activity and patient motion. Acquisition should be immediately repeated if necessary. Stress and rest reconstructed images should not contain artifacts (see below) and be properly aligned and count-normalized for comparison. ECG-gated images should be reviewed in cine display after reconstruction and checked for "flickering" effect produced by count loss in the last images of the cardiac cycle due to variable R-R interval or other ECG-gating related artifacts. QC of gated studies should also include the analysis of the volume curve which should be of expected shape (starting and ending at diastole), and checking for extracardiac activity that could be interpreted as belonging to the cardiac wall by the edge-tracking algorithm.
