Quality control (QC) of single-photon emission computed tomography (SPECT) and positron emission tomography (PET) nuclear cardiology procedures is a multiple-step process that begins before the patient enters the laboratory, continues during the acquisition, and after the patient leaves the laboratory. SPECT QC and PET QC require the close attention of all personnel and physicians that are involved with the laboratory. Requirements for imaging systems QC are based on Nuclear Regulatory Commission (NRC) requirements, agreement state requirements, accepted imaging guidelines, and the Intersocietal Accreditation Commission Nuclear/PET (IAC Nuclear/PET), and American College of Radiology (ACR) Accreditation standards.1–8 The terminology used, requirements, and frequency may vary slightly between the standards, the model of scanners, and the original equipment manufacturers (OEM) recommendations. However, the basic premise of why it is necessary is the same in all situations: to ensure adequate camera performance, identify any potential sources of error or artifact within an acquisition, and ultimately provide the patient and referring physician with the best quality information possible. If QC procedures are not followed, it may lead to an equivocal or falsely interpreted study, which may result in increased downstream costs as well as poor outcomes. This chapter will review the routine QC procedures performed by the nuclear medicine technologist before, during, and after the acquisition of a nuclear cardiology study. Additional QC procedures and calibrations that are performed by camera service engineers or medical physicist will not be discussed.
There are several required and recommended equipment QC procedures that should be performed on each imaging system.2,5–31 The frequency of the procedures may vary among equipment manufacturers; however, all are important to ensure proper system performance (Tables 5-1 and 5-2). These tasks consist of daily, weekly, monthly, and quarterly system testing.
Energy peaking (photopeak analysis) should be performed daily to verify that the camera is counting photons using the correct energy.2,3,9–11 Each imaging system should be checked before use to ensure that the camera peaking electronics are functioning properly, that the energy window has not drifted, and that the energy spectrum is in appropriate shape.
During the procedure, the pulse height analyzer’s energy window should be manually or automatically placed over the correct photopeak energy. It is recommended that no greater than a 20% energy window be used in order to obtain the most accurate peak energy.2,9–11 If the test is performed intrinsically, a point source should be placed at least 1.5 m away from the surface of the camera detector. If performed extrinsically, a sheet source should be used. In either case, the source should be enough to flood the entire field of view (FOV).
Verifying the photopeak daily will help prevent artifacts that may occur due to inappropriate photons entering the acquisition and degrading image quality. An off-centered photopeak may also result in poor count statistics, which will result in a poor-quality image.2–4,9–11 If dual-isotope procedures are being performed, on some older imaging systems it may be necessary to perform this procedure between each acquisition of thallium-201 (Tl-201) and technetium-99m (Tc-99m) isotopes. It is also necessary to repeat this procedure whenever the camera is powered down for any reason during the day.
A daily uniformity flood should be performed to analyze system performance and to ensure the sensitivity response of the system is uniform across the detector surface.2–15 This is performed by exposure of the detector surface to a radioactive source. The recommended method is to perform this procedure intrinsically using a Tc-99m point source of approximately 100 to 500 μCi in ≤0.5 mL of volume. The point source should be placed in the center and at a distance of approximately five useful fields of view (UFOV) away from the detector surface. An acquisition should be performed for approximately 2 to 5 million counts using a 20% energy window.2,9,11,15 This may also be performed extrinsically using a 57-cobalt sheet source. This method is performed frequently on dual-head camera systems because the acquisition can be performed on both detectors at the same time. It is important to remember, however, that during the uniformity analysis of extrinsic floods, the outer 10% to 20% of the FOV should not be considered due to possible edge packing.2,9,11,15
After the acquisition of the flood field uniformity, the image should be evaluated visually, and a computerized analysis should be performed to measure the performance of the system. Central FOV and UFOV parameters should be <5% based on standards; <3% is preferred. This analysis should be performed following manufacturers’ protocols and will be specific to each imaging system. If nonuniformities are detected, the system should not be used until service is performed. Severe artifacts, such as malfunctioning photomultiplier tubes, will be easily detected (Fig. 5-1). Smaller abnormalities may be more difficult to detect, however. These small, undetected nonuniformities may produce artifacts within patient acquisitions and result in a misinterpretation of the study.2,9–15 It should be noted that the daily uniformity flood is not the same as the high count extrinsic uniformity flood that is completed on a monthly or per OEM recommendations to correct for detector and collimator nonuniformities.2,5–8
A center of rotation (COR) evaluation may be performed weekly or monthly (depending on the manufacturer) in order to ensure and maintain the detector’s electronic matrix alignment.2,5–15 This is the x-axis position of the actual axis of rotation as seen by the image matrix.2 If a COR error occurs, it may produce what has been characterized as a “doughnut”-shaped or “tuning fork” artifact (Fig. 5-2).2 This artifact may appear similar to that caused by patient motion on the perfusion acquisition, but occurs on every patient study. This error becomes more pronounced if the deviation widens, particularly >2 pixels in a 64 × 64 matrix. Smaller deviations may not produce an artifact but could result in decreased spatial resolution and image contrast.2,9,11,15
Several manufacturers have recommended camera-specific protocols to follow when performing COR procedures. Most recommend using a 500 to 750 μCi point source placed off center in the FOV at approximately 4 to 8 in away from the detector surface.2,9,11,15 The procedure is then performed using similar parameters as a standard SPECT acquisition. An analysis of this acquisition should be performed, and for any misalignment of >0.5 pixels of the x-axis, the COR should be recalibrated (Fig. 5-3).
System spatial resolution and linearity evaluation should be performed weekly. This procedure is to document spatial resolution over time, as well as to evaluate the detector’s ability to produce straight lines.2,9–11,15 This should be performed intrinsically using a radioactive point source and a test phantom. This evaluation should not be performed extrinsically because the patterns of the lead bars in the phantom and the lead septa of the collimator may interfere with one another causing artifacts.
There are several bar phantoms available commercially that may be used for this test. The most commonly used are the parallel-line-equal-spaced (PLES), orthogonal hole, and four-quadrant phantoms.2,5–15 The four-quadrant phantom has four sections of differing thickness lead bars that are equally spaced. If this type of phantom is used routinely, it should be placed over the detector surface so that each differing size lead bar is in a different position from the previously performed test (rotated 90 degrees from the previously placed position). This will allow the most tightly spaced bars of the phantom to appear over the entire surface of the detector every fifth acquisition. This will provide the most thorough evaluation of the entire detector surface over time.2,9,11,15
Acquisition parameters for the resolution and linearity test should be similar to those used to perform a daily uniformity flood. On completion, the image should be evaluated visually to assess the straightness of the lines produced by the bar phantom and for how well each different-size lead bar is visualized (Fig. 5-4). The test should be stored in an electronic format for comparison of system resolution over time. As a decrease in resolution appears (loss of visualization of individual bars), preventive maintenance should be performed on the system. Manufacturers may supply software to evaluate linearity and resolution that should be used when available.
The Intersocietal Accreditation Commission Nuclear/PET and ACR recommend that SPECT phantoms be performed annually.5,6 This allows the evaluation of an imaging system’s performance and limitations by providing a comparative means to judge previous performance with the most recent phantom acquisition.2,9–11,15 Commercially available multipurpose Plexiglas and water-filled SPECT phantoms have attenuation and scattering properties similar to those of tissue, thus simulating clinical conditions in a three-dimensional (3D) view. This provides a realistic comparison of system performance similar to that in the clinical setting. Therefore, acquisition parameters used should be similar to those of standard SPECT.
Several manufacturers have recently brought solid-state imaging systems to market, such as the D-SPECT (Spectrum Dynamics Medical, Inc.).18–21,23,24 Daily QC procedures are performed with either a Co-57 rod or phantom. Depending on the manufacturer, the camera software performs a pass/fail test for: energy resolution, energy peaking, detector registration, regional/global detector homogeneity, scan sensitivity, FOV, and faulty pixels.18–21,23 If the daily QC fails, the camera service engineer should be contacted for repair.
A blank scan is performed daily prior to any patient testing to analyze system performance and stability.5–8,16,18,24–26 The daily blank scan is the equivalent of the daily uniformity scan for SPECT and allows for detection of any sudden change in system performance such as a module malfunction.
Once a blank scan is acquired, it is processed using standard OEM reconstruction parameters and defaults, then displayed as a two-dimensional (2D) sinogram in gray scale. The blank scan is inspected visually (Fig. 5-5) and quantitatively (Fig. 5-6), if available on the system, for variations in image quality. Hot or cold streaks in the sinogram (Fig. 5-5) are an indication of a detector or block malfunction and identify blocks or modules (buckets) which are more (or less) sensitive than the respective system average.16,18,24,26 It is also important to evaluate the results of the blank scan over time, looking for trends in system performance.25 If blank scan quality is poor, corrective action is required. Updates to the PMT gain calibrations, coincidence timing calibration (CTC) or normalization files, and a repeat blank scan may be necessary.16,18,24,26 If the blank scan fails after this step, then a service call should be placed and the system should be removed from service until repaired. If blank scan is acceptable, it is saved and set as the daily transmission reference scan for corrections when using Ge-68 or Cs-137 sources for attenuation correction.
Singles update gain or Bucket setup test is performed on a weekly basis, or as recommended by the manufacturer, using the Ge-68 rod sources.2,25,27,30,31 The purpose of this calibration is to balance the gain characteristics of the photo multiplier tubes in each block. Singles or Bucket setup will compensate for photo multiplier tube drift caused by time and environmental changes.
The Coincidence Timing Calibration (CTC) scan is performed, at minimum, on a weekly basis, but may be performed daily.2,25,27,31 The CTC scan is completed using Ge-68 rod sources to evaluate the constancy of the timing resolution. The duration of the scan is very short, typically 2 minutes. The calculation of the scan data compensates for differences in event detection hardware and adjusts for timing delays, ensuring events from all blocks are time stamped equally. This scan can be manufacturer and scanner model specific.
A normalization correction is routinely performed monthly or quarterly depending on the scanner’s manufacturer.16,18,24–27,30,31 It should also be acquired after a failed daily blank scan or scanner service. The normalization scan is used to correct for variations in the sensitivity of the blocks or buckets and adjustments in the efficiency in each line of response (LOR) in the sinogram.16,18,24,26,30,31 A poor normalization scan will result in horizontal streaks through the image (Fig. 5-7).
A normalization correction is performed using radioactive rod (pin) sources or a solid phantom source; most commonly used is germanium-68 (Ge-68) source. Normalizations are performed in both the 2D and 3D mode, if available on the scanner.2,24,26,30,31 Normalization scans may be performed by the technologist or the service provider. It is preferable to run the normalization scan overnight as the duration of the scan may range from 30 minutes to 12–18 hours, depending on the system type and age of the source. As the source decays, the normalization data should be monitored to ensure acceptable count statistics and quality transmission maps are obtained. Sources should be routinely replaced based on manufacturer’s recommendations, typically every 12 to 18 months depending on the system in use.
Once complete, the normalization scan is reconstructed and set as a new default based on OEM guidelines. A new blank scan must also be completed and applied each time a new normalization correction is completed.16,18,24,26 Some systems have a pass or fail message upon completion of the scan. If the scan fails, the camera service engineer should be contacted for evaluation and repair.
Well counter correction (WCC) is an image pixel cross calibration to the dose calibrator. WCC uses a water-filled phantom and a known quantity of radioisotope to correlate the measured numerical value in each image pixel to a specific activity measured in physical units.31 The 3D WCC also removes any axial sensitivity variation between images. Depending on the manufacturer, it is recommended that the 2D and 3D well counter acquisitions and corrections be updated quarterly to maintain good match of sensitivity. WCC can be performed by technologists, but are usually performed by service engineers during the quarterly preventative maintenance services.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
