Recognizing and preventing artifacts with SPECT and PET imaging

Patient motion

Motion is a common source of artifact in SPECT studies and can occur from patient motion while being imaged, respiratory motion, or cardiac creep. Frank patient motion can occur in the axial, lateral, or rotational direction. ^, Typically, motion in the axial direction or motion equal to or greater than one pixel is of greater consequence. Additionally, motion during the middle of the acquisition and abrupt motions are more likely to produce significant artifacts than gradual patient motion. Planar raw data, sinograms, and summed images of the planar projections should be carefully reviewed by both the technologist and interpreting physician. Imaging findings include defects on opposite walls, tails of diminished activity, or the so-called hurricane sign. Misalignment of left ventricular segments is also commonly seen, producing a tuning fork artifact ( Fig. 5.1 A–B–C). Cardiac creep consists of a gradual upward motion of the heart in the chest and can be seen when stress images are performed too soon after exercise. Early after exercise, the maximally expanded lung volume shifts the diaphragm and the heart downward at the start of the acquisition, which gradually moves up in the chest as the patient recovers from exercise and the lung volume returns to normal. Much of this can be prevented by waiting at least 15 minutes after an exercise stress test; by counseling the patient before imaging to lie still, not cough, and not fall asleep; and by ensuring the patient’s comfort and shortening imaging times as much as possible. Given that ^99m technetium ( ^99m Tc) has largely replaced ²⁰¹ thallium ( ²⁰¹ Tl) for SPECT myocardial perfusion imaging (MPI), this artifact is now seldom seen because of the typical longer delay from injection to acquisition of ^99m Tc images (15 to 45 minutes) as described in Chapter 4 . Patient comfort should be ensured by carefully placing arms overhead to limit arm motion during image acquisition. Images should be evaluated carefully after each acquisition to assess for motion. Motion correction algorithms are now available in most commercially available SPECT processing software. ^, Although motion correction may reduce artifact severity, image quality may be adversely affected because of a lower count statistics secondary to removal of counts for motion correction (see Fig. 5.1 D). If motion correction does not substantially improve the artifact, images may have to be reacquired.

Attenuation artifacts

Photon attenuation by soft tissue is probably the most common source of artifacts seen on SPECT MPI. Degradation in image quality resulting from exaggerated soft tissue attenuation poses significant diagnostic challenges and is usually magnified in selected populations (e.g., obese people, women). Careful evaluation of rotating planar projection images should be performed for every study to identify potential sources of regional attenuation artifacts. One clue to the diagnosis of attenuation artifacts is that they may not conform to a particular coronary artery distribution. Although attenuation artifacts can affect any myocardial wall, those affecting the anterior or inferior wall are most common. The close proximity to the SPECT detector makes the lateral wall less vulnerable. Nevertheless, it can also be affected if the patient is imaged with the left arm down by the patient side. Computed tomography (CT) or radioactive line source-based attenuation correction can help minimize attenuation artifacts and false-positive findings.

Anterior wall

Breast or other chest wall soft tissue can significantly attenuate photons from the anterior wall in particular but also the lateral and inferior wall, causing artificially reduced counts in those areas. Although more prevalent in women, this type of artifact can also occur in men. If the breast position is consistent on both the stress and rest images, it will cause a similar decrease in counts on both imaging data sets. Normal wall motion and thickening in that area help confirm that the defect is likely secondary to attenuation and not a fixed perfusion defect from a prior infarct. This can be particularly problematic, however, if the breast shadow shifts position between stress and rest imaging. If the attenuation artifact only affects the stress images, it may show as a pseudoreversible defect that can result in a false-positive finding (the so-called shifting breast artifact, Fig. 5.2 ). Images should be carefully reviewed for the size and placement of the breast shadow and to see if any shift may have occurred between rest and stress. The technologists should ensure that the arms are placed above the head in similar positions between rest and stress so that the natural position of the breast will be similar in the two sets of images. Imaging without a bra is most helpful to minimize breast thickness or shifting breast artifacts.

Inferior wall

The left hemi-diaphragm may cause attenuation with decreased counts in the inferior or inferolateral walls and, in particular, in the basal inferior segment. This is most prevalent in obese patients who have an elevated left hemi-diaphragm; ascites; men; or when ²⁰¹ Tl is used. Additionally, men have greater inferior attenuation compared with women. As previously discussed, careful attention to the raw images for the position of the diaphragm, along with assessment of wall motion and thickening in that area, helps to discern if the defect represents an attenuation artifact. In addition to attenuation correction, the use of prone imaging is quite helpful to reduce diagnostic uncertainty from inferior attenuation artifacts, especially when performing stress-first imaging protocols ( Fig. 5.3 ). Prone-only imaging has been evaluated as a diagnostic option, and although it may be more comfortable for some patients, a change in position can result in a change in attenuation from surrounding soft tissue, which has to be considered during image interpretation. It is for this reason that prone imaging is seldom used alone and two-position imaging is recommended to troubleshoot inferior attenuation artifacts.

Increased extracardiac activity

Subdiaphragmatic gastrointestinal (GI) radiotracer activity can significantly affect interpretation of the images and can either increase or decrease counts in the inferior wall of the heart. Significant GI activity immediately adjacent to the heart may cause scatter of photons, which may be misregistered as activity in the inferior left ventricular (LV) wall. This may mask ischemic defects on stress imaging. The additional counts scattered into the inferior wall may now be referenced as “normal,” and the normalization of the remaining myocardial counts to these segments may cause artifactual defects in the anterior or lateral walls. In addition, the contours generated by the processing software to delineate the LV boundaries may now include this adjacent GI activity, which will introduce errors into the calculation of LV function. Additionally, excessive isotope uptake in the liver can induce a ramp filter artifact (discussed later) and produce an inferior wall perfusion defect.

Strategies to reduce artifacts and improve recognition

Awareness of the potential patient-related causes of imaging artifacts is paramount to maximizing the value of radionuclide perfusion imaging. Patients who can exercise should preferentially undergo an exercise rather than pharmacologic stress test. An exercise stress test provides incremental diagnostic and prognostic value and also reduces liver and GI radioisotope uptake. This is because of the increased blood flow to the extremities during exercise and shunting of the blood away from the splanchnic circulation, thereby reducing liver and GI radioactivity. Additionally, among patients referred for vasodilator stress testing, a combination of low-level exercise and pharmacologic stress may also reduce subdiaphragmatic uptake and improve image quality ( Fig. 5.4 ). Having the patient drink a large amount of water to distend the stomach may also displace the bowel away from the heart and improve image quality on repeat acquisition. When increased liver uptake is noted, imaging should be repeated after a 30- to 45-minute delay to allow for the tracer activity to clear from the liver. The use of ^99m Tc-labeled tracers also helps improve image quality and reduce attenuation artifacts.

Careful attention should be paid to patient comfort, arm positioning, and counseling on laying still under the camera before beginning image acquisition. It is imperative to position the patient’s arms above the head consistently during rest and stress imaging. Careful review of the planar rotating images is critical to assess for any notable or moving soft tissue attenuation patterns between rest and stress, patient motion, and extracardiac activity that may lead to artifacts and false-positive studies ( Fig. 5.5 , Video 5.1 ). Although Anger cameras with supine imaging are still the most prevalent SPECT systems, certain cameras allow for upright imaging as well (see Chapter 1 ). Upright imaging significantly improves patient comfort, thereby reducing motion artifacts and reducing the potential for shifting breast artifacts.

Soft tissue attenuation can be minimized by employing attenuation correction by means of a ¹⁵³ gadolinium line source or CT in hybrid SPECT/CT systems. Unfortunately, attenuation-correction images can have their own range of artifacts and need to be reviewed carefully. In particular, transmission/emission misalignment, truncation artifacts, exaggerated apical thinning, and normalization issues are prevalent. ^, If attenuation-correction software is not available, repeat imaging with a change in patient position may need to be employed. For example, this may involve the addition of prone imaging to troubleshoot inferior defects as previously discussed (see Fig. 5.3 ).

Assessment of wall motion and thickening is critical for differentiating between infarction and attenuation artifacts in areas with fixed defects. The presence of normal regional wall motion and thickening in an area with a fixed defect is consistent with an attenuation artifact as opposed to prior myocardial infarction, especially in a patient without a history of prior coronary artery disease (CAD). Importantly, the presence of normal regional wall motion and thickening cannot be used to troubleshoot defects that show apparent reversibility.

Hardware