Artifacts in SPECT radionuclide myocardial perfusion imaging

INTRODUCTION — The value of radionuclide myocardial perfusion imaging (rMPI) with single-photon emission computed tomography (SPECT) for the diagnosis and risk stratification of coronary artery disease (CAD) is well-established. However, the full clinical potential of rMPI has not been realized due to numerous factors that result in image artifacts. These artifacts reduce image quality and increase the risk of misinterpretation of the results. Interpretation of rMPI requires knowledge of common patterns and causes of artifacts.

Soft tissue attenuation is one of the more common causes of rMPI artifacts. Photon attenuation occurs as photon beams experience loss of energy while traversing tissue. Since the heart is surrounded by tissues of varying densities (eg, bone, lungs, and breast), radionuclide imaging of the thorax results in nonuniform myocardial photon activity due to attenuation from these intervening structures. Soft tissue attenuation may result in an apparent fixed or reversible perfusion abnormality or defect. Fixed anatomic structures generally will produce the same attenuation artifact both on stress and rest imaging. Photon attenuation and the resultant artifacts are affected by several factors, including:

●Body habitus

●Depth of the heart in the body

●Male or female gender

●Patient positioning under the camera

●Choice of radioactive tracer (eg, 99m-technetium, thallium-201) and activity of the radioisotope

Sources of attenuation with SPECT rMPI and methods to interpret or eliminate attenuation artifacts will be discussed here. The clinical uses of SPECT rMPI are discussed in detail separately. (See "Overview of stress radionuclide myocardial perfusion imaging".)

SOURCES OF ATTENUATION ARTIFACT — Soft tissue attenuation may result in an apparent focal fixed or reversible perfusion abnormality or defect. Fixed anatomic structures generally will produce the same attenuation artifact both on stress and rest images. This is generally interpreted as a fixed perfusion abnormality, and may be difficult to differentiate from myocardial scar or infarction. However, if the body part causing attenuation (eg, breast, diaphragm, arm, etc) changes its position between stress and rest imaging, the resultant soft tissue attenuation may produce a reversible perfusion defect mimicking myocardial ischemia.

Although there are multiple sources of attenuation artifact, our discussion will focus on the following common sources in detail [1]:

●Diaphragmatic attenuation

●Breast attenuation

●Liver attenuation

●Attenuation due to obesity

●Attenuation due to the patient's arm not being raised overhead (on a supine or prone study)

Diaphragmatic attenuation — Attenuation from the left hemidiaphragm can reduce photon counts in the inferior wall. This artifact results in a perfusion defect in the inferior segments which generally correspond to the posterior descending coronary artery (PDA) territory. Diaphragmatic attenuation is estimated to occur in up to 25 percent of myocardial perfusion studies [2].

Diaphragmatic attenuation is more common in the following settings:

●Male patients

●Obesity

●Abdominal protuberance (eg, ascites or peritoneal dialysis)

●Elevated left hemidiaphragm

●Use of thallium-201 (compared with 99m-technetium) radionuclide

Soft tissue attenuation — Soft tissue attenuation can lead to an apparent reduction in counts in myocardial segments depending upon the location of the soft tissue. This results in apparent perfusion artifacts generally involving the anterior or anterolateral segments, corresponding to in the left anterior descending (LAD) and/or left circumflex coronary artery (LCX) territories [3]. Attenuation artifacts of the anterior and anterolateral walls from breast tissue may occur in up to 40 percent of myocardial perfusion studies in women, but may also occur in men with gynecomastia [4]. Attenuation artifacts of the lateral, inferolateral, and anterolateral walls may also occur in patients with significant obesity [5]. Patients with breast implants require special consideration as artifacts may intensify.

Breast size, breast tissue density, and breast position relative to the heart may affect the extent and severity of soft tissue attenuation. Large, pendulous breasts that tend to lie closer to the lateral chest wall are more likely to cause attenuation of the anterolateral or lateral walls, whereas small or dense breasts that lie over the anterior chest wall may cause attenuation of signal from the anterior or anteroseptal walls. Large, pendulous breasts tend to shift in position between rest and stress imaging, and may result in a reversible defect, which is indistinguishable from ischemia (movie 1).

Liver or bowel attenuation — Soft tissue attenuation due to increased radionuclide tracer in the liver (ie, a "hot" liver) or radiotracer in a loop of bowel adjacent to the myocardium may distort counts detected in the inferior wall, resulting in artifactual perfusion abnormalities in the left circumflex or posterior descending coronary artery territory [6,7].

OUR APPROACH TO REDUCING ARTIFACTS — A variety of artifacts cause an increase in false-positive rMPI studies, resulting in decreased specificity for the diagnosis of CAD. Artifacts may also decrease interpretive certainty of the interpreting clinician and lower the confidence of referring clinicians. False-positive scans may lead to unnecessary downstream testing, including invasive coronary angiography, resulting in increased costs and the potential for patient morbidity. Paradoxically, a high frequency of artifacts may also cause the interpreting clinician to disregard true perfusion abnormalities, assuming an artifact instead, thereby reducing sensitivity of the test. Consequently, patients with CAD may be missed or go undertreated.

Our approach to reducing artifacts is as follows:

●Exercise patients whenever possible, even when combined with vasodilator stress, to reduce extracardiac radiotracer uptake

●Reduce patient motion (eg, shorten acquisition time, ensure patient comfort, explain the importance of not moving during the acquisition)

●Use of Tc99m myocardial perfusion tracers rather than thallium-201

●Re-acquiring images when appropriate (eg, with Tc99m sestamibi and tetrofosmin, images can be re-acquired if there is patient motion of significant extracardiac activity on initial images)

●Use of multiple patient positions as needed (eg, prone or upright in addition to supine)

●Use of attenuation correction algorithms

●Use of newer SPECT technology

STRATEGIES TO REDUCE ATTENUATION ARTIFACT AND/OR IMPROVE ATTENUATION ARTIFACT RECOGNITION — While it is difficult to entirely eliminate the attenuation of photon counts due to soft tissue, with single-photon emission computed tomography (SPECT), soft tissue attenuation can often be recognized and distinguished from true myocardial perfusion defects, thereby improving diagnostic accuracy. Several methods have been developed to improve artifact recognition, with varying degrees of success. These include:

●Review of planar images

●Quantitative analysis

●Improving count statistics

●Higher-energy radioactive tracers (99m-technetium versus thallium-201)

●Prone or upright imaging

●Electrocardiogram (ECG)-gated SPECT imaging

●Attenuation correction

Review of planar images — Review of a cine loop of planar rotating images may identify potential problems due to soft tissue attenuation, prominent patient motion, abnormal lung activity, and other incidental findings (eg, neoplastic lesions). The planar images provide valuable information about the proximity of soft tissue to the myocardial structures to alert the readers of potential areas of soft tissue attenuation [8].

Quantitative analysis — The use of quantitative analysis of perfusion SPECT scans may assist in differentiating attenuation artifacts from true perfusion defects [9]. Normal databases have quantified gender-dependent myocardial perfusion distributions in normal populations [10]. Based upon the degree of variation of the study from expected patterns, abnormality criteria have been developed that may serve as a second opinion in differentiating between coronary disease and attenuation artifact.

Improving isotope count statistics — When prominent soft tissue attenuation is anticipated or encountered, the quality of the imaging can often be improved by prolonging duration of data acquisition. Prolonged acquisition allows for greater isotope detection by the camera, which should result in improved image quality and may reduce the potential for attenuation artifacts [11]. A potential downside to prolonging the acquisition time, however, is that the extended time can lead to patient discomfort and increase the likelihood of motion artifacts.

Advancing technology in newer camera systems, using solid-state conductors with higher energy resolution than traditional SPECT cameras, and innovative designs of the gantry, collimators, and detectors, allows greater sampling of the myocardial region, and thus higher local count sensitivity and faster acquisitions than traditional SPECT systems [12]. These newer systems provide an improvement in spatial resolution and sensitivity. (See "Basic properties of myocardial perfusion agents".)

Use of technetium-based imaging agents — 99m-Technetium has advantages over thallium-201 as a radionuclide tracer for myocardial imaging. 99m-Technetium has higher emission energy (140keV versus 78 keV) and a shorter half-life (six hours versus 72 hours) that permit administration of a higher dose. These qualities result in improved image quality and contrast, and less soft tissue attenuation with reduced patient radiation exposure [13]. Despite this advantage, attenuation artifact continues to be a significant problem with the use of technetium-based agents. A recognized downside to 99m-technetium-based imaging is the increased frequency of liver and gut activity in contrast to thallium-201.

Prone or upright imaging — Supine imaging is standard on most scanners. Depending on the scanner, sequential scans in either the supine and prone, or supine and upright, positions may be performed to allow differentiation of true perfusion defects from attenuation artifacts in systems without attenuation correction hardware.

The use of prone imaging can reduce diaphragmatic attenuation and improve inferior wall image quality. The rationale for prone imaging is that the heart shifts slightly superior and the diaphragm is more inferior in the prone position, increasing the distance between the diaphragm and the left ventricular (LV) inferior wall. Thus, in a patient with an inferior wall perfusion abnormality observed on supine imaging but not on prone imaging, the abnormality should be considered to be due to attenuation, and the study can be considered normal. Prone imaging may occasionally create fixed anterior or lateral defects, which may represent attenuation artifact. However, if the defect is not present on the supine images, it is most likely due to attenuation artifact.

Similarly, the use of upright imaging, available with some modern camera systems, can also reduce diaphragmatic attenuation by shifting the diaphragm inferiorly, thereby reducing potential overlap of the diaphragm and the inferior wall and improving image quality. Similar to the process with prone imaging, a study in which an inferior perfusion abnormality is observed in supine but not upright positioning should be considered due to attenuation artifact. Likewise, defects observed in other walls in one patient position but not in the other may also be interpreted as artifact due to other types of soft tissue attenuation, (eg, breast, adipose tissue).

Several approaches to sequential scans in two positions are available depending on the camera system and the laboratory workflow.

●Most laboratories start with standard supine imaging.

●Some laboratories may perform prone imaging if there is concern regarding inferior wall attenuation. (See 'Diaphragmatic attenuation' above.)

●Other laboratories have algorithms whereby select patients such as those with specific body habitus (eg, obesity or abdominal protuberance) undergo serial supine and prone imaging in addition to those who routinely perform serial supine and prone scans [14-16].

●Some laboratories routinely perform serial two-position imaging for both rest and stress acquisitions, whereas others may perform serial two-position imaging for only the stress acquisition.  

Combined imaging (prone and supine) requires additional camera and technologist time, thus reducing laboratory efficiency. However, modern solid state scanners facilitate fast imaging times and routine serial two-position imaging without significantly impacting efficiency.

ECG-gated SPECT imaging — ECG-gated imaging may help differentiate between myocardial infarction (MI) and attenuation artifact.

●Normal regional function (ie, normal wall motion and myocardial thickening) in an area with a fixed defect on perfusion imaging may be classified as a soft tissue attenuation artifact, whereas abnormal function in conjunction with a similar fixed defect would be compatible with MI or myocardial stunning.

●In contrast, this gating technique is less useful in determining whether a reversible perfusion abnormality is due to coronary artery disease since the gating procedure is acquired 45 to 60 minutes later at such a time that a wall motion abnormality due to ischemia has most often resolved and will not be observed. However, in the presence of a normal resting image, the development of a post-stress defect associated with a wall motion abnormality increases the likelihood of CAD and likely represents myocardial stunning due to severe ischemia.

ECG-gating refers to the process of acquiring images on successive beats at multiple points in the cardiac cycle. Based upon timing related to the QRS complex, a moving image of a sample ventricular contraction is reconstructed. Using this technique, both LV and right ventricular (RV) function can be assessed. ECG-gating requires regular R to R intervals, and the technique is not as reliable in patients with arrhythmias such atrial fibrillation and frequent ectopic beats.

The potential value of ECG-gated imaging has been illustrated in multiple studies. As one example, in a study of 180 consecutive patients with fixed myocardial perfusion defects, among the 102 patients with ECG or clinical evidence of a prior myocardial infarction, 96 percent had abnormal wall motion on ECG-gated imaging, whereas only 23 percent of patients with a fixed defect but no prior evidence of clinical myocardial infarction had abnormal wall motion [17]. In patients with normal function on gated images, 91 percent of the fixed defects could be accounted for by soft tissue attenuation.

Exercise — Generally, patients who can exercise should have exercise rather than pharmacologic stress testing. In addition to the physiologic data (eg, exercise duration, blood pressure/heart rate response) derived from exercise testing, exercise enhances the clearance of radioactive tracers from the hepatobiliary system and from bowel, such that these artifacts are less common with exercise stress than with vasodilator stress. To allow for tracer clearance from the hepatobiliary system and bowel, it is advisable to wait longer (at least 45 minutes) prior to imaging patients after pharmacologic stress testing compared with waiting 15 minutes following exercise. Combining exercise with vasodilator stress reduces liver activity in comparison to vasodilator stress alone and patients may be imaged earlier and may improve image quality [18,19]. (See 'Liver or bowel attenuation' above.)

ATTENUATION CORRECTION — Attenuation correction refers to automated methods that adjust the intensity of the myocardial perfusion image to reflect the estimated magnitude of soft tissue attenuation on different regions of the heart. Attenuation correction methods include either external line-source of radiation or computed tomographic (CT) techniques.

Methods — Artifacts due to soft tissue attenuation may be improved by various techniques of attenuation correction [20]. In order to apply such methods, it is necessary to create a patient-specific attenuation map [21]. An attenuation map represents the degree of photon attenuation across different areas of the thorax.

Transmission imaging is one means of creating an attenuation map. Different camera systems have been designed with various geometric arrangements of transmission sources, including point/line/sheet sources or a scanning line source of external radiation [20,22-27]. The most common configuration for transmission imaging uses a gadolinium-153 scanning line source with dual-head 90-degree detectors [20]. With this method, an external source of radiation (eg, technetium, gadolinium, or x-ray) is positioned on one side of the patient and a detector on the other side [20]. The distance between the source and detector is known, and from the measured intensity data one can calculate the degree of attenuation across the thorax. By obtaining a number of these measures at different angles, a "map" of attenuation across the thorax is generated. The attenuation maps are then integrated with the raw myocardial perfusion (emission) images to reconstruct image data with attenuation correction [22,23].

Hybrid SPECT/CT systems have not only become more common, but they have also evolved and demonstrate a range of capability and integration. A detailed discussion of the types of systems and capabilities is beyond the scope of this review [12]. Regardless of the approach used, it is important to emphasize quality control procedures in the use of CT for attenuation correction, and practical considerations including pitfalls such as truncation of data and misregistration of transmission and emission data, which can create defects indistinguishable from perfusion abnormalities due to underlying CAD. Because of the complexity of current attenuation correction methods in SPECT, we recommend that the non-attenuation-corrected images be interpreted along with the attenuation-corrected images [28-31].

Clinical studies — The combination of attenuation correction with other corrective adjustments, including ECG-gating, results in additive improvements in diagnostic accuracy [32,33]. Attenuation correction improves the relative uniformity of radionuclide tracer distribution in patients with a low likelihood of coronary artery disease (CAD), resulting in an improvement in diagnostic accuracy [28-31,34]. In general, attenuation correction results in fewer false-positive tests, which is demonstrated by improved specificity and higher rates of normal studies in control populations. Furthermore, an added value of the use of CT for attenuation correction is that the presence or absence of coronary artery calcification may aid in SPECT MPI interpretation and frequently changes the behavior of both the treating physician and patient [35,36].

The diagnostic value of attenuation correction is demonstrated by the following observations:

●In a series of 60 patients with angiographic CAD and 59 healthy controls (defined as patients with ≤5 percent likelihood of CAD), the addition of attenuation correction increased the percent of studies in controls that were read as normal (88 to 98 percent) [28]. In the overall population, attenuation correction resulted in a marked increase in specificity, from 46 to 82 percent.

●In a multicenter trial involving 96 patients with angiographic CAD and 88 subjects with low likelihood of CAD, attenuation correction resulted in a significant higher normalcy rates (86 versus 96 percent) [31]. However, sensitivity was not improved with attenuation correction, and for patients with RCA or multivessel disease, sensitivity was slightly lower.

●Attenuation artifact is more prevalent in obese patients. The value of attenuation correction in this population was evaluated in 116 patients (60 with body mass index (BMI) <30 kg/m2, and 56 with BMI ≥30 kg/m2) who underwent SPECT rMPI with ECG-gating followed by coronary angiography within 60 days [32]. The specificity for clinically significant CHD (stenosis ≥70 percent) was significantly improved with attenuation correction, regardless of BMI (79 versus 50 percent in nonobese patients, and 76 versus 41 percent in obese patients).

Attenuation-corrected quantitative analysis — Myocardial perfusion studies are usually interpreted by visual inspection. However, photon count intensity can also be measured by automated quantitative analytic tools. These automated counts are compared against values derived from databases of normal patients. (See 'Quantitative analysis' above.)

The diagnostic accuracy of these databases may be compromised by attenuation artifact. In addition, gender-differences in attenuation artifact contribute to the differences in normal values used in quantitative analyses for men and women. As with visual inspection, the diagnostic accuracy of quantitative analysis is improved with attenuation correction [31,34,37].

A gender-independent normal database for attenuation corrected rMPI has been generated and subsequently validated in separate series of obese patients [34]. In contrast to uncorrected rMPI studies, there was no difference in perfusion distributions between normal men and women with attenuation corrected studies.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Stress testing and cardiopulmonary exercise testing".)

SUMMARY AND RECOMMENDATIONS

●Because the heart is surrounded by tissues of varying densities (eg, bone, lungs, and breast), soft tissue attenuation is one of the most common causes of artifacts seen on radionuclide myocardial perfusion imaging (rMPI). (See 'Introduction' above.)

●There are multiple sources of attenuation artifacts, with the most common clinically relevant sources being diaphragmatic attenuation, breast attenuation, liver attenuation, and attenuation due to obesity. (See 'Sources of attenuation artifact' above.)

●While it is difficult to entirely eliminate the attenuation of photon counts due to soft tissue, with single-photon emission computed tomography (SPECT), soft tissue attenuation can often be recognized and distinguished from true myocardial perfusion defects. Methods employed with SPECT to reduce the impact of attenuation and potential attenuation artifacts include the review of planar images, the ability for quantitative analysis, prolonged acquisition times, prone or upright imaging, ECG-gating, and attenuation correction. (See 'Strategies to reduce attenuation artifact and/or improve attenuation artifact recognition' above.)

●The use of prone or supine imaging, in which the heart shifts, thereby increasing the distance between the diaphragm and the inferior wall, can reduce diaphragmatic attenuation and improve inferior wall image quality. Most laboratories start with standard supine imaging and perform prone imaging if there is concern regarding inferior wall attenuation, although local protocols may vary significantly depending upon the patient population, types of equipment, etc. (See 'Prone or upright imaging' above.)

●ECG-gating refers to the process of acquiring images on successive beats at multiple points in the cardiac cycle. ECG-gated imaging may help differentiate between myocardial infarction and attenuation artifact. Normal wall motion and myocardial thickening in an area with a fixed defect on perfusion imaging is consistent with soft tissue attenuation artifact, whereas abnormal function in conjunction with a perfusion defect would be compatible with myocardial infarction or myocardial stunning. (See 'ECG-gated SPECT imaging' above.)

●Attenuation correction refers to automated methods that adjust the intensity of the myocardial perfusion image to reflect the estimated magnitude of soft tissue attenuation on different regions of the heart. Attenuation correction improves the relative uniformity of radionuclide tracer distribution in patients with a low likelihood of coronary artery disease (CAD), resulting in an improvement in diagnostic accuracy. (See 'Attenuation correction' above.)

ACKNOWLEDGMENT — The editorial staff at UpToDate acknowledge Gary Heller, MD, who contributed to an earlier version of this topic review.