Image Artifacts on Myocardial SPECT

Paul J Roach MB BS

J Anthony Parker MD PhD

November 23, 1993

Case Presentation:

A 53 year old male with a history of inferior myocardial infarction presented with recurrent anginal chest pain.

Findings:

Myocaridial perfusion scintigraphy with Tl-201 showed a mild inferior wall defect on the post exercise images. Reinjection images showed marked abnormality in perfusion with increased uptake in the anterior wall anda more prominent defect in the inferior wall. (Upper left: verticle long axis: upper right: horizontal long axis; lower left: short axis.) Cine display of the raw data showed condiiderable axial movement at the mid-point of acquisition. Repeat study without patient motion show a pattern similar to the stress images.

Discussion:

While SPECT has improved the sensitivity of Tl-201 myocardial scintigraphy in the detection of CAD this is frequently at the expense of specificity. Often correctable, artifacts may be caused by:

1. Soft tissue attenuation

A) Breast. This usually causes defects in the anterior wall but if breasts are pendulous, lateral defects may occur. Mastectomy, augmentation and gynaecomastia in men will complicate interpretation of studies, especially when bullseye reconstructions are used. Transmission CT, markers and cine displays are useful in dealing with this problem. Some centres (including BIH) routinely perform prone SPECT, in an effort to reduce diaphragmatic attenuation. This may lead to small anterior and anteroseptal defects and depending on the system used, the distance between camera and patient may be increased with a resultant fall in resolution.

B) Obesity. Commonly associated with lateral wall fixed defects. Potentiated in the supine position.

C) Inferior wall. Caused by the diaphragm (especially if elevated left hemidiaphragm or ascites).

2. Overlying visceral activity.

Problem since SPECT reconstruction leads to apparent increased count densities in areas underlying visceral activity. As polar maps are normalised to areas with highest activity, bullseye displays may be erroneously interpreted. This problem is most frequently seen after pharmacological testing, in redistribution views and following suboptimal stress. The rotating cine should be carefully viewed to identify this finding.

3. Myocardial Hot spots.

Related to increased myocardial thickness near anterior and posterior papillary muscles and exacerbated by LVH. Seen more frequently in early views. Creates problems for bullseye reconstruction normalisations. All short axis views should be carefully viewed for these artifacts.

4. Apical variants

Usually best identified in the HLA view. May occasionally be displaced laterally or rarely septally. This makes selection of the apex difficult and may consequently lead to bullseye reconstruction errors.

5. Non-coronary disease

A) LBBB - Caused by asymmetrical relaxation of the septum resulting in delayed filling during diastole. Worst if heart rate > 170/min. Typically causes reversible septal defects. ECGs should be viewed for conduction abnormalities.

B) LVH - reduced lateral/septal ratio (1.02 cf 1.17) possibly simulating lateral infarction. Reduce by asking patients for history of congenital or valvular heart disease, hypertension and inspect ECG.

C) Dextrorotation/levorotation. Ask for history and inspect all short axis slices. In these cases, 360 degree SPECT should be performed.

6. Patient motion

Phantom and patient studies have shown that movement of as little as 3.25mm may lead to artifacts and movement greater than 13mm may lead to quantitative errors. Patient motion complicates approximately 25% of clinical studies leading to visible image deterioration in 5%. Motion artifacts appear to be a greater problem with axial than lateral movement and are accentuated if movement occurs in the middle of rotation. Horizontal motion typically causes septal and lateral artifacts and vertical motion may cause anterior and inferior wall artifacts. Motion artefact should be suspected if there is blurring, loss of contrast, hot spots, irregular lumpy distribution, opposing defects and curvilinear extraventricular activity. Comparisons have recently been published in both phantom and clinical studies between different methods for motion correction. Methods described include :
  1. Viewing cine rotation - reliable for detecting axial movement
  2. Chest wall markers - May cause artifacts, Best for multidetector systems.
  3. Cross-correlation computer method - appears to be best method for detecting axial and lateral motion and time motion occurred
  4. Diverging squares approach
  5. Two dimensional fit - best for detecting distance of motion.
By immoblilising the anterior wall of the thorax,prone imaging has been shown to reduce patient motion. Multihead systems and imaging with Tc99m tracers may be more sensitive to patient motion.

7. Upward creep.

During exercise the heart is displaced interiorly against the diaphragm as the thorax expands. During recovery, the heart "creeps" back up in the thorax leading to reversible defects in the inferoseptal wall. By waiting 15 minutes after exercise, these artifacts may be reduced. Cine display or summed images should be viewed to assess for the presence of this phenomenon.

8. Oblique axis and bullseye reconstruction errors

Reconstruction around axes other than the true cardiac long axis may lead to artifacts in the basal regions. This may occur in the presence of apical infarction or an apical cleft (especially if displaced). Similarly, incorrect alignment of stress and redistribution images may lead to errors.

9. Routine SPECT QC

COR and flood field uniformity errors may lead to artifacts in cardiac studies and bullseye reconstructions, the latter typically causing ring artifacts. Improper peaking and filtering may also lead to image distortion.

To reduce artifacts and improve specificity in cardiac SPECT studies, the following areas should be addressed in all patients:

  1. History - hypertension, valvular or congenital disease, breast surgery
  2. Examination - sex, height, weight, bra size, chest wall deformity, abdominal protuberance
  3. ECG - LBBB, LVH
  4. Rotating cine - localised attenuation, motion, upward creep, viscera
  5. Slices - hot spots, apical cleft, ring artifacts, apical position.
  6. QC - COR, uniformity
Accurate attenuation correction (using line sources mounted to single or multiheaded systems), motion correction algorithms, gated SPECT, 99mTc tracers and improved resolution with multidetector systems are all promising techniques which will enhance the specificity and reduce the occurrence of artifacts in cardiac SPECT studies.

References:

*DePuey EG and Garcia EY. Optimal specificity of thallium-201 SPECT through recognition of imaging artifacts J Nucl Med 1989; 441-9

*Manglos SH, Thomas D, Gagne GM et al. Phantom study of breast tissue attenuation in myocardial imaging. J Nucl Med. 1993; 992-6

*Freidman J, Van Train K, Maddahi J et al. "Upward creep" of the heart: A frequent cause of false positive reversible defects during thallium-201 stress-redistribution SPECT. J Nucl Med, 1989; 1719-22

*Germano G, Chua T, Kavanagh, PB et al. Detection and correction of patient motion in dynamic and static myocardial SPECT using a multi-detector camera. J Nucl Med, 1993; 1349-55

*Kiat H, Van train KF, Freidman JD et al. Quantitative stress-redistribution Thallium-201 SPECT using prone imaging. Methodological Development and validation. J Nucl Med 1992; 1509-15

*Cooper JA, Neumann PB, McCandless BR Effect of patient motion on tomographic myocardial perfusion imaging. J Nucl Med 1992; 1566-71

*Roach PJ, Meikle SR, Bailey DB et al. Transmission based quantitative SPECT improves the accuracy of Thallium-201 myocardial scintigraphy. Eur J Nucl Med Aug 1993 (abst)

*Parker JA. Effect of motion on cardiac SPECT imaging. J Nucl Med, 1993; 1355-6

*Botvinick EH, Zhu YY, O'Connell WJ. A quantitative assessment of patient motion an its effect on myocardial perfusion SPECT images. J Nucl Med, 1993;303-10

________________________________________________________

J. Anthony Parker, MD PhD, jap@nucmed.bih.harvard.edu