|
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
|
|
|
|||||||||
|
|||||||||||
|
J Am Coll Cardiol, 2002; 39:999-1004 © 2002 by the American College of Cardiology Foundation |
* Rhode Island Hospital and Brown University, Providence, Rhode Island, USA
Manuscript received July 23, 2001; revised manuscript received December 18, 2001, accepted December 19, 2001.
* Reprint requests and correspondence: Dr. Lynne L. Johnson, Rhode Island Hospital, Main Building, Room 208, 593 Eddy Street, Providence, Rhode Island 02903, USA.
Lynne_Johnson{at}brown.edu
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
BACKGROUND: The need for performing rest in addition to stress perfusion imaging to confirm normalcy is due largely to defects created on the stress images by attenuation artifacts. A low cardiac event rate among patients undergoing stress-only imaging with attenuation correction (AC) would validate the safety of stress-only imaging.
METHODS: Patients with low to medium pretest probability for coronary artery disease (CAD) referred for chest pain evaluation from July 1, 1997 to July 1, 1999, were scheduled for a two-day stress/rest tomographic (single photon emission computerized tomography, or SPECT) perfusion imaging study with Tc-99m sestamibi. Patients were imaged on a variable angle camera with AC using Gd-153 scanning line sources (Vantage ADAC, Milpitas, California). If the stress scan was normal without AC or corrected with AC, the patient did not return for rest scan and was followed.
RESULTS: Seven hundred twenty-nine patients underwent stress-only imaging, and follow-up was obtained on 652 (89%) of those patients. There were 224 males and 428 females with mean age of 52 ± 13 years. Mean follow-up was 22.3 ± 6.4 months. The mean pretest probability was 37 ± 24%. The non-AC images showed breast and/or diaphragmatic attenuation artifacts severe enough to have required the patient to return for rest imaging in 37% of patients, and all corrected completely with AC. During follow-up, there were two noncardiac deaths and no cardiac deaths. There was one myocardial infarction; three patients with progressive unstable angina underwent diagnostic coronary angiography showing significant CAD. The overall cardiac event rate was 0.6%.
CONCLUSIONS: These results support stress-only imaging in patients with low to medium probability for CAD as a safe, time- and cost-efficient imaging modality.
| |||||||||||||||
Recognizing typical locations of attenuation and comparing rest with stress images are methods used to help distinguish artifacts from real defects. These strategies are suboptimal because of variability in position of artifacts from one session to another (5–8). Electrocardiogram (ECG) gated tomographic acquisition with cine display and software providing global and regional left ventricular (LV) function became the standard for differentiating artifact from scar (9). Normal wall motion in walls with mild fixed perfusion defects and overlying soft tissue seen on raw projection images supports soft tissue as source of defect.
A more direct approach to the problem of soft tissue artifacts is attenuation correction (AC). Several approaches have been pursued using different detector configurations and transmission sources (5). A configuration now used by several vendors and employed in this study is a dual detector camera with moving bar sources (10). The purpose of this study was to evaluate the safety of stress-only Tc-99m sestamibi SPECT imaging, employing AC in patients with low to medium probability for CAD referred for chest pain evaluation. If a normal stress image without rest confers an excellent prognosis and the addition of AC helps confirm normalcy for stress images, fewer rest images would be necessary.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Stress test and imaging protocol. Patients either underwent a symptom-limited treadmill stress test using the Bruce protocol or received intravenous dipyridamole if unable to exercise. Imaging for all patients was performed on a dual head variable angle camera with AC using gadolium-153 scanning line sources mounted opposite each detector (Vertex camera, Vantage ADAC, Milpitas, California). The bars move once across the field of view during each step of the tomographic acquisition. The difference in energies between gadolinium (predominant energy 100 keV gamma and 42 to 47 keV X-rays) and technetium-99m (140 keV) allow separation. Further corrections are made for scatter and resolution compensation. High-resolution collimators were used, and a 20% window was set on the 140 keV photopeak of technetium-99m. Imaging was performed over 64 azimuths in step mode using a 64 x 64 x 8 matrix. The transmission mode was continuous and recommended information density was used for azimuth collection time. The transmission scan time was equal to time per azimuth minus one second. For the latest update of this software a QC display checks the attenuation maps for banding, truncation and count deficiencies. Gating was performed in 62% of studies.
Data interpretation. Electrocardiographic evidence of ischemia was defined as at least 0.1 mV of horizontal or downsloping ST-segment depression. The electrocardiographic response was classified as indeterminate in the presence of digoxin, left bundle branch block, significant baseline ST-segment abnormalities or paced rhythm. Patients who had strongly positive ECG treadmill tests (>0.2 mV depression in the absence of left ventricular hypertrophy or baseline abnormalities) were brought back for rest scans, regardless of the results of the stress scans. Scans were interpreted by consensus of two observers. The rotating raw projection images were interpreted for presence and location of attenuation artifacts. Breast attenuation was identified as photopenic shadows overlying either the anteroseptum, anterior wall or anterolateral wall of the left ventricle. Diaphragmatic attenuation was identified as photopenic shadows representing elevated fundus of the stomach overlying inferior surface of the left ventricle in a lateral view. Presence and location of these artifacts were correlated with patient weight, age and gender. The non-AC and AC images were displayed together in row tomographic format. The presence of significant soft tissue attenuation was defined as the following. An inferior wall defect or anterior wall defect >20% reduction from peak counts (by polar map and/or color table) seen on at least two sets of slices comprising more than two slices in each set was identified as potentially due to soft tissue, if the raw projection images confirmed corresponding visible soft tissue planes. This degree of soft tissue artifact was considered severe enough to require the patient to return for rest imaging if only these images were available. If the Vantage corrected for the defect(s), the scan was read as normal but the soft tissue limitations noted. If the non-AC images showed homogeneous tracer uptake in all sets of reconstructed slices, then the scans were interpreted as normal and no soft tissue limitations noted. If the scan showed subdiaphragmatic activity or incomplete correction for attenuation artifact, reduced left ventricular ejection fraction (<50%) on gated study or significant problems with the attenuation maps, the patient underwent rest imaging and was not included in this study.
Patient follow-up. Follow-up letters were sent to all patients near the first anniversary of their test date. Telephone interviews were performed and hospital charts reviewed and primary physicians contacted for confirmation. Outcome events noted were hard end points of cardiac death or recurrent nonfatal acute myocardial infarction and soft end points of hospitalization for chest pain, cardiac catheterization for chest pain or revascularization. Events were confirmed by chart review.
Statistical analysis. Kaplan-Meier cumulative event-free survival curves were generated from the outcome data. The statistical software used was Strata Statistics/Data Analysis Version 6.0 (Strata Corporation, College Station, Texas).
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Exercise testing. Six hundred nine patients performed a symptom-limited exercise treadmill test using a standard Bruce protocol with 12-lead ECG recording. Forty-three patients were unable to exercise or had LBBB and underwent pharmacological stress with dipyridamole. The mean treadmill time was 6 min 42 s. The mean peak heart rate achieved was 145 ± 20 beats/min. Of 609 patients, 469 (77%) had no ST changes and achieved >85% age-predicted heart rate. Sixty-seven of 609 (11%) achieved <85% age-predicted heart rate without ECG changes. Sixteen patients (2%) had ST depression. They were all women, and the ECG findings were thought most likely to represent a false positive response. Forty-seven (7%) had borderline ECG changes, seven had uninterpretable ECG and three had nondiagnostic ECG responses.
Scan results. The non-AC stress perfusion SPECT images showed significant breast or diaphragmatic attenuation artifacts severe enough to require rest imaging to confirm normalcy in 243 (37%) patients (Figs. 1 and 2). Of these defects, 68% were inferior (diaphragm), and 32% were breast (anterior, apical or anterolateral).
|
|
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
Relatively low specificity is a recognized limitation to stress perfusion imaging. The major factor contributing to this reduced specificity is complex attenuation patterns in the thorax, leading to variable attenuation of photons coming from the target organ (heart) (5–8). Breast tissue in women can attenuate photons over the anterior, anterolateral or anteroseptal walls of the left ventricle. The degree of attenuation is greater in younger women with denser breast tissue (11). Breasts can shift between rest and stress imaging, further confounding interpretation. The location of an anterior defect may correspond well to the observed soft tissue plane, but the defect may still hide a true perfusion defect. In men and in some women, the diaphragm rises in the supine position to impinge upon the inferior wall of the left ventricle, decreasing the accuracy of identifying disease of the right coronary artery. Early reports of sensitivity and specificity values from multicenter myocardial perfusion SPECT studies with thallium revealed low specificity values (1,2). Corrected for referral bias, the normalcy rates were higher but continued to be less than sensitivity values, especially for women (3). The higher energy of the technetium labeled perfusion agents will theoretically reduce scatter and attenuation when compared with thallium-201. The improvement in normalcy rates using technetium-99m sestamibi studies has been shown (4). The addition of gating has improved the specificity/normalcy in women (9). The finding of normal anterior wall motion confirms normalcy in the presence of a fixed anterior perfusion defect. This technique to confirm normalcy requires two sets of images: rest and stress. Prone imaging helps prevent the diaphragm from obstructing the heart and improves the specificity in men (12).
Attenuation correction. The approach to AC used in commercially available systems is based on creation of patient-specific attenuation maps using external radiation sources. The detectors simultaneously record emission photons from the injected radiopharmaceutical dose in the patient and transmission photons from a sealed solid source. The key to successful hardware/software design is producing high quality attenuation maps (13). Some of the limitations of early systems were due to poor quality attenuation maps, and many of these have now been overcome. Ficaro et al. (14), using a system with a different configuration than used in the present study, reported an increase in normalcy rates from the low 80% range into the high 90% range in low-likelihood patients. Kluge et al. (15) studied the effect of AC on diaphragmatic attenuation and RCA disease in a small population of males. They found that AC improved normalcy rates to the same extent that Ficaro found. They also found that AC scans detected 100% of catheterization-documented RCA lesions. An early version of the Vantage software was used in a multicenter study and reported by Hendel (16). These investigators found improved specificity in the RCA territory (inferior wall) but reduced sensitivity to detect RCA lesions. The reduced sensitivity was attributed to poor transmission images. Sources of error included low counts, truncation and using filtered back projection reconstruction algorithm. The current version of the software (Vantage Pro, ExSPECT II, ADAC) has been updated to improve the quality of the transmission images to reduce artifacts in the transmission reconstructions and improve reliability of the method. The improved transmission scan reconstruction algorithm includes The Bayesian Iterative Transmission Gradient Algorithm (BITGA) (17,18).
Experience using AC and interpreting scans has also shown that subdiaphragmatic activity can either throw extraneous counts into the inferior wall and/or down scale the anterior wall (19). Because Tc-99m sestamibi is excreted via the enterohepatic route, subdiaphragmatic activity not infrequently is present, especially after pharmacologic stress. This limitation was recognized in the present study, and scans with bright spots of activity immediately adjacent to the heart were not read as normal if an inferior defect appeared to be "corrected." These patients returned for rest scanning and the non-AC images were interpreted.
Prognostic value of normal scan. The good prognostic value of a normal perfusion scan has been documented by several groups of investigators (20–23). Brown et al. (21) reported on the prognostic value of a normal thallium-201 scan, and the Cedars Sinai laboratory reported on the prognostic value of a normal Tc-99m sestamibi scan. This latter group recently reported the incremental prognostic value of dual isotope (rest thallium, stress Tc-99m sestamibi) perfusion scans in patients with low, intermediate and high post-exercise treadmill test likelihood of CAD (20). The patients included in the present study would correspond to the low and intermediate post-exercise treadmill likelihood groups reported in the aforementioned study. There were two hard events in the 440 patients with intermediate post-exercise tolerance test likelihood and three events in our group of 652 patients yielding very similar event rates.
The need to complete imaging in as short a period of time as possible to reduce length of stay for inpatients and reduce patient waiting time and improve laboratory throughput has led to increased use of a dual isotope protocol, which requires rest imaging to be performed first. The results of the present study show that normal stress perfusion images when read alone, without rest images, are safe and confer the same good prognosis as normal rest/stress scans. The addition of AC increases the number of stress scans that can be interpreted as normal.
Study limitations. The follow-up was 89%. We cannot exclude the possibility that the event rate for those patients lost to follow-up was higher than for those followed. We excluded patients with known CAD and those with markedly positive ECG changes by design. We wanted to be conservative in this first study evaluating the safety of stress-only imaging. Based on these results, further evaluation of this protocol in larger patient groups may be indicated.
Conclusions. Patients with low to medium probability for CAD referred for chest pain evaluation underwent stress-only SPECT Tc-99m sestamibi imaging with AC. If the non-AC stress images were interpreted as normal or soft tissue artifacts (present in 37%) corrected completed with attenuation correction then the patient did not undergo rest imaging and was followed. During follow-up there were no cardiac deaths, one myocardial infarction and three patients with progressive unstable angina among the total cohort of 652, giving an overall cardiac event rate of 0.6%. These results support stress-only imaging in patients with low to medium probability for CAD as a safe, time- and cost-efficient imaging modality.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. D. Metz, M. Beattie, R. Hom, R. F. Redberg, D. Grady, and K. E. Fleischmann The Prognostic Value of Normal Exercise Myocardial Perfusion Imaging and Exercise Echocardiography: A Meta-Analysis J. Am. Coll. Cardiol., January 16, 2007; 49(2): 227 - 237. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. W. Hayes, A. De Lorenzo, R. Hachamovitch, S. C. Dhar, P. Hsu, I. Cohen, J. D. Friedman, X. Kang, and D. S. Berman Prognostic Implications of Combined Prone and Supine Acquisitions in Patients with Equivocal or Abnormal Supine Myocardial Perfusion SPECT J. Nucl. Med., October 1, 2003; 44(10): 1633 - 1640. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | SUBSCRIPTIONS | CURRENT ISSUE | PAST ISSUES | CARDIOSOURCE | SEARCH | HELP | FEEDBACK |
