| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Nuclear Medicine |
1 From the Departments of Medicine (Cardiology) (S.B.N., R.H.T., L.K.S., W.T.S., D.W.) and Radiology (Nuclear Medicine) (S.B.N., R.E.C.) and Duke Clinical Research Institute and Drexel School of Medicine (Cardiology) (D.J.), Duke University Medical Center, PO Box 3949, Durham, NC 27710. Received February 13, 2003; revision requested May 7; revision received October 9; accepted November 12. This study was supported in part by an unrestricted grant from Amersham Health. Address correspondence to S.B.N. (e-mail: borge001@mc.duke.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: This study included 1,818 consecutive patients who underwent a rest and stress single photon emission computed tomographic (SPECT) examination with either 99mTc sestamibi (n = 915) or 99mTc tetrofosmin (n = 903) and cardiac catheterization. A clinical index was generated and consisted of clinical and demographic variables. Information concerning death, cardiovascular death, and nonfatal myocardial infarction was 93% complete during the 1.5-year study period. Cox proportional hazards models were generated to help determine the incremental contribution of SPECT sum stress score (SSS) and the imaging agent variable to the clinical index.
RESULTS: Exercise was used for stress testing in 473 (52%) patients who received 99mTc tetrofosmin and 519 (57%) patients who received 99mTc sestamibi (P = .06). Cardiovascular death or myocardial infarction occurred in 130 patients. Resulting P values for 
2 differences between models for the end points of (a) death from any cause, (b) cardiovascular death, and (c) cardiovascular death or myocardial infarction showed that SSS combined with clinical index was a significantly better model than adjusting for only baseline characteristics (P = .001, P < .001, P = .004, respectively). Incremental addition of either 99mTc tetrofosmin or 99mTc sestamibi to those models containing SSS and the clinical index did not show further significant improvement (P = .87, P = .88, and P = .26 for death from any cause, cardiovascular death, and cardiovascular death or myocardial infarction, respectively).
CONCLUSION: This study shows that the type of clinically available 99mTc-labeled myocardial perfusion agents should not affect interpretation of results for risk stratification and prognostic assessment.
© RSNA, 2004
Index terms: Coronary vessels, diseases • Coronary vessels, radionuclide studies, 54.1217 • Coronary vessels, SPECT, 54.12162 • Radionuclides, comparative studies, 54.12162
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Coupled with cardiovascular stress testing, gated single photon emission computed tomography (SPECT) is a noninvasive imaging modality and is frequently used in the evaluation of patients suspected or known to have ischemic heart disease (9). Technetium 99m (99mTc) sestamibi (Cardiolite; Bristol Myers Squibb, North Billerica, Mass) and 99mTc tetrofosmin (Myoview; Nycomed Amersham, Princeton, NJ) are the most commonly used 99mTc agents in clinical practice, with properties that are overall quite similar (10–12). The results of previous studies have demonstrated that both agents have similar accuracy in the detection of coronary artery disease (13–15). Our institution recently reported that 99mTc tetrofosmin improved the efficiency of our nuclear cardiology laboratory by decreasing the time needed to complete each examination and the need to repeat scanning because of extra cardiac activity (16).
Studies have shown that 99mTc sestamibi provided incremental prognostic value over clinical information in the prediction of survival and nonfatal myocardial infarction (17). More recently, in a multicenter registry, Shaw et al (18) showed that a normal 99mTc tetrofosmin study in a group of patients with low to intermediate risk of coronary artery disease has a low event rate, similar to that in 99mTc sestamibi studies; however, limited information is available regarding the prognostic value of a 99mTc tetrofosmin study in a group of patients at intermediate or high risk for coronary artery disease.
On the basis of the literature, given the equivalence in tracers in the detection of coronary artery disease (14,15), we hypothesized that the imaging agents 99mTc tetrofosmin and 99mTc sestamibi would provide equivalent prognostic value after we adjusted for differences in baseline clinical characteristics. Thus, we undertook the present study to determine if there is any difference in the ability of physicians to predict prognosis with 99mTc sestamibi or 99mTc tetrofosmin in a large consecutive series of patients at high risk for coronary artery disease who underwent imaging and who also underwent coronary angiography.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Clinical Information
Clinical characteristics, which consisted of multiple descriptors from each patient’s history and physical examination, were prospectively collected by physicians and physician assistants from the nuclear cardiology laboratory for each patient at the time of cardiac catheterization and were stored in the Duke Databank for Cardiovascular Disease by personnel at the Duke Clinical Research Institute (19–22).
As previously described (8), a clinical index was generated by using a combination of the following clinical and demographic variables: age; sex; type, frequency, and electrocardiogram characterization of chest pain; variables used to assess myocardial damage and infarction history; descriptors of vascular disease; and electrocardiogram myocardial conduction data previously found to be independent predictors of outcome in patients undergoing cardiac catheterization (20).
Exercise Stress Test
Patients who were capable of exercising underwent an exercise treadmill stress test, which was the stress test of choice. Whenever possible, cardiac medications (particularly ß-blockers) were not administered during the 48 hours prior to exercise. Patients exercised according to a standard Bruce protocol with 3-minute stages, unless the physician specifically requested another protocol or believed that a modified Bruce protocol was appropriate. Blood pressure, heart rate, and a continuous 12-lead electrocardiogram were monitored throughout the stress portion of the test and into recovery, until heart rate and blood pressure returned to preprocedure baseline levels. Interpretation of a 12-lead electrocardiogram was performed independent of interpretation of perfusion images. In addition, patients’ symptoms were continually assessed and recorded.
Patients were instructed to inform the personnel administering the test when they approached their functional limit so that the 99mTc radioisotope could be injected approximately 1 minute prior to the conclusion of exercise.
Pharmacologic Stress Test
Patients who were unable to exercise or who had clinical contraindications to exercise underwent a pharmacologic stress test with one of three agents: dobutamine, dipyridamole, or adenosine. The use of different stress agents has been part of our clinical protocol rather than part of the prospective study. For all pharmacologic agents, a continuous 12-lead electrocardiogram was monitored throughout the study, and data was recorded at 1-minute intervals. Heart rate and blood pressure were monitored during the infusion of these agents at 3-minute intervals and during recovery until hemodynamic parameters returned to pretest baseline levels.
Dobutamine.—An intravenous infusion of dobutamine (Eli Lilly, Indianapolis, Ind) was initiated at a rate of 10 µg/kg/min and increased at a rate of 10 µg/kg/min every 3 minutes until a target heart rate of 85% of the age-predicted maximal heart rate was reached. The examination could also be terminated because of patient symptoms, evidence of significant ischemia, or prolonged arrhythmias, or when the dobutamine infusion rate reached a maximum rate of 40 µg/kg/min. Atropine sulfate (
1 mg) was used when the target heart rate was not achieved despite maximum infusion rate of dobutamine. The 99mTc-labeled tracer was injected when one of the previously mentioned end points was reached.
Dipyridamole.—Dipyridamole (Persantin; DuPont Pharmaceuticals, North Billerica, Mass) was administered intravenously at an infusion rate of 0.142 mg/kg/min for 4 minutes. The 99mTc-labeled tracer was administered 2 minutes after completion of the dipyridamole infusion. Intravenous aminophylline (75–250 mg) was used to reverse dipyridamole-induced side effects. No caffeine intake was permitted within 12 hours of the performance of the stress test.
Adenosine.—Adenosine (Adenoscan; Fujisawa Healthcare, Deerfield, Ill) was administered at a rate of 0.142 mg/kg/min for 6 minutes. The 99mTc-labeled tracer was injected 3 minutes after adenosine infusion was started. No caffeine intake was permitted within 12 hours of the performance of the stress test.
Myocardial Perfusion SPECT Protocol
The protocol for performing SPECT myocardial perfusion imaging studies in our institution has been described previously (16). We routinely used more than one camera system in our laboratory, and patients underwent imaging according to camera system availability. The 12-segment model was used because our database supported only 12 segments at the outset of the data collection period. In summary, SPECT data were obtained with two different systems that are used clinically, both of which use a step-and-shoot protocol. The rest images were obtained for 30 seconds per projection, and the stress images were obtained for 20 seconds per projection. The first camera system used 90 projections during a 360° rotation (4° per stop) with a three-headed gamma camera with low-energy high-resolution parallel hole collimators, and a hamming filter (cutoff, 0.6 cycles per centimeter) (Triad; Trionix, Twinsburg, Ohio). The second camera system employed 60 projections during a 180° rotation (3° per stop) by using a fixed 90° two-headed gamma camera with low energy, high resolution, parallel-hole collimators, and a Butterworth filter (cutoff, 0.35 cycles per pixel) and a power of 5.0 (ElScint, Haifa, Israel). Images were reconstructed with filtered backprojection and no attenuation correction.
Thallium 201 (201Tl) was the agent of choice for obtaining rest images as part of our clinical dual isotope protocol. In patients who weighed more than 280 pounds (127 kg), however, either 99mTc tetrofosmin or 99mTc sestamibi was used to obtain rest images. All stress images were obtained with 99mTc tetrofosmin or 99mTc sestamibi. Images were obtained by using a rest and stress same-day protocol, except in very obese patients; in these patients, images were obtained by using the same 99mTc-based radiopharmaceutical and a 2-day protocol. Rest studies were performed with injection of 3 mCi for 201Tl-labeled agents or 10–12 mCi for 99mTc-labeled agents, with SPECT performed 30 minutes later (60 minutes later for 99mTc sestamibi). Exercise or pharmacologic stress tests were performed with injection of 21–36 mCi, and SPECT started 20 minutes later for exercise stress tests and 30 minutes later for pharmacologic stress tests. In patients injected with 99mTc sestamibi during a pharmacologic stress test, a 60-minute waiting period was allowed prior to imaging.
Image Interpretation and Candidate Nuclear Variables
Images were independently interpreted and clinically reviewed by either of two experienced (>15 years) nuclear medicine and nuclear cardiology physicians (S.B.N., R.E.C.). A 12-segment reporting system, which is illustrated in Figure 1, was used to quantify perfusion to various vascular territories and is similar to methods previously described with a 20-segment model (5,6). The 12-segment model was used because our database supported only 12 segments at the outset of the data collection period. The segments are illustrated in Figure 1. The relative perfusion to each segment was also quantified with four grades of perfusion defect, with each assigned a numeric value as follows: 0, no defect; 1, mild defect; 2, moderate defect; and 3, severe defect. A cumulative SPECT sum stress score (SSS) was obtained by summing the score for each of the 12 segments. Thus, a normal study would yield a SSS of 0, while the maximum score possible would be 36 (severe perfusion defect in all 12 segments). Similarly, a SPECT sum rest score (SRS) and sum difference score (SDS), which was the change from stress to rest, were derived. The score variables have been shown to be highly predictive of cardiovascular outcome with a 20-segment model (5,6,17).
|
Follow-up Data
Patients or their relatives were contacted prospectively by research personnel, either with a mailed questionnaire or telephone interview at 6 months and 1 year and at yearly intervals thereafter after cardiac catheterization. The questionnaires and interviews were used to request follow-up information concerning death, rehospitalization, and nonfatal myocardial infarction. Follow-up was 93% complete during the study period. An independent clinical events review committee that did not have knowledge of the patients’ clinical, catheterization, stress test, or perfusion data evaluated the cause of death in patients who died and the cause of nonfatal myocardial infarction in patients who survived. Data acquisition and follow-up techniques have been described previously (20,22).
Statistics
Baseline characteristics of the study patients are presented as percentages for discrete variables and as median, 25th, and 75th percentiles for continuous variables. Pearson 2 tests were used to test for differences in discrete variables for the two imaging agents.
The Wilcoxon rank-sum test was used to determine if there were significant differences between the distributions of continuous variables for the two imaging agents. Linear regression analysis was used to evaluate differences between the two imaging agents in their relationship to SSS, which was the dependent variable. A P value of less than .05 indicated a statistically significant difference between the two imaging agents for all test statistics.
Unadjusted, event-free survival curves stratified by the type of imaging agent used were generated by using Kaplan-Meier survival estimates for the three end points: death from any cause, cardiovascular death, and cardiovascular death or nonfatal myocardial infarction. The follow-up time for each estimate was 1.5 years. The log-rank 2 test was used to evaluate differences between the survival curves.
Cox proportional hazards models were constructed to assess the relationship of baseline clinical characteristics, SSS, and the imaging agent with each of the three outcomes. The model improvement after adding SSS and the tracer variable to the clinical index was evaluated. Similar Cox modeling schemes were used to evaluate the model improvement after incremental addition to the clinical index of SRS and SDS separately, followed by the imaging agent variable.
To determine whether the relationship between SSS and each of the outcomes was different for the two imaging agents, Cox models for each of the separate end points—both unadjusted and adjusted for differences in baseline clinical risk—were generated to test the significance of an interaction term, namely SSS according to the imaging agent variable. In a similar fashion, interactions between the imaging agent variable and SRS and SDS, respectively, were tested with adjusted and unadjusted Cox models. A P value of less than .05 for the interaction term in the model suggests that the imaging agent affects the relationship between SSS and the end point.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
Follow-up Endpoints and Outcome Events
During follow-up, there were 49 deaths in the 99mTc tetrofosmin group and 61 in the 99mTc sestamibi group. Of these deaths, 68 (62%) were classified as cardiovascular. Cardiovascular death or myocardial infarction occurred in 130 patients. The median follow-up time for living patients was 1.5 years (25th percentile, 1.1 years; 75th percentile, 1.5 years) for the patients who received 99mTc tetrofosmin and 1.5 years (25th percentile, 1.5 years; 75th percentile, 1.5 years) for the patients who received 99mTc sestamibi.
The unadjusted overall mortality rate after 1.5 years of follow-up was 7.1%. Mortality rates of 6.5% and 7.5% were observed at 1.5-year follow-up in patients who received 99mTc tetrofosmin and 99mTc sestamibi, respectively, with no significant difference in mortality between the groups over the study period (P = .53, Fig 2). Similarly, there was no significant difference between 99mTc tetrofosmin and 99mTc sestamibi regarding cardiovascular death (4.4% vs 4.6%, respectively, P = .74) and the composite end point of cardiovascular death or myocardial infarction (8.9% vs 7.8%, respectively, P = .48).
|
2 test when SSS was added to the baseline models were significant for death from any cause (P = .001), cardiovascular death (P < .001), and cardiovascular death or myocardial infarction (P = .004), which confirms the prognostic value of SSS in our study population. Inclusion of the imaging agent (99mTc tetrofosmin vs 99mTc sestamibi) in addition to the SSS and clinical index in each of these models did not, however, provide any additional statistically significant information (P = .87, .88, and .26, respectively). P values for the interaction term for SSS with imaging agent interaction in each of these models were not significant, which indicates that the imaging agent does not alter the effect of SSS for death, cardiovascular death, or the composite end point of cardiovascular death or myocardial infarction (P = .43, .51, and .55, respectively). When resting perfusion results as defined by SPECT SRS (reflecting infarction areas), calculated by using the same gradations as the SSS, and changes from stress to rest as defined by the SPECT SDS (reflecting ischemic areas) was analyzed, similar results were obtained (Figs 4, 5).
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
These results show that SSS can be used to predict survival and to provide incremental prognostic information above and beyond clinical data, which confirms the conclusions drawn from the results of previous studies by using nuclear stress test results. Hachamovitch et al (5) used a dual-isotope protocol to demonstrate that SSS provided significant prognostic information when added to clinical data and exercise data in 2,113 patients (P < .001). In a follow-up study of 5,183 patients that included patients with pharmacologic stress and exercise stress, Hachamovitch et al (6) again confirmed their initial findings that SSS provided significant prognostic information beyond clinical variables (P < .001). Vanzetto et al (7) also found that myocardial perfusion imaging with 201Tl provided significant prognostic information in addition to that provided by clinical variables and exercise stress tests (P < .01). The findings of Vanzetto et al (7) included 6 years of follow-up compared with the shorter 1.5 and 1.75 years of follow-up in the studies of Hachamovitch et al (5,6). In our more recent cohort of 3,287 patients at high risk for coronary artery disease (S.B.N., unpublished data, 2004), SSS from myocardial perfusion studies provided substantial prognostic information beyond both clinical data and anatomic descriptions from cardiac catheterization.
Potential Differences and Clinical Impact of Available Myocardial Perfusion Agents
Despite some potential differences between 99mTc tetrofosmin and 99mTc sestamibi in myocardial extraction at high coronary blood flow rates, previous results have shown equivalent sensitivity and specificity in the identification of patients with coronary artery disease, including defect detection and reversibility of defects (14,15). In a recently published report by Soman et al (23), investigators found a substantial difference between 99mTc sestamibi and 99mTc tetrofosmin in the identification of reversible defects in patients with mild to moderate coronary artery disease during pharmacologic stress testing. Several limitations, such as the presence of disease in multiple vessels, lack of specific correlation with vascular territory abnormalities, small sample size, absence of quantitative coronary angiography, wide range of stenosis severity, and no evaluation of patients with normal coronary arteries for specificity, however, preclude definitive conclusions from being drawn on the basis of results of that study. Furthermore, no differences in treatment allocation or outcomes were reported.
Prognostic Value of 99mTc Tetrofosmin versus 99mTc Sestamibi
In the current study, which comes from a single institution, the prognostic value of the two clinically available 99mTc agents was directly compared in a large cohort of patients at high risk for coronary artery disease. Previously reported results in a low-risk cohort of patients have suggested effective risk stratification with 99mTc tetrofosmin (24,25). Because 99mTc tetrofosmin was approved for clinical use more recently than 99mTc sestamibi, the limited availability of prognostic information for this perfusion agent is not surprising.
In a recent large multicenter registry, Shaw et al (18) evaluated the prognostic value of a normal 99mTc-tetrofosmin SPECT study in 4,728 patients at intermediate or low risk for coronary artery disease. In contrast to our present findings, which include 50% of patients with pharmacologic stress and higher risk SSS, only one-third of the patients underwent a stress test with adenosine in the multicenter registry. The authors observed an annualized event rate of 0.6% with metaanalysis. By comparing previously reported outcome data for normal myocardial perfusion studies by using 201T1 and 99mTc sestamibi with the results of Shaw et al, the overall survival rates were again very similar and ranged from 99.3% to 99.7%. Regardless of which tracer is used in low-risk populations, similar annualized event rates of less than 1% have been observed.
Limitations
This study had some unavoidable limitations. We used more than one camera system and a 12-segment model with a four-grade scoring system to describe the extent and severity of total perfusion abnormalities. We routinely used more than one camera system in our laboratory, and patients underwent imaging according to camera system availability. The 12-segment model was used because our database supported only 12 segments at the outset of the data collection period. Currently, the American Society of Nuclear Cardiology recommends a 17-segment model (26) and a five-step scoring system for grading the severity of any perfusion defect. While the recommended model would provide greater ability to define regions of defect within the myocardium and may provide a higher resolution of SSS, the 12-segment scoring systems were powerful, and the use of a 17-segment model would likely improve prognostic abilities with SPECT.
Since patients included in this study underwent both a nuclear stress test and cardiac catheterization, a selection bias for patients with a higher-risk profile is inherent and evidenced by an event rate that was higher than that seen in earlier studies. The higher event rate did, however, increase our power to identify differences between independent variables and strengthens our finding that no difference is documented between 99mTc tetrofosmin and 99mTc sestamibi myocardial perfusion tracers in the prediction of hard events outcome.
Despite previous studies that have identified poststress ejection fraction as a predictor of outcome, a gated SPECT ejection fraction was not incorporated into this analysis (27). At the time of data collection, gated SPECT was not a standard component of the diagnostic study; thus, it was not included in the database. Given previous results, it is likely that gated SPECT ejection fraction could provide further incremental prognostic value beyond the clinical variables and perhaps beyond the perfusion score, particularly for the prediction of death and cardiovascular death (28). The purpose of this study, however, was to compare tracers and their effect on estimating outcomes on the basis of perfusion abnormalities.
Finally, the relationship between mortality and perfusion scores may have been minimized by means of revascularization procedures being preferentially performed in patients with positive stress tests. The finding that perfusion scores (SSS, SRS, and SDS) are substantial predictors of outcome in the face of revascularization only supports the use of these scores in predicting prognosis, even in patients who are undergoing revascularization. Furthermore, the similar percentage of revascularization between imaging agent cohorts suggests that treatment is independent of tracer type.
In conclusion, for the physician who refers patients for nuclear cardiology testing, the results of SPECT myocardial perfusion imaging provide important information that should influence the clinician in the decision-making process regarding appropriate therapy options, as well as help providers and patients understand the risk for future clinical events. Along with previously published reports and multicenter registry trial results, the findings of this study should reassure clinicians that the type of clinically available 99mTc-labeled myocardial perfusion agents used in nuclear cardiology examinations should not change the interpretation of the results for risk stratification and prognostic assessment. Furthermore, the development of clinical guidelines that address 99mTc-labeled myocardial perfusion imaging studies should be tracer independent.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, S.B.N.; study concepts, S.B.N., D.W., W.T.S., D.J.; study design, S.B.N., R.H.T., L.K.S., D.W.; literature research, D.W., W.T.S.; clinical and experimental studies, S.B.N., R.E.C., D.W.; data acquisition, S.B.N., D.W., L.K.S., R.H.T., W.T.S.; data analysis/interpretation, S.B.N., D.W., L.K.S., R.H.T., D.J.; statistical analysis, L.K.S., R.H.T.; manuscript preparation, S.B.N., R.H.T., L.K.S., D.W., W.T.S., R.E.C.; manuscript definition of intellectual content, editing, revision/review, and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. P. Piccini, J. R. Horton, L. K. Shaw, S. M. Al-Khatib, K. L. Lee, A. E. Iskandrian, and S. Borges-Neto Single-Photon Emission Computed Tomography Myocardial Perfusion Defects Are Associated With an Increased Risk of All-Cause Death, Cardiovascular Death, and Sudden Cardiac Death Circ Cardiovasc Imaging, November 1, 2008; 1(3): 180 - 188. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M Hawkins, A F Scarsbrook, S Pavlitchouk, N R Moore, and K M Bradley Detection of an occult thymoma on 99Tcm-Tetrofosmin myocardial scintigraphy Br. J. Radiol., April 1, 2007; 80(952): e72 - e74. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
