Author + information
- Received May 24, 1996
- Revision received September 11, 1996
- Accepted September 23, 1996
- Published online January 1, 1997.
- Raymond Taillefer, MDA,*,
- E.Gordon DePuey, MDB,
- James E Udelson, MD, FACCC,
- George A Beller, MD, FACCD,
- Yves Latour, MDA and
- François Reeves, MDA
- ↵*Dr. Raymond Taillefer, Department of Nuclear Medicine, Hotel-Dieu de Montreal, 3840 St-Urbain, Montreal, H2W 1T8, Canada.
Objectives. This prospective study was conducted in 115 women to directly compare the sensitivity and specificity of thallium-201 (Tl-201), technetium-99m (Tc-99m) sestamibi perfusion and Tc-99m sestamibi electrocardiographic (ECG)-gated single-photon emission computed tomographic (SPECT) studies for detection of coronary artery disease (CAD).
Background. Data on the comparative diagnostic accuracy of Tl-201 and Tc-99m sestamibi perfusion imaging for the detection of CAD, specifically in women, are very limited.
Methods. Eighty-five patients with suspected CAD, scheduled for coronary angiography, and 30 volunteers with a pretest likelihood of ≤5% for CAD were evaluated. Within 1 week, each patient underwent Tl-201 and Tc-99m sestamibi SPECT imaging procedures (both perfusion and gated SPECT imaging). Treadmill stress testing was used in 78 patients and dipyridamole in the remaining 37 patients. All images were interpreted by three observers in a blinded manner (consensus reading). Technetium-99m sestamibi SPECT studies were read without and then with ECG gating. Technetium-99m sestamibi gated SPECT studies were used to differentiate scar tissue from soft tissue attenuation artifact.
Results. The overall sensitivities for detecting ≥50% and ≥70% stenoses were 75.0% and 84.3%, respectively, for Tl-201, and 71.9% and 80.4%, respectively, for Tc-99m sestamibi perfusion studies (p = 0.48). The specificity for lesions ≥50% was 61.9% for Tl-201 and 85.7% for Tc-99m sestamibi perfusion (p = 0.07), whereas for lesions ≥70% it was 58.8% for Tl-201 and 82.4% for Tc-99m sestamibi perfusion (p = 0.01). When the 34 patients with a normal coronary angiogram were added to the group of 30 normal volunteers, the “specificity” for lesions ≥70% was 67.2% for Tl-201, 84.4% for Tc-99m sestamibi SPECT perfusion (p = 0.02) and 92.2% for Tc-99m sestamibi gated SPECT (p = 0.0004).
Conclusions. Both Tl-201 SPECT and Tc-99m sestamibi SPECT perfusion studies had a similar sensitivity for the detection of CAD in women. However, Tc-99m sestamibi SPECT perfusion imaging shows a significantly better specificity, which is further enhanced by the use of ECG gating.
(J Am Coll Cardiol 1997;29:69–77)>
Cardiovascular disease is a major health problem in women, causing death in over 500,000 women each year in the United States. Coronary artery disease (CAD) is responsible for over half of all coronary deaths (). Although the rate of heart disease is declining in North America, it still remains the principal cause of death among women, outranking stroke, lung cancer and even breast cancer (). Despite these statistics, CAD continues to be underdiagnosed and undertreated in women compared with men ().
Considering the significant prevalence of CAD in women, the use of adequate noninvasive diagnostic and prognostic evaluation procedures becomes very important. Given the high false positive rate of the stress electrocardiogram (ECG) in women ([3–7]), a more accurate noninvasive diagnostic procedure adapted to this specific patient population is essential. Several studies, performed predominantly in men, have shown that the sensitivity and specificity of radionuclide myocardial perfusion imaging during treadmill exercise or after pharmacologic vasodilation is superior to treadmill exercise ECG testing alone ([8–11]). However, surprisingly, few specific data exist on the clinical value of exercise or pharmacologic radionuclide myocardial perfusion scintigraphy in the noninvasive diagnosis of CAD in women. Furthermore, most of the studies involve a relatively small number of patients using planar qualitative thallium-201 (Tl-201) imaging (). This type of imaging procedure has also been associated with a variable incidence of breast soft tissue attenuation artifacts caused by both attenuation of the 73 to 80-keV photons of Tl-201 and the low energy scatter (). The use of a higher energy radionuclide may substantially reduce the problem of soft tissue attenuation artifacts on myocardial perfusion studies.
The relatively recent introduction of technetium-99m (Tc-99m) sestamibi may potentially improve the diagnostic accuracy of myocardial perfusion imaging in women, considering the physical properties of Tc-99m and also the possibility of performing ECG-gated single-photon emission computed tomography (SPECT). The simultaneous acquisition of Tc-99m sestamibi perfusion and function (gated SPECT) studies has also been helpful in determining the cause of a fixed myocardial perfusion defect by differentiating a scar tissue from a soft tissue attenuation artifact (). At present, scant data are available dealing with the direct comparison between Tl-201 and Tc-99m sestamibi imaging, specifically in women and on the supplementary value of gated SPECT Tc-99m sestamibi studies in this patient population. Accordingly, the purpose of this study was to compare the sensitivity and the specificity of Tl-201 and Tc-99m sestamibi imaging in the detection of CAD in women and to determine if gated SPECT Tc-99m sestamibi studies can improve the diagnostic accuracy of radionuclide myocardial perfusion in women using this imaging agent.
1.1 Patient group.
The patients for this prospective study consisted of a group of 85 consecutive women with known or suspected CAD scheduled for coronary angiography. A second group of 30 women with a ≤5% likelihood of CAD, based on probability analysis of CAD (), were also studied. Patients with the following inclusion criteria were enrolled: 1) All patients had to be scheduled for coronary angiography or had to have a previous coronary angiogram within 2 months of their participation in the study, providing that no change in cardiac status occurred between the times of the imaging procedures and coronary angiography. 2) The referring physician agreed with study participation. The stress imaging procedure that led to the decision to proceed with coronary angiography was not included in this study; the Tl-201 and Tc-99m sestamibi studies were performed for this study after the angiographic decision had been made. 3) Patients agreed to participate and signed an informed consent approved by the Institutional Review Board of Hôtel-Dieu de Montréal. Patients were excluded from the study if they had undergone previous coronary artery bypass grafting or percutaneous transluminal coronary angioplasty, had unstable angina, a recent (<6 months) myocardial infarction, symptomatic valvular heart disease, congestive heart failure or other contraindications of having either a treadmill stress test or pharmacologic vasodilation with dipyridamole. The 30 normal volunteers with a ≤5% likelihood of CAD were chosen according to clinical criteria and a negative ECG treadmill stress test. This group had myocardial perfusion scintigraphy using treadmill as the stress modality. These 30 subjects also signed informed consent statements approved by the same Institutional Review Board.
1.2 Study design.
Once patients and volunteers had given their consent and were enrolled in the study, they were submitted to two SPECT studies within 1 week: one with Tl-201 and one with Tc-99m sestamibi using a 2-day protocol. The order of the tests was randomly assigned. The interval between the Tl-201 and one of the Tc-99m sestamibi studies was at least 2 days to avoid any contamination of the Tl-201 study from a previous Tc-99m sestamibi study. All studies were performed at the Department of Nuclear Medicine of Hôtel-Dieu de Montréal. The choice of the stress modality for a given patient—treadmill stress test or pharmacologic vasodilation with dipyridamole—was determined by agreement between both the referring cardiologist and the nuclear medicine physician.
1.3 Exercise testing.
1.3.1 Treadmill stress test.
Patient preparation was identical for both Tl-201 and Tc-99m sestamibi studies. Whenever possible, beta-blockers, calcium channel antagonists and nitrates were discontinued for 24 to 48 h before the treadmill stress test. Patients fasted for at least 8 h before testing. All patients underwent symptom-limited treadmill exercise testing using the Bruce protocol. A dose of 3.0 to 3.3 mCi of Tl-201 or a dose of 25 to 30 mCi of Tc-99m sestamibi was injected intravenously at peak exercise, and the patients exercised for an additional 1 to 2 min. The test was terminated when there was ischemic ST segment depression ≥2 mm, significant arrhythmia, moderate or severe angina pectoris, significant hypotension, excessive fatigue or shortness of breath or achievement of 100% of the maximal predicted heart rate. Special care was given to obtain a similar level of exercise for both Tl-201 and Tc-99m sestamibi studies.
1.3.2 Pharmacologic vasodilation with dipyridamole.
In patients who were unable to perform or achieve an adequate level of exercise (i.e., <85% of the maximal predicted heart rate), a dipyridamole infusion was used instead of the treadmill stress test. Patients were instructed to fast after midnight and were told specifically to avoid methylxanthine-containing medications and alimentary products such as tea, coffee, chocolate or soft-drinks. With the patient in the supine position, baseline heart rate and blood pressure were recorded. An intravenous line with normal saline solution was inserted with a 20-gauge cannula into an antecubital vein. Dipyridamole was infused at a rate of 0.142 mg/kg body weight per min over 4 min, and the vital signs and ECG were recorded each minute. After the infusion, the patient stood up and walked in place for 2 min. At that point, a dose of Tl-201 (3.0 to 3.3 mCi) or Tc-99m sestamibi (25 to 30 mCi) was injected as a compact bolus. The patient continued to walk in place for another 2 min. During each study, aminophylline (125 mg) was available to reverse adverse effects of dipyridamole, if necessary.
1.4 Thallium-201 SPECT imaging.
All patients were scanned with the same dual-detector SPECT system (DST, Sopha Medical) equipped with low energy, parallel-hole, all-purpose collimators. Thallium-201 tomographic images were obtained within 10 min of the injection of the radiopharmaceutical using a 180° circular orbit, from 45° right anterior oblique to 45° left posterior oblique, 32-frame step-and-shoot, 50 s/frame, with the patient in the supine position. Redistribution imaging was performed 4 to 6 h later using the same acquisition measurements.
1.5 Technetium-99m sestamibi SPECT imaging.
A 2-day Tc-99m sestamibi injection protocol was used. For the rest studies, tomographic images were obtained 60 to 90 min after the intravenous injection of 25 to 30 mCi of Tc-99m sestamibi (reconstituted from Cardiolite kits made by DuPont Merck). The same dual-detector SPECT system as for Tl-201 imaging was used for Tc-99m sestamibi studies with the same type of orbit and projections (32 frames, 50 s/frame). A high resolution collimator was used. Single-photon emission computed tomographic imaging was performed 30 to 60 min after the injection of 25 to 30 mCi of Tc-99m sestamibi at peak stress or 60 to 90 min after the dipyridamole infusion. The same imaging variables as for the rest study were used with the addition of ECG-gated projection for tomographic reconstruction. At each projection, a total of eight individual ECG-gated frames per cardiac cycle were acquired.
1.6 Image reconstruction.
Thallium-201 and Tc-99m sestamibi images were processed on a NXT Sopha computer. For Tl-201 imaging, reconstruction was performed using a Butterworth filter with a cutoff frequency of 0.35 Nyquist and an order of 5, whereas a Butterworth filter with a cutoff frequency of 0.60 Nyquist and an order of 5 was used for Tc-99m sestamibi perfusion imaging and a Hann filter for ECG-gated SPECT imaging.
1.7 Image analysis.
A total of 230 sets of stress-rest or stress-redistribution images (one Tl-201 and one Tc-99m sestamibi study for each of the 85 patients and 30 volunteers) were analyzed together by three experienced observers (E.G.D., J.E.U., G.A.B.) who had no knowledge of the clinical history (patient vs. volunteer), type of stress test (exercise vs. dipyridamole), results of stress test or coronary angiography or radiopharmaceutical agent used (Tl-201 vs. Tc-99m sestamibi). Disagreements were resolved by consensus.
For each set of images, the observers graded the image quality (four categories: poor, fair, good, excellent), rendered a final patient diagnosis from the results of myocardial perfusion studies (three categories: normal, ischemia, scar) and determined the diagnostic certainty of the overall study (five categories: definitely normal, probably normal, equivocal, probably abnormal, definitely abnormal). For each study, the observers evaluated three short-axis slices and one vertical long-axis slice available in three types of image displays: color and black-and-white displays and studies on x-ray films. The basal and mid short-axis slices were divided into six segments each, whereas the apical short-axis slice was divided into three segments and the mid-vertical long-axis slice was divided into two segments (anteroapical and inferoapical), yielding a total of 17 segments per patient. The degree of radiotracer uptake for each of the 17 segments was semiquantitatively assessed using a 5-point scoring system (0 = absent uptake; 1 = severely reduced uptake; 2 = moderately reduced uptake; 3 = slightly reduced uptake; 4 = normal uptake). The observers had access to polar maps and to eight short-axis, eight vertical long-axis and eight horizontal long-axis slices. A perfusion abnormality on the stress image showing partial or complete normalization on the rest or delayed image was designated as a reversible abnormality or ischemia, whereas a perfusion abnormality that remained unchanged on the rest or delayed image was considered as a fixed lesion or scar. A localized and fixed defect receiving a grade of 3 was considered to be a normal variant or artifact. On a second, separate reading session, two of the observers (E.G.D., J.E.U.) analyzed the results of gated SPECT Tc-99m sestamibi studies. They had no knowledge of the results of both Tl-201 and Tc-99m sestamibi perfusion studies. For each 115 gated SPECT Tc-99m sestamibi studies, color cine displays were available in the three axes (short-axis, vertical long-axis and horizontal long-axis). The observers assessed the wall motion and the wall thickening using a 5-point scoring system (0 = akinesia; 1 = severe hypokinesia; 2 = moderate hypokinesia; 3 = slight hypokinesia; 4 = normal wall motion). The results of the gated SPECT studies were added to those of the perfusion studies when a fixed abnormality was detected on Tc-99m sestamibi perfusion studies. In these cases, a fixed perfusion defect associated with a normal wall motion study was read as a soft tissue attenuation artifact, and the final result was a normal study. In contrast, a fixed perfusion defect with abnormal wall motion on gated imaging was interpreted as a positive perfusion study consistent with myocardial scar.
1.8 Coronary angiography.
Coronary angiography was performed using the Judkins technique with visualization of each major coronary artery in multiple orthogonal projections. All angiograms were interpreted semiquantitatively by two experienced observers who had no knowledge of the results of myocardial perfusion imaging studies. Two thresholds were used to define an angiographically significant stenosis: ≥50% or ≥70% reduction in lumen diameter of one or more major coronary arteries.
1.9 Statistical analysis.
All data are expressed as mean value ± SD. Treadmill stress test and dipyridamole data were assessed using the paired Student ttest. The degree of agreement between Tl-201 and Tc-99m sestamibi studies was assessed with a kappa statistic. Data from the three techniques (Tl-201, Tc-99m sestamibi perfusion and Tc-99m sestamibi gated SPECT) were compared on the basis of their sensitivity and specificity using log linear analysis, allowing pairwise comparisons of the three techniques. A p value <0.05 was considered statistically significant. The Bonferroni correction was applied to keep the alpha level at 0.05 (revised alpha = 0.016).
2.1 Patient group.
The study group consisted of 85 women with or suspected of having CAD and 30 female volunteers with a ≤5% likelihood of CAD. Forty-eight patients (56%) underwent treadmill stress testing, and the remaining 37 patients (44%) had pharmacologic stress testing with dipyridamole. All 30 volunteers had a treadmill stress test. The 48 patients who had a treadmill stress test had a mean age of 58 ± 8 years (range 41 to 72), a mean weight of 66.6 ± 11.3 kg (range 46 to 85) and a mean height of 156.9 ± 5.9 cm (range 145 to 170), whereas the group of patients having dipyridamole had a mean age of 63 ± 11 years (range 37 to 79), a mean weight of 67.5 ± 14.8 kg (range 48 to 109) and a mean height of 156.0 ± 5.6 cm (range 145 to 169). The group of volunteers had a mean age of 45 ± 12 years (range 26 to 62), a mean weight of 64.4 ± 9.1 kg (range 54 to 82) and a mean height of 161.6 ± 6.1 cm (range 152 to 170). Thirty-nine patients were hypertensive (27 in the treadmill stress test group, 12 in the dipyridamole group), 12 had a documented myocardial infarction or a history of myocardial infarction (8 in the treadmill group and 4 in the dipyridamole group); 22 had known diabetes mellitus (11 in the treadmill group and 11 in the dipyridamole group); 35 had known hyperlipidemia (18 in the treadmill and 17 in the dipyridamole group); and 28 were cigarette smokers (16 in the treadmill and 12 in the dipyridamole group).
Using a threshold of 50% for coronary stenosis, coronary angiography was normal in 21 patients, showed single-vessel disease in 31 patients, double-vessel disease in 25 and triple-vessel disease in 8 patients. When the criteria for significant stenosis was ≥70%, coronary angiography was normal in 34 patients, showed single-vessel disease in 34, double-vessel disease in 12 and triple-vessel disease in 5.
For patients undergoing treadmill stress testing, no significant statistical differences were observed in rest and exercise heart rates, maximal systolic blood pressure, rate-pressure product, percent maximal predicted heart rate, exercise work load and the number of clinically and electrically positive stress tests between Tl-201 and Tc-99m sestamibi studies (Table 1). In patients who had a dipyridamole study, there were no significant statistical differences in rest and maximal heart rate, rest and maximal blood pressure and changes in heart rate and blood pressure before and after the administration of dipyridamole (Table 2) between Tl-201 and Tc-99m sestamibi studies.
Table 3summarizes the image quality rating for Tl-201 and Tc-99m sestamibi scintigraphic studies. The quality of the images was assessed by using a simple, semiquantitative scoring system (poor = 1 point; fair = 2 points; good = 3 points; excellent = 4 points). The quality of Tc-99m sestamibi images was significantly better than the quality of Tl-201 images. The “average” quality index of Tl-201 was 2.9 ± 0.6 and that of Tc-99m sestamibi was 3.5 ± 0.5 (p < 0.0001).
The final patient diagnoses for both Tl-201 and Tc-99m sestamibi are shown in Table 4. The agreement between the two studies was 79.1% (91 of 115 patients, kappa = 0.64). Thallium-201 studies were normal in 49 patients, showed transient defects in 53 and fixed defects in 13, whereas Tc-99m sestamibi studies were normal in 64 patients, showed transient defects in 44 and fixed defects in 7.
2.2 Sensitivity of Tl-201 and Tc-99m sestamibi.
Fig. 1and Fig. 2illustrate the sensitivity of Tl-201 and Tc-99m sestamibi for the detection of CAD in patients with coronary artery stenosis ≥50% and ≥70%, respectively. No statistically significant difference between the two types of studies in detection of single-, double- and triple-vessel disease for both stenosis severities was seen.
2.2.1 Stenosis ≥50%.
The overall sensitivity (both treadmill and dipyridamole) for the detection of CAD (≥50% stenosis) was 75.0% for Tl-201 and 71.9% for Tc-99m sestamibi. In the group of patients undergoing treadmill stress testing, the sensitivity of Tl-201 was 68.7% (11 of 16 patients), 78.6% (11 of 14) and 100% (2 of 2) for detection of single-, double- and triple-vessel disease, respectively, whereas for Tc-99m sestamibi these values were 62.5% (10 of 16 patients), 78.6% (11 of 14) and 100% (2 of 2), respectively. For the dipyridamole group, the sensitivity of Tl-201 was 73.3% (11 of 15 patients), 72.7% (8 of 11) and 83.3% (5 of 6), and that of Tc-99m sestamibi was 73.3% (11 of 15 patients), 63.6% (7 of 11) and 83.3% (5 of 6) for the detection of single-, double- and triple-vessel disease, respectively.
2.2.2 Stenosis ≥70%.
The overall sensitivity of Tl-201 was 78.6% (22 of 28 patients) and 91.3% (21 of 23) for the treadmill and dipyridamole groups, respectively. For Tc-99m sestamibi these values were 75.0% (21 of 28 patients) and 87.0% (20 of 23), respectively. In the group of patients having the treadmill stress test, the sensitivity of Tl-201 was 75.0% (12 of 16 patients), 77.8% (7 of 9) and 100% (3 of 3) for the detection of single-, double- and triple-vessel disease, respectively, whereas for Tc-99m sestamibi these values were 68.8% (11 of 16 patients), 77.8% (7 of 9) and 100% (3 of 3), respectively. For the dipyridamole group, the sensitivity of Tl-201 was 88.9% (16 of 18 patients), 100% (3 of 3) and 100% (2 of 2), and that of Tc-99m sestamibi was 83.3% (15 of 18), 100% (3 of 3) and 100% (2 of 2) for the detection of single-, double- and triple-vessel disease, respectively.
2.3 Normalcy rate of Tl-201 and Tc-99m sestamibi.
The “normalcy” rate, defined as the percentage of normal or true negative studies in the group of normal volunteers with a ≤5% likelihood of CAD, was determined for both Tl-201 and Tc-99m sestamibi. Seven patients had a positive Tl-201 study and four patients had a positive study on Tc-99m sestamibi imaging. The normalcy rate was 76.7% (23 of 30 patients) for Tl-201 and 86.7% (26 of 30) for Tc-99m sestamibi.
2.4 Specificity of Tl-201 and Tc-99m sestamibi.
2.4.1 Perfusion study.
Fig. 3illustrates the specificities of Tl-201 and Tc-99m sestamibi SPECT perfusion imaging for the detection of CAD in patients with ≥50% and ≥70% stenosis. Using a threshold of ≥50% to define significant coronary stenosis, there was no significant difference in specificity between Tl-201 and Tc-99m sestamibi, although Tc-99m sestamibi was more specific than Tl-201, with a borderline probability value for the group of patients having a treadmill stress test (p = 0.07) and for the entire group (treadmill and dipyridamole, p = 0.07). When the threshold of ≥70% was used, Tc-99m sestamibi was significantly more specific than Tl-201 for the whole group (p = 0.01) and for the group of patients undergoing treadmill testing (p = 0.02). No significant difference between Tl-201 and Tc-99m sestamibi was seen in the dipyridamole group.
2.4.2 Perfusion and function studies.
Fig. 4illustrates the specificity of Tl-201, Tc-99m sestamibi perfusion and Tc-99m sestamibi gated SPECT studies for both patients without significant CAD (21 patients using the threshold of 50% and 34 patients using the threshold of 70% stenosis) and the group of 30 normal volunteers, yielding a total of 51 subjects and 64 subjects for stenosis ≥50% and ≥70%, respectively. Although there was no significant statistical difference between the Tc-99m sestamibi perfusion and gated SPECT groups, the use of gated SPECT significantly improved the specificity of Tc-99m sestamibi perfusion when compared with Tl-201. In the group of patients with ≥50% stenosis, the specificity was 70.6% for Tl-201 SPECT perfusion imaging, 86.3% for Tc-99m sestamibi SPECT perfusion imaging (p = 0.05) and 94.1% for Tc-99m sestamibi gated SPECT imaging (p = 0.002 in comparison to Tl-201 SPECT imaging). For ≥70% stenosis, the specificity was 67.2% for Tl-201, 84.4% for Tc-99m sestamibi perfusion (p = 0.02) and 92.2% for Tc-99m sestamibi gated SPECT imaging (p = 0.004 compared with Tl-201).
Considering a sample of 64 observations, a 17% difference in sensitivity could have been detected with a power level of 0.80, at alpha = 0.05. With regards to the determination of specificity, the power level associated with the available sample of 21 subjects was limited to 0.57 at alpha = 0.05 when analyzing the observed difference of 24% between the two techniques. Such a difference could be declared significant with a sample size of at least 39 observations.
2.4.3 False positive studies.
Using the threshold of 70% for significant stenosis, there was a total of 21 false positive results for Tl-201 (7 normal volunteers and 14 patients) and 10 false positive results for Tc-99m sestamibi perfusion studies (4 normal volunteers and 6 patients). From the 21 false positive studies on Tl-201, 9 patients had fixed defects with a final diagnosis of scar and 12 patients had partially or completely reversible defects with a final diagnosis of ischemia. Sixteen patients had defects corresponding to the left anterior descending coronary artery vascular territory (anterior, septal or anterolateral walls of the left ventricle) and five had myocardial defects corresponding to the right coronary artery vascular territory (inferior or inferolateral walls). For the 10 false positive results on Tc-99m sestamibi perfusion study, five patients had fixed defects and five had partially or completely reversible defects. Eight of these patients had defects corresponding to the left anterior descending coronary artery vascular territory and two had defects corresponding to the right coronary artery vascular territory. The five patients with fixed defects on Tc-99m sestamibi perfusion study had normal wall motion on gated SPECT, explaining the increased specificity of gated SPECT. For the other five patients, although gated SPECT studies showed normal wall motion, it was not useful for improving the specificity, because these patients had either partially or completely reversible myocardial perfusion defects.
3.1 Thallium-201 imaging in women.
Over the last two decades, the diagnostic accuracy of radionuclide myocardial perfusion imaging using either treadmill exercise or pharmacologic stress testing has been shown to be superior to treadmill exercise electrocardiography alone. However, the data in the majority of the studies showing that planar or SPECT myocardial perfusion scintigraphy is a good noninvasive diagnostic procedure for the detection of CAD have been obtained in men (). Few specific data exist regarding the clinical utility of radionuclide myocardial perfusion imaging in the noninvasive detection of CAD in women. Furthermore, most of these studies performed in women involved a relatively small number of patients using qualitative planar Tl-201 imaging. The reported sensitivity for the detection of CAD with this procedure ranged from 71% to 86% in four different studies ([17–20]). The specificity varied from 81% to 91%. Although several studies have shown the advantages of Tl-201 SPECT imaging over planar studies, very few have provided significant data on improved specificity in women. Fintel et al. () reported a sensitivity of 76% and a specificity of 95%. Another study using quantitative analysis, however, showed a similar sensitivity but a lower specificity of 65% ().
As in men, several factors may affect the diagnostic accuracy of radionuclide myocardial perfusion imaging in women. These include the type of patient population, achievement of adequate heart rate response, type of image acquisition, the way imaging data are analyzed, criteria for image interpretation and the degree of soft tissue attenuation artifacts identified. Breast attenuation artifacts were recognized early after the introduction of Tl-201 planar imaging. Friedman and co-workers () were among the first to consider the effect of the attenuation of overlying breast tissue on the specificity of the test. When fixed anterolateral wall perfusion defects were considered to represent breast attenuation artifacts instead of real perfusion abnormalities, the specificity increased from 88% to 97%. Hung et al. () reported the same observation with an improved specificity from 81% to 91%. Different strategies have been suggested to improve the specificity of Tl-201 imaging in women, but factors such as the size, the position, the density and the change in position of the breast between two imaging sessions may still limit test performance.
3.2 Advantages of Tc-99m sestamibi imaging.
The use of Tc-99m sestamibi, a relatively new myocardial perfusion imaging agent, may also be helpful in decreasing the effects of soft tissue attenuation artifacts. Although this type of artifact is less frequently seen with Tc-99m sestamibi, it remains a potential source of misinterpretation of the test (). Despite the higher energy level of Tc-99m compared with Tl-201, tissue attenuation with Tc-99m is decreased by only ∼15% ([23, 24]). However, one of the major advantages of Tc-99m sestamibi over Tl-201 is the high count density of the perfusion images, which allows high quality images (superior to Tl-201 studies as shown in this study) and gated images synchronized to the patient’s ECG. This allows assessment of myocardial perfusion and ventricular function simultaneously. It has been suggested that attenuation artifacts may be more clearly recognized with this technique when there is normal wall thickening and wall motion in the presence of a fixed perfusion defect ().
3.3 Results of the present study.
The results of our study show that Tl-201 SPECT and Tc-99m sestamibi SPECT imaging had similar sensitivities for the detection of CAD in women. These results are consistent with those reported in large series comparing the two agents in both men and women ([25–30]). However, the specificity of Tc-99m sestamibi perfusion imaging was significantly higher than that of Tl-201. The addition of the ECG-gated imaging to the Tc-99m sestamibi SPECT perfusion study further improved the specificity of this procedure in patients showing fixed defects on perfusion imaging. These latter data confirm, using coronary angiography for the first time as the reference standard, that the presence of preserved wall thickening and normal wall motion on Tc-99m sestamibi gated SPECT imaging in the setting of a fixed defect, indeed, diminishes the false positive rate of perfusion imaging in women. It should be noted, however, that in cases where a partially or completely reversible defect is detected on perfusion imaging, the gated SPECT study cannot be used to improve the specificity of the test. This is because regions corresponding to reversible myocardial ischemia are expected to exhibit preserved systolic thickening at rest.
One interesting finding of our study, as reported by other investigators (), was that soft tissue attenuation artifacts were not only confined to the anteroseptal or anterolateral walls (breast artifacts), but were also seen in projections of the inferior wall. This suggests that diaphragmatic soft tissue attenuation artifacts may occur in women as in men, particularly in obese patients. Electrocardiographic-gated SPECT studies can be helpful in recognizing both types of soft tissue attenuation artifacts in women.
3.4 Study limitations.
This study was performed in a selected group of patients since the group was composed of patients referred for coronary angiography. Thus, these results may not be applicable to a general patient population. Nevertheless, inclusion of catheterized patients was necessary to validate the results of the different radionuclide imaging procedures. Quantitative analysis of SPECT imaging, in comparison to a gender-specific normal data base, might potentially improve the overall diagnostic accuracy of radionuclide perfusion imaging. Such quantitative analysis was not applied in this study. Instead, experienced observers from three different institutions interpreted the studies by consensus, employing visual analysis of tomographic images. Perhaps the application of quantitative image analysis, added to the results of the ECG-gated SPECT study, might further improve the diagnostic accuracy of SPECT perfusion imaging. The effect of quantitation with reference to gender-specific normal data bases cannot be inferred from the current results; however, the one reported study using quantitative analysis of SPECT thallium imaging in women demonstrated a specificity of 65% (). More data are needed to better evaluate the value of quantitative gated SPECT.
Thallium-201 and Tc-99m sestamibi myocardial perfusion scintigraphic studies have similar sensitivities in the detection of CAD in women. However, the specificity of Tc-99m sestamibi is significantly higher than that of Tl-201. This specificity is further enhanced by the use of ECG-gated SPECT Tc-99m sestamibi imaging.
We gratefully acknowledge the valuable help of Linda P. Giering, PhD, Micheline Lauzon, André Gagnon, CNMT, Carole Benjamin, CNMT, Sylvie Lajeunesse, CNMT, Marc Dumont (statistical analyses), the audiovisual service, all the referring cardiologists and the nuclear medicine technologists of the Hôtel-Dieu de Montréal Hospital.
☆ This study was partially supported by a research grant from DuPont Merck, Billerica, Massachusetts.
- coronary artery disease
- electrocardiogram, electrocardiographic
- single-photon emission computed tomography
- Received May 24, 1996.
- Revision received September 11, 1996.
- Accepted September 23, 1996.
- The American College of Cardiology
- Wenger NK
- Detry JMR,
- Kapita BM,
- Cosyns J,
- Sottiaux B,
- Brasseur LA,
- Rousseau MF
- Guiteras Val P,
- Chaitman BR,
- Waters DD,
- et al.
- Dash H,
- Massie BM,
- Botvinick EH,
- Brundage BH
- Pohost GM,
- Zir LM,
- Moore RH,
- McKusick KA,
- Guiney TE,
- Beller GA
- Brown KA,
- Boucher CA,
- Okada RD,
- et al.
- Wackers FJT
- DePuey EG,
- Rozanski A
- Hung J,
- Chaitman B,
- Lam J,
- et al.
- Melin J,
- Wijns W,
- Vanbutsele RJ,
- et al.
- Fintel DJ,
- Links JM,
- Brinker JA,
- Frank TL,
- Parker M,
- Becker LC
- Chae SC,
- Heo J,
- Iskandrian AS,
- Wasserleben V,
- Cave V
- Watson DD,
- Smith WH
- Kahn J,
- Mc Ghie I,
- Akers M,
- et al.
- Iskandrian AS,
- Heo J,
- Kong B,
- Lyons E,
- Marsch S
- Inglese E,
- Galli M,
- Parodi O,
- et al.