The aim of this study was to assess the effect of attenuation correction (AC) on left ventricular (LV) volumes and LV transient ischemic dilatation (TID) during dual-isotope single-photon emission computer tomographic (SPECT) myocardial perfusion imaging (MPI). Ninety-six patients (mean age 58 ± 11 years, 15% women, 38 patients completed exercise and 58 dipyridamole pharmacologic stress tests) assessed for known or suspected coronary artery disease underwent dual-isotope thallium-201 rest and technetium-99m sestamibi stress SPECT MPI with computed tomography–based AC. The TID ratio was calculated separately for non-AC and AC SPECT MPI studies as the ratio of the LV endocardial volume at stress divided by LV endocardial volume at rest. The mean and range of the gated LV ejection fraction during exercise and pharmacologic stress was 54 ± 12% (29% to 80%) and 58 ± 12% (27% to 80%), respectively. In the exercise stress group, the same mean LV endocardial volumes in non-AC and AC stress (76.4 ± 30 and 76.5 ± 28) and rest (66.3 ± 26 and 66.4 ± 24) studies were found (p = 0.90). There was no statistical difference between the mean exercise TID ratio in non-AC and AC studies (1.27 vs 1.31, respectively, p = 0.10). The same mean LV endocardial volumes in non-AC and AC in pharmacologic stress (79.9 ± 42 and 80 ± 41) and rest (71.4 ± 41 and 72.3 ± 37), respectively, were found (p = 0.50). There was no statistical difference between the mean dipyridamole TID ratio in non-AC and AC studies (1.20 vs 1.17, respectively, p = 0.10). In conclusion, LV volumes and TID indexes obtained on SPECT MPI with exercise or pharmacologic stress using dipyridamole are not affected by AC.
Transient ischemic dilatation (TID) of the left ventricle during stress myocardial perfusion imaging (MPI) is a marker that has been shown to improve the diagnostic accuracy of MPI for the extent and severity of coronary artery disease. The presence of TID is a high-risk parameter indicating the need for further assessment of patients with suspected coronary artery disease. The validation of TID has been performed before the implementation of the use of attenuation correction (AC) for routine single-photon emission computed tomographic (SPECT) MPI studies. AC has been associated with improved diagnostic accuracy due to its ability to decrease the impact of tissue attenuation. However, we observed that AC processing may change the cardiac geometry (length of endocardial or epicardial cardiac borders), left ventricular (LV) end-diastolic and end-systolic volumes and may therefore potentially lead to changes in TID values. Although a single study showed that AC reveals gender-related differences in the normal values of TID index, another recent study revealed good agreement for measurements of the LV ejection fraction by AC gated SPECT imaging and non-AC gated SPECT imaging. Whether TID indexes are calculated more accurately after AC processing appears to be unclear. The aim of the present study was to assess whether computed tomography–based AC has any influence on LV volumes and TID values.
Methods
This study was a retrospective analysis of 96 patients (82 men, mean age 58 ± 11 years) referred for SPECT MPI evaluation for known or suspected coronary artery disease. All patients underwent dual-isotope thallium-201 rest and technetium-99m sestamibi stress testing and computed tomographic (CT) coronary angiography from May 2005 to February 2006. Patients with acute coronary syndromes, significant arrhythmias, iodine allergies, major hypersensitivity, renal disease with serum creatinine >1.5 mg/dl, immune disorders, and malignancies were not included in the study population. The study was approved by the institutional ethics (Helsinki) committee. Thirty-eight patients completed maximal exercise stress tests, and the remaining 58 patients performed pharmacologic stress tests with additional low-grade exercise when possible.
Before treadmill exercise testing, patients were instructed to withdraw long-acting nitrates for 6 hours, calcium channel blockers for 24 hours, and β-blocking agents for 48 hours. Multistage treadmill exercise was performed using the Bruce protocol. Twelve-lead electrocardiograms were recorded at rest, every 3 minutes thereafter, at peak exercise, and at each minute during the 5-minute recovery time. Before pharmacologic stress testing, patients were instructed to stop caffeine consumption for 24 hours and xanthine-containing medications for 36 hours. For the pharmacologic stress tests, dipyridamole 0.56 mg/kg was infused intravenously over a 4-minute period with the patient in the sitting position, after which patients were encouraged to exercise on a treadmill for 3 to 4 minutes unless contraindicated. Intravenous aminophylline was used for reversal of dipyridamole adverse effects. Twelve 12-lead electrocardiograms were obtained at baseline before dipyridamole infusion, every 3 minutes thereafter for an 8-minute period, at peak stress, and at each minute during the 5-minute recovery time. Stress-induced chest pain (angina) and electrocardiographic changes of myocardial ischemia, defined as the presence of either ≥0.1-mV horizontal or down-sloping or ≥0.15-mV up-sloping ST-segment depression 80 ms after the J point during the exercise or recovery, were recorded.
The studies were performed using a hybrid SPECT imaging/CT coronary angiography prototype (Infinia gamma camera and LightSpeed CT; GE Healthcare, Milwaukee, Wisconsin) that comprised a dual-head variable-angle gamma camera and a 16-slice CT scanner, spatially aligned to enable sequential acquisition of the SPECT and computed tomography coronary angiographic studies. The gamma camera component is equipped with low-energy, high-resolution parallel-hole collimators, with the detectors at 90° to each other. Processing of SPECT data was performed on a Xeleris workstation (GE Healthcare), and processing of the CT coronary angiographic data was performed on an Advantage Windows 4.2P workstation (GE Healthcare).
SPECT MPI studies were performed using a same-day rest and stress dual-isotope protocol. Sixty projections were acquired over a 180° orbit into a 64 × 64 matrix, with a pixel size of 6.8 mm and a time per projection of 20 seconds. For the rest study, 92.5 MBq (2.5 mCi) thallium-201 was administered, with imaging starting <15 minutes after the intravenous injection. Energy window width setting of 30% and 20% for the peaks at 70 and 167 keV, respectively, were used. The stress test was performed after the rest imaging. An activity of 740 MBq (20 mCi) technetium-99m sestamibi was injected at peak exercise or pharmacologic stress. Stress electrocardiography–gated SPECT imaging was performed 30 to 60 minutes after the tracer administration. Energy window width was set to 20% at the 140-keV peak. The rest and stress SPECT studies were followed by a low-dose (30 mA, 140 keV) CT scan over the area of the heart, as defined by the technologist, used for correcting the emission data for photon attenuation. SPECT MPI data were reconstructed with ordered subsets expectation maximization iterative reconstruction for non-AC and AC studies. The attenuation maps were computed from the dedicated low-dose CT AC scans acquired and registered with the emission data. As an initial step, the operator performed a visual analysis to ensure that the reconstructed emission volume (without AC) was properly aligned with the CT AC volume. Subsequently, the Xeleris application proceeded according to the following steps: (1) Reformat the CT slices into a volume having the same voxel and matrix size as the SPECT projections. (2) Convert the Hounsfield units of the reformatted CT slices into a linear attenuation coefficient corresponding to the energy of the emission isotope and depending on the effective energy spectrum of the CT device. For CT values <0, materials are assumed to have energy dependence similar to water. CT values >0 are treated as being a mixture of bone and water. (3) Apply a Gaussian filter to finally obtain attenuation maps with a resolution similar to that of SPECT data. The ordered subsets expectation maximization reconstruction uses 2 iterations and 10 subsets, with a postreconstruction low-pass Butterworth 3-dimensional filter. For the rest study, a critical frequency of 0.35 Nyquist with an order of 10 was used, and for the stress study, 0.4 Nyquist with an order of 10 was used. The same filtering scheme was applied for data reconstructed with and without AC. No scatter or depth-dependent resolution corrections were applied. One-pixel-thick oblique tomographic slices were generated and displayed in a standard format as short-axis, vertical, and horizontal long-axis slices. Cardiac gating was performed using the detection of the R wave for the monitoring of different phases of the cardiac cycle as 8 frames/cycle. The acceptance window for the RR interval was set to 20%. Image reconstruction and processing by use of the ordered subsets expectation maximization method were performed with a Xeleris workstation.
TID was defined as the ratio of LV endocardial volume at stress divided by LV endocardial volume at rest. It was calculated using Myovation software (GE Healthcare) with automatic measurement of the LV endocardial volume derived separately from the endocardial contour of the left ventricle at stress and endocardial contour of the left ventricle at rest from the ungated images. The software allows manual correction to optimize myocardial contour delineation. The same calculations of the TID ratios were performed automatically for corrected and uncorrected data. For study purposes, all TID images were double-checked by a technologist and a nuclear cardiologist, and in 5 studies, mild manual correction of LV contours was needed.
In immediate sequence with the SPECT study and without moving the patient from the scanning table, CT coronary angiography was performed. The acquisition parameters included a rotation time of 0.5 seconds, tube voltage of 120 kV, and tube current of 420 mA, acquired for 0.625- or 1.25-mm slice thickness, depending on the axial scan range (e.g., larger for post–coronary artery bypass graft patients) and breath hold duration capability. The scanner automatically sets the helical pitch in the range 0.275:1 to 0.325:1, as a function of the heart rate. Contrast material was delivered intravenously by a dual automatic injector, with 85 ml of Ultravist 370 (Bayer AG, Munich, Germany) at 4 ml/s followed by 50 ml of saline flush at 3 ml/s. A single-segment or multisegment reconstruction method was applied, depending on the patient’s heart rate, resulting in slices of 0.625- or 1.25-mm thickness. Immediately upon completion of the scan, a late diastole phase (75% of R-R time) was automatically reconstructed. Intravenous or oral metoprolol was used to slow the heart rate as needed. After the release of the patient, off-line retrospective reconstruction was applied to obtain the phases from 0% to 90% of RR intervals, in steps of 10%, and the 45% and 75% phases. The CardIQ application of the Advantage Workstation (GE Healthcare) was used to obtain the following items from the multiphase CT coronary angiographic image: (1) 3-dimensional rendering of the myocardial walls, coronary arteries, and coronary bypass grafts and (2) tracking and segmental analysis of coronary arterial tree and bypass grafts.
The results of CT coronary angiography were interpreted regarding the major coronary arteries and/or the coronary bypass grafts. Severe and extensive coronary artery disease was defined by CT coronary angiographic data as the presence of either a left main coronary artery stenosis >50%, a proximal left anterior descending coronary artery stenosis ≥90%, or a stenosis ≥90% in 2 or 3 coronary vessels or in ≥1 bypass graft. Mild to moderate coronary artery disease was defined as the presence of >70% but <90% luminal diameter stenosis in ≥1 coronary vessel or bypass graft or ≥90% stenosis in 1 vessel other than the proximal left anterior descending coronary artery. The absence of significant coronary artery disease was defined as <70% stenosis in all native vessels or bypass grafts.
Continuous variables are expressed as mean ± SD. Chi-square analysis was applied for the comparison of categorical variables. A p value ≤0.05 was considered to represent statistical significance. Receiver-operating characteristic analysis was used to evaluate the best cut-off point for the best TID performance indexes. The sensitivity and specificity of TID for the identification of severe and extensive coronary artery disease were calculated and evaluated separately for the non-AC and AC studies.
Results
SPECT MPI and CT coronary angiographic studies and the medical records of 96 patients were retrospectively evaluated. Patients’ characteristics, risk factors, and long-term medical treatment are listed in Table 1 . Of the entire study population, 85% were men, 65% had histories of myocardial infarction and/or coronary bypass grafts, 67% had hypertension, and 32% had diabetes. The numbers of patients with diabetes and hypertension and those with previous myocardial infarction were significantly higher in the dipyridamole group (p <0.01; Table 1 ). Fifty-nine patients (61%) had normal myocardial perfusion, 13 (14%) had fixed perfusion defects consistent with scar, and 24 patients (25%) had reversible perfusion defects indicating ischemia. Normal or nonsignificant coronary artery disease was demonstrated in 41 patients (43%), mild to moderate coronary artery disease in 25 patients (26%), and severe extensive coronary artery disease in 30 patients (31%). None of the patients was found to have isolated left main coronary disease. In the dipyridamole group, the number of patients with normal or nonsignificant coronary artery disease was significantly smaller than in the exercise group (p <0.005; Table 2 ).
| Variable | Exercise Group | Dipyridamole Group | p Value |
|---|---|---|---|
| (n = 38) | (n = 58) | ||
| Age (years) | 58 ± 12 (24–82) | 59 ± 10 (39–79) | 0.65 |
| Women | 3 (8%) | 11 (19%) | 0.13 |
| Body mass index (kg/m 2 ) | 27 ± 5 (19–43) | 29 ± 3 (22–37) | 0.02 |
| Diabetes mellitus | 6 (16%) | 25 (43%) | 0.005 |
| Hypertension ⁎ | 20 (53%) | 44 (76%) | 0.018 |
| Smokers | 19 (50%) | 37 (64%) | 0.18 |
| Peripheral vascular disease | 3 (8%) | 8 (58%) | 0.37 |
| Dyslipidemia † | 29 (76%) | 48 (83%) | 0.43 |
| Typical angina | 8 (21%) | 17 (30%) | 0.36 |
| Previous myocardial infarction | 12 (32%) | 29 (50%) | 0.07 |
| Previous coronary bypass | 8 (21%) | 13 (22%) | 0.02 |
| Aspirin | 24 (63%) | 53 (91%) | 0.0007 |
| β blockers | 23 (60%) | 45 (83%) | 0.07 |
| Angiotensin-converting enzyme inhibitors | 16 (42%) | 20 (34%) | 0.45 |
| Statins | 26 (68%) | 47 (81%) | 0.15 |
⁎ Systolic blood pressure >140 mm Hg and/or diastolic blood pressure >90 mm Hg and/or use of antihypertensive drugs.
† Low-density lipoprotein cholesterol >130 mg/dl and/or high-density lipoprotein cholesterol <40 mg/dl and/or triglycerides >200 mg/dl and/or use of statins and fibrates.
| Variable | Exercise Group | Dipyridamole Group | p Value |
|---|---|---|---|
| (n = 38) | (n = 58) | ||
| Normal perfusion | 24 (63%) | 35 (60%) | 0.39 |
| Fixed defect | 5 (13%) | 8 (14%) | 0.46 |
| 1 reversible defect | 5 (13%) | 9 (16%) | 0.37 |
| ≥2 reversible defects | 4 (11%) | 6 (10%) | 0.49 |
| Gated LV ejection fraction (%) | 54 ± 12 (29–80) | 58 ± 12 (27–80) | 0.82 |
| Nonsignificant CAD | 21 (55%) | 20 (34%) | 0.02 |
| Mild to moderate CAD | 7 (19%) | 18 (32%) | 0.16 |
| Severe and extensive CAD | 10 (26%) | 20 (34%) | 0.19 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
