A significant amount of data has accumulated over the past 20 years, demonstrating the value of exercise or pharmacologic stress non-invasive cardiovascular imaging techniques for the diagnostic and prognostic assessment of patients with suspected or known cardiovascular disease.1-8 Stress can be induced by either multistage exercise testing that is symptom-limited or by pharmacologic means with either a vasodilator (e.g. dipyridamole or adenosine) or an inotropic agent (e.g. dobutamine). The combination of low-level exercise with vasodilator stress radionuclide myocardial perfusion imaging enhances both the quality of images (diminished visceral tracer uptake) and reduces vasodilator-induced side effects.
Detection of Coronary Artery Disease
Radionuclide Single-photon Emission Computed Tomography Stress Perfusion Imaging
The sensitivity of radionuclide exercise myocardial perfusion imaging for detection of coronary artery disease (CAD) averaged 87% in 33 studies pooled from the literature.8 Specificity averaged 73% with a normalcy rate significantly higher at 91%. The normalcy rate is a variable that reduces the effect of referral bias inherent in specificity determinations that require cardiac catheterization in normal coronary angiograms for the gold standard for true negative test results. This is because the preponderance of referrals to the catheterization laboratory among patients undergoing stress perfusion imaging are those who have abnormal scans. Thus, a bias is inherent in the fact that more false positive patients are referred for coronary angiography than true negative patients. The normalcy rate is defined as the percentage of patients who have <5% pretest likelihood of CAD and who have normal myocardial perfusion studies. The sensitivity and specificity of exercise myocardial perfusion imaging are higher than the sensitivity and specificity of electrocardiographic treadmill testing alone.9 When patients with resting ST-segment depression are excluded, the sensitivity and specificity of the exercise echocardiogram (ECG) stress test were only 67% and 84%, respectively.10
With respect to myocardial perfusion imaging employing the single-photon emission computed tomography (SPECT) technique, the major problem is with specificity. This is predominantly due to attenuation artifacts (breast attenuation in women and posterobasal attenuation) that may be misconstrued as defects attributed to underlying coronary artery stenoses. The use of technetium-99m (99mTc)-labeled perfusion agents (e.g. sestamibi and tetrofosmin) combined with ECG-gating substantially improves the specificity of SPECT perfusion imaging compared with ungated thallium-201(201Tl) scintigraphy.11
ECG-gated SPECT permits the differentiation of perfusion abnormalities due to soft tissue attenuation versus those that are due to myocardial scar. Non-reversible defects from stress to rest corresponding with normal regional wall thickening or wall motion represent attenuation artifacts, whereas non-reversible defects associated with regional dysfunction are most likely secondary to underlying myocardial scar. In the study by Smanio et al.11 comprising 285 patients undergoing 99mTc-sestamibi SPECT imaging, the number of borderline interpretations was reduced from 89 to 29, with the percentage of interpretations designated as definitely normal having increased from 74% to 93% upon review of the gated SPECT images.
Another approach to reducing attenuation artifacts is the application of attenuation-correction software, where transmission data from attenuation maps are constructed to correct emission data for photon attenuation. This is currently undertaken using an X-ray source and a CT scanner that is embedded in a hybrid manner with a SPECT camera.12 Like ECG-gated SPECT, attenuation-correction imaging may improve the identification of more definitely normal studies and reduce the degree of uncertainty characterized by interpretations often designated as 'borderline normal' or 'borderline abnormal.'13 In another study in obese patients,14 attenuation-corrected 99mTc-sestamibi SPECT improved specificity from 50% to 79%.
For patients who are unable to exercise, pharmacologic stress perfusion imaging is an alternative approach.The sensitivity and specificity for dipyridamole or adenosine stress perfusion imaging are comparable with values derived from the exercise imaging studies.8 Vasodilator stress should also be used in patients who have a resting left bundle branch block (LBBB) pattern. Exercise imaging in patients with LBBB yields a high percentage of false positive defects in the intraventricular septum.
Dobutamine infusion as the stressor is preferred for patients who have bronchospasm or a history of asthma, or in those who have consumed caffeine within eight hours of testing. In a pooled analysis of more than 1,200 patients from 13 studies in the literature,15 dobutamine stress had a sensitivity of 85% and a specificity of 82%.
Recent interest has been generated regarding the use of positron emission tomography (PET) with rubidium-82 (82Rb) for detection of CAD in conjunction with vasodilator stress. Advantages of PET over SPECT for detecting CAD include the standard ability to perform attenuation correction and a high spatial and contrast resolution permitting detection of small perfusion defects with greater specificity. PET scanners are now most often manufactured with CT scanners, permitting the simultaneous assessment of myocardial physiology and anatomy.16 A useful approach that is now being evaluated clinically for enhancing sensitivity and specificity for detection of CAD is the performance of CT angiography with 82Rb PET imaging.17 In one study, three-vessel disease was better identified using quantitative 82Rb PET compared with standard stress perfusion imaging in which only relative tracer uptake was evaluated.18Figure 1 shows a fusion image where the zone of normal 82Rb uptake (red) is superimposed on a normal multislice CT angiogram displayed volumetrically.19 Fusion imaging is an important advance in that some patients with three-vessel disease may have exhausted flow reserve in the coronary territories of all stenoses and exhibit 'balanced ischemia.ÔÇÖ Such patients might show uniform uptake of a tracer like 82Rb but show multiple coronary stenoses on CT angiography.
Stress Echocardiography for Detection of CAD
Stress echocardiography can also be performed using either exercise or pharmacologic stress using stressors such as dobutamine.1,6,7 In a pooled analysis in the literature, the weighted mean sensitivity, specificity, and overall accuracy for exercise echocardiography were 86%, 81%, and 85%, respectively.20 For dobutamine stress, the respective values were 82%, 84%, and 83%. The sensitivity of exercise echocardiography may be diminished if only submaximal exercise heart rates are attained or if acquisition of ultrasound images are delayed in the post-exercise period.6
Prognostic Applications of Non-invasive Stress Imaging Techniques
Stress Myocardial Perfusion SPECT Imaging
Perhaps one of the most valuable clinical applications of stress perfusion imaging is identifying the low-risk patient presenting with undiagnosed chest pain and for whom a normal perfusion scan would yield a good prognosis. Prognostic data in the literature pertaining to the outcome of patients with normal stress perfusion scans suggest that the subsequent cardiac death or infarction rate in patients with no post-stress perfusion abnormalities averages 0.6% per year.4
This pooled analysis comprised 39,173 patients who had a normal or low-risk scan and were followed for an average of 2.3 years. The myocardial imaging SPECT variables associated with high-risk CAD are as follows:
- multiple perfusion abnormalities in more than one coronary supply region reflective of multivessel disease;
- an ischemic defect size of >20% of the left ventricle as determined from quantitative analysis of stress and rest perfusion images;
- stress-induced transient ischemic left ventricular (LV) cavity dilation in which the LV cavity appears larger on stress compared with rest images;
- multiple regional wall thickening or wall motion abnormalities even in the absence of defects; and
- an LV ejection fraction of <40%. If 201Tl stress imaging is undertaken, an increased lung/heart thallium uptake ratio is a high-risk variable.
In a pooled analysis of 69,655 patients in studies where follow-up was obtained in patients with normal or abnormal scans, those with high- or moderately high-risk SPECT findings had an annual event rate of 5.9% per year compared with 0.85% per year in the low-risk scan group.4 The assessment of regional wall thickening or wall motion on post-stress SPECT studies has shown a value of increasing the identification of patients with three-vessel disease.21
Perhaps one of the most potent predictors of underlying high-risk CAD and a marker of a high event rate is transient ischemic dilation (TID) of the LV cavity as pointed out above.This TID occurs on both exercise and vasodilator stress SPECT images and is most likely secondary to severe subendocardial ischemia and not due to an enlargement of the entire left ventricle (e.g. increase in end-diastolic and end-systolic volumes).TID is an independent prognostic variable.22Figure 2 shows the stress and rest short-axis 99mTc-sestamibi tomograms in a patient with TID and multiple reversible defects.
Patients with normal or abnormal SPECT pharmacologic stress perfusion scans have a worse prognosis compared with patients who have the same perfusion patterns on exercise SPECT studies.23 In this analysis of a large number of patients, the annual cardiac death rate was 0.15% per year in the 5,000 patients with normal exercise perfusion scans compared with a 0.8% annual mortality in the 1,600 patients who had normal pharmacologic stress studies. Similarly, patients who had abnormal pharmacologic stress perfusion images had an annual mortality rate of 5.8% per year compared with a 1.7% annual mortality rate in patients with abnormal exercise images.
This difference is surely related to the difference in patient populations who are referred for pharmacologic stress imaging compared with those referred for exercise stress testing. Obviously, patients referred for pharmacologic stress imaging are most often older, have more co-morbidities, such as chronic pulmonary or peripheral vascular disease, and are those tested soon after an acute coronary event. Stress myocardial perfusion imaging has significant incremental prognostic value over and above the prognostic information obtained from clinical and standard ECG stress test variables (e.g. Duke treadmill score).24
Special populations of patients undergoing stress perfusion imaging deserve to be mentioned. Stress perfusion imaging has been shown to be extremely valuable in assessing prognosis in patients with diabetes.4,25,26 For diabetics with either normal scans or abnormal scans, the future cardiac death/infarction rate is substantially higher than the non-diabetic patients with similar perfusion scan findings (see Figure 3).4 Diabetic women with a high-risk scan have the highest annual cardiac death or myocardial infarction (MI) rate (>10%), which appears to be almost twice the event rate observed for diabetic men with a high-risk scan (see Figure 3).4 As previously mentioned, diabetic women with a normal myocardial perfusion scan have an annual event rate of 3%.4 This is a higher combined event rate than seen in diabetic men with a normal scan.
In a multicenter trial reported by Giri et al.,25 women who had inducible ischemia on SPECT studies in the coronary supply regions of ≥2 major coronary vessels had a 40% death or infarction rate after three years compared with 21% for men. Diabetic women with normal or abnormal adenosine SPECT images also have a higher future cardiac mortality rate than men with the same findings.26 Asymptomatic diabetics with abnormal perfusion scans also have a significant subsequent cardiac event rate.27,28
African-Americans and Hispanic patients with moderate or severely abnormal SPECT scans have substantially higher annual cardiac event rates compared with Caucasian, non-Hispanic patients.29 In that study, the annualized risk-adjusted death rates associated with a post-stress ejection fraction of <45% were 2.7% for Caucasian, non-Hispanic patients versus 8.0% and 14.0% for African-American and Hispanic patients, respectively. Thus, these data add to the body of evidence suggesting that ethnic minority patient populations have a worsening outcome related to cardiovascular disease.
Renal disease patients who are on hemodialysis and undergo renal transplantation are a particularly high-risk subgroup for future coronary events. Patel et al.30 reported that the event-free survival after renal transplantation was significantly worse in patients who had an abnormal stress myocardial perfusion scan prior to transplantation compared with patients with normal scans.
Stress perfusion imaging is one of the tests that can be used for pre-operative risk stratification in patients scheduled to undergo non-cardiac surgery.31 The patients who benefit most from pre-operative stress imaging prior to non-cardiac surgery are those with aortic or peripheral vascular disease who are scheduled to undergo a major vascular procedure such as an abdominal aortic aneurysm resection. Diabetic patients with peripheral vascular disease may be at a particularly high risk for subsequent cardiac events. What is not yet clear is whether coronary revascularization prior to the non-cardiac operation reduces the risk of intra-operative death or infarction. Certainly, patients with extensive ischemia on pre-operative pharmacologic stress imaging deserve coronary angiography for identification of high-risk anatomy such as left main or proximal multivessel CAD.
Stress myocardial perfusion imaging is often used for determining which patients with CAD benefit most from revascularization compared with medical therapy, even if minimally symptomatic. Although no randomized studies are published, Hachamovitch et al.,32 using a propensity analysis algorithm, compared the short-term survival benefit associated with revascularization with medical therapy in patients with no prior documented CAD who underwent stress SPECT perfusion imaging. They showed that patients with >20% of the left ventricle rendered ischemic on stress SPECT imaging benefited most from revascularization compared with medical therapy. Women who exhibited >20% ischemia on stress SPECT imaging had a 17% annual mortality rate with medical therapy compared with 4.4% with revascularization.
Evaluation of regional and global LV function on rest and stress ECGs may provide significant prognostic information. A large area of asynergy is predictive of future cardiac events. Such physiologic information, obtained from regional and global ventricular function at peak exercise or peak pharmacologic stress, provides incremental value to clinical ECG stress variables as well as more prognostic data than can be obtained simply from the determination of resting LV function.6,33-35
In a pooled analysis comprising 27 stress echocardiographic studies involving 6,799 patients with a mean follow-up of 27 months, those with a normal stress ECG had a cardiac death or non-fatal MI rate of 0.9% per year.36 Patients with intermediate- or high-risk stress ECGs were shown to have an annual cardiac event rate of 3.1% and 5.2% per year, respectively.37
Thus, as observed with stress perfusion imaging, stress echocardiography can also risk-stratify patients into those with a low cardiac event rate and those with an increased event rate. The latter is based on the extent of inducible wall motion abnormalities. Dobutamine stress echocardiography has been shown to improve risk stratification in patients undergoing pre-operative risk assessment prior to non-cardiac surgery.38
For the future, other technologies may emerge as being competitive or synergistic with the non-invasive techniques discussed in this article. Perhaps the most promising is CT angiography employing multislice CT scanning systems using multidetector arrays at fast rotation times.39 It is likely that CT angiography will complement, rather than replace, stress SPECT perfusion imaging or stress echocardiography. It may become most useful in lower risk populations, such as patients presenting to an emergency department with atypical chest pain and a normal resting ECG. A completely normal non-invasive CT angiogram would exclude CAD as the cause of the chest pain syndrome. Nevertheless, combining CT angiography and SPECT or PET should yield information pertaining to coronary anatomy and physiology simultaneously. This combined assessment in selected patients would improve the overall non-invasive assessment of patients with suspected CAD. Perhaps CT angiography would be indicated in patients who have an unexpectedly normal perfusion scan but a high likelihood of underlying multivessel disease. Conversely, some patients with a low pretest likelihood of CAD might demonstrate a perfusion defect on stress imaging considered to perhaps be a probable false positive test result. Rather than sending such a patient to invasive coronary angiography, a normal non-invasive CT angiogram would be sufficient to rule in or exclude CAD as causing that defect.
Finally, cardiac magnetic resonance imaging is being evaluated for the detection of coronary artery stenoses using quantitative kinetic assessment of the wash-in of contrast medium.40,41 This technique is based on traditional first-pass perfusion measurement where less contrast appears in zones that are supplied by stenotic arteries.
SPECT and PET imaging, as well as contrast echocardiography, may be used for molecular imaging where probes are administered in order to detect functional pathophysiologic entities such as an inflamed atherosclerotic plaque, apoptosis of myocytes in heart failure, and increased macrophage density in a potentially vulnerable plaque, as well as angiogenesis or gene expression.44-44