Appropriate use of non-invasive testing for diagnosis of stable coronary artery disease

Background

When assessing a patient with clinically suspected stable coronary artery disease (SCAD), two questions are in order to establish a diagnosis: 

1. Is the patient eligible for non-invasive testing  
2. If so, what is the most appropriate non-invasive modality? 

Initial diagnostic assessment of individuals with suspected SCAD has clinical evaluation as cornerstone (Fig.1) (1).

• Strong clinical evidence of SCAD with comorbidities and quality of life making revascularisation an unlikely option call for limited testing and initiation of medical therapy as part of a conservative approach - even if the diagnosis has not been confirmed by specific testing. 
• Typical angina and depressed left ventricular ejection fraction (LVEF) <50% should prompt direct referral for ICA as part of a more aggressive approach.
• Pre-test probability (PTP) of SCAD, based on simple clinical characteristics that include chest pain classification, age, and gender will determine the clinical likelihood that SCAD is present and thus determine eligibility for specific non-invasive testing.
• Sensitivity - the true positive rate - of most modalities is around or below 85% and specificity - the true negative rate - is approximately or below 85%, thus a key message is that specific non-invasive testing to establish diagnosis of SCAD is only recommended for patients with intermediate clinical likelihood of SCAD (15-85%), but not for individuals with higher PTP >85% or lower PTP <15%.

Gender-related implications of the estimated PTP values are shown in Table 1.

1. Men with suspected SCAD always have a PTP >15%, even in the low-risk constellation of young age (30-39) and non-anginal chest pain. 
2. Only women with clinical suspicion of SCAD are likely to have a truly low (<15%) clinical PTP and can thus be managed as not having SCAD without further specific testing. 
3. Only elderly men (aged >70) with typical angina have a PTP that surpasses the threshold of 85%, thus establishing diagnosis of SCAD without the need for specific testing.
4. Female patients cannot reach a clinical PTP of SCAD that is high enough to justify deferral of confirmatory non-invasive testing.    

I - Non-invasive tests

Non-invasive modalities for diagnosis of SCAD are either functional tests - non-imaging stress tests or stress tests combined with imaging - that detect inducible myocardial ischemia, or modalities that visualise coronary anatomy.

A) Non-invasive testing of myocardial ischemia

Non-imaging functional testing

Although ECG is a time-honored, simple, generally safe, and is the least costly of non-imaging tests, it has a sensitivity (true positive rate) of 45-50% and specificity (true negative rate) of 85-90%, suggesting a higher efficacy for exclusion rather than confirmation of diagnosis of SCAD.
It performs well at intermediate PTPs - between 15–65% - in patients with normal (interpretable) resting ECGs without ST–T abnormalities.

• Stress testing is preferable over ECG exercise testing in cases of 1) inability to exercise to an adequate level of workload due to orthopedic or other disabling comorbidities 2) resting ECG features such as left bundle branch block, ventricular paced rhythm, pre-excitation syndromes, repolarisation abnormalities due to left ventricular hypertrophy and digoxin use that preclude identification of ischemia.
• ECG is absolutely contraindicated in case of suspected SCAD including severe symptomatic aortic stenosis, uncontrolled arrhythmias causing symptoms or hemodynamic instability, decompensated heart failure and acute non-cardiac disorders that may affect exercise performance or be aggravated by exercise (e.g., infection, thyrotoxicosis).
• Anti-ischemic medicationsreduce heart rate and myocardial workload and can thereby lead to false negative findings. ECG exercise testing is also less useful than stress imaging for localisation of distribution and assessment of extent of ischemia, and it has a lower diagnostic value in women (2,3). 

Stress imaging 

The addition of imaging to exercise testing provides incremental benefit for accurate diagnosis of obstructive CAD with an acceptable increase in cost (4, 5); however, whether this diagnostic performance translates into clinical outcomes is not yet established (6).
Advantages of stress imaging include 1) accurate assessment in the presence of resting ECG abnormalities, 2) option of using pharmacological testing in patients who cannot exercise adequately, and 3) the ability to quantify and localise areas of ischemia. 
The diagnostic endpoint of stress imaging for detection of inducible ischemia is either left ventricular (LV) wall motion [stress echocardiography; stress cardiac magnetic resonance (CMR)], or myocardial perfusion [myocardial perfusion scintigraphy using single photon emission computed tomography (SPECT) or positron emission tomography (PET); perfusion CMR].
Stress imaging may be performed either with exercise or pharmacologically.

• In general, exercise stress imaging is preferred over pharmacological testing if a patient is able to exercise to a sufficient cardiac workload feasible, because exercise can provide higher physiological stress and better correlation between a patient’s symptoms and physical work capacity than what would be achieved by pharmacological testing. 

Stress echocardiography

The sensitivity and specificity of stress echocardiography are 80-85% and 80-88%, respectively.
The diagnostic endpoint is wall thickening and new or worsening wall motion abnormalities with stress.
Additional criteria for an abnormal stress echocardiogram are dilatation and decline of global LV systolic function with stress. Intravenous contrast agents can improve endocardial border delineation and thus, diagnostic accuracy (7).
In patients unable to exercise adequately, the pharmacological agent of choice is the inotropic agent dobutamine. Dobutamine however, is contraindicated in the presence of severe hypertension, hemodynamically significant LV outflow obstruction, or sustained ventricular arrhythmias. Vasodilators such as dipyridamole are alternative (second-line) pharmacological agents.  
Advantages of stress echocardiography include 1) wide availability, 2) relatively low cost, and 3) lack of ionising radiation. Stress echocardiography provides valuable information regarding the distribution and extent of myocardial ischemia. Additionally, stress echocardiography is unique to provide clinically valuable insights regarding hemodynamics during stress (e.g., exercise parameters of diastolic function and valvular heart disease), and it has better spatial resolution compared with SPECT that allows for detection of subendocardial ischemia. Presence of resting regional wall abnormalities and inter-observer variability may limit its diagnostic accuracy.
Suboptimal image quality may be obtained as a result of patient body habitus (obesity), lung disease, or respiratory motion. 

Myocardial perfusion scintigraphy using SPECT

Technetium 99m is used more commonly than thallium 201 for SPECT. The sensitivity and specificity of exercise stress SPECT are 73-92% and 63-87%, respectively. The respective figures for vasodilator stress SPECT are 90% and 75-87%. In comparison, SPECT has better sensitivity whereas stress echocardiography has a higher specificity for diagnosis of SCAD. Reduced regional trace uptake during stress compared with preserved perfusion at rest is the hallmark of reversible myocardial ischemia; transient ischemic dilatation at stress and extensive stress-induced wall motion abnormalities are additional markers of significant CAD. 

Pharmacological stress SPECT

The agents of choice are vasodilators: adenosine, dipyridamole, or regadenoson. Dobutamine is preferred when vasodilators are contraindicated, such as in the presence of bronchospastic airway disease or severe obstructive pulmonary disease, significant hypotension, sick sinus syndrome, a high degree atrio-ventricular block, current use of aminophylline, and caffeine ingestion within 12 hours before testing. 
Advantages of SPECT over stress echocardiography include satisfactory quality imaging despite presence of lung disease, and reliable assessment of ischemia in the presence of resting wall motion abnormalities. If left bundle branch block or ventricular paced rhythm is present, exercise stress MPI may result in septal artifacts (false-positive septal defects) due to dyssynchronous septal motion, thus reducing diagnostic specificity (8). In these cases, vasodilators instead of exercise stress imaging (even if the patient can exercise) result in fewer false-positive results by reducing tachycardia and attenuating dyssynchrony-related artifacts.
Limitations of SPECT include patient exposure to non-trivial radiation, relatively long acquisition protocols and high cost, poor spatial resolution that limits detection of subendocardial ischemia, and artifacts related to respiratory motion, breast tissue, or sub-diaphragmatic attenuation. Because relative perfusion is assessed, SPECT has reduced sensitivity for detecting diffuse ischemia due to left main disease or 3-vessel disease.

Myocardial perfusion scintigraphy using PET

Although less well studied than SPECT, PET has better diagnostic accuracy for detection of CAD, including in women (9). Advantages of PET over SPECT include lower radiation exposure, higher resolution, and fewer attenuation artifacts that allow for better image quality even in obese patients. In addition, PET can uniquely quantify blood flow, thus allowing detection of microvascular angina. The main limitations of PET are limited availability and increased cost. 

Cardiac magnetic resonance stress imaging

The imaging endpoint of CMR for diagnosis of SCAD depends on the stress agent: CMR may detect either ischemia-induced wall motion abnormalities (dobutamine stress CMR) or myocardial perfusion (vasodilator stress CMR).
Dobutamine stress CMR has an estimated sensitivity and specificity of 79-88% and 81-91%, respectively, for detection of obstructive CAD and can be useful in patients with suboptimal image quality on stress echocardiography (e.g., poor acoustic window due to obesity or lung disease) (10).
Perfusion CMR has a sensitivity and specificity of 67-94% and 61-85%, respectively, and its accuracy is comparable to nuclear perfusion imaging (11, 12). Quantitative CMR perfusion measurements correlate well with fractional flow reserve measurements of hemodynamically significant coronary stenoses (13).
Advantages of CMR for perfusion imaging include lack of radiation, high spatial resolution, ability to perform absolute quantification of perfusion, limited operator dependence, signal characteristics that are largely independent of the patient’s body habitus, and additional information on cardiac structure and function provided in a comprehensive CMR study.
Limitations pertain mainly to high cost, limited availability and expertise, and limited functional analysis in the presence of arrhythmias.   

B) Non-invasive assessment of coronary anatomy

Improvements in temporal and spatial resolution of coronary computed tomography angiography (CTA) have enabled evaluation of coronary arteries with high image quality (14). Coronary CTA has a sensitivity of 95-99% and specificity of 64-83% for detection of obstructive CAD. The diagnostic performance is higher in patients within the lower range of intermediate PTP (15-50%) for SCAD, and it is therefore most useful for ruling out rather than confirming diagnosis of SCAD. Diagnostic accuracy appears to be similarly high among women (15).
Strict patient selection is crucial for satisfactory image quality and diagnostic accuracy. Coronary CTA should only be considered in individuals without severe obesity, with adequate breath holding capabilities, an Agatston calcium score <400, in sinus rhythm and with a heart rate <65 beats per minute (possibly achieved by administration of beta blockers). A major advantage of coronary CTA in appropriately selected patients is the very high negative predictive value that can reassure physicians to safely defer performance of ICA. Disadvantages of the method include relatively limited (yet growing) availability, high radiation, and reduced image quality in the presence of arrhythmias or elevated heart rate.   

II - Test-selection process 

As outlined above, available non-invasive modalities for diagnosis of SCAD differ distinctly in terms of diagnostic accuracies, predictive values, and costs.
In patients with clinically suspected SCAD, three key question need to be answered to determine patient suitability for non-invasive testing, type of stress (exercise vs. pharmacologic), and functional vs. anatomic testing endpoint:

1. What is the clinical PTP? 
2. Can the patient exercise adequately? 
3. Are there specific conditions precluding ECG interpretation, or patient characteristics (e.g., body habitus, comorbidities, or current medications) limiting the performance or interpretability of available imaging modalities? Local availability and expertise, radiation exposure, and the need for concomitant assessment of hemodynamics or valvular disease are also important factors affecting appropriate test selection in a given individual (16).  

Non-invasive testing is not recommended in patients who are deemed eligible for direct referral for ICA or unsuitable for specific cardiac testing (Figure 1). Furthermore, non-invasive modalities are not recommended for the purpose of establishing diagnosis in individuals with low (<15%) or high (>85%) PTP of SCAD.

• ECG stress test is recommended as the initial diagnostic step in patients with a PTP 15-65% of SCAD, provided that the patient can exercise adequately. 
• Stress imaging (either exercise or pharmacological) is preferred over ECG stress test if local expertise and availability permit and recommended overall as the initial test option for diagnosis of SCAD. 
• Stress imaging is definitely recommended and preferred over ECG exercise testing in specific patient groups, including patients within the higher range of intermediate PTP (66-85%); patients with LVEF <50% and atypical angina; and those with resting ECG abnormalities preventing accurate interpretation of stress-induced ECG changes. 

When functional stress imaging is chosen, selection of the exact modality for a given patient is based on specific characteristics that may adversely affect the performance or interpretability of a test. Basic principles regarding suitability of certain patient groups for certain testing modes are summarised below. It is important to remember, however, that for many patients there may not be a single appropriate test, but rather more than 1 or even all tests may be reasonably well suited. In the latter case, physician judgment and available local expertise should determine the correct test for an individual patient (17).

• Symptom-limited exercise is generally the preferred mode of stress in patients who can exercise to a satisfactory workload. Pharmacologic stress testing with imaging is typically performed in patients unable to exercise adequately.   
• In patients with LBBB, ventricular pacing, or significant resting wall motion abnormalities, vasodilator SPECT should be preferred (even if the patient can exercise), unless vasolilators are contraindicated.  
• In patients without significant resting wall motion abnormalities, either stress echocardiography or stress SPECT may be considered as the initial stress test, based primarily on local availability and expertise. It should be noted, however, that higher sensitivity may be more important that higher specificity, and stress SPECT may therefore be preferred, in patients with higher PTP of CAD. In contrast, higher specificity may be more important in patients with lower PTP, and thus stress echocardiography may be preferred in those patients. Stress PET combines high sensitivity and specificity, but it is limited by reduced availability and high cost.

The value of non-invasive anatomic evaluation of CAD as an alternative to functional testing is also emphasised if certain strict prerequisites are met.

• Coronary CTA should be considered as an alternative to stress imaging to rule out SCAD in patients with low intermediate PTP 15-50% (class IIa/C indication), provided that there is adequate technology, sufficient local expertise, and patients are considered suitable candidates, i.e., patient characteristics make a fully diagnostic CTA scan highly probable. 
• Coronary CTA should also be considered in patients within the lower range of intermediate PTP after an inconclusive exercise ECG test or stress imaging test, or in patients with contraindications to stress testing if ICA would otherwise be performed to rule out CAD. In patients with unclear results with any functional test or with coronary CTA, a second imaging test may be required to establish diagnosis according to patient characteristics and preference (Fig. 1). 

Conclusions 

In individuals with clinically suspected SCAD, possible strategies for initial diagnosis may range from no specific testing to non-invasive cardiac testing to direct referral for ICA.
Non-invasive testing to establish diagnosis of SCAD is appropriate for patient with an intermediate pre-test probability of disease.
Appropriate selection of the exact non-invasive modality that is most suitable for a given patient is based on clinical characteristics in combination with local resources and specific features of different testing modalities.      

Table 1. Clinical pre-test probabilities of SCAD in patients with stable chest pain symptoms. Adapted from (1).   

Fig. 1: Diagnostic triage of patients with suspected coronary artery disease.