Background
Current decisions to refer for angiographic coronary assessment are based on pain character, risk scores, stress testing, and occasionally calcium scoring. Carotid plaque has emerged as an effective vascular biomarker, but the cost and time of a full carotid ultrasound examination are disadvantageous. Focused vascular ultrasound (FOVUS) is a rapid limited assessment of carotid plaque that can be conducted by non-vascular-trained operators. The objective of the study was to determine the test characteristics of FOVUS for the assessment of significant coronary atherosclerosis in symptomatic patients referred for cardiac assessment.
Methods
In this prospective study, FOVUS was performed in 208 outpatients at low to intermediate risk undergoing same-day angiography. Carotid artery maximal plaque height was measured in each participant. A previously established receiver operating characteristic curve determined that a value of ≥1.5 mm was the threshold for significant angiographic coronary artery disease. FOVUS scan results, alone or combined with stress testing, were analyzed for the prediction of significant coronary artery disease.
Results
The negative predictive value and sensitivity of plaque height alone by FOVUS were found to be 77% and 93%, respectively. Adding the FOVUS scan result to stress testing significantly increased the negative predictive value and sensitivity of these traditional risk stratification tools.
Conclusions
Rapid carotid plaque height measurement by FOVUS enhanced atherosclerosis risk prediction in patients referred for cardiac assessment. Rapid plaque quantification had good negative predictive value and high sensitivity alone or in combination with stress testing. FOVUS may serve as a potential point-of-care ultrasound tool in the integrated assessment of cardiac pain.
Coronary angiography, an invasive procedure with low but important risks, remains the clinical standard for the diagnosis of coronary artery disease (CAD) in symptomatic patients. Recent studies have shown that up to 40% of angiographic examinations may have normal findings, resulting in low-risk patients’ being unnecessarily referred for angiography. In this era of cost restraints, there are calls for enhanced screening strategies with better negative predictive value (NPV) to minimize unnecessary procedures and better identify symptomatic patients with significant disease. On the other hand, a screening test for CAD must maximize the true-positive rate and have a cut point with high sensitivity to avoid missing this potentially fatal disease.
A formal, full carotid ultrasound test is deemed a class D indication by the US Preventive Services Task Force in asymptomatic patients as a screening tool for endarterectomy. However, this guideline is limited to the assessment of asymptomatic patients in the context of stroke and does not consider evidence demonstrating the utility of carotid plaque assessment as a vascular biomarker for cardiac risk screening. In a recent position paper by the European Society Cardiology Working Group on Peripheral Circulation, carotid ultrasound and plaque assessment was compared with established vascular biomarkers and recommended to be a highly effective method of cardiac risk screening given its high incremental value, clinical utility, and ease of use. Compared with coronary artery calcium scoring, carotid ultrasound is less costly, carries lower risk, and is highly correlated with coronary calcium. Carotid plaque burden is more closely related to coronary disease and more predictive of cardiovascular risk than carotid intima-media thickness (IMT), and it has previously been shown (along with plaque height) to predict coronary stenosis more accurately than IMT. Plaque height has previously been reported to predict cardiovascular risk. Some practitioners use real-time carotid atherosclerosis assessment as an office-based tool to integrate with traditional risk scoring and to educate and motivate patients if subclinical disease is present. Accordingly, the American Medical Association recently approved a category 1 reimbursement code for carotid IMT and plaque scanning.
To take advantage of the incremental value of carotid plaque assessment for cardiac risk screening without the associated cost of formal diagnostic carotid ultrasound, we hypothesize that a focused vascular ultrasound (FOVUS) protocol may serve as a low-cost screening method with high sensitivity and/or NPV for ruling out clinically significant coronary atherosclerosis in symptomatic patients. In this proof-of-concept study, we define the test characteristics of FOVUS for predicting important coronary atherosclerosis in a symptomatic population presenting for cardiac assessment.
Methods
Human and Animal Rights and Informed Consent
Informed consent was obtained for all participants in this study for FOVUS and the use of data for publication. This study conforms to the ethical guidelines of the 1975 Declaration of Helsinki, as reflected in a priori approval by the institution’s human research committee. The Queen’s University Health Sciences Research Ethics Board approved this study. The study was designed and implemented using the 14-item original Quality Assessment of Diagnostic Accuracy Studies tool to ensure the quality of diagnostic accuracy studies.
FOVUS
Focused cardiac ultrasound (FCU) is defined by the American Society of Echocardiography guidelines as a cardiac ultrasound using a protocol limited to a few specific views and conducted by operators without formal echocardiography training, usually using handheld or portable devices. The FOVUS protocol proposed is analogous to FCU in that it is conducted by non-vascular-trained operators and is limited to a few views (long- and short-axis views of the right and left carotid bulb) to rapidly screen for the presence of plaque. A single, simple measurement (maximal plaque height in the carotid bulb from the long axis) is conducted ( Figure 1 ). The FOVUS protocol does not include assessment of IMT, Doppler velocity, and calculation of percentage stenosis as described by the American Society of Echocardiography guidelines for a formal full diagnostic carotid ultrasound study.
The FOVUS scan was approximately 7 min in length and was limited to the long- and short-axis views of the right and left carotid bulb only. As a proof-of-concept study, scans were conducted by a single sonographer who received 15 hours of informal training in the FOVUS protocol and was not accredited in vascular imaging. Specifically, this was a typical midcareer registered cardiac sonographer (∼10 years’ experience) who conducts echocardiographic examinations on a daily basis but conducts no vascular imaging and has no formal vascular training. Scans were conducted using a vascular device (Vivid E9 Cardiovascular Ultrasound System, GE Healthcare, Mississauga, Canada) equipped with an 9L transducer for two-dimensional imaging. Scans were interpreted by a non-vascular-trained cardiologist. Additionally, images obtained were reviewed by a specialist with extensive expertise in vascular imaging and full carotid ultrasound performance (M.F.M.).
For each participant, the maximal plaque height (thickness) in the left and right carotid artery bulb was measured as previously described. Receiver operating characteristic curve analysis plaque height cutoff values specific to the total enrolled population ( n = 318) have been previously established. In this study, the previously derived threshold of ≥1.54 mm was used to indicate angiographically significant CAD. These values were nearly identical to the Atherosclerosis Risk in Communities definition for plaque height (≥1.5 mm). Patients were thus stratified for CAD using a plaque height ≥1.5 mm and were considered to have positive FOVUS results. All procedures were conducted and interpreted by investigators blinded to the other procedure results.
Study Design
This was a prospective study designed from its inception to test the incremental value of FOVUS to stress testing for the assessment of atherosclerosis. All stress tests in question were critical components of the patients’ index assessments leading to angiographic evaluation. Enrollment in the study occurred only after referral for angiography in order to generate a clinical standard (finding of atherosclerosis present or not).
Study Population
Three hundred eighteen outpatients for clinically indicated coronary angiography were approached prospectively from the Cardiac Catheterization Laboratory at Kingston General Hospital during a 6-month period in 2011 for FOVUS. In the participants recruited ( n = 208), indications for coronary angiography included angina, abnormal results on resting or exercise electrocardiography (ECG), and/or positive results on functional imaging stress tests. Inclusion criteria were age >18 years and no clinical contraindications to angiography or FOVUS. Recruited participants underwent same-day angiography and FOVUS to assess for plaque only. Participants who had complete scans and who had undergone stress tests within 3 months before angiography were studied.
Patients’ pretest likelihood (probability) of CAD was calculated on the basis of age, sex, and type of angina, according to the American Heart Association combined Forrester-Diamond and Coronary Artery Surgery Study data. If records had Canadian Cardiovascular Society angina pectoris grades of 1 to 3, patients were classified as having typical angina. To be conservative, patients with no grades (0) or who were not experiencing chest pain at referral were classified as having atypical angina. Symptoms other than chest pain requiring cardiologic assessment and leading to both invasive and noninvasive investigation included dyspnea, syncope, and palpitation (53 of 208 patients [25%]).
Coronary angiography was performed using the standard Judkins method with a GE System 2000 (GE Healthcare, Milwaukee, WI) by experienced interventional cardiologists. Luminal narrowing of coronary arteries was analyzed using a 16-segment model to produce an overall score, as previously defined. Angiographic scoring was as follows: 0 = no or minimal disease (0%-19% narrowing in any segment), 1 = mild disease (20%-49% narrowing in any segment), 2 = moderate disease (luminal narrowing of at least one segment of 50%-69%), and 3 = severe disease (≥70% narrowing within any segment of the main branches of the coronary artery or ≥50% in the left main coronary artery). Patients were grouped as having significant CAD (angiographic score of 2 or 3) or insignificant CAD (angiographic score of 0 or 1).
Stress Testing
Patients with stress tests (ECG, echocardiography, or nuclear stress testing) within the 3 months before angiography and FOVUS were analyzed ( n = 208). Test results were classified as either positive (high risk) or negative (low risk). Patients with inconclusive stress test results ( n = 2) were grouped with those with negative stress test results, whereas patients with intermediate stress test results ( n = 2) were grouped with those who were reported to have positive or abnormal results. For cases in which more than one stress test was conducted, if at least one test had positive results, patients were assigned to the high-risk group.
At our center, an inconclusive result is one that cannot be defined as either positive or negative by the interpreter because of factors such as inadequate visualization of the endocardium and failure to achieve target heart rate. An intermediate result was defined as one that was equivocal for ischemia, area of myocardium not corresponding to a coronary distribution, development of arrhythmia, augmentation of a hypokinetic segment with exercise, which despite achieving target heart rate and a study of adequate quality prevents a confident interpretation of a positive result. The stress tests were all performed by one of eight physicians in our group who interpret ECG or the imaging-based stress tests. Physicians interpreting the stress tests were unaware as to which patients would proceed to angiography and hence be potentially recruited for the FOVUS scan.
In this cohort, patients with negative stress test results were referred for clinically indicated angiography because of ongoing symptoms. The decision to refer to angiography was made by the patient’s treating physician in full context of the presenting symptoms and medical history.
Statistical Analysis
Statistical analyses were conducted using JMP version 11.0.0 (SAS Institute Inc, Cary, NC) unless otherwise specified. Contingency tables were used to determine the positive predictive value, NPV, sensitivity, and specificity of FOVUS and of the stress test in relation to the angiographic score for significant CAD (≥50% stenosis in at least one vessel, score 2 or 3). Stress test and FOVUS results, alone or combined, were analyzed for accuracy in stratifying patients for significant CAD. The Fisher’s exact test (two tailed) was used to compare nominal variables, and two-sample t tests (two tailed, unequal variances) were used for continuous variables. The McNemar test (exact significance two tailed) for two related samples was used to compare test sensitivities for accurate prediction of significant CAD, using SPSS Statistics version 22.0 (IBM, Armonk, NY). Statistical significance was accepted at α <0.05. Interrater reliability for carotid maximum plaque height was assessed from the images of a subset of 21 patients (both carotid arteries) rated by two experienced imaging technicians. Intrarater reliability was estimated from the same 21 patients measured by the same imaging technicians on two occasions. The Shrout-Fleiss correlation was used to assess the reliability of agreement.
Results
Study Population
Of the 318 patients enrolled, 208 patients had received at least one stress test before angiography and were analyzed as the study sample. The study sample was composed of 70% men and included a high percentage of patients with hypertension (69%), hyperlipidemia (72%), and family history of CAD (84%) ( Table 1 ). The group with significant CAD had a higher percentage of hyperlipidemia, known CAD, history of myocardial infarction, and smoking ( P < .02 for all). The group with significant CAD also had a significantly higher percentage on angiotensin-converting enzyme inhibitors and statins ( P < .001 for all). Pretest probabilities of CAD were not significantly different between patients with and those without CAD (81% vs 76%, respectively, P = .07). Of the 208 patients with stress tests, 64 underwent stress echocardiography, 136 stress ECG, and 45 nuclear stress tests.
| Variable | All patients ( n = 208) | With CAD ≥50% stenosis ( n = 140) | Without CAD <50% stenosis ( n = 68) | P |
|---|---|---|---|---|
| Men | 146 (70%) | 111 (79%) | 35 (51%) | <.001 ∗ |
| Age (y) | 63 ± 11 | 64 ± 10 | 60 ± 12 | .03 ∗ |
| Caucasian | 206 (99%) | 139 (99%) | 67 (99%) | .55 |
| Diabetes | 50 (24%) | 38 (27%) | 12 (18%) | .17 |
| Hypertension | 144 (69%) | 99 (71%) | 45 (66%) | .52 |
| Hyperlipidemia | 149 (72%) | 108 (77%) | 41 (60%) | .01 ∗ |
| Current tobacco use | 33 (16%) | 28 (20%) | 5 (7%) | .02 ∗ |
| Family history of CAD | 175 (84%) | 119 (85%) | 56 (82%) | .69 |
| Known CAD † | 77 (37%) | 70 (50%) | 7 (10%) | <.001 ∗ |
| History of MI | 33 (16%) | 30 (21%) | 3 (4%) | .001 ∗ |
| BMI (kg/m 2 ) | 30 ± 6 | 30 ± 6 | 30 ± 6 | .65 |
| Creatinine (μmol/L) | 89 ± 47 | 93 ± 55 | 80 ± 16 | .009 ∗ |
| Pretest probability of CAD (%) | 79 ± 17 | 81 ± 16 | 76 ± 18 | .07 |
| Medication use | ||||
| Aspirin | 130 (63%) | 89 (64%) | 41 (60%) | .65 |
| β-blockers | 84 (40%) | 63 (45%) | 21 (31%) | .07 |
| ACE inhibitors | 82 (40%) | 66 (47%) | 16 (24%) | .001 ∗ |
| Statins | 115 (55%) | 90 (64%) | 25 (37%) | <.001 ∗ |
| Diabetic medications | 40 (19%) | 32 (23%) | 8 (12%) | .063 |
| Angiotensin receptor blockers | 29 (14%) | 16 (11%) | 13 (19%) | .14 |
| Anticoagulant agents | 12 (6%) | 7 (5%) | 5 (7%) | .53 |
∗ Significant difference between patients with and those without CAD.
† The percentage of patients with known CAD was determined via self-report data (i.e., patients identified with known CAD were those who had been previously diagnosed verbally or suspected from a doctor or as noted from previous stress testing).
FOVUS
Maximum plaque height measured by FOVUS was calculated for patients with or without CAD and compared between patients with positive or negative stress test results ( Table 2 ). Irrespective of the stress test result, the mean plaque height in the CAD group was significantly higher than in patients without CAD ( Table 2 ). The NPV and sensitivity of FOVUS for significant CAD were found to be 77% and 93%, respectively, higher than stress testing alone for this group.
| With CAD ≥50% stenosis ( n = 140) | Without CAD <50% stenosis ( n = 68) | P | |
|---|---|---|---|
| All patients ( n = 208) | 140 (67%) | 68 (33%) | |
| FOVUS mean plaque height (mm) | 2.59 ± 0.89 | 1.67 ± 1.00 | <.001 ∗ |
| Stress test result positive ( n = 158) | 104 (66%) | 54 (34%) | |
| FOVUS mean plaque height (mm) | 2.63 ± 0.93 | 1.65 ± 0.99 | <.001 ∗ |
| Stress test result negative ( n = 50) | 36 (72%) | 14 (28%) | |
| FOVUS mean plaque height (mm) | 2.48 ± 0.76 | 1.76 ± 1.08 | .03 ∗ |
∗ Significant difference between patients with or without CAD.
Next, we studied the test characteristics of FOVUS by stratifying patients as having positive or negative stress test results. We found that 66% of patients with positive results (104 of 158) had significant CAD. In the group with positive results, maximum plaque height was significantly higher if they had significant CAD ( P < .001).
Of the 50 patients with negative stress test results who still underwent angiography, 36 (72%) were diagnosed with significant CAD. Of these 36 patients, 34 were found to have plaque ≥1.5 mm. Thus, the presence of plaque ≥1.5 would have predicted significant CAD in 94% of patients with false-negative stress test results in this cohort. As expected, plaque height was significantly greater in patients with CAD (2.48 mm) compared with those without CAD (1.76 mm) ( P = .03). Maximum predicted heart rate was achieved in the majority of cases.
Significant plaque was found bilaterally in 105 patients. As expected there were several-fold more patients with bilateral plaque in patients with significant CAD ( n = 85) compared with no or mild disease ( n = 20). The presence of bilateral plaque may be an indicator of CAD severity.
Finally, we studied the test characteristics of FOVUS in combination with stress testing. We assessed the utility of an integrated approach to predicting CAD ( Table 3 ) using the angiographic score as the clinical standard. When comparing the different types of stress tests alone (i.e., stress echocardiography, exercise ECG, and nuclear stress), nuclear stress testing had the highest sensitivity (83%) but the lowest NPV (17%) ( Table 3 ). When combining stress test results with the plaque height obtained by FOVUS, the sensitivity increased significantly no matter what type of stress test was performed. Sensitivity increased from 59% to 100%, 74% to 98%, and 83% to 97% when plaque height was added to stress echocardiography, exercise ECG, and nuclear testing, respectively. Similarly, the NPV of stress echocardiography and exercise ECG increased from 29% to 100% and 36% to 78%, respectively. Although the results indicated an NPV of 0% when plaque height was added to nuclear stress testing, this value is driven by the fact that all of the 45 patients who underwent nuclear testing were stratified as having significant CAD on the basis of angiographic scores, so there were no true negatives in that group. Overall, the combination with the best profile considering sensitivity, NPV, specificity, positive predictive value, and likelihood ratios was the combination of stress echocardiography with FOVUS to measure plaque height ( Table 3 ).
| Test | n | Test result | With CAD ≥50% stenosis | Without CAD <50% stenosis | PPV | NPV | Sensitivity | Specificity | LR+ | LR− | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Plaque height by FOVUS | 208 | ≥1.5 mm (+) | TP | 130 | FP | 34 | 79% | 77% | 93% | 50% | 1.86 | 0.14 |
| <1.5 mm (−) | FN | 10 | TN | 34 | ||||||||
| Stress echocardiography | 64 | + | TP | 24 | FP | 16 | 60% | 29% | 59% | 30% | 0.84 | 1.36 |
| − | FN | 17 | TN | 7 | ||||||||
| Exercise ECG | 136 | + | TP | 67 | FP | 33 | 67% | 36% | 74% | 28% | 1.04 | 0.90 |
| − | FN | 23 | TN | 13 | ||||||||
| Nuclear test | 45 | + | TP | 24 | FP | 15 | 62% | 17% | 83% | 6% | 0.88 | 2.76 |
| − | FN | 5 | TN | 1 | ||||||||
| Stress echocardiography + plaque height | 64 | + | TP | 41 | FP | 20 | 67% | 100% | 100% | 13% | 1.15 | 0.00 |
| − | FN | 0 | TN | 3 | ||||||||
| Exercise ECG + plaque height | 136 | + | TP | 88 | FP | 39 | 69% | 78% | 98% | 15% | 1.15 | 0.15 |
| − | FN | 2 | TN | 7 | ||||||||
| Nuclear test + plaque height | 45 | + | TP | 28 | FP | 16 | 64% | 0% | 97% | 0% | 0.97 | ∗ |
| − | FN | 1 | TN | 0 | ||||||||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
