Medical Policy
Medical Policy
Policy Num: 11.003.021
Policy Name: Gene Expression Testing in the Evaluation of Patients with Stable Ischemic Heart Disease
Policy ID: [11.003.021][Ar B M- P- ][2.04.72]
Last Review: April 20, 2021
Next Review: Policy Archived
Issue: April:2021
| Population Reference No | Populations | Interventions | Comparators | Outcomes |
|---|---|---|---|---|
| 1 | Individuals:
| Interventions of interest are:
| Comparators of interest are:
| Relevant outcomes include:
|
Expression levels of various genes in circulating white blood cells or whole blood samples have been reported to discriminate between cases of obstructive coronary artery disease and healthy controls. Multiplex gene expression testing has been combined with other risk factors to estimate the likelihood of obstructive coronary artery disease in patients who present with stable ischemic heart disease. These tests have the potential to improve the accuracy of predicting coronary artery disease. A commercially available test, Corus CAD, has been developed for this purpose without diabetes or inflammatory conditions.
For individuals who have suspected stable ischemic heart disease without diabetes or inflammatory conditions who receive gene expression testing, the evidence includes retrospective case-control and prospective cohort studies. Relevant outcomes are overall survival, disease-specific survival, test validity, change in disease status, morbid events, and resource utilization. The diagnostic pathway for coronary artery disease includes information from medical history, along with age and sex, stress testing, and imaging. Newer noninvasive methods are being tested, such as gene expression testing. It is not clear how the Corus CAD gene expression test fits in the current diagnostic pathway and how results would be used to change current guideline-based risk stratification before and/or after other noninvasive testing. Results of 2 validation studies (Personalized Risk Evaluation and Diagnosis In the Coronary Tree [PREDICT], Coronary Obstruction Detection by Molecular Personalized Gene Expression [COMPASS]) have reported that the test may improve coronary artery disease prediction beyond the Diamond-Forrester prediction model. In the COMPASS study, the sensitivity and negative predictive value of the Corus CAD score in diagnosing obstructive coronary artery disease was superior to myocardial perfusion imaging in patients referred for myocardial perfusion imaging testing. However, in that study, the reported sensitivity of myocardial perfusion imaging was considerably lower than that generally reported in the literature. Neither PREDICT nor COMPASS used the guideline definition of obstructive coronary artery disease as the reference standard and had relatively few patients at intermediate risk based on clinical prediction rules. The sensitivity and negative predictive value of clinical models were not reported. An analysis of a cohort from the PROspective Multicenter Imaging Study for Evaluation of chest pain (PROMISE) trial including patients with an intermediate pretest probability of obstructive coronary artery disease confirmed a high, negative predictive value for the Corus CAD score. The test also has been shown to have some predictive ability of future revascularization; too few major cardiac events have been observed during the limited duration of follow-up to assess predictive ability for that outcome. Evidence for the Corus CAD score has not directly demonstrated that the test is clinically useful and a chain of evidence cannot be constructed to support its utility. The evidence is insufficient to determine the effects of the technology on health outcomes.
The objective of this evidence review is to determine whether gene expression testing in patients with stable ischemic heart disease improves the net health outcome compared with standard clinical evaluation, including established noninvasive testing.
Gene expression testing in the evaluation of patients with stable ischemic heart disease is considered investigational for all indications, including but not limited to prediction of coronary artery disease in stable, nondiabetic patients.
Please see the Codes table for details.
BlueCard/National Account Issues
Some Plans may have contract or benefit exclusions for genetic testing.
Heart disease is the leading cause of death in the United States, accounting for approximately one-third of all deaths in people over age 35.1, The death rate is higher in men compared with women, and in blacks compared with whites but lower in Hispanic populations compared with blacks and whites. The most common form of heart disease is ischemic heart disease, also known as coronary artery disease.
Angina is the first symptom of coronary artery disease in approximately 50% of patients. However, women and the elderly are more likely to present with atypical symptoms such as nausea, vomiting, gastric discomfort, or atypical chest pain, which makes diagnosis more challenging.2,
Patients with signs and symptoms of obstructive coronary artery disease may be evaluated with a variety of tests according to prior risk. Coronary angiography is the criterion standard for diagnosing obstructive coronary artery disease but it is invasive and associated with a low but finite risk of harm. Coronary angiography also has a relatively low yield. In a study of nearly 400,000 patients without known coronary artery disease undergoing elective coronary angiography, approximately 38% were positive for obstructive coronary artery disease (using the coronary artery disease definition, ³50% stenosis of the diameter of the left main coronary artery or ³70% stenosis of the diameter of a major epicardial or branch vessel >2.0 mm in diameter) and 41% if using the broader definition (³50% stenosis in any coronary vessel).3, Thus, methods of improving patient risk prediction before invasive coronary angiography are needed.
In an initial proof-of-principle study of the Corus CAD score in patients referred for invasive coronary angiography, Wingrove et al (2008) evaluated 27 cases (96% symptomatic) with and 14 controls without angiographically defined coronary artery disease for expression of genes that differed significantly between the 2 groups, selecting 50 genes.4, To that, authors added 56 genes selected from relevant literature reports and evaluated the expression of these 106 genes in an independent set of 63 cases and 32 controls, resulting in the selection of 14 genes that independently and significantly discriminated between groups in multivariable analysis. The significance of 11 of these 14 genes was replicated in the third set of 86 cases and 21 controls. Expression of the 14 genes was proportional to maximal coronary artery stenosis in the combined cohort of 215 patients.
In 2011, Elashoff et al described the final Corus CAD score development.5, Investigators conducted 2 successive case-control gene expression discovery studies using samples from independent cohorts. Cases were angiographically defined as 75% or greater maximum stenosis in 1 major vessel, or 50% or greater in 2 vessels, and controls defined as less than 25% stenosis in all major vessels. Of clinical factors, diabetes had the most significant effect on gene expression; in the first case-control study in symptomatic patients (CATHeterization GENetics; n=195), expression of 42 genes in nondiabetic patients and 12 genes in diabetic patients were found to (p<0.05) discriminate significantly between cases and controls with no overlap. As a result, the second case-control study, in a subset of 198 patients from the prospective Personalized Risk Evaluation and Diagnosis In the Coronary Tree study, and final development of the assay was limited to nondiabetic patients (62% symptomatic). The participants were 76% male and 89% white. Final variable selection comprised the expression of 20 coronary artery disease associated genes, 3 normalization genes, and terms for age and sex. The majority of the selected genes were immune and inflammatory-related. All terms were incorporated into an algorithm that resulted in an obstructive coronary artery disease score ranging from 1 to 40.
Regulatory Status
CardioDX has closed operations and the Corus CAD test is no longer available.
This evidence review was created in June 2011 and has been updated regularly with searches of the MEDLINE database. The most recent literature update was performed through January 15, 2020.
Evidence reviews assess whether a medical test is clinically useful. A useful test provides information to make a clinical management decision that improves the net health outcome. That is, the balance of benefits and harms is better when the test is used to manage the condition than when another test or no test is used to manage the condition.
The first step in assessing a medical test is to formulate the clinical context and purpose of the test. The test must be technically reliable, clinically valid, and clinically useful for that purpose. Evidence reviews assess the evidence on whether a test is clinically valid and clinically useful. Technical reliability is outside the scope of these reviews, and credible information on technical reliability is available from other sources.
Population Reference No. 1 Policy Statement
The 2012 joint guidelines by the American College of Cardiology Foundation and 6 other medical associations on the diagnosis of stable ischemic heart disease provides details on the diagnostic pathway for evaluation and treatment of heart disease. The pathway is summarized in Figure 1 and in the following paragraphs. When patients present with signs and symptoms of obstructive coronary artery disease, the estimated risk (or pretest probability) of obstructive coronary artery disease is estimated using clinical characteristics such as age, sex, type of angina symptoms, smoking, and other comorbidities (eg, diabetes, hyperlipidemia).2,6, The guidelines provide a table of pretest probabilities of coronary artery disease by age, sex, and type of angina adapted from the Diamond-Forrester tool.2, For example, a woman aged 30 to 39 with nonanginal chest pain has a 4% pretest probability of coronary artery disease and a man aged 60 to 69 with typical anginal chest pain has a 94% pretest probability of coronary artery disease.
For patients initially assessed at low-risk (<10% pretest probability of obstructive coronary artery disease, no further testing is generally needed, and the patient can be observed and treated with medical therapy.2, Patients at high-risk of obstructive coronary artery disease may proceed to coronary angiography if the symptoms or findings suggest a high-risk lesion.
The classification of intermediate-risk varies in the literature but is frequently defined as a pretest probability between 10% and 90%. In patients with an intermediate pretest probability of obstructive coronary artery disease, noninvasive diagnostic methods, such as exercise or pharmacologic stress tests with or without imaging methods such as myocardial perfusion imaging, or coronary computed tomographic angiography may be recommended. The noninvasive testing used depends on patient characteristics such as the ability to exercise, electrocardiographic results, and other comorbidities as well as local expertise, availability of the testing modality, and patient preference. Some noninvasive imaging methods have potential risks of exposure to radiation and contrast material. After noninvasive testing, patients initially classified as having an intermediate pretest probability of obstructive coronary artery disease are further risk-stratified based on the estimated risk of coronary event or death using clinical data and results of noninvasive testing. The 2012 American College Cardiology Foundation joint guidelines also provide risk stratification following noninvasive testing.2,For example, severe stress-induced left ventricular dysfunction (peak exercise left ventricular ejection fraction <45% or drop in left ventricular ejection fraction with stress ≥10%) indicates a high (>3%) annual risk of death or myocardial infarction; a 1-mm ST-segment depression occurring with exertional symptoms indicates an intermediate (1% to 3%) annual risk of death or myocardial infraction; a normal stress or no change of limited resting wall motion abnormalities during stress indicates a low-risk (<1%) annual risk of death or myocardial infraction. Patients at high-risk of coronary event or death following noninvasive testing may proceed to coronary angiography.
CardioDx, the manufacturer of the gene expression score (Corus CAD), has stated the test “complements and improves the current noninvasive assessment” of suspected obstructive coronary artery disease. The manufacturer-supported registry collects data in the primary care setting and a decision impact study using registry data has suggested that the test may be used to identify stable, nonacute outpatients presenting with symptoms suggestive of obstructive coronary artery disease who can safely forgo referral to cardiology or advanced cardiac testing.7, Other studies have been performed in patients who have been referred for invasive angiography and myocardial perfusion imaging.
The question addressed in this evidence review is: Does gene expression testing in patients with stable ischemic heart disease improve the net health outcome compared with standard clinical evaluation?
The following PICO was used to select literature to inform this review.
The intended population is patients with suspected ischemic heart disease with stable angina. The manufacturer states that appropriate patients are those who do not have diabetes, without systemic infectious or systemic inflammatory conditions, and who are not currently taking steroids, immunosuppressive agents, or chemotherapeutic agents. The intended use population might be all such patients or a subset of them identified by risk stratification, depending on exactly how the test fits into the diagnostic pathway.
A gene expression score classifier (Corus CAD) has been developed based on expression levels derived from the previously described studies, in whole blood samples, of 23 genes plus patient age and sex. This information is used in an algorithm to produce a score from 1 to 40, with higher values associated with a higher likelihood of obstructive coronary artery disease. A score of less than 15 has been used to indicate a low-risk of obstructive coronary artery disease.
Blood for the test is collected using a routine blood draw and stored between 2° and 10°C for up to 1 day before shipping to the CardioDx Commercial Laboratory, which is certified by Clinical Laboratory Improvement Amendments (CLIA) and accredited by the College of American Pathologists. The results are available within a few days.
The intervention of interest for assessing validity would be Corus CAD score added to current risk prediction models..
The comparator would be clinical risk prediction models alone that estimate the pretest probability of obstructive coronary artery disease (eg, Diamond-Forrester). Noninvasive testing would be a comparator for determining whether a patient would be referred for coronary angiography.
The reference standard for diagnosing obstructive coronary artery disease is coronary angiography with obstructive coronary artery disease defined as any stenosis 50% or greater in the left main coronary artery or 70% or greater in any other coronary artery according to joint guidelines from the American College of Cardiology Foundation, the American Heart Association, and the Society for Cardiovascular Angiography and Interventions.8, However, this is also an imperfect reference standard for the outcome of a cardiac event.
Beneficial outcomes resulting from a true-negative test result are avoiding unnecessary subsequent testing. Harmful outcomes resulting from a false-positive test result are unnecessary noninvasive and invasive testing or receiving unnecessary treatment. Harmful outcomes resulting from a false-negative test result are increased risk of cardiovascular events and death.
In Figure 1, (ie, a triage “rule-out” test), the test would need to identify precisely a group of patients that could safely forgo additional noninvasive testing; therefore, the sensitivity, negative predictive value and negative likelihood ratio are key test performance characteristics.
The time period of interest for measuring the diagnostic performance is the time to obstructive coronary artery disease diagnosis. For assessing cardiovascular outcomes, 2.5 years is consistent with the PROspective Multicenter Imaging Study for Evaluation of chest pain (PROMISE) trial, which compared diagnostic strategies for coronary artery disease.9,
There are 3 core characteristics for assessing a medical test. Whether imaging, laboratory, or other, all medical tests must be:
Technically reliable
Clinically valid
Clinically useful.
Assessment of technical reliability focuses on specific tests and operators and requires review of unpublished and often proprietary information. Review of specific tests, operators, and unpublished data are outside the scope of this evidence review, and alternative sources exist. This evidence review focuses on the clinical validity and clinical utility.
A test must detect the presence or absence of a condition, the risk of developing a condition in the future, or treatment response (beneficial or adverse).
Characteristics and results of clinical validity studies evaluating the performance of the Corus CAD score for diagnosing obstructive coronary artery disease are shown in Tables 1 and 2. Four studies reported the performance characteristics for Corus CAD for diagnosing obstructive coronary artery disease. Voora et al (2017), (PROMISE) was the largest study and it used the American Heart Association definition for obstructive coronary artery disease.10, In this population of patients referred for nonurgent, noninvasive testing, the sensitivity was 73% (95% confidence interval, 64% to 81%), the negative likelihood ratio was 0.56 (95% confidence interval, 0.42 to 0.77), and the negative predictive value was 94% (95% confidence interval, 92% to 96%). The Rosenberg et al (2010), (Personalized Risk Evaluation and Diagnosis In the Coronary Tree [PREDICT])11, and Thomas et al (2013),(Coronary Obstruction Detection by Molecular Personalized Gene Expression [COMPASS])12, studies used a broader definition of obstructive coronary artery disease and enrolled few patients at intermediate risk (18% and 17%, respectively) based on clinical risk prediction rules. The sensitivities were 85% (95% confidence interval, 79% to 90%) and 89% (95% confidence interval, 78% to 95%) in PREDICT and COMPASS, respectively while the negative predictive value rates were 83% (95% confidence interval, 77 to 89) and 96% (95% confidence interval, 93% to 99%). The thresholds used to identify obstructive coronary artery disease were not clear in Ladapo et al (2017).7, The studies are described in more detail in the following paragraphs.
Corus CAD score was validated in the prospective multicenter PREDICT study (2010) in which blood samples were collected from 526 nondiabetic patients who did not have systemic infectious or inflammatory conditions and who were not receiving immunosuppressive or chemotherapeutic agents with a clinical indication for coronary angiography but no known previous myocardial infraction, revascularization, or obstructive coronary artery disease (71% symptomatic).11,This is the same cohort from which the second assay development case-control cohort was drawn.5, Patients were sequentially allocated to development and validation sets. The development cohort was 58% male and 87% white. The validation cohort is described in the tables. Investigators defined obstructive coronary artery disease as 50% or greater stenosis in 1 or more major coronary arteries on quantitative coronary angiography, which they stated corresponded to 65% to 70% stenosis on clinical angiography. PREDICT compared the predictive accuracy of the gene expression score test with clinical predictors and myocardial perfusion imaging stress testing. A 2014 follow-up publication, including patients from the gene discovery and algorithm development cohorts in combination with the validation cohort (n=1038), reported similar performance.13,
In another follow-up from PREDICT, Lansky et al (2012) found that the Corus CAD score was an independent predictor of coronary artery disease in multivariate analysis, with odds ratios of 2.53 (p=0.001) for the total study population and 1.99 (95% confidence interval, 1.35 to 2.96; p=0.001) and 3.45 (95% confidence interval, 1.97 to 5.91; p=0.001) for males and females, respectively.14, In this analysis, myocardial perffusion imaging was not associated with any measures of coronary artery disease in the general population or when stratified by sex.
Thomas et al (2013) assessed the clinical validity and utility of the Corus CAD score for detection of obstructive coronary artery disease in symptomatic, nondiabetic patients without inflammatory conditions in a multicenter, prospective study, COMPASS.12, Obstructive coronary artery disease was defined as 50% or greater stenosis in 1 or more major coronary arteries on quantitative coronary angiography. The COMPASS sample base differed from the PREDICT sample by including patients who had received a referral for myocardial perfusion imaging but had not been referred for invasive coronary angiography. Myocardial perfusion imaging positive participants underwent invasive coronary angiography based on clinician judgment, and all other participants received coronary computed tomographic angiography. Of 537 enrolled patients, only 431 (80%) were evaluable, primarily due to refusal to undergo invasive coronary angiography or coronary computed tomographic angiography. The performance characteristics for myocardial perfusion imaging (core-lab) in this population were also provided as follows: sensitivity, 36% (95% confidence interval, 24% to 50%); specificity, 90% (95% confidence interval, 87% to 93%); positive predictive value, 41% (95% confidence interval, 28% to 56%); and negative predictive value, 88% (95% confidence interval, 84% to 92%). The sensitivity of myocardial perfusion imaging in COMPASS was lower than generally reported in the literature. In 2013, Ladapo et al reported simulation analyses demonstrating how referral bias could have influenced the performance characteristics that have been reported in the literature.15,
Voora et al (2017) evaluated the Corus CAD score in a cohort from the PROMISE trial funded by National Heart, Lung, and Blood Institute.10, PROMISE was a randomized controlled trial (RCT; 2015) that enrolled 10,003 outpatients who were randomized to functional (ie, exercise, echocardiographic, or nuclear stress testing) or anatomic (ie, computed tomographic angiography) diagnostic testing.16, Patients were symptomatic and at increased risk for coronary artery disease based on age and/or the presence of coronary artery disease risk factors, and presented with symptoms suggestive of obstructive coronary artery disease. An ancillary analysis of PROMISE patients was supported in part by the manufacturer and included 2,370 PROMISE patients without diabetes who were not on anti-inflammatory medications and who had samples in the biorepository of sufficient quality for analysis. The definition of obstructive coronary artery disease was 70% or more stenosis in a major coronary artery or 50% or more left main stenosis using computed tomographic angiography data.
Several studies have evaluated Corus CAD in a cohort of patients from A Registry to Evaluate Patterns of Care Associated with the Use of Corus CAD in Real World Clinical Care Settings(PRESET) registry. The PRESET registry is funded by the manufacturer. This registry enrolled patients from 21 primary care practices in the United States between August 2012 and August 2014. Patients had nonacute chest pain and typical or atypical symptoms of obstructive coronary artery disease without history of myocardial infraction or revascularization, diabetes, suspected acute myocardial infraction, high-risk unstable angina pectoris, New York Heart Association class III or IV heart failure symptoms, cardiomyopathy with an ejection fraction of 35% or less, severe cardiac valvular diseases, current systemic infectious or inflammatory condition, or recent treatment with an immunosuppressive or chemotherapeutic agent. A report by Ladapo et al (2017) is primarily focused on physician decision-making but includes a table of the Corus CAD score and advanced cardiac testing results for obstructive coronary artery disease in 84 patients.7, Therefore, those data are included in the following tables. Subsequent reports focused on adults aged 65 and older (n=176) and women of all ages (n=288) with stable symptoms suggestive of obstructive coronary artery disease, showing higher referral rates for patients with a higher Corus CAD score.17,18,
| Study | Study Populationa | Design | Reference Standard for Obstructive CAD | Threshold Score for Positive Corus CAD Score Test | Timing of Reference and Corus CAD Score Tests | Blinding of Assessors | Comment |
| Rosenberg et al (2010)11, PREDICT |
| Prospective | ≥50% stenosis in ≥1 major coronary arteries by quantitative CA | 14.75 | Blood samples drawn before CA | Yes |
|
| Thomas et al (2013)12, COMPASS |
| Prospective | ≥50% stenosis in ≥1 major coronary arteries by quantitative CA or CCTA | 15 | Blood samples drawn before MPI and CA | Yes |
|
| Voora et al (2017)10,PROMISE |
| Nonconcurrent, prospective | ≥70% stenosis in a major coronary artery or ≥50% left main stenosis using CCTA | 15 | Blood samples drawn before CA | Yes |
|
| Ladapo et al (2017)7, PRESET |
| Prospective | Cardiac stress test or ICA (thresholds NR) | 15 | Blood samples drawn before further testing | NR |
|
CA: coronary angiography; CAD: coronary artery disease; CCTA: coronary computed tomographic angiography; COMPASS: Coronary Obstruction Detection by Molecular Personalized Gene Expression; ICA: invasive coronary angiography;MPI: myocardial perfusion imaging; NHLBI: National Heart, Lung, and Blood Institute; NR: not reported; PREDICT: Personalized Risk Evaluation and Diagnosis In the Coronary Tree; PRESET: A Registry to Evaluate Patterns of Care Associated with the Use of Corus CAD in Real World Clinical Care Settings;PROMISE: PROspective Multicenter Imaging Study for Evaluation.a In all studies, patients were nondiabetic, without inflammatory conditions, and were not receiving immunosuppressive or chemotherapeutic agents.
| Study | Initial N | Final N | Excluded Samples | Prevalence of Obstructive CAD | Sensitivity (95% CI), % | Specificity (95% CI), % | PPV (95% CI), % | NPV (95% CI), % | AUC |
| Reference standard: ≥50% stenosis in ≥1 major coronary arteries by quantitative CA | |||||||||
| Rosenberg et al (2010)11, PREDICT | 649 | 525 |
| 37% | 85 (79 to 90)a | 43 (38 to 49)a | 46 (41 to 52)a | 83 (77 to 89)a | 0.70 (NR) |
| Thomas et al (2013)12, COMPASS | 537 | 431 |
| 15% | 89 (78 to 95)b | 52 (47 to 57)b | 24 (19 to 30)b | 96 (93 to 99)b | 0.79 (0.72 to 0.84) |
| Reference standard: ≥70% stenosis in a major coronary artery or ≥50% left main stenosis using CCTA | |||||||||
| Voora et al (2017)10, PROMISE | 2370 | 1137 |
| 10% | 73 (64 to 81)a | 48 (45 to 51)a | 14 (11 to 17)a | 94 (92 to 96)a | 0.63 (0.57 to 0.68) |
| Reference standard: cardiac stress test or ICA (thresholds NR) | |||||||||
| Ladapo et al (2017)7, PRESET | 126 | 84 |
| 12% | 100 (59 to 100)a | 18 (10 to 28)a | 14 (7 to 25)a | 100 (66 to 100)a | NR |
AUC: area under the curve; CA: coronary angiography; CAD: coronary artery disease; CI: confidence interval; CCTA: coronary computed tomography angiography; COMPASS: Coronary Obstruction Detection by Molecular Personalized Gene Expression; CTA: computed tomography angiography; ICA: invasive coronary angiography; MPI: myocardial perfusion imaging; NPV: negative predictive value; NR: not reported; PPV: positive predictive value; PREDICT: Personalized Risk Evaluation and Diagnosis In the Coronary Tree.PRESET: A Registry to Evaluate Patterns of Care Associated with the Use of Corus CAD in Real World Clinical Care Settings; PROMISE: PROspective Multicenter Imaging Study for Evaluation.a CIs not reported in publication; calculated based on data provided.b The performance characteristics for MPI (core-lab) in this population were also provided: sensitivity, 36% (95% CI, 24% to 50%); specificity, 90% (95% CI, 87% to 93%); PPV, 41% (95% CI, 28% to 56%); and NPV, 88% (95% CI, 84% to 92%).
Relevance, design and conduct limitations in the studies are described in Tables 3 and 4.
| Study | Population | Intervention | Comparator | Outcomes | Duration of Follow-Up |
| Rosenberg et al (2010)11, PREDICT | 2. Test use in current diagnostic pathway unclear4. Study only includes patients referred for ICA and only 18% of patients were at intermediate risk5. Racial minorities were not well-represented | None noted | 2. Used broad obstructive CAD definition | 3. Diagnostic performance characteristics not provided for clinical risk models; performance characteristics by sex not provided | None noted |
| Thomas et al (2013)12, COMPASS | 2. Test use in current diagnostic pathway unclear4. Only 17% of patients were at intermediate risk5. Racial minorities were not well-represented | None noted | 2. Used broad obstructive CAD definition | 3. Diagnostic performance characteristics not provided for clinical risk models performance characteristics by gender not provided | None noted |
| Voora et al (2017)10, PROMISE | 2. Test use in current diagnostic pathway unclear5. Racial minorities were not well-represented | None noted | 3. Performance characteristics for comparators not provided | 3. Diagnostic performance characteristics calculated based on data provided; performance characteristics not provided for clinical risk models; performance characteristics by sex not provided | None noted |
| Ladapo et al (2017)7, PRESET | 2. Test use in current diagnostic pathway unclear | None noted | 1. Thresholds for diagnosis not given | 3. Diagnostic performance characteristics not provided for clinical risk models; performance characteristics by sex not provided | None noted |
| Key | 1. Intended use population unclear2. Clinical context for test is unclear3. Study population unclear4. Study population not representative of intended clinical use5. Study population is subpopulation of intended use | 1. Classification thresholds not defined2. Version used unclear3. Not version currently in clinical use | 1. Classification thresholds not defined2. Not compared to credible reference standard3. Not compared to other tests in use for same purpose | 1. Study does not directly assess a key health outcome2. Evidence chain or decision model not explicated3. Key clinical validity outcomes not reported (sensitivity, specificity, predictive values)4. Reclassification of diagnostic or risk categories not reported5. Adverse events of the test not described (excluding minor discomforts and inconvenience of venipuncture or noninvasive tests) | 1. Follow-up duration not sufficient with respect to natural history of disease (TP, TN, FP, FN cannot be determined) |
CAD: coronary artery disease; COMPASS: Coronary Obstruction Detection by Molecular Personalized Gene Expression; FN: false negative; FP: false positive; ICA: invasive coronary angiography; PREDICT: Personalized Risk Evaluation and Diagnosis In the Coronary Tree; PRESET: A Registry to Evaluate Patterns of Care Associated with the Use of Corus CAD in Real World Clinical Care Settings; PROMISE: PROspective Multicenter Imaging Study for Evaluation; TN: true negative; TP: true positive.
| Study | Selection | Blinding | Delivery of Test | Selective Reporting | Completeness of Follow-Up | Statistical |
| Rosenberg et al (2010)11, PREDICT | None noted | None noted | None noted | None noted | None noted | 1. CIs not reported, calculated based on data provided |
| Thomas et al (2013)12, COMPASS | None noted | None noted | None noted | None noted | 2. 90 patients with negative MPI refused CTA and were excluded; no description of these patients was provided | None noted |
| Voora et al (2017)10, PROMISE | None noted | None noted | None noted | None noted | None noted | 1. CIs not reported, calculated based on data provided2. No comparison to noninvasive testing provided |
| Ladapo et al (2017)7, PRESET | None noted | 1. Blinding not reported | None noted | None noted | None noted | 1. CIs not reported, calculated based on data provided2. No comparison to noninvasive testing provided |
| Key | 1. Selection not described2. Selection not random nor consecutive (ie, convenience) | 1. Not blinded to results of reference or other comparator tests | 1. Timing of delivery of index or reference test not described2. Timing of index and comparator tests not same3. Procedure for interpreting tests not described4. Expertise of evaluators not described | 1. Not registered2. Evidence of selective reporting3. Evidence of selective publication | 1. Inadequate description of indeterminate and missing samples2. High number of samples excluded3. High loss to follow-up or missing data | 1. CIs and/or p values not reported2. No statistical test reported to compare to alternatives |
CAD: coronary artery disease; CI: confidence interval; COMPASS: Coronary Obstruction Detection by Molecular Personalized Gene Expression; CTA: computed tomography angiography; MPI: myocardial perfusion imaging; PREDICT: Personalized Risk Evaluation and Diagnosis In the Coronary Tree;PRESET: A Registry to Evaluate Patterns of Care Associated with the Use of Corus CAD in Real World Clinical Care Settings; PROMISE: PROspective Multicenter Imaging Study for Evaluation.
Net reclassification for the Corus CAD score compared with other tests for the diagnosis of obstructive coronary artery disease was performed in Rosenberg et al (2010)11, and Thomas et al (2013)12, and are shown in Table 5 below. In Rosenberg et al (2010), the Corus CAD, Diamond-Forrester, and expanded clinical model scores were prospectively categorized as low (0% to <20%), intermediate (≥20% to <50%), or high (≥50%) risk for obstructive coronary artery disease. Myocardial perfusion imaging results were categorized as negative (no defect or possible fixed or reversible defect) or positive (fixed or reversible defect). In Thomas et al (2013), Corus CAD scores were categorized as low (≤15), intermediate (16-27), and high (≥28). The Diamond-Forrester and Morise scores were categorized as low (<15%), medium (≥15 to ≤50%), or high likelihood (>50%). It was not clear how the cutoffs were chosen in Thomas et al (2013).
As described in the Clinical Context section of this review, the pretest probability cutoffs from clinical models used for risk stratification vary in the literature, but intermediate risk frequently ranges from 10% to 90%. Net reclassification using this cutoff has not been reported.
| Author (Year) | Net Reclassification Improvementa for Corus CAD score vs. Second Modality (95% CI) | ||||
| Myocardial Perfusion Imaging | |||||
| Site-Read | Core-Lab | Diamond-Forrester | Morise | Expanded Clinical Model | |
| Rosenberg et al (2010)11, PREDICT | 21% (NR) | NR | 20% (NR) | NR | 16% (NR) |
| p | <0.001 | <0.001 | <0.001 | ||
| Thomas et al (2013)12,COMPASS | 26% (NR) | 11% (NR) | 28% (NR) | 60% (NR) | NR |
| p | NR | NR | NR | NR | NR |
CI: confidence interval; NR: not reported; CAD: coronary artery disease; COMPASS: Coronary Obstruction Detection by Molecular Personalized Gene Expression; PREDICT: Personalized Risk Evaluation and Diagnosis In the Coronary Tree.a Net reclassification improvement quantifies the difference between the proportion of patients correctly reclassified from an incorrect initial classification and the proportion incorrectly reclassified from a correct initial classification.
Voros et al (2014) pooled results from PREDICT and COMPASS to compare Corus CAD score with computed tomography imaging for detecting plaque burden (coronary artery calcium), and luminal stenosis.19, Six hundred ten patients, 216 from PREDICT (19% of enrolled patients) and 394 from COMPASS (73% of enrolled patients), who had undergone coronary artery calcium scoring, computed tomographic angiography, and Corus CAD score were included. Mean age was 57 years; 50% were female, and approximately 50% used statin medication. Prevalence of obstructive CAD (≥50% stenosis) was 16% in the PREDICT cohort (patients referred for coronary angiography) and 13% in the COMPASS cohort (patients referred for myocardial perfusion imaging). In linear regression analyses, Corus CAD scores statistically and significantly correlated with coronary artery calcium (r=0.50), the number of arterial segments with any plaque (r=0.37), overall stenosis severity (r=0.38), and maximum luminal stenosis (r=0.41) (all p<0.01), but the strength of the correlations was modest. Several Corus CAD score cutoffs were explored (eg, to maximize diagnostic accuracy). Results using a cutoff of 15 points are shown in Table 6. For detecting luminal stenosis of 50% or greater, the Corus CAD score positive predictive value and negative predictive value were 23% and 95%, respectively. For detecting clinically significant coronary artery calcium (≥400), the Corus CAD score positive predictive value and negative predictive value were 14% and 97%, respectively. Limitations of the study included a lack of clinical outcomes (eg, survival, morbidity) and lack of comparison with coronary artery calcium and computed tomographic angiography for predicting these outcomes (ie, incremental Corus CAD score predictive value was not assessed).
| Outcome | Corus CAD AUROC (95% CI) | Diamond-Forrester AUROC (95% CI) | Sensitivity, % | Specificity, % | PPV, % | NPV, % |
| Plaque burdena | ||||||
| CAC >0 | 0.75 (0.71 to 0.79) | 0.65 (0.61 to 0.69) | 71 | 62 | 65 | 68 |
| CAC ≥400 | 0.75 (0.68 to 0.82) | 0.61 (0.53 to 0.69) | 84 | 49 | 14 | 97 |
| Luminal stenosis by CTA | ||||||
| ≥50% | 0.75 (0.70 to 0.80) | 0.65 (0.59 to 0.71) | 84 | 51 | 23 | 95 |
| ≥70% | 0.75 (0.67 to 0.83) | 0.63 (0.53 to 0.73) | 90 | 48 | 8 | 99 |
Adapted from Voros et al (2014).19,AUROC: area under the receiver operating characteristic curve; CAC: coronary artery calcium; CAD: coronary artery disease; CI: confidence interval; CTA: computed tomography angiography; NPV: negative predictive value; PPV: positive predictive value.a Long-term outcomes are generally excellent for patients with CAC >0 and substantially worse for patients with CAC >400.
The diagnostic pathway for coronary artery disease includes information from medical history, along with age and sex, stress testing, and imaging. It is not clear how the Corus CAD gene expression test fits in the current diagnostic pathway and how results would be used to change current guideline-based risk stratification before and/or after other noninvasive testing. Results of 2 validation studies (PREDICT, COMPASS) have reported the test may improve coronary artery disease prediction beyond the Diamond-Forrester prediction model. In the COMPASS study, the sensitivity and negative predictive value of the Corus CAD score in diagnosing obstructive coronary artery disease was superior to myocardial perfusion imaging in patients referred for myocardial perfusion imaging testing. However, in that study, the reported sensitivity of myocardial perfusion imaging was considerably lower than that generally reported in the literature. Neither PREDICT nor COMPASS used the guideline definition of obstructive coronary artery disease as the reference standard and had relatively few patients at intermediate risk based on clinical prediction rules. The sensitivity and negative predictive value of clinical models were not reported. An analysis of a cohort from the PROMISE trial including patients with an intermediate pretest probability of obstructive coronary artery disease confirmed a high negative predictive value for the Corus CAD score.
The test excludes patients with diabetes, acute and chronic inflammatory conditions, and such patients are expected to be common among those being evaluated for obstructive coronary artery disease. Thus applicability to clinical practice may be narrow. Although the test is marketed as a sex-specific test, performance characteristics by sex and age were not provided. One study reported that the Corus CAD score was associated with obstructive coronary artery disease in both men (OR=1.99; 95% CI, 1.35 to 2.96) and women (OR=3.45; 95% CI, 1.97 to 5.91). The gene selection, algorithm development, and validation studies have been performed in populations that were approximately 90% white.
Net reclassification has been reported comparing the Corus CAD score with other clinical prediction tools and myocardial perfusion imaging. While the pretest probability cutoffs from clinical models used for risk stratification vary in the literature, intermediate-risk frequently ranges from 10% to 90% and net reclassification using this cutoff has not been reported.
Publications from 4 of the previously described studies have reported performance of the Corus CAD score in the prognosis of cardiovascular events. Table 7 summarizes the results. Rosenberg et al (2012) published a follow-up report from PREDICT on the association between Corus CAD score and subsequent major adverse cardiac events, including myocardial infraction, stroke/transient ischemic attack, all-cause mortality, and coronary revascularization.20,
Thomas et al (2013) patients were followed for 6 months after Corus CAD testing, with 420 of 431 completing follow-up.12, Major adverse cardiac events (nonfatal myocardial infraction, stroke/transient ischemic attacks, or all-cause mortality) and revascularization events were recorded. Only 2 major adverse cardiac events occurred.
Voora et al (2017) included analysis of 2,370 PROMISE patients with samples in the biorepository who were followed for a median of 25 months.10, The association between the Corus CAD score and a composite outcome of death, myocardial infraction, revascularization, or unstable angina was statistically significant after adjustment for the Framingham Risk Score. The association was driven primarily by the revascularization component. When revascularization was removed from the composite, there was no longer a significant association between the Corus CAD score and the outcome after adjusting for the Framingham Risk Score. A low Corus CAD score was associated with a low-risk (1.6%) of revascularization and a negative predictive value of 98% (confidence interval not reported).
Ladapo et al (2018) and Gul et al (2019) evaluated the association between Corus CAD scores and cardiovascular events at 12 months in elderly adults (n=176) and women (n=288) from the PRESET registry.17,18, In adults 65 years of age or older the incidence of major adverse cardiovascular events or revascularization was 0% in patients with a low Corus CAD score and 10% in patients with a higher Corus CAD score (p=0.04). In the cohort of women of all ages, the incidence of major cardiac events was not statistically different between women with a low Corus CAD score (1.3%) and those with a higher Corus CAD score (4.2%, p=0.16).
| Author | N | Event | Incidence | Sens (95% CI) | Spec (95% CI) | PPV (95% CI) | NPV (95% CI) | Association (95% CI) |
| Rosenberg et al (2012)20, | 1160 | 12-mo MACEa | 1.5 | 82 (NR) | 34 (NR) | 1.8 (NR) | 99 (NR) | OR=2.41 (0.74 to 10.5) |
| 12-mo MACEa or revascularizations | 25 | 86(NR) | 41(NR) | 33(NR) | 90(NR) | OR=4.32 (3.02 to 6.25) | ||
| Thomas et al (2013)12, | 420 | 6-mo revascularizations or MACEa | 6.7 | 96 (NR) | NR | NR | 99 (NR) | NR |
| Voora et al (2017)10, | 2370 | Death, MI, or UA with median 25-mo follow-up | 2.6 | NR | NR | NR | NR | HR=0.98 (0.52 to 1.87)b |
| Death, MI, UA, or revascularization with median 25-mo follow-up | 6.0 | NR | NR | NR | NR | HR=1.70 (1.10 to 2.64)b |
Values are percent unless otherwise indicated.CI: confidence interval; CAD: coronary artery disease; MACE: major adverse cardiac events; MI: myocardial infarction; NPV: negative predictive value; NR: not reported; OR: odds ratio; HR: hazard ratio; PPV: positive predictive value; Sens: sensitivity; Spec: specificity; UA: unstable anginaa MACE included MI, stroke/transient ischemic attack, all-cause mortality.b Adjusted for Framingham Risk Score.There is less evidence on the association between the Corus CAD score and cardiovascular events. The available evidence provides a preliminary indication that a Corus CAD score of 15 or less identifies a group unlikely to require revascularization within 2 years. No data was given regarding which revascularizations were planned versus emergent; eg, information is needed describing how many revascularizations were performed to alleviate symptoms, for progression to unstable angina, or to decrease the risk of cardiac outcomes such as death, heart failure, or myocardial infraction. More data are needed on coronary events other than revascularizations. Notably, confidence intervals for performance characteristics are lacking in these studies.
There is uncertainty regarding the role of the test in the diagnostic pathway. The proposed strategy for integrating the results of the test with current guidelines for risk stratification before and/or after other noninvasive testing is not clear. The diagnostic strategy incorporating the Corus CAD test should be explicitly described so that it is clear which existing data are relevant for evaluating the proposed use. Proposed changes in stratification compared with existing guidelines are needed so that net reclassification analyses compared with guideline-recommended stratification can be constructed. Decision models of a strategy incorporating the Corus CAD score into the guideline recommendations would be useful.
The Corus CAD score is correlated with the presence of obstructive coronary artery disease. The PREDICT and COMPASS studies reported that the gene expression score is superior to the Diamond-Forrester model and to myocardial perfusion imaging for predicting obstructive coronary artery disease. However, the available studies do not specify the use of the test in the guideline, recommended diagnostic pathway for stable ischemic heart disease. Therefore, it is not possible to make conclusions about clinical validity. Performance characteristics by sex and age were reported from a safety analysis of registry data. A high Corus CAD score was associated with adverse cardiac events in older adults (both men and women), but this association was not statistically significant when assessed in the cohort of women.
A test is clinically useful if the use of the results informs management decisions that improve the net health outcome of care. The net health outcome can be improved if patients receive correct therapy, or more effective therapy, or avoid unnecessary therapy, or avoid unnecessary testing.
Direct evidence of clinical utility is provided by studies that have compared health outcomes for patients managed with and without the test. Because these are intervention studies, the preferred evidence would be from RCTs.
There is no direct evidence from RCTs.
Indirect evidence on clinical utility rests on clinical validity. If the evidence is insufficient to demonstrate test performance, no inferences can be made about clinical utility.
To develop a chain of evidence or a decision model requires explication of the elements in the model and evidence that is sufficient to demonstrate each of the links in the chain of evidence or the validity of the assumptions in the decision model. A chain of evidence or decision model must be constructed so to permit a comparison between a diagnostic strategy including Corus CAD testing and a strategy of no Corus CAD testing. The Corus CAD test is associated with the diagnosis of obstructive coronary artery disease. The Corus CAD test classifies patients into clinically credible diagnostic groups (low- and high-risk of obstructive coronary artery disease) that were defined a priori and evaluated in prospective studies. However, it is not clear how the test fits in the current diagnostic pathway and how results would be used to change current guideline-based risk.
Patients managed without the Corus CAD test should be evaluated according to established guidelines for the noninvasive evaluation of patients with stable ischemic heart disease.2, Studies examining patient outcomes of Corus CAD testing have primarily analyzed changes in physician management as an outcome.
The Investigation of a Molecular PersonAlizedCoronary Gene Expression Test (IMPACT)-CARDiology Practice Pattern study (2013) compared a prospective cohort with matched historical controls to evaluate whether the Corus CAD test altered cardiologist evaluation and clinical management of coronary artery disease.21, Coronary artery disease was categorized by authors as no coronary artery disease (0% stenosis), coronary artery disease with 50% or less stenosis, or coronary artery disease with more than 50% stenosis. Eighty-eight patients were enrolled and 83 included in the final analysis. The matched cohort comprised 83 patients selected with similar distributions of age, sex, and clinical risk factors evaluated at a participating clinic within the past 3 to 30 months. Diagnostic testing plans were changed for 58% of patients in the prospective cohort (95% CI, 46% to 69%; p<0.001) with a greater reduction in testing intensity (39%) compared with increased testing intensity (19%). Compared with the historical control group, the prospective cohort had a 71% reduction in overall diagnostic testing (p<0.001).
IMPACT-Primary Care Practice Pattern (PCP) (2014) evaluated whether having the Corus CAD altered primary care providers’ diagnostic evaluation and clinical management of stable, nonacute, nondiabetic patients presenting with coronary artery disease symptoms.22, Nine primary care providers at 4 centers evaluated 261 consecutive patients, 251 (96%) of whom were eligible for participation. Clinicians documented their pretest impressions and recommendations for further evaluation and management on a clinical report form. All patients underwent Corus CAD testing. The primary outcome was the change in patient management between preliminary and final treatment plans. Diagnostic testing plans were changed for 58% of patients, with reductions in testing intensity more common (64%) than increases (34%; p<0.001). No study-related major adverse cardiac events were observed in 247 (98%) patients who had at least 30 days of follow-up.
The Enhanced Assessment of Chest Pain and Related Symptoms in the Primary Care Setting Through the Use of a Novel Personalized Medicine Genomic Test (REGISTRY 1)study (2015) assessed the impact of having the Corus CAD on patient management decisions by examining the association between Corus CAD results and posttest referral patterns.23, Primary care practitioners at 7 centers evaluated 342 stable, nonacute, nondiabetic patients presenting with CAD symptoms. All patients underwent Corus CAD testing. Of 167 patients with low (≤15) Corus CAD score, 10 (6%) were referred for further cardiac evaluation compared with 122 (70%) of 175 patients in the high Corus CAD score group (p<0.001). Over a mean follow-up of 264 days, there were 5 major adverse cardiac events, 2 in the low Corus CAD score group and 3 in the high Corus CAD score group. Of 21 patients who underwent elective invasive coronary angiography, 1 (50%) of 2 in the low Corus CAD score group and 8 (42%) of 19 in the high Corus CAD score group had obstructive findings.
Ladapo et al (2015) pooled results for women who participated in the IMPACT-PCP (n=140) and REGISTRY 1 (n=180) studies to evaluate the impact of Corus CAD score on further cardiac evaluation (n=320).24, Referral rate for further cardiac evaluation was 4% for women with low Corus CAD score (n=248) versus 83% for women with elevated Corus CAD score (n=72).
The Ladapo et al (2017) analysis of the 566 patients from the PRESET registry (described previously) evaluated the association between the Corus CAD score and cardiac referrals (referral to cardiology or further cardiac testing).7, Ten percent (26/252) of low Corus CAD score patients were referred versus 44% (137/314) of high Corus CAD score patients. After adjusting for age, sex, body mass index, smoking status, hypertension, and dyslipidemia, the association between Corus CAD score and referral rate remained statistically significant (odds ratio = 0.15; 95% confidence interval, 0.10 to 0.24; p<0.001). With 1 year of follow-up, major adverse cardiac events and revascularizations were noted in 3 (1.2%) of 252 low Corus CAD score patients and 14 (4.5%) of 314 high Corus CAD score patients (p=0.03).
There are no rigorous studies comparing clinical outcomes for patients managed using Corus CAD testing with alternative methods for stable ischemic heart disease (ie, no direct evidence that the test is clinically useful). Currently, it is unclear whether a chain of evidence can be constructed because of the lack of evidence on use of the test in the intermediate-risk population.
For individuals who have suspected stable ischemic heart disease without diabetes or inflammatory conditions who receive gene expression testing, the evidence includes retrospective case-control and prospective cohort studies. Relevant outcomes are overall survival, disease-specific survival, test validity, change in disease status, morbid events, and resource utilization. The diagnostic pathway for coronary artery disease includes information from medical history, along with age and sex, stress testing, and imaging. Newer noninvasive methods are being tested, such as gene expression testing. It is not clear how the Corus CAD gene expression test fits in the current diagnostic pathway and how results would be used to change current guideline-based risk stratification before and/or after other noninvasive testing. Results of 2 validation studies (Personalized Risk Evaluation and Diagnosis In the Coronary Tree [PREDICT], Coronary Obstruction Detection by Molecular Personalized Gene Expression [COMPASS]) have reported that the test may improve coronary artery disease prediction beyond the Diamond-Forrester prediction model. In the COMPASS study, the sensitivity and negative predictive value of the Corus CAD score in diagnosing obstructive coronary artery disease was superior to myocardial perfusion imaging in patients referred for myocardial perfusion imaging testing. However, in that study, the reported sensitivity of myocardial perfusion imaging was considerably lower than that generally reported in the literature. Neither PREDICT nor COMPASS used the guideline definition of obstructive coronary artery disease as the reference standard and had relatively few patients at intermediate risk based on clinical prediction rules. The sensitivity and negative predictive value of clinical models were not reported. An analysis of a cohort from the PROspective Multicenter Imaging Study for Evaluation of chest pain (PROMISE) trial including patients with an intermediate pretest probability of obstructive coronary artery disease confirmed a high, negative predictive value for the Corus CAD score. The test also has been shown to have some predictive ability of future revascularization; too few major cardiac events have been observed during the limited duration of follow-up to assess predictive ability for that outcome. Evidence for the Corus CAD score has not directly demonstrated that the test is clinically useful and a chain of evidence cannot be constructed to support its utility. The evidence is insufficient to determine the effects of the technology on health outcomes.
| Population Reference No. 1 Policy Statement | [ ] MedicallyNecessary | [x] Investigational | [ ] Not Medically Necessary |
In 2012, the American Heart Association (AHA) released a policy statement on genetics and cardiovascular disease.25, Gene expression testing is not specifically mentioned. Generally, the AHA supported recommendations issued in 2000 by a now defunct Advisory Committee to the U.S. Department of Health and Human Services, which stated: “No test should be introduced in the market before it is established that it can be used to diagnose and/or predict a health-related condition in an appropriate way.”26,
In 2017, the AHA released a scientific statement on the expressed genome in cardiovascular diseases and stroke.27, The statement summarized the clinical validity and utility evidence for the Corus CAD score, stating “…the Corus CAD test is a clinically available diagnostic test that has been evaluated, has been deemed to be valid and useful.…”
In 2012, the joint guidelines of the American College of Cardiology Foundation and 6 other medical societies for the diagnosis and management of patients with stable ischemic heart disease did not mention the gene expression score.2, The 2014 update to these guidelines also did not mention the gene expression score.6,
Not applicable.
Medicare National Coverage
There are no Medicare national coverage determinations for Corus CAD testing to predict coronary artery disease. In 2019, Noridian MolDX rescinded coverage of Corus CAD and issued a non-coverage determination.28,29,
| Codes | Number | Description |
|---|---|---|
| CPT | 81493 | Coronary artery disease, mRNA, gene expression profiling by real-time RT-PCR of 23 genes, utilizing whole peripheral blood, algorithm reported as a risk score |
| 81599 | Unlisted multianalyte assay with algorithmic analysis: use for similar products | |
| ICD-10-CM | Investigational for all relevant diagnoses | |
| Z15.89 | Genetic susceptibility to other disease | |
| Z13.6 | Encounter for screening for cardiovascular disorders | |
| ICD-10-PCS | Not applicable. ICD-10-PCS codes are only used for inpatient services. There are no ICD procedure codes for laboratory tests. | |
| Type of Service | Laboratory | |
| Place of Service | Outpatient |
Some modifiers
| Date | Action | Description |
|---|---|---|
| 04/20/2021 | Annual Review/Policy Archived | The Corus CAD test is no longer available. Policy Archived. |
| 04/28/2020 | Review due to MPP | Policy updated with literature review through January 15, 2020; references updated. Policy statement unchanged. |
| 08/02/2019 | Annual Reviews | No changes |
| 01/13/2017 | ||
| 12/02/2016 | ||
| 08/12/2015 | ||
| 06/11/2015 |