Loading...

    Explore
    • Policy Monitoring
    • Live Policy Feed
    • Billing Codes
    • Healthcare Providers
    • Health Insurance Payers
    • States
    • Service Lines
    • Taxonomy Codes
    Lookup tools
    • Billing Code Lookup
    • CPT Modifier Lookup
    • Provider Taxonomy Code Lookup
    • Healthcare Provider & Facility Lookup
    • Health Insurance Payer Lookup
    • Payer Contact Directory

    Policy & Benchmark Tools

    • Prior Authorization Lookup
    • Coverage Criteria Lookup
    • Timely Filing Lookup
    • Billing Code Reimbursement Rates
    • Payer Comparison Dashboard
    • Facility Comparison Dashboard
    • All tools

    Company

    • About OpenPayer
    • Data Sources
    • Request a Policy / Feedback
    • Trek Health
    OpenPayerby Trek Health
    LinkedInYouTube

    © 2026 Trek Health. All rights reserved.

    OpenPayer is built and maintained by Trek Health.

    v1.0.31
    OpenPayerPolicy UpdatesCardiologyAetna SAECG AI‑ECG Coverage Update | OpenPayer
    Ask Trek AI about this page
    CurrentAetnaPolicy 0664

    Signal-Averaged Electrocardiography (SAECG) and Artificial Intelligence Algorithmic Electrocardiogram for Cardiovascular-Related Diseases

    Customize your policy alerts

    Sign up for Aetna Policy 0664 alerts

    Get alerted when Policy 0664 changes without checking for updates manually.

    Monitor payer policy activity

    Receive email alerts as soon as relevant payer policies are published or updated.

    Defines Aetna's coverage stance for SAECG and assistive/artificial intelligence algorithmic electrocardiogram analyses for detection and risk stratification of cardiovascular-related diseases; affects providers submitting claims to Aetna for these tests.

    Policy Summary
    PayerAetna
    PolicySignal-Averaged Electrocardiography (SAECG) and Artificial Intelligence Algorithmic Electrocardiogram for Cardiovascular-Related Diseases
    Policy CodePolicy 0664
    Change TypeNo material change
    Effective DateMay 2, 2003
    Last ReviewMar 16, 2023
    Next Review DateJul 12, 2024
    Key ActionVerify coverage and prior-authorization requirements with Aetna before submitting claims for CPT codes 0764T, 0765T, or 93278, which are listed as not covered for the indications in this bulletin.
    SourceLink

    OpenPayer is powered by Trek Health's payer performance platform. Trek continuously ingests, validates, and normalizes Transparency in Coverage data alongside payer policies and other commercial payer data to create a structured payer intelligence foundation. OpenPayer uses this foundation to deliver personalized search results, dynamically generated policy pages, and tailored policy monitoring based on each user's payers, specialties, billing codes, and areas of interest. The same intelligence powers broader payer performance workflows, including reimbursement benchmarking, contract evaluation, payer negotiations, and financial decision-making.

    POLICY UPDATE CHANGES

    No material clinical or coverage changes in this revision.

    3Not-covered procedures
    >=1Not-covered CPTs
    0.90AIeECG detection AUC
    102Meta-analysis studies
    2024Next review

    Coverage Determinations

    Experimental and Investigational — Not Covered

    Aetna considers the following procedures experimental and investigational because effectiveness has not been established:

    Experimental/Investigational Procedures: 1. Signal-averaged electrocardiography (SAECG) because no prospective clinical studies have demonstrated the utility of this testing in improving clinical outcomes; 2. Remote algorithmic analysis of electrocardiographic-derived data (Premier Heart's Multifunction Cardiogram (MCG); also known as 3DMP Computerized EKG System) because clinical value in managing persons suspected of having significant coronary artery disease has not been established; 3. Assistive artificial intelligence algorithmic electrocardiogram assessment (e.g., MyoVista Wavelet ECG Cardiac Testing Device) for evaluation of cardiac dysfunction because the clinical value of this approach has not been established.

    Exact wording from policy.

    Evidence-based coverage considerations

    Evidence-based stance and applicability summary from document excerpts

    MCG/3DMP diagnostic utility: MCG/3DMP devices have variable diagnostic performance; studies report missed significant lesions and false positives, and prospective clinical studies are necessary to demonstrate clinical utility. In the MED-FIT study MCG severity score showed low diagnostic performance (AUC 0.51-0.58) with low sensitivity and specificity using a cut-off of 4.0.

    Suggests not supported as a replacement for angiography or standard diagnostic pathways.

    SAECG in arrhythmia substrate and risk: SAECG-derived metrics (examples: electrical regression after RF ablation, number of abnormal SAECG criteria in ARVC, LSDf in ICD patients, P wave SAECG measures in ESUS) have been associated with outcomes in small observational studies, but sample sizes are limited and further validation is needed; utility for routine clinical decision-making is unproven.

    Findings are hypothesis-generating; prospective validation required.

    AI-ECG models: AI models applied to ECGs (including for atrial fibrillation and other cardiovascular conditions) show high pooled sensitivity and specificity in meta-analyses (eg, AF detection: sens ~94.8%, spec ~96.96%), and deep neural networks report high AUROC; however, performance varies by dataset and many studies have methodological limitations — clinical utility and impact on patient outcomes require external validation and prospective trials.

    Dataset quality and external validation are concerns; insufficient evidence to support routine clinical use for detection of cardiac dysfunction.

    CPT codes 0764T, 0765T, and 93278 are listed in this Clinical Policy Bulletin as not covered for the indications addressed. These codes correspond to assistive algorithmic electrocardiogram risk-based assessments (0764T, 0765T) and signal-averaged electrocardiography (SAECG) (93278). Providers should bill using standard ECG CPT codes (for example, 93000–93010) when appropriate and verify coverage before submitting claims that include the not-covered codes to avoid denial.

    Computer-enhanced analyses marketed as Multifunction Cardiogram (MCG / 3DMP) are not supported as replacements for angiography or as primary diagnostic tests for coronary artery disease. Published evaluations found missed significant lesions and false positives, and prospective studies to validate clinical utility are lacking. In particular, analyses from MED-FIT reported poor diagnostic performance with low area under the curve (AUC 0.51–0.58) and low sensitivity/specificity using the reported severity cut-off, indicating the test is insufficiently accurate to select patients for definitive testing such as angiography.

    Systematic reviews of AI-enabled ECG approaches to detect left ventricular systolic dysfunction (LVSD) report promising diagnostic accuracy (for example, median AUC ~0.90), but importantly no studies have examined the effect of these tests on treatment decisions or patient outcomes. Prospective, randomized trials are needed to demonstrate clinical impact, cost-effectiveness, and whether use of these algorithms improves management or outcomes before routine coverage can be recommended.

    Signal-averaged electrocardiography (SAECG) is considered experimental and investigational in this bulletin because no prospective clinical studies have demonstrated that SAECG improves clinical outcomes. As a result, SAECG is not supported as medically necessary for the indications addressed in this policy.

    Routine use of Multifunction Cardiogram (MCG/3DMP) or assistive AI-algorithmic ECG for detection of coronary artery disease or cardiac dysfunction is not supported by sufficient evidence of clinical utility. Existing studies have methodological limitations, variable performance, and lack prospective outcome data; therefore these technologies should not replace established diagnostic pathways outside of research or validated clinical programs.

    Coding and Billing References

    Not covered / Experimental CPT codes referencedCPTNot Covered
    0764TAssistive algorithmic electrocardiogram risk-based assessment for cardiac dysfunction (eg, low-ejection fraction, pulmonary hypertension, hypertrophic cardiomyopathy); related to concurrently performed electrocardiogram (List separately in addition to code for primary procedure).
    0765TAssistive algorithmic electrocardiogram risk-based assessment related to previously performed electrocardiogram.
    93278Signal-averaged electrocardiography (SAECG) with or without ECG.
    Standard ECG CPT codes referencedCPT
    93000Electrocardiogram, routine ECG with at least 12 leads; with interpretation and report.
    93010Electrocardiogram, tracing only, without interpretation and report, or interpretation and report only.
    ICD-10 codes referencedICD-10
    I05.0-I52Chronic rheumatic heart disease, hypertensive disease, ischemic heart disease, diseases of pulmonary circulation, and other forms of heart disease (ICD-10 range referenced).
    AI ECG - temporary Category III CPTCPT
    Category III CPT codesNew Category III CPT codes for novel assistive AI algorithmic ECG risk assessment (AMA announcement)
    LSDf threshold
    LSDf threshold (Layered Symbolic Decomposition frequency)13.25% — selected ROC threshold shown to predict worse survival and ventricular arrhythmia in a pilot ICD cohort (statistically significant in Kaplan-Meier and multivariate Cox analyses).
    Comparator variables reportedLSDf compared against LVEF (28.5%), QRS duration (100 ms), age, sex, NYHA class, and anti-arrhythmic use in multivariate analysis.
    Study limitations notedPilot ICD cohort was small and highly selected; cause of death not adjudicated and SAECG not routine in many settings.
    MCG cut-off (MED-FIT)
    MCG severity score cut-offSeverity score cut-off of 4.0 used in the MED-FIT study (primary cut-off evaluated against FFR ≤0.80 and angiographic stenosis ≥50%).

    Provider Billing, Documentation, and Authorization Actions

    Prior Authorization

    Coverage / prior-authorization coding guidance: not-covered codes vs standard ECG codes

    CPT codes 0764T, 0765T, and 93278 are identified in this Clinical Policy Bulletin as not covered for the indications listed; when appropriate, use standard ECG CPT codes 93000–93010 and verify coverage/prior authorization requirements with Aetna before submitting claims.

    • 0764T, 0765T, 93278 are listed as not covered for indications in this bulletin.
    • Standard ECG CPT codes referenced: 93000–93010.
    • Providers should verify coverage/prior authorization requirements with Aetna prior to claim submission.
    Prior Authorization

    Prior authorization may apply to Category III AI-algorithmic ECG codes

    The policy notes that novel assistive AI algorithmic ECG procedures have recently been assigned Category III CPT codes by the AMA; prior authorization or payer review may be required for these investigational/limited-use technologies.

    • AMA issued new Category III CPT codes for assistive AI algorithmic ECG risk assessment (announcement dated July 7, 2022).

    Background and Scope

    Signal-averaged electrocardiography (SAECG) is a computerized analysis of ECG segments designed to detect ventricular late potentials, which have been associated with an increased risk of ventricular tachyarrhythmias and sudden cardiac death. The technique evaluates low-amplitude, high-frequency components of the QRS to identify late potentials, but evidence shows limited prospective validation and uncertain impact on clinical management, supporting the policy’s investigational stance.

    Definitions and Key Terms

    Signal-averaged electrocardiography (SAECG)
    DefinitionComputerized analysis of segments of a standard electrocardiogram to detect ventricular late potentials that may indicate risk of ventricular tachyarrhythmias and sudden cardiac death.
    Clinical intentUsed to identify electrophysiologic substrate (ventricular late potentials) associated with arrhythmic risk after MI or in cardiomyopathy.
    Performance characteristicsData show relatively consistent high negative predictive values but poor positive predictive values and variable sensitivity/specificity in studied populations.
    Assistive artificial intelligence algorithmic electrocardiogram
    DefinitionAI-based algorithmic assessment of ECG data (examples referenced include MyoVista Wavelet ECG) intended to assess cardiac dysfunction; novel assistive devices/algorithms are discussed in the policy.
    Regulatory/coding note

    Policy Revision History

    2024-07-12Next review date

    Scheduled next clinical policy review date.

    2023-03-16Last reviewLatest

    Policy last reviewed on this date; content and evidence were evaluated.

    2003-05-02Effective date

    Policy originally became effective.

    Policy Summary
    PayerAetna
    PolicySignal-Averaged Electrocardiography (SAECG) and Artificial Intelligence Algorithmic Electrocardiogram for Cardiovascular-Related Diseases
    Policy CodePolicy 0664
    Change TypeNo material change
    Effective DateMay 2, 2003
    Last ReviewMar 16, 2023
    Next Review DateJul 12, 2024
    Key ActionVerify coverage and prior-authorization requirements with Aetna before submitting claims for CPT codes 0764T, 0765T, or 93278, which are listed as not covered for the indications in this bulletin.
    SourceLink
    On This Page
    Coverage DeterminationsCoding and Billing ReferencesProvider Billing, Documentation, and Authorization ActionsBackground and ScopeDefinitions and Key TermsPolicy Revision History
    Diagnostic performance at cut-offUsing cut-off 4.0, sensitivity/specificity for functional ischemia were 32%/67% and for angiographic stenosis 37%/72% (AUCs low: 0.51–0.58).
    Study contextSingle-center prospective MED-FIT study; 91 patients with adequate-quality MCG data analyzed; prevalence of functional ischemia ~42.7%.
    P wave duration threshold
    P wave duration (PWD) threshold>= 135 ms — prolonged PWD predicted recurrent embolic stroke more robustly than SCAF detection in an ESUS cohort.
    Study population125 consecutive ESUS patients followed with Holter and ECG monitoring; 25.6% diagnosed with subclinical AF during follow-up; stroke recurrence in 17.6%.
    Recommendation from authorsPWD may be a useful biomarker to predict SCAF and identify patients at higher risk for recurrent embolic stroke; further research needed.
  1. Insufficient evidence of clinical utility is stated in the policy, and payer prior authorization may apply.
  2. Note

    Policy dates and administrative links — verify operational requirements with Aetna

    Policy effective and review dates and links to Aetna resources are provided in the policy history; the document does not enumerate specific procedural prior authorization rules in these sections—providers should consult Aetna resources for administrative requirements.

    • Effective: 05/02/2003; Last review: 03/16/2023; Next review: 07/12/2024.
    • Links provided for Review History and Definitions for additional operational details.
    Note

    Provider action: confirm coverage before ordering or billing investigational ECG tests

    Highlight: Follow the policy's not-covered determinations and verification steps before ordering or billing experimental ECG-based tests to avoid denials.

    • Signal-averaged ECG and certain assistive AI/MCG procedures are considered experimental/investigational.
    • Confirm coverage and prior authorization requirements with Aetna before testing or billing.
    Step Therapy

    Use standard diagnostic pathways for suspected CAD before relying on ECG-based signal analysis

    When evaluating suspected coronary artery disease (CAD), follow standard diagnostic pathways—clinical assessment, conventional ECG, stress testing, and angiography as indicated—before relying on ECG-based signal analysis devices as substitutes.

    • 3DMP/MCG performance is variable and prospective studies are needed; it should not replace angiography.
    • Follow established diagnostic evaluation (history/physical, standard 12‑lead ECG, appropriate stress testing, and invasive angiography when indicated).
    Note

    Administrative action: verify payer-specific prior authorization and documentation rules

    Administrative note placeholder: verify any payer-specific prior authorization, coverage, or documentation procedures with Aetna prior to ordering investigational ECG-based services.

    • Policy history and links are provided; do not assume absence of other administrative requirements.
    • Contact Aetna for operational/prior authorization specifics.
    Billing Rule

    Documentation and coding: use standard ECG CPT codes 93000–93010; avoid experimental codes for covered indications

    When billing ECG-related services, use standard ECG CPT codes 93000–93010 where appropriate; claims billed with the listed experimental/investigational codes may be processed as not covered under this bulletin.

    • Standard codes: 93000 (ECG with interpretation/report) through 93010 (tracing only or interpretation/report only).
    • Avoid billing CPTs 0764T, 0765T, or 93278 for covered indications addressed by this policy.
    Documentation Required

    Documentation: record SAECG-derived metrics and comparator clinical variables

    If SAECG metrics are used for risk stratification, document the specific SAECG-derived values (for example, LSDf percentage) and comparator clinical variables such as LVEF, QRS duration, presenting symptoms, and any device therapy.

    • Record the exact LSDf value and the 13.25% threshold if used in analysis.
    • Include comparator data: LVEF, 12‑lead QRS duration, symptoms, and device therapy status.
    Documentation Required

    Documentation: record P wave SAECG measures and rhythm monitoring strategy for ESUS evaluations

    For stroke/ESUS evaluations using P wave SAECG, document measured P wave SAECG values (including P wave duration) and the rhythm monitoring strategy employed (Holter/ECG schedule) used to detect subclinical atrial fibrillation.

    • Document P wave duration and note if PWD ≥ 135 ms was observed.
    • Document rhythm monitoring plan and follow-up interval (Holter/ECG schedule or other monitoring).
    Note

    Policy review and effective dates — consult linked resources for definitions and history

    Policy history and review dates are listed in the Policy History section; consult the linked Definitions and Review History for additional operational and definitional details.

    • Effective date: 05/02/2003; Last review: 03/16/2023; Next review: 07/12/2024.
    • Links available for Review History and Definitions for further reference.
    Denial Risk

    Denial risk: claims with CPT 0764T, 0765T, or 93278 for listed indications may be denied

    Claims submitted with CPT codes 0764T, 0765T, or 93278 for indications listed in this Clinical Policy Bulletin are identified as not covered and may be denied.

    • Do not submit these CPT codes for the listed indications without prior verification of coverage.
    • Denials may occur if investigational/not-covered CPT codes are billed for these indications.
    Denial Risk

    Risk: devices/tests with poor or unproven accuracy (eg, 3DMP/MCG) may cause missed diagnoses or false positives

    Devices or tests with poor or unproven diagnostic accuracy (for example, 3DMP/MCG with missed significant lesions and false positives) may lead to missed diagnoses or false positives and could result in denial when used to replace definitive testing such as angiography.

    • 3DMP missed some significant angiographic lesions and produced false positives in studies.
    • Do not use such devices as a substitute for angiography or established diagnostic pathways without validated evidence.
    Denial Risk

    Risk: MCG or assistive AI‑algorithmic ECG lacks sufficient evidence and may lead to noncoverage

    Use of multifunction cardiogram (MCG) or assistive AI-algorithmic ECG for detection of cardiac dysfunction lacks sufficient evidence of clinical utility according to the policy; reliance on these modalities for diagnostic decision-making may lead to noncoverage or denial.

    • Policy considers MCG/3DMP and assistive AI algorithmic ECG investigational due to insufficient evidence of clinical value.
    • Claims relying on these technologies for diagnosis or management decisions may be denied.
    Note

    Administrative note: confirm other payer-specific prior-authorization requirements despite policy silence

    Policy history and review dates are provided; absence of an explicit prior-authorization requirement within this policy document does not eliminate other administrative or payer-specific prior authorization obligations—confirm with Aetna.

    • The policy lists effective and review dates and provides links for definitions and review history.
    • Providers should verify whether separate Aetna prior authorization requirements apply.
    AMA has issued new Category III CPT codes for novel assistive AI algorithmic ECG risk-assessment procedures (announced July 7, 2022).
    Coverage stanceAetna considers assistive AI algorithmic ECG assessment investigational because clinical value has not been established.
    Assistive AI algorithmic ECG (AI-driven analytic)
    DefinitionAn AI-driven analytic applied to ECG waveforms to generate a risk-based assessment for cardiac dysfunction or other cardiovascular conditions.
    Coding noteNew temporary Category III CPT codes have been issued for these novel assistive AI algorithmic ECG risk assessments (AMA announcement).
    Evidence summarySystematic reviews report many AI-ECG models with high diagnostic accuracy for some tasks (eg, AF detection), but overall clinical utility and outcome impact remain unproven and require prospective validation.
    Multifunction cardiogram (MCG / 3DMP)
    DefinitionComputer-enhanced analysis of digital 12-lead ECG waveforms (frequency-domain/power spectral analysis) intended to detect coronary artery disease; examples include Premier Heart's 3DMP/MCG systems.
    Regulatory noteSome MCG/3DMP devices received FDA clearance (510[k]) and a CPT code was issued in 2010, but diagnostic accuracy and clinical utility remain inadequately validated.
    Performance concernsStudies (eg, Weiss et al., MED-FIT) reported missed significant lesions and low AUC (MED-FIT AUC 0.51–0.58), raising concern about use in lieu of angiography.
    LVSD / Left ventricular systolic dysfunction
    DefinitionLeft ventricular systolic dysfunction (LVSD) — commonly defined as reduced left ventricular ejection fraction (LVEF).
    Clinical relevanceAI-enabled ECG (AIeECG) algorithms have been evaluated as opportunistic screening tools to detect LVSD (median AUC ~0.90 in systematic review).
    Evidence gapNo studies examined effect on treatment or patient outcomes; clinical utility requires prospective evaluation.
    AIeECG (artificial intelligence-enabled electrocardiogram)
    DefinitionArtificial intelligence-enabled electrocardiogram (AIeECG) — algorithmic analysis of ECG signals using machine learning/deep learning techniques to detect or risk-stratify cardiovascular conditions.
    Performance evidenceMeta-analyses and systematic reviews have reported high diagnostic accuracy for tasks like AF detection and LVSD screening, but study quality and external validation vary.
    Policy stanceAlthough promising, AIeECG/assistive AI algorithmic ECG lack sufficient evidence of clinical utility for the indications addressed and are considered investigational by Aetna.