Response to Comments: Artificial Intelligence Enabled CT Based Quantitative Coronary Topography (AI-QCT)/Coronary Plaque Analysis (AI-CPA)

The following comment was submitted to Palmetto GBA, CGS, Noridian, NGS and WPS:

A comment letter was received from the American College of Cardiology (ACC) indicating support of the proposed LCD and submits additional comments for consideration.

The ACC agrees with the eligibility criteria within the Proposed LCD but states that the coverage should include both acute and stable chest pain. They refer to the AHA/ACC/Multi-society guidelines for evaluation of chest pain and cite:

• For intermediate-high risk patients with stable chest pain and no known CAD, CCTA is effective for diagnosis for CAD, for risk stratification, and for guiding treatment decisions (1A recommendation).
• For intermediate-risk patients with acute chest pain and no known CAD eligible for diagnostic testing after a negative or inclusive evaluation for ACS, CCTA is useful for exclusion of atherosclerotic plaque and obstructive CAD (1A recommendation).
• For symptomatic patients with known nonobstructive CAD who have stable chest pain, CCTA is reasonable for determining atherosclerotic plaque burden and progression to obstructive CAD and guiding therapeutic decision making (2A recommendation).

They explain, “no known CAD” is an overly broad exclusion and it would be more appropriate to use the wording, “no prior coronary revascularization and no previously documented obstructive CAD.”

They state that patients with “a prior chest CT (performed for other clinical indications) or coronary artery calcium testing demonstrating coronary calcifications or a prior coronary CT angiogram (CCTA)” demonstrating the presence of coronary plaques should not be excluded as quantifying plaque burden and characteristics may offer a potential role in patient management.

They recommend expansion of coverage to CAD-RAD 1 representing minimal stenosis (1-24%) since patients may have minimal stenosis, but still have a large amount of plaque placing them at risk for future cardiac events. “Patients with CAD RADS 1 who have significant plaque would benefit from aggressive medical interventions, as supported by the CAD-RADS consensus statement. Additionally, the ACC/AHA Chest Pain Guideline recommends reporting plaque burden in patients with non-obstructive CAD, which includes CAD-RADS 1 patients. Furthermore, the DECODE trial showed that 40% of CAD-RADS 1 patients changed treatment based on AI-CPA alone.”

They state there may be a benefit for plaque analysis for CAD-RADS 4 “as the amount of residual plaque is a strong determinant of prognosis and can be helpful in guiding patient treatment. This is particularly important as CAD-RADS 4 may be assigned to a patient that has stenosis in a distal vessel or a branch vessel. In the DECODE trial, 93.8% of CAD-RADS 4 patients changed treatment based on AI-CPA information.” They recommend expansion to CAD-RADS 4.

They suggest listing the specific contraindication to CCTA rather than “no contraindications to CCTA.”

The College agrees that AI-QCT/AI-CPA is not to be used with normal CCTA results (CAD-RADS=0 or no plaque disease). They agree that limitation for CAD-RAD 5 and recent MI are appropriate. They also agree it should not be done in conjunction with invasive coronary catheterization, unstable coronary symptoms, or serial imaging for surveillance. They suggest there may be a role for monitoring treatment impact in the future.

They recommend the “definition of atherosclerotic plaque, both calcified and non-calcified, should be defined by the FDA-approved approach being utilized and indicated in SCCT’s expert consensus document, Standards for Quantitative Assessments by Coronary Computed Tomography Angiography (CCTA).”

They conclude with a discussion on cost and pricing of AI software-based services.

Thank you for your comments. Three references not included in the proposed LCD were included with the letter and added to the LCD.

Based on the potential improved diagnostic capabilities and role for change in management based on the evidence, the use of AI-QCT/AI-CPA is determined to be reasonable for stable chest pain with no known CAD. This is supported by a 1A recommendations in the AHA/ACC/Multi-Society Guidelines. The LCD has been updated and rationale added to Analysis of Evidence section.

The commenters state that “no known CAD” is overly broad and provide a suggested alternative. However, the current AHA/ACC/ASE/CHEST guidelines utilize the term “no known CAD.” If the terminology changes in the future and the guidelines are updated accordingly, the LCD can also be updated to reflect this evolution.

The policy does not have any limitation that prohibits quantifying plaque burden after other studies have been conducted; except patients must meet the criteria for CCTA (see CCTA policy) and cannot be performed in conjunction with invasive coronary catheterization.

The commenters request expansion to CADS-RAD 1, CADS-RAD 4 and patients with known nonobstructive CAD with stable chest pain. The policy has been expanded to include CADS-RAD 1. Literature shows 40% of major adverse cardiovascular events (MACE) events occur in patients with <25% diameter stenosis, suggesting minimal plaque burden may predispose MACE. High-risk plaque features are strong predictors of adverse events, even in patients with nonobstructive CAD, as demonstrate by the PARADIGM study and PROMISE trial. Plaque analysis offers the potential to not only identify severity of stenosis but plaque burden and characterizations that may identify higher risk patients and provide the opportunity for more aggressive intervention as reported in the DECODE and CERTAIN studies. Literature has identified that escalated interventions are correlated with improved patient outcomes.

CAD-RADS 4 and known non-obstructive stable chest pain was not added to coverage due to lack of evidence to support clinical utility in this group. This population has already been identified at higher risk and more aggressive interventions recommended. There is no evidence to support improved outcomes as a result of management changes related to AI-QCT/AI-QPA in those with CADS-RAD 4 or above or known obstructive CAD. Functional assessment is a management option for CAD-RADS 4A however AI-QCT/AI-CPA is not included under the recommended options for functional testing. While there may be a benefit to understanding the plaque burden within this population, how it correlates with patient management and outcomes is not established. Future research may provide evidence and understanding of potential roles of AI-QCT/AI-CPA in these populations and can be considered through the LCD reconsideration process.

Contraindications to CCTA are listed in each MAC’s CCTA policies. To avoid any potential conflicts between policies, clarification to refer to the MAC’s CCTA policy for a full list of contraindications has been added to the LCD.

The definitions of calcified and non-calcified plaque per the SCCT Consensus document (2024) have been added to the LCD. The prior definitions are also retained as these are the definitions that utilized in most of the existing literature to date.

Pricing and cost are beyond the scope of the LCD, which is on coverage of this technology only.

The following comment was submitted to Palmetto GBA, CGS, Noridian, NGS and WPS:

The Society of Cardiovascular Computed Tomography (SCCT) supports the draft LCD and offers 2 recommendations:

1. Addition of ‘stable, symptomatic patients’
2. Addition of CAD-RADS 1 & CADS-RADS 4

They provide a summary from “CAD-RADS™ 2.0 - 2022 Coronary Artery Disease-Reporting and Data System” comparing CAD-RADS 0-4 in patients presenting with stable and acute chest pain. This included “Management Considerations” for each CAD-RADS group including a breakdown by plaque burden denoted as P1-P4 descriptors. They explain that plaque burden can alter management plans and “enhance the quality of care for patients with CAD, leading to better patient outcomes and more effective use of healthcare resources.”

Thank you for your comments and submitted literature. Any new literature not in the proposed LCD has been added. See Comment and Response 1.

With CAD-RADS 1 and stable chest pain the management considerations from “CAD-RADS™ 2.0 - 2022 Coronary Artery Disease-Reporting and Data System” are:

Consider non-atherosclerotic causes of symptoms.

• P1: Risk factor modification and preventive pharmacotherapy.
• P2: Risk factor modification and preventive pharmacotherapy.
• P3 or P4: Aggressive risk factor modification and preventive pharmacotherapy.

For CAD-RADS 1 and acute chest pain management consideration is:

• P1 or P2: Referral for outpatient follow-up for risk factor modification and preventive pharmacotherapy.
• P3 or P4: Referral for outpatient follow-up for aggressive risk factor modification and preventive pharmacotherapy.

Given a change in management and potential for improved patient outcomes, this coverage has been expanded to include stable chest pain and CAD-RADS 1.

With CAD-RADS 4 management was not changed based on plaque burden. The recommendations for CAD-RADS 4 are “P1, P2, P3, or P4: Preventive management, including aggressive preventive pharmacotherapy.” Therefore, it is not clear how plaque burden alters patient management or improves outcome in this group and therefore coverage is not expanded. Future evidence can be considered via the LCD Reconsideration process.

The following comment was submitted to Palmetto GBA, CGS, Noridian, NGS and WPS:

A comment letter was received from Cleerly who “applaud” the decision for coverage of AI-QCT and request expansion of coverage to those with stable chest pain and CAD-RADS1. They explained this is consistent with the 2021 American College of Cardiology/American Heart Association (ACC/AHA) Guidelines.

The letter concludes with a thank you “for determination of coverage for the clinically valuable AI-QCT procedure. We respectfully request that the coverage policy eligibility reflect that of evidence-based professional societal guidelines derived from large-scale landmark multicenter and randomized controlled trials performed over the last 2 decades. Your dedication and expertise in the field of cardiology care are invaluable and we deeply appreciate your time, effort, and commitment to continuing the best care for patients. Together, we can make significant strides towards advancing cardiology care and improving patient outcomes.”

Thank you for your comments and the submitted literature. New literature has been reviewed and added to the LCD. We agree with expansion of coverage for stable chest pain and CAD-RADS 1. These changes have been made in the LCD. See Comments and Responses 1 and 2.

The following comment was submitted to Palmetto GBA, CGS, Noridian, NGS and WPS:

A comment letter was received commending the proposed LCD, “We greatly appreciate the evidence review on this topic and we are supportive of the direction of the Proposed LCD and Draft Article”. The commenter provides some recommendations for revision:

1. Request expansion of coverage to include stable chest pain, CAD-RADS 1 and CAD-RADS 4.
2. Recommend limitation #5 be revised to remove high grade stenosis (>70%) and CAD-RADS 4.
3. Request modifications to the definition section:
   • Revise the definition of “Artificial Intelligence Enabled CT Based Quantitative Coronary Topography (AI-QCT)/Coronary Plaque Analysis (AI-CPA)” as follows: “Artificial intelligence application to imaging obtained through coronary computed tomography (CT) scans to quantify and characterize coronary atherosclerotic plaque to assess severity of coronary disease calculate coronary artery dimensions and degree of stenosis per vessel and coronary plaque composition and burden.”
   • Remove “Coronary Plaque Analysis (CPA),” “Quantitative Coronary Plaque Analysis (QCPA)” and “Quantitative Coronary Topography (QCT)” as defined terms.
4. Request the addition of 3 articles to the evidence review (Narula 2024, Nurmohamed 2024, Dundas 2024).
5. Request multiple ICD-10 codes to be added to the billing and coding article.

The commenter requests clarification be made including:

In #2 of the Coverage Indications section, add language making clear that “no known coronary artery disease (CAD)” means that the patient did not have known CAD prior to the CCTA imaging that is related to the covered AI-QCT/AI-CPA service in the Proposed LCD. Ensure limitations are clear on what is not reasonable and necessary. Add software is FDA cleared or approved. Provider qualifications should align with that required to CCTA.

Thank you for your comments and support of the policy. The policy has been expanded to cover stable chest pain and CAD-RADS 1 but not CAD-RADS 4 or known CAD as explained in Comments and Responses 1 and 2. There is no revision to limitation #5.

There was no supporting evidence to support a change in definition. These definitions were derived from the literature as well as subject matter experts input obtained during the contractor advisory meeting and were felt to best represent the current technology. Therefore, no changes to these definitions were made.

The 3 submitted articles have been added to the LCD.

We do not agree with the additional ICD-10 codes. These codes may represent appropriate codes to determine the need for a CCTA study, but in order to proceed to AI-QCT/AI-CPA the CCTA must have abnormal findings consistent with the coverage requirements in the policy and related billing and coding article.

Suggested clarifications have been made. We did not change provider requirements as this is determined by the training program and credentialing bodies. The existing language states those reading the test need to be appropriately trained to do so and that should be part of their training experience.

The following comment was submitted to Palmetto GBA, CGS, Noridian, NGS and WPS:

A comment supports the LCD and provides the following suggestion: To remove the word acute from chest pain and add obstructive as a descriptor of CAD.

Thank you for your comments and support of the policy. We will retain terminology to align with clinical trials and AHA/ACC Guidelines for clarity and specific population for which the evidence supports at this time.

The following comment was submitted to Palmetto GBA, CGS, Noridian, NGS and WPS:

A comment letter was received stating, “We support the policy as a whole and believe it will greatly benefit patients and healthcare providers alike.” The commenter suggests inclusion of stable chest pain and to remove the limitation “not in conjunction with invasive coronary catheterization.” They cite the CT-PERICISION registry for supporting evidence.

They request to modify the title to be vendor agnostic and avoid potential confusion stating: “We recommend that the LCD title be revised to be product-agnostic rather than naming specific brand-specific products currently available on the market. This approach will ensure there is no confusion about the intent of the policy and clarify that the LCD encompasses all relevant technologies, rather than being confined to specific product names. Additionally, it will ensure that all products will be evaluated based on functionality and compliance with existing CPT codes, rather than brand-specific considerations. Furthermore, maintaining a product-agnostic title will prevent potential confusion among providers who might otherwise misinterpret the policy as being applicable only to specific products.”

Thank you for your comments and support of the policy. The policy has been expanded to cover stable chest pain. See Comments and Responses 1 and 2.

The title of the LCD is not specific to any product currently on the market. The title and definition for AI-QCT/AI-CPA was derived from the clinical literature and input from subject matter experts during the CAC meeting. The LCD specifies, “Software to perform AI-CPA must be FDA cleared or approved.” No suggested alternative title was provided.

Thank you for providing the supporting literature for expansion of coverage to inclusion in conjunction with invasive coronary catheterization. While this study does provide clinical validity that CTA derived non calcified plaque version may aid in visually identified side branch occlusion. This study does not specify the patient population for which this technology would be applicable or how this improves clinical outcomes. If additional literature further clarifies the role of combining these tests, it can be submitted through the LCD reconsideration process.

The following comment was submitted to Palmetto GBA, CGS, Noridian, NGS and WPS:

A comment letter supports the policy and requests expansion to stable chest pain and mild stenosis (1-25%) who have a high cardiac risk profile or symptoms suggestive of ischemia. They cite the ROMICAT II trial for support in acute chest pain, CONFIRM with stable chest pain, PARADIGM for association of high-risk plaque with increase in MACE.

Thank you for your comments. The policy has been expanded to include stable chest pain and CAD-RADS 1. See Comments and Responses 1 and 2.

The following comment was received from multiple stakeholders:

Multiple commenters cite multiple studies demonstrating that medical intervention can reduce plaque burden. One commenter explains this is significant, as evidence that a reduction of as little as 1% can lead to a 20% reduction in MACE. They also discuss that the SCOT-HEART trial shows that high amounts of plaque are associated with a 7.3 times higher risk of heart attack. Comments are supported by multiple papers including Dawson 2022, Bhindi 2022, Williams 2021, Dundas 2024, Gulanti 2021, SCOT-HEART, and ADVANCE registry.

Thank you for your comments and supporting evidence. We agree that coverage should include stable chest pain and the policy has been updated to reflect this. See Comments and Responses 1 and 2. Dawson was not added to the LCD since this was a review paper, but all other literature has been added to the LCD if it was not already cited.

The following comment was received from multiple stakeholders:

Multiple providers from across the country submit overall support of the policy and request expansion to stable chest pain and CAD-RADS 1. They refer to the 2021 ACC/AHA Chest Pain Guidelines, CAD-RADS 2.0-2022 Consensus document, SCOT-HEART, PARADIGM, PROMISE, REVEALPLAQUE, CREDENCE, CONFIRM and DECODE trials for supporting literature.

A provider cites Kim, Lipkin, Nurmohamed and DECODE studies for support. They state: “Aside from the immediate improved patient outcome of prevention of an invasive procedure and potential complications, which are increased in Medicare patients, we know from the attached PARADIGM Study, that optimal medical therapy has been shown to slow overall plaque progression, increase plaque calcification, and decrease high-risk plaque characteristics. Converting high-risk, non-calcified plaque to low-risk, calcified plaque and slowing disease progression decreases the likelihood of future cardiac events, which offers long-term improved patient outcomes as well.”

Most commenters limited their requests to stable chest pain and CAD-RADS 1 but several also request expansion to CAD-RADS 4, and 1 commenter requested CAD-RADS 1-5. Some commenters recommend expansion to those status post coronary intervention of bypass grafting.

Thank you for your comments. The policy will be expanded to cover CAD-RADS 1 and stable CAD, but not CAD-RADS 4. See Comments and Responses 1 and 2. All supporting literature has been added to the policy if it was not already cited. For definitions see Comment and Response 4. If additional literature is developed that supports expansion to use in status post coronary intervention of bypass grafting that can be submitted via the LCD reconsideration process.

A provider encourages coverage for stable chest pain and that the knowledge base in this area continues to grow. In addition to established societal guidelines covering patients with stable chest pain, recent data continue to point to benefits of plaque analysis in patients with stable CAD. PREVENT for example (Lancet May 2024) shows significant benefits (reduction in MACE) with preventive intervention to patients with stable and acute chest pain who are found to have non-flow limiting vulnerable plaques.

Thank you for your comments and submission of new literature which has been added to the LCD. See Comments and Responses 1 and 2.

The following comment was submitted to Palmetto GBA, CGS, Noridian, NGS and WPS:

A commenter from the state of Florida supports coverage of AI-QCT. They request expansion of coverage to include those with 1-25% stenosis and stable chest pain with low or intermediate pre-test probability of CAD. Supporting references provided.

This policy is proposed by 5 Medicare Administrative Contractors (MACs) and does not include the MAC for the state of Florida. We recommend contacting them directly with your concerns regarding this topic.

The following comment was submitted to Palmetto GBA, CGS, Noridian, NGS and WPS:

A patient shares their story in which he attributes AI-QCT/AI-CPA to saving his life. He shares his support of coverage for this potentially lifesaving test for Medicare.

Thank you for sharing your story and support of the policy.

The American College of Radiology (ACR) comments supporting the LCD and recommending extending coverage to stable chest pain and CAD-RAD 1. They cite the 2021 ACC/AHA Guidelines, SCOT-HEART, CAD-RADS 2.0 and PARADIGM study for supporting evidence.

Thank you for your comments. The policy has been expanded to cover stable chest pain and CAD-RADS 1. See Comments and Responses 1 and 2.

The following comment was received from multiple stakeholders:

Multiple providers write in support of the policy. They state this technology enhances their ability to care for their patients. One such commenter explains: “plaque analysis enables him to provide better diagnosis for his patients by cutting the downtime for analysis from hours to minutes and results in more accurate and comprehensive evaluation of the anatomy. He is able to determine which lesions are excrement and help triage patients who need catheterization avoiding unnecessary downstream test. It helps him to treat patients with more plaque that are at higher risk of plaque rupture and subsequent MI with more aggressive medical therapy. He explains where he practices has some of the worst rates of coronary artery disease associated with risk factors such as diabetes, obesity, smoking and this aids in the care of this high-risk population. He states companies such as “Cleerly, Elucid, and Heartflow have a robust evidence base and expertise in this area and offer plaque metrics that would benefit from having paid coverage.”

Another commenter explains the value of earlier diagnosis and treatment of CAD. “As an invasive cardiologist, I see the value in coronary CT angiography as a gatekeeper to invasive cardiac catheterization and coronary angiography. The ISCHEMIA study demonstrated that 80% of patients randomized to CT angiography only and treated with medical therapy did equally as well as those randomized to early invasive angiography and revascularization. I believe that CT angiography will provide greater detail of the vasculature in terms of calcification, vessel diameter, lesion length, and other characteristics that can be helpful in patients undergoing invasive angiography and revascularization.”

Another provider states: “During my lifetime there have been many advances. Cardiac CTA is 1 of them. CTFFR is another; however, the most important is probably for us to be able to determine who is at risk for a cardiac event and who is not. Plaque analysis has that potential. We know very well that lesions that appear to be significant may not effect/ block the flow of blood and vice versa meaning there are lesions that appear to be only moderately obstructive but are active and at high risk of acute thrombosis in the future. Having coverage for plaque analysis is important for the care of our patients.”

“In all the years that I have been practicing medicine, I have never seen a technology as useful as CCTA with AI-QCT. This will allow cardiologists for the first time to treat, adjust treatment and reduce cardiac events. I applaud the recent decision by the local MAC to extend coverage to Artificial Intelligence Enabled CT-Based Quantitative Coronary Topography (AI-QCT) for coronary plaque analysis.”

Thank you for your comments and support of the policy. We also thank providers caring for high-risk patients and the research being done to advance cardiac care.

A provider recommends that the definitions of atherosclerotic plaque should align with the newly published consensus statement “Standards for quantitative assessments by coronary computed tomography angiography (CCTA): An expert consensus document of the society of cardiovascular computed tomography (SCCT).” They also request expansion of coverage to stable chest pain.

Thank you for your comments. The policy has been expanded to cover stable chest pain and the definitions revised. See Comments and Responses 1 and 2