Response to Comments: Non-Invasive Fractional Flow Reserve (FFR) for Stable Ischemic Heart Disease

The following comment was received from multiple stakeholders:

We have worked with HeartFlow in the development of our program over the past 2 years. As part of the process, we have sought out experts in the field to ensure we are sending appropriate studies of excellent quality for analysis.

As part of this process, we invested in a higher quality scanner, which allows us to achieve much more flexibility in scanning parameters and obtain excellent CT coronary imaging results at higher heart rates. This is important to understand because if the CT quality is sufficient, then the FFR-CT values are accurate and enable us to incorporate the results of these patients with stable angina in a meaningful way. Our acceptance rate for analysis is currently 92%, which is above the national average.

The use of FFR-CT helps enable physicians to choose appropriate treatment strategies (medical vs. intervention). In studies, the use of FFR-CT has changed patient management in 63% of cases (Fairbarn, et al. Eur Heart J. 2018). It has also been shown that we can safely avoid invasive coronary angiography in > 60% of cases where invasive study was initially recommended (Douglas, et al. Eur Heart J. 2015. 36;470:3359-67) Long term outcomes have also demonstrated that use of FFR-CT is a safe strategy and that clinical pathways which incorporate FFR-CT can decrease long term health care cost.

During research studies, FFR-CT has demonstrated higher diagnostic performance when compared to standard coronary CTA, SPECT and PET for vessel level ischemia (PACIFIC FFR-CT study). Thus, when establishing policies, we want to ensure that appropriate patients are financially covered.

We have several concerns with the proposed LCD including the following:

• Restrictions on single use vessel disease
• Indications which would be most appropriately placed in a different LCD
• Indications related to coronary CT angiography image quality
• Restrictions on use of FFR-CT in severe 3 vessel disease

We do feel that there are incomplete indications for the LCD. The policy should include the right coronary artery. Coverage criteria should at minimum include the proximal and mid portions of the left anterior descending artery and left circumflex artery. However, by CT criteria, we should be able to interpret any coronary segment > 2 mm, which would be more inclusive. We also recommend removing the referral to the dominant system.

Single Vessel Disease: The proposed LCD seeks to limit utilization of FFR-CT in patients with single vessel disease and insufficient response to optimal medical therapy (OMT). This restriction should relate to appropriate use of revascularization instead of restricting use of a diagnostic test. The information obtained from FFR-CT can help guide the treatment decision for medical therapy vs. need for revascularization.

We are in agreement that it is reasonable not to cover patients who are uncooperative or unable to sustain breath holds for 5-10 seconds. However, this is a clinical judgement call at the time of the scan. In regards to allergies or intolerance to the contrast, there are many times when allergies listed are not true allergies and the patient is able to tolerate contrast well. This is a physician judgment call. Each of these issues, as well as renal insufficiency are related to restrictions for a baseline cardiovascular CT angiogram and have no relation to the acquisition or performance of the FFR-CT service. Once the images are acquired, the decision to proceed with FFR-CT is independent of these factors.

FFR-CT in relation to CCTA:

1. Severe obesity (BMI > 35 kg/m²): FFR-CT is accurate if quality of scan passes the standards regardless of BMI. With newer CT technology, we can enhance image quality by adjusting radiation parameters and injecting the contrast at higher rates. In addition, weight distribution could be more abdominal in nature, which does not adversely affect coronary images. When reviewing patients, our goal is to only proceed with CT if we think we can get a diagnostic study. Macron, et al. (Archives of Cardiovascular Diseases Supplements, 2016;(8):1-22) demonstrated that they were able to analyze 97% of vessels in CT patients with mean BMI of 36 kg/m². In addition, other imaging modalities, i.e., SPECT, have a higher false positive rate in the obese patients secondary to attenuation artifact. Thus, CT coronary angiography may still be a better option for the obese patient.
2. Bare metal stent: Heart flow can process FFR-CT results on patients who have had stents to one vessel. However, they can only interpret the vessels without stent. HeartFlow has stated that there are 3 scenarios where FFR-CT cannot be utilized (and therefore would be appropriate to not cover):
   1. Stent is present in the left main coronary artery
   2. A stenosis > 30% in the left main coronary artery and 1 or more stent in the left system
   3. Stents present in 2 or more coronary systems
3. Prosthetic valves: Some patients with valves may not have had significant disease prior to valve surgery. While there is some beam hardening artifact from the valves, these artifacts can be minimized with use of certain scanning parameters and reconstruction algorithms. It is not prohibitive for us to interpret coronary arteries on these studies.
4. Heart rate arrhythmias > 65 beats per minute: Current high-end CT scanners are able to scan studies effectively at much higher HR. Macron, et al. have shown that they could analyze 97% of vessels on studies with a mean HR of 91. In addition, SCCT has stated that "heart rates > 80 bpm used to represent a relative contraindication for CCTA because of high incidence of motion artifacts. However, ongoing hardware and software improvements, such as dual-source CT and wide-detector scanner, now allow the imaging of patients with higher and irregular heart rates with good imaging success. The limitations in regard to higher heart rates depend on the equipment used, and scan settings have to be adjusted accordingly." In regards to FFR-CT, the accuracy is still good if it passes the quality process regardless of the patient's heart rate, although the goal would still be to aim for a HR < 65 bpm.
5. Extensive coronary calcification (Agatston score > 1000): In essence, the distribution of the calcium and the size of the vessels is much more important than the total score. This is why most physicians will personally review the high score and determine if it is acceptable to proceed with the coronary CTA. SCCT guidelines recommend that "the decision to proceed with CCTA in the presence of high coronary calcium score should be left to the discretion of the referring and attending physician (Abbara, et al. JCCT, 2009. 3(3):190-204). The NXT Calcium paper (Norgaard, et al. JACC Cardiovasc Imaging. 2015. 8(9):1045-1055) has also shown that FFR-CT provided a higher diagnostic performance and discrimination for ischemia over a wide range of coronary calcification severity and that the performance of FFR-CT was superior to CCTA alone regardless of calcium score.
6. Renal insufficiency (i.e., GFR < 60ml/min/1.73 m²): Every CT department will have contrast policies. If the GFR is at a certain level, there still may be an ability to proceed with the test with hydration or holding certain medications. This decision should be left to discretion of interpreting/referring MDs when deciding the risk/benefit of the study compared to other imaging modalities. This does not impact on the use of CT-FFR but is a clinical decision by the department if it is safe to proceed with the CTA.
7. Severe aortic stenosis: While we are often being asked to review the coronary arteries on some of the TAVR CT angiograms, we do not administer NTG to these patients. If there are borderline lesions, it is reasonable to recommend another diagnostic modality as FFR-CT requires administration of NTG. A more relevant determination would be to provide coverage of FFR-CT in patients who were able to receive NTG rather than restricting disease states.
8. Known severe 3-vessel disease(> 50% in all 3 major vessels): This is an area of strength compared to standard of care (SPECT). With SPECT, you are at risk of having balanced ischemia and reporting a normal nuclear scan without evidence of ischemia. However, we know from the FAME trial data that some 50% blockages are hemodynamically significant whereas other 70% blockages are not. The advantage to using FFR-CT is that the interventionalist can assimilate the data to enhance his decision-making process since FFR-CT is accurate regardless of disease burden if the examination is a good study. In the ADVANCE registry (Eur Heart Journal 2018), 23% of patients were recommended for revascularization. Using FFR-CT, 22% were reclassified to medical therapy alone. The SYNTAX Ill trial (Collet, et al. JACC, 2018. 71(24):2756-2769) demonstrated that FFR-CT had good accuracy in detecting functionally significant lesions in patients with 3 vessel CAD. Noninvasive functional Syntax score (FSS) using FFR-CT reclassified 305 of patients from high and intermediate SS tertiles to the low risk tertile.
9. Post coronary artery bypass surgery, evaluation of venous grafts: HeartFlow currently does not recommend FFR-CT be performed in this patient population.
10. ACS patient: Currently the appropriate use criteria indicates that CTA is appropriate in low to intermediate risk patients. Our policy is to proceed if they are low to intermediate risk with one negative troponin. Given that we are not completing a stress test on these patients, we feel that one troponin is sufficient.

Again, thank you for the opportunity to provide feedback. Several of the current restrictions in the proposed LCD will limit the ordering of FFR-CT and place financial burden on providers and patients. Access to FFR-CT can lead to fewer unnecessary diagnostic catheterizations making it a cost-effective diagnostic test. We feel that the appropriate use of cardiovascular CT and FFR-CT will help transform the care of our patients.

Thank you for your comments we will address each one below.

Single Vessel disease: We agree with the removal of the single vessel disease and the reference to optimal medical therapy and have removed these requirements.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing, therefore, does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Prosthetic valves: Strict adherence to careful patient selection contributed significantly to demonstrating the net health benefit for FFRCT. No supporting research was submitted with these comments. Pivotal trials for FFRCT consistently excluded patient populations with prosthetic valves and extensive coronary calcification from study. Furthermore, these limitations are reported by the manufacturer in FDA filings. Due to the unavailability of U.S.-based, high-quality, peer-reviewed published clinical research in patient populations with extensive calcification or prosthetic valves, we must conclude that it is still investigational and, therefore, unreasonable to consider for coverage at this time. Upon availability of this evidence, as well as removal of this limitation from the FDA-validated filings, we can reconsider this coverage limitation upon request. The limitation will remain as written.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Severe aortic stenosis: Thank you for the comment we agree with maintaining this restriction.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Post coronary artery bypass surgery: Thank you for your comment, we agree.

ACS (acute coronary syndrome): While we realize this is an emerging area with ongoing studies, the current FDA indication as well as sufficient well-designed studies are not available to support this being a reasonable and necessary indication for FFRCT at the current time. We will be happy to reevaluate if such studies are published and supplied for review.

I am a cardiologist specializing in interventional cardiology and all types of cardiac imaging including echocardiography, nuclear cardiology, and cardiovascular computed tomography.

I wanted to share my thoughts on Palmetto's proposed Local Coverage Determination (LCD) on FFRCT. I hope that you will take my comments into consideration as you finalize the coverage indications. My specific comments are regarding the indications which are considered not reasonable and therefore not covered. These restricted indications include the following per your draft policy document:

• Severe obesity (BMI >35 kg/m²)
• Bare metal intracoronary stent
• Prosthetic valves
• Heart rate or arrhythmia> 65 beats/min at the time of image acquisition
• Allergy or intolerance to iodinated contrast material
• Extensive coronary calcification by plain film or high calcium scores with prior Agatston score> 1000
• Renal insufficiency (i.e., estimated glomerular filtration rate< 60 mL/min/1.73m²)
• Uncooperative patient
• Patient unable to sustain breath-hold for 5-10 seconds
• Heart transplantation

My overall comments regarding FFRCT based on the literature and my experience in over 100 patients is as follows. The overall accuracy of FFRCT compared to other non-invasive cardiac tests is high in my experience with my patients and the literature bears this out. The NXT trial showed a per-vessel accuracy of 86% compared to invasive FFR in 254 patients. The PACIFIC study demonstrated 87% per vessel accuracy compared to invasive FFR, higher than SPECT, PET and CCTA. I have come to rely on FFRCT for clinical decision making given its high accuracy for predicting lesion specific ischemia. It is my understanding that FFRCT processing involves a quality assurance process that evaluates each patient dataset for image quality and suitability for calculation of FFRCT. To maintain the accuracy of the test the manufacturer does not provide FFRCT results, and does not charge providers, if the image quality is inadequate. This is done regardless of patient characteristics. Also, the manufacturer has worked with our imaging team to optimize our CCTA protocols to ensure that the appropriate patients can be processed for FFRCT. Thus, I feel confident in using FFRCT in patients with some of the exclusions below.

My specific thoughts regarding patients with BMI > 35 is as follows - The image quality in these obese patients depend on CT scan protocols, expertise of the imaging team and the scan equipment. We have upgraded our scanner to the latest generation 256 slice Philips iCT scanner and trained our team on this CT platform. We find that we can get good quality images in these patients with the right patient specific tailoring of protocols, and with the use of modern reconstruction methods. Similar results have been reported in the literature by Macron, et al. who showed that physicians were able to successfully analyze 97% of vessels in patients with a mean BMI of 36.0 ± 5.9 kg/m² using CCTA. I would recommend allowing imaging for patients with a BMI up to 45 kg/m².

The alternative to a CCTA first pathway is to do stress tests on these patients with suspected coronary artery disease. It has been shown that stress tests result in very high rates of unnecessary ICAs [Patel paper], whereas CT with FFRCT as needed has been shown to reduce unnecessary ICAs significantly. The reduction of unnecessary ICA is especially relevant for morbidly obese patients as the rate of vascular access-site complications for patients undergoing ICA is highest in these patients. Also, as stated above, the provider of FFRCT has an acceptance process, and acceptance criterion that is based on image quality of the patient's CCTA data and provide results only if the images are good enough to calculate FFRCT. The results from NXT and PACIFIC show that FFRCT has high accuracy for cases that are accepted for processing by the manufacturer, regardless of patient BMI.

I also disagree with the exclusion of patients with HR > 65 as I find that it is possible to get good image quality at higher heart rates with modern CT scanners. The SCCT guidelines state clearly that with modern scanner technologies like dual source, wide detectors, and improved software it is possible to successfully image patients with high and irregular heart rates. The literature also supports this. A study from Macron, et al. showed that readers were able to successfully read 97% of vessels utilizing CCTA in patients with a mean heart rate of 91.4 ± 48.9 bpm.

I use CT and FFRCT on many patients with heart rates > 65 BPM who are suitable for CCTA, have symptoms and suspicion of coronary artery disease. In most of these patients I am able to get high quality CCTA and FFRCT results. My understanding is that the provider of FFRCT builds the anatomic model using a combination of multiple cardiac phases, which makes it possible to compute FFRCT results in many cases with cardiac motion. For a minority of patients in which FFRCT results cannot be calculated, the manufacturer of FFRCT technology does not provide results and does not charge our institution. I would recommend modifying the language to allow FFRCT for patients with heart rates up to 90 bpm.

Next, I think that the exclusion of patients with coronary stents is unnecessary and should be removed. I have many patients with known coronary artery disease and prior revascularization with a high risk of disease progression and often my clinical question is in the vessel that does not have the stent. The manufacturer of FFRCT does not provide results in the vessels that do contain a stent enabling me to use the functional information from non-stented vessels in decision making.

Finally, I would like to argue against the proposal to exclude patients that have severe 3-vessel disease. The SYNTAX III trial has shown that FFRCT has good accuracy in detecting functionally significant lesions in patients with 3-vessel CAD, and that Noninvasive functional Syntax score (FSS) using FFRCT reclassified 30% of patients from the high and intermediate risk to the low risk group. Similarly, the ADVANCE registry shows that FFRCT changed the management plan for 22.9% of the patients, with 22.3% reclassified from revascularization to medical therapy. In my view FFRCT provides relevant functional information that changes patient management in patients with 3 vessel disease, and this results in more appropriate treatment of these patients. I would like to use FFRCT in these patients.

Thank you for your comments we will address each specific concern.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Heart rate >65: We agree that the Macron study, as well as improvements in imaging technology, are satisfactory to remove the requirement. Also, this restriction is more related to CCTA acquisition, and we added a requirement for a technically adequate CCTA study as a requirement for FFR.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing, therefore, does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Thank you for the opportunity to comment on the draft LCD and provide additional insights that will ensure appropriate access to FFRct for Medicare beneficiaries who present with symptoms suggestive of stable ischemic heart disease (SIHD). In short, HeartFlow would like to comment on the following restrictions in the proposed LCD:

1. Severe obesity (BMI >35 kg/m²) – There is ample clinical evidence that accurate FFRct results can be obtained in patients with a BMI much greater than 35 kg/m². We suggest that Palmetto remove this restriction from the final LCD.
2. Bare metal intracoronary stent – Accurate FFRct results can be provided, even for patients with intracoronary stents (although not for the specific vessels which contain stents). There are some circumstances, described in the body of our comments, under which FFRct results cannot be provided. We suggest that Palmetto amend the final LCD language to align with these restrictions.
3. Heart rate or arrhythmia > 65 beats/min at the time of image acquisition – There is ample clinical evidence that accurate FFRct results can be obtained in patients with heart rates much greater than 65 beats/min. We suggest that Palmetto remove this restriction from the final LCD.
4. Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000 – There is ample clinical evidence that accurate FFRct results can be obtained in patients with high calcium scores. There is no known cutoff value above which FFRct accuracy declines. We suggest that Palmetto remove this restriction from the final LCD.
5. Known severe 3-vessel disease (> 50% stenosis is all 3 major vessels) – FFRct is accurate regardless of patient disease burden and there is clinical evidence demonstrating that FFRct provides additional information to inform decision-making for patients with severe coronary disease and 3-vessel disease. We suggest that Palmetto remove this restriction from the final LCD.

Background

HeartFlow is transforming the way cardiovascular disease is diagnosed and treated. The FFRct service is the first available non-invasive solution that enables a physician to evaluate more accurately whether a patient has significant SIHD based on both anatomy and physiology.

As described in the ‘Summary of Evidence’ section of the proposed LCD, FFRct allows physicians to choose safely the most appropriate treatment for their patients. Use of FFRct has been demonstrated to change patient management in 63% of cases and has allowed safe avoidance of invasive coronary angiography (ICA) in > 60% of cases where ICA was initially intended for patient with stable symptoms of SIHD. Long term (1-5 years) outcomes have been studied and reported in over 6,000 patients indicating the safety of patient management strategies that incorporate FFRct. Finally, clinical pathways that utilize FFRct significantly reduce healthcare costs.

Rationale for Removing or Modifying Coverage Restrictions in the Proposed LCD

The FFRct Analysis involves creation of a patient-specific anatomical model of the coronary arteries based on coronary CT angiogram (CCTA) images. Accuracy of the anatomic model is dependent on the ability to: (a)define the lumen boundaries of the entire coronary tree, (b) see all major branches, and (c) segment the myocardium. HeartFlow has a validated CT image quality process that assesses the quality of incoming CCTA images. This process ensures that the above anatomical features can be resolved adequately to enable computation of FFRct. CCTA datasets that do not meet the quality criteria are rejected, FFRct results are not produced, and neither the clinician nor patient is charged. The image quality process has been validated by the FDA and has demonstrated the ability to produce FFRct accuracy in line with our FDA label.

Multiple clinical trials have been conducted using this CCTA acceptance process and all have demonstrated that FFRct has high and consistent accuracy compared to the gold standard of invasive FFR. Investigators in the NXT study demonstrated high per-vessel diagnostic performance [Accuracy 86%, Sensitivity 84%, specificity 86%] for FFRct compared to invasive FFR. The PACIFIC FFRct study showed that FFRct demonstrated higher diagnostic performance than standard CCTA, SPECT, and PET for vessel-specific ischemia.

This information is shared to demonstrate that regardless of the underlying characteristics of the patient or CCTA image itself (e.g., BMI > 35 kg/m², presence of intracoronary stents, heart rate > 65 beats/min, extensive coronary calcification, severe 3-vessel disease, etc.), accurate FFRct results can be provided as long as the CCTA images pass HeartFlow’s quality criteria. As a reminder, there is no charge for FFRct for CCTA exams that fail the quality process. The image quality process and HeartFlow’s ability to provide accurate FFRct results independent of patient and CCTA characteristics form the basis for our arguments to remove the restrictions in the proposed LCD.

1. Severe obesity (BMI >35 kg/m²)

As stated above, HeartFlow can provide accurate and actionable FFRct results regardless of a patient’s BMI if their CCTA scan passes our image quality process. A recent publication based on data from the ADVANCE registry showed that BMI was not a statistically significant predictor of CT image quality rejection, despite inclusion of a subjects with a wide range of BMI (BMI: 26.3 ± 4.8 kg/m²). The ADVANCE registry examined the impact of CCTA and the FFRct Analysis on clinical decision making, safety, and patient outcomes in over 5,000 patients across Europe, North America, and Japan.

Investigators have also demonstrated the ability to acquire high quality CCTA images regardless of BMI in independent publications. A study from Macron, et al. showed that physicians were able to successfully analyze 97% of vessels utilizing CCTA in patients with a mean BMI of 36.0 ± 5.9 kg/m². Another publication from Mangold, et al. demonstrated that CCTA provides high diagnostic accuracy in both obese and non-obese patients: there was no significant difference in diagnostic accuracy for obese (≥ 30 kg/m²) and non-obese (< 30 kg/m²) patients. Although the CCTA exams in these two studies were not sent for FFRct Analysis, we are confident the exams would pass our image quality process and accurate FFRct results could be provided. Restricting access for Medicare beneficiaries based only on BMI rather than the image quality of the CCTA exam would unnecessarily restrict access to FFRct. Were the proposed LCD to take effect, obese patients would likely be exposed to unnecessary ICAs which pose a greater risk of vascular access-site complications. We recommend that the BMI restriction be removed from the final LCD and that Palmetto rely on our image quality process to determine whether FFRct is feasible and reasonable.

2. Bare metal intracoronary stent

HeartFlow can process FFRct results for patients regardless of whether they have a bare metal or drug-eluding intracoronary stent. FFRct results will be returned for all vessels which do not have an intracoronary stent, enabling the physician to determine the functional significance of SIHD in vessels which have not been previously stented. The vessel with the stent will be “grayed out,” and calculated FFR values will not be provided.

HeartFlow cannot provide FFRct results for patients when the following stent and disease status are present:

1. A stent is present in the left main coronary artery
2. There is a stenosis > 30% in the left main coronary artery and 1 or more stent in the left system
3. There are stents present in 2 or more coronary systems (e.g., the left circumflex artery and the left anterior descending artery)

We recommend that the intracoronary stent restriction in the final LCD be amended to align with the situations under which HeartFlow cannot provide FFRct results.

3. Heart rate or arrhythmia > 65 beats/min at the time of image acquisition

As stated above, HeartFlow can provide accurate and actionable FFRct results regardless of a patient’s heart rate or the presence of arrhythmia at the time of image acquisition if their CCTA scan passes our image quality process. Investigators in both the NXT trial and the PACIFIC FFRct trial demonstrated the ability to obtain high quality CCTA images that led to accurate FFRct results regardless of patient heart rate. Patients in the NXT trial had a mean heart rate of 63 ± 10 beats/min and patients in the PACIFIC FFRct trial had a mean heart rate of 57.8 ± 7.7 beats/min. These trials demonstrate HeartFlow’s ability to produce accurate FFRct results across a wide range of heart rates and CT scanner technology.

Over the past decade CT scanner technology has continued to advance and now provides significant functionality that is focused on imaging the heart and coronary arteries. The Society of Cardiovascular Computed Tomography (SCCT) states that “heart rates > 80 bpm used to represent a relative contraindication for CCTA because of high incidence of motion artifacts. However, ongoing hardware and software improvements, such as dual-source CT and wide-detector scanner, now allow the imaging of patients with higher and irregular heart rates with good imaging success. The limitations in regard to higher heart rates depend on the equipment used, and scan settings have to be adjusted accordingly.”

While these newer technologies do provide advantages and, in general, improved CCTA image quality for patients with higher heart rates, they are not essential to perform FFRct. Figure 1 demonstrates CCTA acceptance rates for over 13,000 patients in our regular clinical environment. Although increasing heart rate is associated with a lower acceptance rate, patients with a heart rate > 65 beats/min still enjoy an acceptance rate ranging from 91% to 74% (for patients with a heart rate of 86-90 beats/min). Were the proposed LCD to take effect as written, these patients would lose access to FFRct and could face unnecessary ICAs and the associated costs and risks. We recommend that the heart rate restriction be removed from the final LCD and that Palmetto rely on our image quality process to determine whether FFRct is feasible and reasonable.

4. Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000

As stated above, HeartFlow can provide accurate and actionable FFRct results regardless of a patient’s calcium burden if their CCTA scan passes our image quality process. In fact, a recent publication from investigators involved in the NXT trial demonstrated that FFRct has similar accuracy regardless of a patient’s calcium burden. Using data from the NXT trial, investigators divided the patients into four quartiles based on their Agatston score and reviewed FFRct accuracy across quartiles. The results showed “there was no statistically significant difference in per-patient or per-vessel diagnostic accuracy, sensitivity, and specificity of FFRct across Agatston Score quartiles”. The investigators also reported that the accuracy of FFRct for patients in the highest quartile of Agatston score was “equal to or superior to diagnostic performance characteristics reported for most conventional noninvasive ischemia testing modalities when compared with directly measured FFR.”

Because of the accuracy of FFRct is consistent regardless of a patient’s calcium score and there is no known upper limit for which FFRct results cannot be processed, we recommend that the Agatston score restriction be removed from the final LCD and that Palmetto rely on our image quality process to determine whether FFRct is feasible and reasonable.

5. Known severe 3-vessel disease (> 50% stenosis is all 3 major vessels)

FFRct accuracy has been tested and is high across a wide range of disease burdens and presentations, including severe 3-vessel disease. FFRct has also proven to be clinically effective and lead to change in treatment strategy for these patients. Data from the ADVANCE registry showed that physicians change treatment decisions for patients initially triaged to coronary artery bypass surgery (CABG) based on CCTA results once FFRct information is made available: 1.4% (n=12) of these patients had their treatment decision changed from CABG to medical therapy and 4.4% (n=38) were shifted to stent placement (PCI).

Investigators in the SYNTAX II trial looked at how availability of functional information provided by FFRct changed treatment decisions for patients with 3-vessel CAD. The investigators used the Syntax Score (SS), an objective assessment of the severity of a patient’s SIHD, to understand how FFRct information can lead to a change in treatment strategy. The investigators concluded that the non-invasive functional information provided by FFRct reclassified 30% of patients from high and intermediate-risk SS to low-risk SS, enabling less invasive treatment. Previous trials have shown that for patients with 3-vessel disease and a SS > 22, CABG is the optimal treatment strategy while for patients with 3-vessel disease and a SS < 22, both CABG and PCI provide similar outcomes. The addition of FFRct information to enable the functional evaluation of a patient’s 3-vessel disease can lead to a change in treatment strategy for up to 30% of patients.

Because of the clinical efficacy and utility of FFRct for patients with severe 3-vessel disease we recommend that the restriction on these patients be removed from the final LCD.

Conclusion

We appreciate your attention to this important matter and your consideration of our comments.

Thank you for your comments we will address each concern individually below.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing, therefore, does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Heart rate >65 and arrhythmia: We agree that the Macron study, as well as improvements in imaging technology, are satisfactory to remove the requirement. Also, this restriction is more related to CCTA acquisition, and we added a requirement for a technically adequate CCTA study as a requirement for FFR.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Prosthetic valves: Strict adherence to careful patient selection contributed significantly to demonstrating the net health benefit for FFRCT. No supporting research was submitted with these comments. Pivotal trials for FFRCT consistently excluded patient populations with prosthetic valves and extensive coronary calcification from study. Furthermore, these limitations are reported by the manufacturer in FDA filings. Due to the unavailability of U.S.-based, high-quality, peer-reviewed published clinical research in patient populations with extensive calcification or prosthetic valves, we must conclude that it is still investigational and, therefore, unreasonable to consider for coverage at this time. Upon availability of this evidence, as well as removal of this limitation from the FDA-validated filings, we can reconsider this coverage limitation upon request. The limitation will remain as written.

Thank you for the opportunity to comment on the proposed LCD for the Non-Invasive Fractional Flow Reserve technology, hereafter referred to as FFRct, which refers to FFR values which are acquired from a previously conducted coronary CT angiogram (CCTA). There are several proposed restrictions in the LCD with which we take great concern, and which would inadvertently restrict access for the Medicare beneficiaries who would benefit from FFRct the most.

The first issue in the proposed LCD with which we take issue is indication #5, which states that FFRct "may be reasonable and necessary" for patients with "single vessel disease (not proximal LAD or proximal left dominant circumflex artery) with >50% stenosis AND insufficient response to optimal medical therapy (OMT)." Our opinion is that the requirement for a patient to have failed OMT prior to undergoing the FFRct service is not based on clinical evidence and should be removed from the final LCD. It appears that Palmetto may have taken this language from various society guidelines on appropriate indications for revascularization, not diagnosis of SIHD.

The 2012 ACC guidelines for the diagnosis and management of patients with SIHD lists two indications (with a class I or IIA recommendation) for patients with single-vessel disease not involving a 50% Left Main or 70% Non-Left Main stenosis. For these lesions, the patient must have either failed OMT or have contraindications to OMT for revascularization to be guideline-supported. The recent ACC appropriate use criteria for revascularization has similar indications. In order for revascularization of single-vessel disease not involving the proximal Left Main or Circumflex artery to be "always appropriate," the patient must be on OMT. Of note, there are several "may be appropriate" indications for revascularization for patients with FFR-positive (FFR ≤ 0.80) disease who are not on OMT.

However, the point is that these are guidelines and appropriate use criteria for revascularization, and do not pertain to a diagnostic service such as FFRct. When a patient presents with symptoms suggestive of SIHD and their CCTA shows single-vessel disease > 50% stenosis (not in the proximal left main or circumflex artery), the physician then asks whether this stenosis is causing the patient's symptoms, not whether revascularization is appropriate. In this clinical situation FFRct can provide the answer the clinician seeks and help guide appropriate therapy (whether OMT or further investigation). Because this restriction does not apply to a diagnostic service, we ask that the OMT requirement be struck from the final LCD.

The next issue in the proposed LCD has to do with Palmetto's assertion that "FFRct testing is not reasonable and therefore not covered" for patients with "known severe 3-vessel disease (> 50% stenosis in all 3 major vessels." There is ample clinical evidence that FFRct has high clinical value in these patients and we ask that this restriction be removed from the final LCD.

Over the past decade, the syntax score (SS) has developed as a tool to help clinicians decide between PCI and bypass grafts (CABG) as the most appropriate course of treatment for severe 3- vessel disease. The SS is an anatomic and functional assessment of SIHD that has been shown to triage patients to the most appropriate care. Patients with a low risk SS (S≤22) are known to have similar outcomes with either PCI or CABG. Patients with a high risk SS (≥23) were shown to have better outcomes with CABG instead of PCI. A recent evaluation of the non-invasive functional SS in the SYNTAX lll trial demonstrated that use of FFRct reclassified 30% of patients from high to low risk, enabling physicians to potentially choose PCI as the treatment strategy. This less invasive treatment option is preferred by patients, comes with less risk, and is less costly. For this reason, we ask that the restriction on severe 3-vessel disease be removed from the final LCD.

Finally, we would like to comment on several of the restrictions for which Palmetto considers "FFRct testing is not reasonable and therefore not covered." It is our opinion that the following restrictions have nothing to do with the performance and interpretation of the FFRct service, and may be better placed on the LCD for the CCTA service:

• Severe obesity (BMI >35 kg/m²)
• Prosthetic valves
• Heart rate or arrhythmia > 65 beats/min at the time of image acquisition
• Allergy or intolerance to iodinated contrast material
• Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000
• Renal insufficiency (i.e., estimated glomerular filtration rate< 60mUmin/1.73m²)
• Uncooperative patient
• Patient unable to sustain breath-hold for 5-10 seconds
• Severe aortic stenosis

These restrictions are related to a physician's decision whether to order a CCTA for the initial evaluation of suspected SIHD and many are directly addressed in the (now outdated) SCCT guidelines for performance of CCTA. It is our opinion that these restrictions should be considered (with comments from the community) for inclusion in Palmetto's LCD on the CCTA service. As mentioned, several of these restrictions are out of date and do not pertain to modern CT scanners with updated technology.

In relation to the restriction on patients with severe obesity (BMI >35 kg/m²), the SCCT guidelines offer advice on how increasing tube voltage may be necessary to achieve acceptable image quality but does not discuss a hard-upper limit on BMI. Similarly, regarding heart rate restrictions SCCT is of the opinion that new CT machines with dual-source CT or wide-detector scanners allow the imaging of patients with higher and irregular heart rates with good success. Because of the technology updates in CT equipment over the past decade and of the irrelevance of these restrictions to the performance of FFRct, we ask that these restrictions be removed from the final LCD.

Thank you again for your consideration of our comments and concerns.

Thank you for your comments, we will address each individual issue below.

Single vessel disease with OMT: We agree with the removal of the single vessel disease and the reference to optimal medical therapy and have removed these requirements.

Severe 3 vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Prosthetic Valve and severe aortic stenosis: Strict adherence to careful patient selection contributed significantly to demonstrating the net health benefit for FFRCT. No supporting research was submitted with these comments. Pivotal trials for FFRCT consistently excluded patient populations with prosthetic valves and severe aortic stenosis from the studies. Furthermore, these limitations are reported by the manufacturer in FDA filings. Due to the unavailability of U.S.-based, high-quality, peer-reviewed published clinical research in patient populations with prosthetic valves and severe aortic stenosis, we must conclude that its use with these conditions is still investigational and, therefore, unreasonable to consider for coverage at this time. Upon availability of this evidence, as well as removal of these limitation from the FDA-validated filings, we can reconsider these coverage limitations upon request. The limitations will remain as written.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

The Society of Cardiovascular Computed Tomography (SCCT) is appreciative of the opportunity to comment on Palmetto’s proposed LCD on Non-Invasive FFR (referred to as FFRct from this point forward) for Stable Ischemic Heart Disease. Representing 2,600 members, SCCT is a community of physicians, scientists and technologists advocating for research, education and clinical excellence in the use of cardiovascular computed tomography (CCT). Our society mission is to improve health outcomes through effective use of cardiovascular CT. We are committed to delivering high-value clinical care for the millions of people with suspected coronary artery disease (CAD). Many SCCT members have experience with Fractional Flow Reserve Computed Tomography (FFRct) in both the Hospital Outpatient Prospective Payment System (OPPS) and Physician Fee Schedule (PFS) settings.

SCCT’s comments are focused around three main areas within the LCD:

1. Restrictions more appropriate for coronary CT angiography (CCTA) than Non-Invasive Fractional Flow Reserve (FFR)
2. Restriction on use of FFRct in single-vessel disease
3. Restriction on use in severe 3-vessel disease

Comments on “Restrictions more appropriate for coronary CT angiography (CCTA) than Non-Invasive FFR.”

The SCCT perceives many of the coverage restrictions within this LCD as more applicable to CCTA. We suggest the following be removed from DL38278, as they pertain to performance of the CCTA and not the performance of the Non-Invasive FFR:

• Uncooperative patient
• Patient unable to sustain breath-hold for 5-10 seconds
• Allergy or intolerance to iodinated contrast material
• Renal insufficiency (i.e., estimated glomerular filtration rate < 60mL/min/1.73m²)

The above proposed restrictions are not related to the acquisition or performance of the FFRct service. These factors are related to a physician’s decision whether to order CCTA for the initial evaluation of suspected cardiac symptoms and many are directly addressed in the SCCT guidelines for performance of CCTA. The acquisition decisions by an individual doctor will vary widely, for example, based upon the severity of the allergy and the GFR related to the need for the CCTA, as well as based on the available technology at a particular site. Regardless, these are irrelevant to the FFRct, as only cases acquired with CCTA can be subsequently referred for FFRct, based upon the resultant image quality.

Several other restrictions within this LCD are related to CCTA image quality, which should be removed. These include:

• Heart rate or arrhythmia > 65 beats/min at the time of image acquisition
• Severe obesity (BMI >35 kg/m²)
• Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000
• Bare metal intracoronary stent
• Prosthetic valves

Regarding heart rate control, improvements in CT scanner hardware and software components have enabled high quality images, even in patients with high heart rates. SCCT guidelines for the performance of CCTA state that heart rates > 80 bpm are used to represent a relative contraindication for CCTA due to high incidence of motion artifacts. Wide detector and dual-source CT scanners now allow for successful imaging of patients with irregular and higher heart rates. The Macron study found; physicians were able to successfully analyze 97% of vessels utilizing CCTA in patients with a mean heart rate of 91.4 ± 48.9 beats/min². Additionally, The NXT trial demonstrated high FFRct accuracy despite high heart rates (63 ± 10 bpm). The SCCT perceives this restriction as dated and not applicable to current scanners. It is recommended the restriction be removed.

In the same vein of technology advances improving CT scanner capabilities, SCCT guidelines states how increasing tube voltage may be required to achieve acceptable image quality but does not report a hard-upper limit on BMI. Several publications address the performance of CCTA in obese patients. The ADVANCE registry indicated that high BMI was not a statistically significant predictor of CCTA quality rejection and that many patients with a BMI > 35 were successfully able to undergo the FFRct service. The PLATFORM study demonstrated utilization of FFRct led to a significant reduction in angiograms performed (61%) which may allow for improved outcomes for the severely obese. Lastly, Mangold, et al. found CCTA provided high diagnostic accuracy in both non-obese and obese patients. Based on the data provided as well as technology updates in CT equipment, the SCCT recommends the restriction on patients with BMI >35 kg/m² be removed.

Next, we would like to address the restriction of “Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000.” While calcification of the coronary arteries can cause artifacts, there is no set cut-off for which a CTA and FFRct exam cannot be performed. SCCT guidelines recommend that “the decision to proceed with CCTA in the presence of high coronary calcium score should be left to the discretion of the referring and imaging physician1.” In a post-hoc analysis of the NXT trial, which examined the accuracy of FFRct compared to invasively acquired FFR values, the investigators looked at FFRct accuracy across four qualities of calcification (Q1: 0-26, Q2: 27-147, Q3: 148-415, Q4: 416-3599). The authors found “there was no statistically significant difference in per- patient or per-vessel diagnostic accuracy, sensitivity and specificity of FFRct across Agatston score (AS) quartiles” and concluded that “FFRct per-patient and per-vessel diagnostic accuracy (74% and 83%, respectively), sensitivity (88% and 82%), and specificity (68% and 84%) in the highest AS quartile were equal to or superior to diagnostic performance characteristics reported for most conventional noninvasive ischemia testing modalities when compared with directly measured FFR.” Provided the high performance of FFRct regardless of Agatston Score, SCCT recommends this restriction on calcium score be removed from the final LCD.

The presence of coronary stents and prosthetic valves does not preclude performance of FFRct. They only preclude the performance of quality CCTA to the extent that these factors prevent adequate imaging of the lumen of the lumen of the coronary arteries and prevent the reading physician from interpreting the presence of SIHD. There is never a restriction based upon prosthetic heart valves, as these typically do not cause a scatter artifact and sit sufficiently far from the coronaries to avoid issue. In regards to FFRct, HeartFlow can process FFRct results even for patients who have a stent in their coronary anatomy. While FFRct results are not provided for the coronary system which contains a stent, FFRct results are provided for all coronary systems which do not contain a stent, allowing functional information to be utilized in clinical decision making. There are specific circumstances for which HeartFlow cannot provide FFRct results, including when a stent is present in the left main coronary artery, stenosis >30% present in the left main coronary artery and 1 or more stent in the left system and if there are stents present in 2 or more coronary systems. It is the opinion of the SCCT, that the restriction on coronary stents and prosthetic valves be removed entirely.

The SCCT will now comment on restriction on use of FFRct in single-vessel and severe 3-vessel disease.

DL38278 cites Non-Invasive Fractional Flow Reserve may be reasonable and necessary “for patients with single vessel disease (not proximal LAD or proximal left dominant circumflex artery) with >50% stenosis AND insufficient response to optimal medical therapy (OMT).” The 2012 ACC guidelines for the diagnosis and management of patients with SIHD lists two indications (with a class I or IIA recommendations) for patients with single-vessel disease not involving a ≥ 50% Left Main or ≥ 70% Non-Left Main stenosis. For these lesions, the patient must have either failed OMT or have contraindications to OMT for revascularization to be guideline-supported.

Additionally, the recent ACC appropriate use criteria for revascularization in patients with SIHD outlines clinical scenarios for when it is appropriate to revascularize patients. For revascularization of single vessel disease not involving the proximal Left Main or Circumflex artery to be “always appropriate,” the patient must be on OMT. Of note, there are several scenarios for which it “may be appropriate” to revascularize patients with FFR-positive (FFR< 0.80) disease who are not on OMT.

The SCCT would like to note, the above guidelines and AUC are for revascularization decision-making and do not pertain to FFRct. Provided this, access to a diagnostic test which can support physicians in determining if a specific lesion is causing ischemia should not be restricted. The information able to be obtained from FFRct can allow physicians to determine if additional information is needed or that OMT is appropriate.

We recommend the requirement for a patient to have failed OMT prior to undergoing FFRct be removed from the final LCD.

DL38278 proposes “FFRct testing is not reasonable and therefore not covered” for patients with “known severe 3-vessel disease (> 50% stenosis in all 3 major vessels).” There is significant clinical evidence supporting FFRct’s value in this patient subset. In the ADVANCE trial investigators found 22.9% of the subjects (n=943) were referred for revascularization after CCTA and 22.3% of these patients were reclassified to medical therapy alone after FFRct analysis (PCI to medical therapy, 20.9% (n=198) and CABG to medical therapy, 1.4% (n=12)). The syntax score (SS), is an objective assessment of the severity of a patient’s SIHD and has developed into a tool supporting clinicians decide between PCI and bypass grafts (CABG) as the most appropriate course of treatment for severe 3-vessel disease.

The Syntax Score (SS) supports the cardiac team in determining the best course for revascularization, either PCI or CABG. Patients with a low risk SS (<22) were shown to have similar outcomes with CABG or PCI while patients with an SS >23 had better outcomes with CABG after 5-year follow up. The SYNTAX II trial examined the use of CT-FFR in patients with 3 vessel CAD, and investigators found that availability of FFRct allowed for the stratification of patients with severe 3-vessel anatomic disease to a lower risk group for which PCI was determined to be safe. This less invasive treatment option is preferred by patients, comes with less risk and has a lower associated cost compared to invasive treatments. SCCT recommends, for these reasons in conjunction with the clinical evidence above, the restriction on severe 3-vessel disease be removed.

SCCT appreciates Palmetto’s consideration of SCCT’s comments. There is international momentum in adopting CCTA and CT-FFR for diagnosing CAD; this is evidenced by recent United Kingdom, Japanese, and European guidelines that classify CCTA evidence as Level 1 for patients with suspected CAD. A key component of this pathway is CT-FFR for assessing how to manage patients with intermediate disease. With the growing momentum in adopting this pathway, SCCT requests Palmetto consider the comments made within this letter and those from other entities to develop a final LCD which ensure adequate access to the CT-FFR pathway for Medicare beneficiaries.

References were provided for review.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing therefore does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Single vessel disease with OMT: We agree with the removal of the single vessel disease and the reference to optimal medical therapy and have removed these requirements.

The local coverage determination support for the use of HeartFlow to assess the coronary fractional flow reserve on coronary CTA datasets looks quite good. There are a few things where I would suggest minor changes— for the most part related to exclusions. For instance, a lot of the exclusions are based on prior CTA guidelines (BMI, arrhythmia, fast heart rates) that are just not as big a problem with faster, more powerful scanners. Most importantly, if the scan quality is insufficient for analysis, it will be rejected by HeartFlow — at no charge to Medicare.

The contrast allergy issue is also one that can be addressed (by steroid pretreatment) and would also be a problem for cardiac cath, the primary alternative. I would make that a "risk-benefit" question for the ordering physician rather than an absolute exclusion.

Regarding calcifications— I can tell you from personal experience that— while we have limitations with CCTA when calcifications are extensive, we also know that these are most profound in the higher-risk patient. Moreover, I have seen many cases where there was a significant stenosis in a non-calcified segment that Cor CTA detected. So, again, I wouldn't make this an absolute exclusion.

Regarding impaired renal function— it is becoming increasingly clear that contrast induced nephropathy from current IV contrast agents is actually significantly overstated. The problem is that there are few controlled studies of what happens to high-risk patients who do not receive contrast, with comparison to those of similar degree of illness who do receive contrast. The ones that have been done (analyzing 56,000 patients) have not shown any significant increased risk from IV contrast (McDonald, Radiology 2016). At MCV, we don't withhold contrast unless the GFR is <30, and even then we will do a risk benefit analysis and may go lower. Keep in mind that the alternative to IV contrast and CCTA is likely Intra-arterial contrast and invasive coronary catheterization— With Contrast!

Thanks for your consideration of these points.

Thank you for your comments, we will address each individually below.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

We would like to share our experiences with Fractional Flow Reserve Computed Tomography (FFRCT) and how the proposed LCD will impact the care we provide our Medicare patients.

FFRCT is a noninvasive diagnostic tool for use in diagnosis and management of coronary artery disease (CAD) and is a reliable, noninvasive alternative to invasive coronary angiogram for clinically stable, symptomatic patients and its use is supported by several large volume prospective and retrospective studies. In our respective practices, the use of FFRCT has allowed for improved efficiency in patient diagnosis and next step patient management. It is through the use of FFRCT performed in conjunction with coronary CTA that we are able to differentiate patients who can benefit from invasive coronary angiography and interventional management from those who need medical management. We have personally cared for patients in whom coronary CTA with FFRCT is being used to manage patient disease in lieu of invasive coronary angiography and also have patients in whom CTA has been able to identify disease that was undiagnosed or mis-diagnosed with other diagnostic exams. We have also been able to confirm for patients and providers when CAD is not the cause of a patient’s symptoms. The incorporation of FFRCT into our work practice has allowed us to differentiate hemodynamically significant and non-hemodynamically significant CAD and refer for the most appropriate treatment without the need of additional testing.

Below we reply to on specific indications:

We agree with the limitation that patient post coronary artery bypass surgery should not undergo FFRCT as it is not yet covered under the current FDA labeling for FFRCT. The ask for the removal of use FFRCT in patients with single vessel disease (not proximal LAD or proximal left dominant circumflex artery) with >50% stenosis AND insufficient response to optimal medical therapy (OMT) as the limitation related to OMT would limit the ability to identify the CAD is hemodynamically significant in patients who may have newly identified with CAD and not yet on medical therapy. Restricting utilization of FFRCT for these patients may lead physicians to order an additional test to determine if the lesion is causing ischemia, with added time and cost to the patient, health system and added risk to the patient. We recommend that the requirement for OMT be removed from this indication in order to enable efficient care.

There are also several indications on the list of proposed coverage restrictions that are not related to the performance of FFRCT and that should be moved to the existing Palmetto LCD on CCTA.

These include:

• Allergy or intolerance to iodinated contrast material
• Renal insufficiency (i.e., estimated glomerular filtration rate < 60mL/min/1.73m²)
• Uncooperative patient
• Patient unable to sustain breath-hold for 5-10 seconds

We would ask for reconsideration of the above as depending on CT hardware and software and patient specific indications as all of the above situations may be able to be addressed for patients allowing them to undergo a coronary CTA with FFRCT when the benefit may outweigh the risks.

Additionally, there are several indications on the list of proposed coverage restrictions which are related to CCTA image quality but do not preclude performance of FFRCT. We believe these indications should be removed from the final LCD.

These include:

• Severe obesity (BMI >35 kg/m²)
• Prosthetic valves
• Heart rate or arrhythmia > 65 beats/min at the time of image acquisition
• Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000

While some of the factors listed above (severe obesity, heart rate > 65, etc) could result in suboptimal image quality, this is not directly related to ability to calculate FFRCT. If all of the coronary arteries have clearly defined lumen boundaries, accurate FFRCT can be achieved. If the CCTA exam does not meet the stated requirements for FFRCT processing, FFRCT is not generated and the health care system is not charged. HeartFlow’s (Heartflow Inc Redwood City, CA) quality standards have led us to experience ~6% rejection rate and we recommend that the above restrictions be removed from the final LCD.

Work by Macron, et al. demonstrated that high quality CCTA results were obtained in patients with a mean BMI of 36.0 ± 5.9 kg/m² and 97% of vessels in these patients were able to be analyzed . Similarly, investigators in the ADVANCE registry report that BMI was not a statistically significant predictor of CCTA quality rejection. Macron, et al. demonstrated that high quality CCTA exams can be performed in patients with a high or irregular heart rate with a 97% rate of success. This was achieved despite a mean heart rate of 91.4 ± 48.9 bpm². Advances in CT hardware and software have and continue to provide improvements in image quality for CTA.

Similarly, the Agatston score is a parameter of the entire coronary artery calcified plaque burden but it is not a reliable predictor of CCTA diagnostic ability. Description of focal or diffuse areas of high-volume calcified plaque which may be an issue for coronary CTA performance are not identified in the Agatston score and patients with a high Agatston score and large sized coronary vessels may have adequate luminal assessment despite the calcified plaque. Accurate FFRCT can be achieved independent of the degree of calcium as noted by investigators in the NXT trial that reported that “there was no statistically significant difference in per-patient or per-vessel diagnostic accuracy, sensitivity, and specificity of FFRct across Agatston Score quartiles.”

For patients with a prosthetic valve, the valve will not interfere with quality imaging of the coronary arteries as we ask that this indication be removed. Similarly, the presence of a bare metal (or drug-eluding) coronary stent does not preclude performance of FFRCT. Therefore, we believe this restriction should be removed.

Thank you again for the opportunity to provide feedback on the proposed LCD and for taking our recommendations into consideration.

References were provided for review.

Thank you for your comments we will address each one below.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Single vessel disease with OMT: We agree with the removal of the single vessel disease and the reference to optimal medical therapy and have removed these requirements.

Prosthetic valves: Strict adherence to careful patient selection contributed significantly to demonstrating the net health benefit for FFRCT. No supporting research was submitted with these comments. Pivotal trials for FFRCT consistently excluded patient populations with prosthetic valves and extensive coronary calcification from study. Furthermore, these limitations are reported by the manufacturer in FDA filings. Due to the unavailability of U.S.-based, high-quality, peer-reviewed published clinical research in patient populations with extensive calcification or prosthetic valves, we must conclude that it is still investigational and, therefore, unreasonable to consider for coverage at this time. Upon availability of this evidence, as well as removal of this limitation from the FDA-validated filings, we can reconsider this coverage limitation upon request. The limitation will remain as written.

I write on behalf of the state chapter of the ACC and on behalf of myself. We greatly appreciate the chance to comment on this proposed LCD for Non-Invasive Fractional Flow Reserve.

There are several areas in the LCD on which I would like to focus:

• Restrictions on use in single vessel disease
• Indications which would be most appropriately placed in a different LCD
• Indications related to coronary CT angiography (CCTA) image quality
• Restrictions on use of FFRct in severe 3-vessel disease
• Documentation requirements which should be modified

Restriction on use in single vessel disease which should be removed

The proposed LCD seeks to limit utilization of FFRct in patients with single vessel disease to those with > 50% stenosis and “insufficient response to optimal medical therapy (OMT).” I would propose that this restriction to patients that have failed OMT comes from appropriate use criteria (AUC) related to revascularization and is not relevant to a diagnostic modality. This requirement should be removed.

The recent AUC on coronary revascularization in patients with SIHD outlines clinical scenarios for which it is appropriate to revascularize patients. For single vessel disease that does not involve the proximal LAD or Circumflex artery, use of OMT is required in order for revascularization to be considered “Always Appropriate.” It should be noted that there are several scenarios for which it “May be Appropriate” to revascularize similar disease for patients not on OMT. However, as stated above, this AUC pertains to revascularization decision-making and should not be used to restrict access to a diagnostic test that seeks to answer the question of whether a specific lesion is causing ischemia. Based on the answer to this question a physician may decide that they need more information to determine appropriate treatment or that OMT is the correct course of action. Therapeutic AUC should not be applied to restrict diagnostic decision-making and therefore this portion of the LCD should be modified to remove the OMT requirement. Physicians need to perform FFRct in many situations before deciding appropriate management decisions.

Indications which would be most appropriately placed in a different LCD

There are several coverage restrictions in the proposed LCD for FFRct that appear to be related to perceived coverage restrictions for CCTA. These include:

1. Allergy or intolerance to iodinated contrast material
2. Renal insufficiency (i.e., estimated glomerular filtration rate < 60 mL/min/1.73m²)
3. Uncooperative patient
4. Patient unable to sustain breath-hold for 5-10 seconds

It is my strong opinion that these proposed restrictions are not related to the acquisition or performance of the FFRct service. Once the CCTA images are acquired, the decision to order a functional analysis of coronary blood flow via FFRct will be made independent of these four factors. These factors may play an important role in a physician’s decision whether to order CCTA to evaluate suspected SIHD, and they may fit better in a Palmetto LCD on the CCTA service. As an experienced provider of CCTA services, I will assure that often these exact limitations do indeed cause a physician not to proceed with CCTA. However, once the images are obtained, they have absolutely nothing to do with the ability to obtain FFRct images.

Indications related to coronary CT angiography (CCTA) image quality which should be removed from the LCD

There are several coverage restrictions in the proposed LCD for FFRct that are related to image quality of the CCTA and its perceived effect on the accuracy and clinical utility of FFRct. These include:

1. Severe obesity (BMI >35 kg/m²)
2. Bare metal intracoronary stent
3. Prosthetic valves
4. Heart rate or arrhythmia > 65 beats/min at the time of image acquisition
5. Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000

I would argue that these criteria do not apply to current generation scanners aimed at image acquisition of the heart (> 64 slice scanners). Contemporary scanners contain improved hardware and software components, such as dual-source CT and wide-detector scanners, which allow quality imaging of the most difficult patient anatomies and situations. These proposed restrictions should be removed or broadly expanded in order to ensure adequate access for Medicare beneficiaries.

There are multiple publications that speak to the performance of CCTA in both the obese and high heart rate populations. A recent study from Mangold et al. examined that accuracy of CCTA in obese (≥ 30 kg/m²) and non-obese (< 30 kg/m²) patients. The investigators determined that there was no statistically significant difference in overall per-patient sensitivity, specificity, positive predictive value, negative predictive value, and accuracy and concluded that CCTA “provides high diagnostic accuracy in non-obese and obese patients.” Another publication from Zimmerman, et al. evaluated how accuracy was affected by both obesity and heart rate control. Four heart rate groups were included in the trial (< 70, 70-79, 80-89, ≥ 90 bpm) and physicians graded image quality based on their ability to visualize the artery. The authors found that 99% of coronary arteries were of diagnostic quality and concluded that “diagnostic quality [CCTA] examinations can be obtained without premedication regardless of body size.”

The presence of coronary stents, prosthetic valves, and extensive coronary calcification only preclude the performance of quality CCTA to the extent that these factors prevent adequate imaging of the lumen of the coronary arteries and prevent the reading physician from interpreting the presence of SIHD. It is the previously reported opinion of the ACC and the Society of Cardiovascular Computed Tomography (SCCT) that the decision to proceed with CCTA “in the presence of high coronary calcium score should be left to the direction of the referring and attending physician.” The presence of other traditional obstacles, including the presence of stents or prosthetic valves, should be handled in a similar manner. If the performing and interpreting physician are confident in their ability to perform and read CCTA in the presence of these or other factors, then the decision to perform CCTA should be left in their hands.

Restrictions on the types of physician who can perform CCTA and FFRct are handled in the “Credentialing and Accreditation Standards” section of the LCD and ensure that patients with these challenging conditions are treated by the appropriate provider.

Finally, there is evidence from the recent NXT trial that examined the accuracy of FFRct in patients with a wide range of coronary calcium scores. The authors of this study concluded that “there was no statistically significant difference in per-patient or per-vessel diagnostic accuracy… of FFRct across Agatston Score quartiles” and that FFRct diagnostic performance was “equal to or superior to diagnostic performance characteristics reported for most conventional noninvasive ischemia testing modalities.”

Restrictions on use of FFRct in severe 3-vessel disease which should be removed from the LCD

The proposed LCD also proposes to limit FFRct utilization in patients with known “severe 3-vessel disease (> 50% in all 3 major vessels)” ostensibly because it is a foregone conclusion that coronary artery bypass grafting (CABG) is necessitated by patients whose CCTA finds such disease. However, there is mounting clinical evidence, investigated in the three SYNTAX family of randomized controlled trials, that PCI, or stenting, is a viable option for patients with 3-vessel disease that is determined to be not functionally significant via FFR information.

In these trials the Syntax Score (SS), an objective assessment of the severity of a patient’s SIHD, aides the heart team in determining the best course for revascularization, either PCI or CABG. Patients with a low risk SS (≤ 22) were shown to have similar outcomes with CABG or PCI while patients with an SS ≥ 23 had better outcomes with CABG after 5-year follow up. Use of FFRct to determine a patient’s functional SS in the SYNTAX III trial reclassified 30% of patients to a low risk SS, potentially enabling a less invasive (PCI) form of treatment.

Documentation requirements which should be modified

There are several “Documentation Requirements” in the proposed LCD which would place an undue burden on providers and inadvertently restrict access to the FFRct service for Medicare beneficiaries. I suggest that the following requirements be removed from the final LCD in order to ensure parity with other services used to diagnose and treat suspected SIHD:

1. Heart rate at time of image acquisition
2. Body mass index

While these are appropriate considerations for CT Angiography acquisition, it will not necessarily be easy to document these with the FFRct report. Again, once the quality check has allowed FFRct to be produced, they are are not clearly relevant for FFRct.

Conclusion

I would propose that all of these restrictions will markedly limit the ordering of FFRct. From a practical standpoint, as soon as these tests are ordered, providers or their facilities are responsible for payment to an external vendor for this analysis. With a lengthy list of requirements (all four points noted above), there will be severe reluctance to ever order this test as providers will be responsible for payments even if ultimately denied by Medicare. Given that there is great hope that FFRct will provide cost savings through fewer invasive studies, these restrictions that would potentially stop the use of the technology do not seem consistent with goals expressed.

Palmetto’s consideration of the comments in this letter is appreciated.

Documentation requirements: We recognize that the heart rate at the time of image acquisition is best applied to CCTA, rather than FFRCT and is removed. Upon receipt of evidence to support the use of FFRCT in Medicare beneficiaries with BMI greater than 39 kg/m², we can reconsider this coverage limitation upon request. The requirement for documentation of BMI will remain unchanged.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Single vessel disease with OMT: We agree with the removal of the single vessel disease and the reference to optimal medical therapy and have removed these requirements.

Prosthetic valves: Strict adherence to careful patient selection contributed significantly to demonstrating the net health benefit for FFRCT. No supporting research was submitted with these comments. Pivotal trials for FFRCT consistently excluded patient populations with prosthetic valves and extensive coronary calcification from study. Furthermore, these limitations are reported by the manufacturer in FDA filings. Due to the unavailability of U.S.-based, high-quality, peer-reviewed published clinical research in patient populations with extensive calcification or prosthetic valves, we must conclude that it is still investigational and, therefore, unreasonable to consider for coverage at this time. Upon availability of this evidence, as well as removal of this limitation from the FDA-validated filings, we can reconsider this coverage limitation upon request. The limitation will remain as written.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing therefore does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

The Georgia Chapter American College of Cardiology would like to thank you for the opportunity to comment on Palmetto’s proposed Local Coverage Determination (LCD) on Non-Invasive Fractional Flow Reserve (FFRct) for Stable Ischemic Heart Disease (SIHD). As the professional home for the entire cardiovascular care team, our mission is to promote highest standards in the practice of cardiovascular medicine/surgery in the state of Georgia and maximize the quality of cardiovascular care received by the residents of the State.

In support of this mission. We are compelled to raise the following areas of concern in regards to the above referenced LCD as currently proposed:

• Incomplete indications
• Restrictions on single vessel disease
• Indications which would be most appropriately placed in a different LCD
• Indications related to coronary CT angiography (CCTA) image quality
• Restrictions on use of FFRct in severe 3-vessel disease
• Refinement to documentation requirements

Incomplete Indications

Several criteria for coverage of FFRct in stable symptomatic patients are listed in the draft LCD. Gaps and odd criteria exist within that list of five criteria. First, the policy should include the right coronary artery which is missing. Second, coverage criteria for the left anterior descending coronary artery and the left circumflex disease are limited to proximal portions of the arteries. This should be changed to include both proximal AND mid-portions of the arteries. Third, the indication for left circumflex disease appears to be limited to the dominant system. Dominant should be removed so that coverage is simply for left circumflex disease.

Single Vessel Disease

The proposed LCD seeks to limit utilization of FFRct in patients with single vessel disease to those with >50% stenosis and insufficient response to optimal medical therapy (OMT). This restriction to patients that have failed OMT is referenced from the Appropriate Use Criteria (AUC) related to revascularization and is not relevant to a diagnostic modality, therefore it should not be utilized to restrict access to a diagnostic test that determines whether a specific lesion is causing ischemia. Depending on the results, a physician may decide they need additional information to determine appropriate treatment or that OMT is the correct course of action. The recent AUC on coronary revascularization in patients with SIHD outlines clinical scenarios for which it is appropriate to revascularize patients. For single vessel disease that does not involve the proximal LAD or Circumflex artery, use of OMT is required in order for revascularization to be considered “Always Appropriate”. Therapeutic AUC should not be applied to restrict diagnostic decision making. The Georgia chapter requests that Palmetto revise this portion of the LCD to remove the OMT requirement.

Indications which would be most appropriately placed in a different LCD

There are several coverage restrictions in the proposed LCD for FFRct that appear to be related to perceived coverage restrictions for CCTA. These include:

• Allergy or intolerance to iodinated contrast material
• Renal insufficiency (i.e., estimated glomerular filtration rate < 60mL/min/1.73m²)
• Uncooperative patient
• Patient unable to sustain breath-hold for 5-10 seconds

These restrictions are not related to the acquisition or performance of the FFRct service. Once the CCTA images are acquired, the decision to order a functional analysis of coronary blood flow via FFRct will be made independent of these four factors. These factors may play an important role in a physician’s decision whether to order a CCTA to evaluate suspected SIHD and should be considered as part of the existing Palmetto LCD on the CCTA service, not in this LCD.

FFRct in relation to image quality of CCTA

There are several coverage restrictions in the proposed LCD for FFRct that are related to image quality of the CCTA and a perceived effect on the accuracy and clinical utility of FFRct. These include:

1. Severe obesity (BMI >35 kg/m²)
2. Bare metal intracoronary stent
3. Prosthetic valves
4. Heart rate or arrhythmia > 65 beats/min at the time of image acquisition
5. Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000

The mentioned criteria do not apply to current generation scanners aimed at image acquisition of the heart (>64 slice scanners). Contemporary scanners contain improved hardware and software components, such as dual-source CT and wide-detector scanners, which allow quality imaging of the most difficult patient anatomies and situations. Our chapter asks for these restrictions to be removed or broadly expanded in order to ensure adequate access for Medicare beneficiaries.

There are multiple publications that speak to the performance of CCTA in both the obese and high heart rate populations. A recent study from Mangold, et. Al examined that accuracy of CCTA in obese (≥ 30 kg/m²) and non-obese (< 30 kg/m²) patients. The investigators determined that there was no statistically significant difference in overall per-patient sensitivity, specificity, positive predictive value, negative predictive value, and accuracy and concluded that CCTA “provides high diagnostic accuracy in non-obese and obese patients.” Another publication from Zimmerman, et al. evaluated how accuracy was affected by both obesity and heart rate control. Four heart rate groups were included in the trial (< 70, 70-79, 80-89, ≥ 90 bpm) and physicians graded image quality based on their ability to visualize the artery. The authors found that 99% of coronary arteries were of diagnostic quality and concluded that “diagnostic quality [CCTA] examinations can be obtained without premedication regardless of body size.”

The presence of coronary stents, prosthetic valves, and extensive coronary calcification only preclude the performance of quality CCTA to the extent that these factors prevent adequate imaging of the lumen of the coronary arteries and prevent the reading physician from interpreting the presence of SIHD. It is the current position of the state and national ACC that the decision to proceed with CCTA in the presence of high coronary calcium score should be left to the direction of the referring and attending physician. The presence of other traditional obstacles, including the presence of stents or prosthetic valves should be handled in a similar manner. If the performing and interpreting physician are confident in their ability to perform and read the CCTA in the presence of these or other factors, then the decision to perform the CCTA should be left up to the attending physician.

Evidence from the recent NXT trial that examined the accuracy of FFRct in patients with a wide range of coronary calcium scores and concluded that there was no statistically significant difference in per-patient or per-vessel diagnostic accuracy…of FFRct across Agatston Score quartiles. It was also found that FFRct diagnostic performance was equal to or superior to diagnostic performance characteristics reported for most conventional noninvasive ischemia testing modalities.

FFRct in Severe 3-Vessel Disease

The proposed LCD limits FFRct utilization in patients with known severe 3-vessel disease (>50% in all 3 major vessels) referencing a forgone conclusion that coronary artery bypass grafting (CABG) is necessitated by patients whose CCTA finds such disease. There is substantial clinical evidence investigated in the three SYNTAX family of randomized controlled trials, that PCI or stenting is a viable option for patients with 3-vessel disease that is determined to be not functionally significant via FFR information.

In these trials, the Syntax Score (SS) and objective assessment of the severity of a patient’s SIHD, aides the heart team in determining the best course for revascularization, either PCI or CABG. Patients with a low risk SS (≤ 22) were shown to have similar outcomes with CABG or PCI while patients with an SS ≥ 23 had better outcomes with CABG after 5-year follow up. Use of FFRct to determine a patient’s functional SS in the SYNTAX III trial reclassified 30% of patients to a low risk SS, potentially enabling a less invasive (PCI) form of treatment.

We urge Palmetto to remove restrictions on use of FFRct in severe 3-vessel disease.

Documentation Requirements

There are several “Documentation Requirements” in the proposed LCD which would place and undue burden on providers and inadvertently restrict access to the FFRct service for Medicare beneficiaries. We ask that the following requirements be removed from the final LCD in order to ensure parity with other services used to diagnose and treat suspected SIHD.

• Heart rate at time of image acquisition
• Body mass index

Conclusion

Several of the restrictions found within the proposed LCD will limit the ordering FFRct and place financial burden on providers on those who do. Access to FFRct can lead to fewer unnecessary diagnostic catheterizations making it a cost-effective, noninvasive diagnostic test.

Documentation requirements: We recognize that the heart rate at the time of image acquisition is best applied to CCTA, rather than FFRCT and is removed. Upon receipt of evidence to support the use of FFRCT in Medicare beneficiaries with BMI greater than 39 kg/m², we can reconsider this coverage limitation upon request. The requirement for documentation of BMI will remain unchanged.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Single vessel disease with OMT: We agree with the removal of the single vessel disease and the reference to optimal medical therapy and have removed these requirements.

Prosthetic valves: Strict adherence to careful patient selection contributed significantly to demonstrating the net health benefit for FFRCT. No supporting research was submitted with these comments. Pivotal trials for FFRCT consistently excluded patient populations with prosthetic valves and extensive coronary calcification from study. Furthermore, these limitations are reported by the manufacturer in FDA filings. Due to the unavailability of U.S.-based, high-quality, peer-reviewed published clinical research in patient populations with extensive calcification or prosthetic valves, we must conclude that it is still investigational and, therefore, unreasonable to consider for coverage at this time. Upon availability of this evidence, as well as removal of this limitation from the FDA-validated filings, we can reconsider this coverage limitation upon request. The limitation will remain as written.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing therefore does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Incomplete indications: We agree with your recommendations and have a bullet for right coronary artery. Secondly, we added proximal and mid portions of the left anterior descending and left circumflex arteries. Lastly, we removed the term “dominant” from the indications.

FFRct is a groundbreaking technology for the diagnosis and treatment of coronary artery disease (CAD). There is a plethora of published studies that have led to nationwide commercial adoption of this technology and coverage by most commercial and government payers. FFRct provides important diagnostic information non-invasively which prevents many patients from undergoing an unneeded invasive coronary angiography (ICA) which leads to a decrease in costs and increase in quality of life for our patients. I am pleased to be able to comment on Palmetto’s proposed Local Coverage Determination (LCD) on this exciting diagnostic and hope my comments are helpful as you finalize the coverage indications.

First, I would like to comment in general on the indications which are considered “not reasonable and therefore not covered.” After general comments I will follow-up with specific comments on several of these restrictions with clinical evidence supporting their removal. These restricted indications include:

• Severe obesity (BMI >35 kg/m²)
• Bare metal intracoronary stent
• Prosthetic valves
• Heart rate or arrhythmia > 65 beats/min at the time of image acquisition
• Allergy or intolerance to iodinated contrast material
• Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000
• Renal insufficiency (i.e., estimated glomerular filtration rate < 60mL/min/1.73m²)
• Uncooperative patient
• Patient unable to sustain breath-hold for 5-10 seconds
• Heart transplantation
• Recent prior myocardial infarction within 30 days
• Suspicion for acute coronary syndrome (where acute myocardial infarction or unstable angina have not been ruled out)
• Severe aortic stenosis
• Known severe 3-vessel disease (> 50% stenosis is all 3 major vessels)
• Post coronary artery bypass surgery, evaluation of venous grafts

It is my understanding from the manufacturer of the FFRct service that the accuracy of FFRct compared to invasively measured FFR is independent of CCTA characteristics. The manufacturer has an FDA-validated quality assurance process that analyzes the quality of submitted CCTA exams and rejects cases for which they cannot guarantee accuracy. If the manufacturer determines the CCTA exam has sufficient quality, then accuracy is in-line with their FDA label. The accuracy of the FFRct service has been proven in multiple clinical trials; the NXT trial examined the accuracy of the FFRct service in 254 patients and found that FFRct has per-vessel accuracy of 86% compared with 65% for CCTA . Similarly, the investigators in the PACIFIC study demonstrated that FFRct has higher diagnostic performance for detecting CAD than CCTA, SPECT (nuclear stress testing) and Positron Emission Tomography (PET) for vessel-specific ischemia . This evidence demonstrates that FFRct information is accurate and useful for clinical decision-making regardless of the underlying CCTA characteristics.

In regard to the restriction on severe obesity (BMI > 35 kg/m²), I would point to the results of the ADVANCE registry which indicated that high BMI was not a statistically significant predictor of CCTA quality rejection and that many patients with a BMI > 35 were successfully able to undergo the FFRct service . Also, it is well known from the PLATFORM study that utilization of FFRct can lead clinicians to cancel up to 61% of planned ICAs. The rate of vascular access-site complications for patients undergoing ICA is the highest for morbidly obese patients. If FFRct can help these patients avoid an unnecessary invasive procedure, then the service should be made available for them. Finally, in m practice at VCU we frequently image patients with a BMI up to 45 kg/m² and obtain diagnostic results. This represents a more reasonable upper limit for obese patients and allows physician discretion in deciding which test is best for that particular patient’s condition and history. I would also like to point out that the restriction on Heart rate or arrhythmia > 65 beats/min is unnecessarily restrictive. In their latest guideline document on the performance of CCTA, the Society of Cardiovascular Computed Tomography (SCCT) states that “heart rates > 80 bpm used to represent a relative contraindication for CCTA because of high incidence of motion artifacts.

However, ongoing hardware and software improvements, such as dual-source CT and wide-detector scanner, now allow the imaging of patients with higher and irregular heart rates with good imaging success. The limitations regarding higher heart rates depend on the equipment used, and scan settings have to be adjusted accordingly.” The restriction as written would be appropriate for imaging equipment from 15 years ago. With the recent technological advances highlighted in the SCCT guideline and advances in image acquisition practice, imaging patients with heart rates up to 90 beats/min is entirely reasonable and these patients should have access to the FFRct service. Finally, I would again point to the accuracy results from the NXT and PACIFIC study which demonstrated high accuracy of FFRct compared to invasive FFR for patients with the following, high heart rates:

• The NXT study: 63 ± 10 bpm
• PACIFIC study: 57.8 ± 7.7 bpm

I also disagree with the restriction on patients with a high calcium score (Agatston score (AS) > 1000). Once again, the SCCT guidelines recommend that “the decision to proceed with CCTA in the presence of high coronary calcium score should be left to the discretion of the referring and attending physician.”6 This should also apply to the physician’s decision whether to send the CCTA exam for FFRct analysis. Another study performed on data from the NXT trial demonstrated that FFRct has similar, high accuracy regardless of the patient’s AS. The authors specifically noted that “there was no statistically significant difference in per-patient or per-vessel diagnostic accuracy, sensitivity, and specificity of FFRct across Agatston Score quartiles” and concluded that FFRct provides high per-patient and per-vessel diagnostic performance and discrimination for ischemia over a wide range of coronary calcification severity . The figure below demonstrates the accuracy of FFRct across AS even up to 3599. This demonstrates that the manufacturer of FFRct can provide accurate results even in the presence of extremely high (>3500) AS.

Finally, I would like to comment on the proposed restriction for FFRct utilization in patients with known, severe 3-vessel disease. There is important data from the ADVANCE registry that demonstrated 22.3% of patients who were initially referred for revascularization based on CCTA results had their final treatment plan changed to medical therapy (MT) after availability of FFRct data. This includes 198 patients (20.9%) whose treatment plan changed from PCI to MT and 12 patients (1.4%) whose treatment plan changed from bypass surgery (CABG) to MT. Without the availability of FFRct these patients may have been subject to an unnecessarily invasive procedure that provides little clinical benefit. FFRct provides important functional information that guides physicians to the appropriate treatment plan for their patients. Even for patients with severe anatomic 3-vessel disease, the functional information FFRct provides can change the final treatment decision. This led to significant cost savings of over $3,000 per-patient in the PLATFORM trial. Based on the importance of FFRct to decision-making in these patients, I would recommend that the restriction on patients with sever 3-vessel disease be striked in its entirety from the final LCD.

References were provided for review.

Thank you for your comments we will address each of your suggestions below.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate and arrhythmias: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

We have been building our cardiac CTA volume, and plan to increase our cardiac CTA volume further once we are FFRct capable. Our experience with cardiac CTA has been very positive, and we have been able to provide definitive coronary assessment on the majority of patients who have gotten this test. This has allowed us to limit additional cardiac testing (nuclear stress testing and heart catheterizations) for these patients. Without FFRct capability, we are limiting cardiac CTA to be performed on patients with a low to low-intermediate (<50%) pretest probability of coronary artery disease (CAD). With FFRct capability we will be able to expand cardiac CTA to be the recommended test to patients with intermediate and even some patients with intermediate-high pretest probability of CAD. We believe that this will replace nuclear stress testing and diagnostic heart catheterizations for a large portion of our patients.

In reviewing the proposed LCD, we agree with several of the criteria listed. Specifically we agree with the first 4 indications for FFRct:

1. Left main disease with intermediate coronary stenosis (lumen diameter reduction 30-70%)

2. Proximal left anterior descending coronary artery (LAD) disease with intermediate coronary stenosis

3. Proximal left dominant circumflex disease with intermediate coronary stenosis

4. Two- or three-vessel disease with intermediate coronary stenosis in at least 2 vessels

Regarding the 5th indication:

5. Single vessel disease (not proximal LAD or proximal left dominant circumflex artery) with >50% stenosis AND insufficient response to optimal medical therapy (OMT)

We feel that the restriction to document insufficient response to optimal medical therapy should be removed. The provider has ordered a coronary CTA to evaluate for chest pain. We should therefore provide them with as much information as possible to help them manage the patient. If the lesion is non-obstructive by FFRct (FFR >0.80), this can provide reassurance to the provider and patient, and allow them to focus on other etiologies of chest pain. However, if the lesion is obstructive by FFRct, this will answer the question for the provider and patient of what is causing the patient’s chest pain and allow the provider to adequately manage the patient. We do agree that the patient should not go for heart catheterization and percutaneous intervention before they have been trialed on OMT, but the same restriction should not also be in place for FFRct.

Regarding the list of situations where FFRct testing is not reasonable and not covered, we recommend the following:

• Severe obesity (BMI >35 kg/m²)
  • FFRct has been shown to be accurate in patients as long as the imaging passes the CT quality process HeartFlow conducts on every scan they receive. This was shown in the ADVANCE registry and through real world experience at high volume centers. We think that BMI should be a consideration when ordering a coronary CTA, but if the images are felt to be of high quality by the reading physician and acceptable quality by Heartflow, the BMI should not be a limitation to FFRct. We feel that the BMI restriction should be removed.
• Bare metal intracoronary stent
• Prosthetic valves
  • We agree with the restriction on patients with prosthetic valves and prior stents.

• Heart rate or arrhythmia > 65 beats/min at the time of image acquisition
  • With newer scanners with improved temporal resolution, image quality remains excellent at higher heart rates. The protocol can be adjusted at the scanner to include the end-systolic phase of the cardiac cycle to allow for accurate imaging in patients with tachycardia or arrhythmias. As above for BMI, we strongly feel that FFRct should be ordered based on imaging quality (as determined by the reading physician and the HeartFlow quality check), and not pre-test factors. We feel that the restriction on heart rate should be removed.
• Allergy or intolerance to iodinated contrast material
  • This would be a potential contraindication for a CTA , but if the CTA is performed, there is no reason that the scan cannot be sent for FFRct analysis if there is a lesion of uncertain significance. We feel that this restriction should be removed.

• Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000
  • As above, HeartFlow conducts a quality/feasibility analysis on every scan they receive. In studies with high calcium burden that pass this analysis, FFRct analysis has been shown to be accurate in both real world experience and clinical trials. We feel that a calcium score restriction should be removed from the LCD.
• Renal insufficiency (i.e., estimated glomerular filtration rate < 60 mL/min/1.73m²)
  • If the CTA is performed, there is no reason it should not be sent for FFRct based on the patient’s kidney function. FFRct provides no additional risk to the patient and requires no additional contrast. We feel that this restriction should be removed from the LCD.

• Uncooperative patient
  • Again, this should be based on image quality, and not image acquisition variables. We feel that this restriction should be removed from the LCD.
• Patient unable to sustain breath-hold for 5-10 seconds
  • Again, this should be based on image quality, and not image acquisition variables. We feel that this restriction should be removed from the LCD.

• Heart transplantation
  • We agree with this restriction.
• Recent prior myocardial infarction within 30 days
  • We agree with this restriction.

• Suspicion for acute coronary syndrome (where acute myocardial infarction or unstable angina have not been ruled out)
  • We agree with this restriction.
• Severe aortic stenosis
  • We agree with this restriction.

• Known severe 3-vessel disease (> 50% stenosis is all 3 major vessels)
  • We agree with this restriction – these patients should undergo left heart cath for further evaluation. Post coronary artery bypass surgery, evaluation of venous grafts

Thank you.

Thank you for your comments and particularly your agreement with some parts of the policy. We will address each of your suggested changes below.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Single vessel disease with OMT: We agree with the removal of the single vessel disease and the reference to optimal medical therapy and have removed these requirements.

Prosthetic valves: Strict adherence to careful patient selection contributed significantly to demonstrating the net health benefit for FFRCT. No supporting research was submitted with these comments. Pivotal trials for FFRCT consistently excluded patient populations with prosthetic valves and extensive coronary calcification from study. Furthermore, these limitations are reported by the manufacturer in FDA filings. Due to the unavailability of U.S.-based, high-quality, peer-reviewed published clinical research in patient populations with extensive calcification or prosthetic valves, we must conclude that it is still investigational and, therefore, unreasonable to consider for coverage at this time. Upon availability of this evidence, as well as removal of this limitation from the FDA-validated filings, we can reconsider this coverage limitation upon request. The limitation will remain as written.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing therefore does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

I am writing to provide comments on Palmetto’s proposed LCD on the FFRct Analysis, titled Non-Invasive Fractional Flow Reserve (FFR) for Stable Ischemic Heart Disease (DL38278). The FFRct Analysis has become an integral part of the workup at our facility for patients who present with symptoms suggestive of coronary artery disease (CAD). The clinical evidence supporting use of a pathway consisting of coronary CT angiography (CCTA) and the FFRct Analysis is strong; recent data from a randomized controlled trial demonstrated that a pathway utilizing CCTA led to a 41% lower rate of cardiovascular death or heart attack compared to patients who underwent traditional forms of stress testing.

For patients in whom additional information is needed by the clinician to make a diagnosis, FFRct has demonstrated the highest diagnostic performance to identify CAD compared to other tests including nuclear stress testing (SPECT), the test that is most commonly performed . For these patients a single healthcare system interaction will provide all the information required for physicians to select appropriate treatment. This means patients can spend less time receiving medical care and more time knowing they received the right treatment.

1. The list of indications for which “FFRct testing is not reasonable and therefore not covered” should be removed from the LCD.

HeartFlow, the vendor that provides the FFRct Analysis, performs an independent quality assessment of submitted CT images which is not dependent on any of the restrictions proposed in the LCD (e.g, Severe Obesity (BMI > 35 kg/m²), Heart rate or arrhythmia > 65 beats/min). During this CT assessment, the team at HeartFlow reviews and scores every major cardiac vessel for image quality. If they can discern the boundaries of all the coronary arteries, then accurate FFRct results can be produced. While many of the restrictions proposed in the LCD may lead to higher CT image quality and a higher likelihood of acceptance by HeartFlow, none of these restrictions is practical or necessary. If HeartFlow cannot process CT images and produce FFRct results, then the case will not be completed and Medicare will not be charged. In vast majority of cases diagnostic quality of these cases is significantly higher than alternative such as stress testing. Based on the published literature we see real-world acceptance rates of ~90% regardless of the underlying patient and CT characteristics. It is my firm belief that the restrictions outlined in the draft LCD would unnecessarily restrict access to the FFRct Analysis and cause many Medicare beneficiaries to undergo less accurate diagnostic testing and ultimately unnecessary invasive procedures. These restrictions should be removed and the quality of the underlying CT images (as assessed by HeartFlow) should be the determining factor in whether the FFRct Analysis is warranted.

In case this argument is not sufficiently persuasive, I do have specific comments, supported by literature, on why several restrictions should be removed:

2. Severe Obesity (BMI > 35 kg/m²)

The CCTA and FFRct Analysis pathway is a more accurate option than traditional stress testing for severely obese patients. In the severely obese population, the image quality with SPECT is better when using exercise as opposed to pharmacologic stress; however, only 25% of extremely obese patients are able to adequately exercise. Therefore, most patients require pharmacologic stress, yet optimal dosing in obesity has not been completely elucidated and remains an area of controversy. In addition, obesity leads to decreased image quality and frequent artifacts leading to high rate of equivocal studies and false positive studies. The accuracy of SPECT imaging in these patients is adversely affected by suboptimal image quality and alternative imaging methods are often necessary.

As stated, CCTA is a superior alternative for these patients. We often perform CCTA for pre-op clearance for gastric bypass surgeries and are able to obtain optimal image quality, even in these patients. For the severely obese patients who present with suspected CAD it would be ill-informed to deny them the FFRct Analysis, if needed. If I can obtain a quality CCTA for these patients, withholding the FFRct Analysis from them simply because of their high BMI would actually worsen care by constricting their treatment options.

3. Heart rate or arrhythmias > 65 beats/min at the time of image acquisition

For these patients I think that a statement from the Society of Coronary Computed Tomography sums up the clinical utility of CCTA for patients with high heart rates. They state that “heart rates > 80 bpm used to represent a relative contraindication for CCTA because of high incidence of motion artifacts. However, ongoing hardware and software improvements, with very low temporal resolution, such as dual-source CT and wide-detector scanner, now allow the imaging of patients with higher and irregular heart rates with good imaging success.” I would echo the sentiment expressed by the society; with modern day scanners (64-slice or greater) temporal resolution is high enabling quality image acquisition even in patients with heart rates > 80 beats/min. My suggestion, were this restriction not eliminated completely, would be to allow scans with heart rates up to 90 bpm.

As a final note, the clinical evidence has revealed the accuracy of the FFRct Analysis even in patients with high heart rates. In the NXT trial, the results of which led to FDA clearance of the FFRct Analysis, patients had a heart rate of 63 ± 10 bpm. The PACIFIC trial, another study that found the FFRct Analysis had the highest accuracy compared to other tests, included patients with a heart rate of 57.8 ± 7.7. These two trials demonstrate that accurate FFRct results can be delivered regardless of a high or unstable heart rate.

4. Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000

I also suggest complete removal of the restriction on high coronary calcium scores. These are the patients for whom the FFRct Analysis is perfectly suited. In the past high calcium scores led to lesions severity overestimation on coronary CCTA, however CT FFR completely changed the game! In situations where reading the CCTA is made difficult because of significant calcification, FFRct data can help clarify the diagnosis and treatment plan. As relayed above, HeartFlow can generate FFRct results as long as they can identify the boundaries of the coronary arteries. The advanced image analysis algorithms available to HeartFlow makes this process much more accurate and repeatable even compared to the trained physician eye. I recently received accurate FFRct results for a patient who had a calcium score of 3400! Without the FFRct Analysis, this patient would have been sent directly to the cath lab resulting in an unnecessary and expensive procedure, putting the patient at risk.

5. Bare metal intracoronary stent

Finally, I would suggest removal of the restriction on patients with bare metal stents. From conversations with HeartFlow, my understanding is that they are able to produce FFRct results regardless of whether the patient has a bare metal or drug eluting stent. FFRct results will be provided for all coronary arteries that do not contain a stent, allowing functional information to be utilized when making a treatment decision. The only downside is that the vessel which contains the stent will be grayed out in the final analysis, meaning that FFRct values will not be available for that vessel. Despite this limitation, the FFRct Analysis still holds great clinical utility in evaluating the functional significance of a patient’s disease and whether they require revascularization.

Thank you again for the opportunity to comment on the proposed LCD and to share some of my clinical experience using the FFRct technology.

References were provided for review.

Thank you for your comments we will address each point below.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing therefore does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Prosthetic Valve, Severe aortic stenosis, Prior CABG, Heart transplantation, recent MI and Acute coronary syndrome: Strict adherence to careful patient selection contributed significantly to demonstrating the net health benefit for FFRCT. No supporting research was submitted with these comments. Pivotal trials for FFRCT consistently excluded patient populations with these conditions from the studies. Furthermore, these limitations are reported by the manufacturer in FDA filings. Due to the unavailability of U.S.-based, high-quality, peer-reviewed published clinical research in patient populations with these conditions, we must conclude that its use with these conditions is still investigational and, therefore, unreasonable to consider for coverage at this time. Upon availability of this evidence, as well as removal of these limitation from the FDA-validated filings, we can reconsider these coverage limitations upon request. The limitations will remain as written.

Palmetto's proposed LCD would put significant and inappropriate restrictions on the application of FFRcr in patient populations where we see clear and incremental value over CT or other non-invasive tests, supported by clinical data demonstrating the utility of a CT and FFRct pathway for patients. Specifically, the four most concerning proposed restrictions are:

1. Severe obesity (BMI >35 kg/m²). We commonly acquire high quality and diagnostic scans on patients with BMI in the range of 35-49, and sometimes even higher. HeartFlow is a unique test by requiring an image quality review for every CT scan sent for analysis, where HeartFlow's technology and process determine if the image quality is sufficient to provide high quality FFRcr consistent with published clinical study data. They quantitatively score every scan, and if CT quality is not sufficient, then they reject the case, do not provide FFRct, and there is no charge for the provider, payor, or patient.

While higher BMI may create possible challenges for CT and noise artifact, it also creates significant challenges with other non-invasive tests. A sub­analysis of PROMISE in 8,889 patients demonstrated that increasing levels of obesity adversely affect the diagnostic yield of Nuclear Stress MPI more than CTA. Restricting CTA and FFRct may lead to more testing of other modalities like MPI with worse outcomes. While the authors showed that the yield for patients undergoing CTA as the initial test did not differ based on a BMI cutoff of 35, the yield of CTA was still only 50-60%, which is where FFRct is necessary to avoid unnecessary invasive catheterization (ICA). The multi­ center PLATFORM study compared usual care with the use of FFRcr. For 61% of the patients, the CCTA / FFRct pathway found no significant obstructive CAD and therefore allowed cancellation of an ICA which would have otherwise been required.

Success rate of scanning patients with high BMI may vary depending on site-specific protocols and scanner technologies. Using our newer generation scanners designated for coronary CTA, we can achieve very high acceptance rates with HeartFlow. Even for institutions with older scanner technology, HeartFlow's CT quality criteria will prevent FFRct analysis from being performed or billed on CT scans with sub-par quality. I would urge Palmetto to consider this approach as a dear advantage compared to other tests that are reimbursed without quality requirements in place. Subsequently, I would ask that Palmetto eliminate the BMJ restriction entirely.

2. Heart rate > 65 or arrhythmia during scan. The topic of heart rate (HR) is similar to obesity. HeartFlow analyzes the quality of each image, and if excessive motion artifact caused by HR is an issue, then FFRct will not be completed. In fact, motion artifact is the most common reason that HeartFlow will reject CT studies. Artifact is the potential issue, not HR. While increased HR is correlated with lower acceptance rate for HeartFlow, restricting access to patients with HR >65 would limit a large population who could benefit from FFRct. The data in the chart shows that HR in the range of 66-75 can still be accepted 84-86% of the time, and HR from 76-90 can be accepted 73-77% of the time. Considering that nobody pays for FFRct when it cannot be analyzed, these patients deserve a chance to benefit from FFRct before considering alternative test strategies. PLATFORM demonstrated CT and FFRct helped avoid unnecessary JCA in 61% of patients, regardless of HR. PLATFORM had no HR restrictions.

Since we use dual-source CT with high temporal resolution, we can obtain good quality images on patients with higher heart rates, atrial fibrillation, or PACs. We see overall acceptance rates in the high 90% range despite often dealing with these challenging populations. HeartFlow also provides significant guidance and training to customers for optimizing CT protocols to help improve acceptance rates. Without HeartFlow, we would need to refer more of these patients to ICA, increasing our rate of unnecessary ICA. I would suggest that Palmetto remove the HR restriction in the final LCD.

3. CKD defined as GFR <60. This is a very restrictive cut point for contemporary CT practices, and we scan coronary CTAs on many patients with chronic kidney disease and GFR down to 30 without issues. The application of FFRct has no fundamental connection to patient kidney function or tolerance to contrast. If we determine that coronary CTA is medically safe for a patient, the application of FFRct has zero incremental contrast or risk for the patient Additionally, a recent study of >20,000 patients showed that intravenous contrast material administration was not associated with excess risk of AKI, dialysis, or death, even among patients with comorbidities reported to predispose them to nephrotoxicity. Alternative options, such as referring patient to ICA or other forms of imaging with contrast, may offer increased risk for patients with chronic kidney disease compared to CTA and FFRct. I would suggest that Palmetto remove the GFR restriction from the LCD as FFRct requires no additional contrast or patient visit.

4. History of stents and preceding calcium score > 1000. Regarding stents, HeartFlow can accept and process FFRct for patients in any vessel without a stent, and they have clearance from the FDA to provide FFRct in non-stented vessels. There are clinical scenarios where patients have stenoses in non-stented vessels, and coronary CTA and FFRct can be used to non-invasively assess the functional significance of these vessels. When the non-stented vessel is the primary clinical question, we do not have a better non-invasive option than CT and FFRct. Thus, I would suggest that Palmetto remove this restriction as well.

Regarding calcium scores, there is no known limitation for FFRct related to a calcium score. HeartFlow may reject a case if image quality is impacted by excessive calcium completely obscuring the lumen for significant portions of the coronary arteries, but this local determination of interpretable image quality is unrelated to the calcium score, which is a global measure of calcium. Many patients with calcium scores >1000 have a very interpretable and high-quality CT because the calcium is well distributed throughout the coronary arteries. A sub-analysis of NXT demonstrated no difference in diagnostic accuracy for patients with calcium scores <400 compared to 400-3,599 in accepted CT scans. Restricting FFRct in patients with high calcium scores will lead to unnecessary ICA since 38% of patients with high CAC in this study have FFRct >0.80 compared to only 15% with a stenosis <50%, which would mean referrals to ICA in this sub-population would increase more than 2x without FFRct.

Based on capabilities of our protocols and new generation scanners, this determination should be at the discretion of the reader seeing the coronary arteries and HeartFlow's image quality scoring, rather than using binary cut­ points of calcium scores or history of stents.

In conclusion, we have seen important patient benefits and clinical impact by incorporating FFRct into our assessment of patients with symptomatic, stable CAD. Overall, we have been very responsible and diligent in our patient selection for FFRct, sending fewer than 20% of all coronary CTA for FFRct analysis. The additional proposed restrictions would reduce our selected group of patients where we see the strongest benefits. If these proposed restrictions remain, then I would certainly expect an increase in referrals to ICA or increased test layering, resulting in an outcome of increasing our rate of non-obstructive findings during ICA. In addition to the lower quality care of unnecessary ICA, we expect these restrictions would result in higher overall costs per patient for CMS.

References were provided for review.

Thank you for your comments we will address each item individually below.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing therefore does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Through my practice of using coronary CTA, I have been exposed to the Non-invasive Fractional Flow Reserve or FFRCT. FFRCT is an innovative technology for the diagnosis and treatment of coronary artery disease (CAD). FFRCT provides important diagnostic information non-invasively which prevents many patients from undergoing an unneeded invasive coronary angiography which leads to reduced costs, better outcomes, and improved quality of care. Thank you for the opportunity to comment on Palmetto’s proposed Local Coverage Determination (LCD) on this non-invasive diagnostic test.

I would like to comment on the proposed contraindications that I feel are unreasonable and should be modified or removed in the final LCD.

These restricted indications include:

• Severe obesity (BMI >35 kg/m²)
• Bare metal intracoronary stent
• Prosthetic valves
• Heart rate or arrhythmia > 65 beats/min at the time of image acquisition
• Allergy or intolerance to iodinated contrast material
• Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000
• Known severe 3-vessel disease (> 50% stenosis is all 3 major vessels)

Severe Obesity BMI > 35kg/m²

Based on the over 250 published studies on Coronary CTA and HeartFlow - FFRCT, I would like to highlight there is no substantive evidence that a BMI > 35 kg/m² leads to a reduction in image quality or inaccurate FFRCT results. One of the trials associated with Coronary CTA quality is a study from Macron, et al. showing that clinicians were able analyze 97% of vessels utilizing Coronary CTA in patients with a mean BMI of 36.0 kg/m². Another trial from Mangold, et al. concluded that Coronary CTA provided high diagnostic accuracy in both non-obese and obese patients. In addition, there has been tremendous improvements in the quality of CT scanners which has allowed for excellent image quality for larger patients.

The rate of vascular access site complications for patients undergoing invasive angiograms is greatest within the morbidly obese population. The PLATFORM trial showed that FFRCT led to a significant reduction in angiograms performed (61%) which may improve outcomes for the severely obese. I would recommend increasing the BMI restriction to at least 45 kg/m².

Bare metal intracoronary stent

Heartflow can process FFRCT analysis whether the patient has a bare metal or drug eluting stent. FFRCT results will not be provided for the coronary system which contains a metallic stent. FFRCT results will be returned for all coronary systems which do not contact a stent, enabling functional information to be utilized in decision making for these systems.

Heartflow cannot provide FFRCT results for patient under the following scenarios circumstances listed below.

1. A stent is present in the left main coronary artery
2. There is a stenosis >30% in the left main coronary artery and 1 or more stent in the left system
3. There are stents present in 2 or more coronary systems.

I recommend removing the restriction on the presence of stents entirely. If Palmetto desires to modify the restrictions in the final LCD, I would recommend modifying to the manufacturer’s list of limitations, provided above.

Prosthetic Valves

The presence of a prosthetic valve in a patient’s heart will not affect the quality of a Coronary CTA study since the valve will not impinge on my ability to read the amount of disease present in the coronary arteries. Similarly, the presence of a valve will not hinder HeartFlow from performing the FFRct analysis. As long as HeartFlow has a clear view of the coronary arteries they can accurately calculate FFRCT values. This restriction should be removed from the final LCD.

Heart rate or arrhythmia > 65 beats/min at the time of image acquisition

Improvements in CT scanner technology have enabled high quality images to be successfully obtained, even in patients with high heart rates. Again, I would like to reference the Macron study that showed physicians were able to successfully analyze 97% of vessels utilizing Coronary CTA in patients with a mean heart rate of 91.4 ± 48.9 beats/min. SCCT guidelines for the performance of Coronary CTA state that heart rates > 80 bpm used to represent a relative contraindication for Coronary CTA because of high incidence of motion artifacts. That being said, dual-source CT and wide detector scanners now allow for successful imaging of patients with irregular and higher heart rates. This has been demonstrated in the recent NXT trial which demonstrated high FFRCT accuracy despite high heart rates (63 ± 10 bpm). I would recommend that the heart rate restriction be removed entirely from the final LCD.

Extensive Coronary Calcification >1000

The NXT Calcium paper (Norgaard, et al.) states that there was no statistically significant difference in per-patient or per-vessel accuracy, sensitivity, and specificity of FFRCT across Agatston Score quartiles. The authors of the paper concluded that FFRCT provided high per patient and per vessel diagnostic performance and discrimination for ischemia over a wide range of coronary calcification severity. Also, the diagnostic performance of FFRCT was superior to coronary CTA stenosis interpretation regardless of calcium level. Lastly, SCCT guidelines recommended the decision to proceed with Coronary CTA in the presence of high coronary calcium score should be left to the discretion of the referring and attending physician. Given the high performance of FFRCT regardless of Agatston Score I would recommend that this restriction be removed from the final LCD.

Known severe 3 vessel disease

Regardless of the severity of a patient’s disease the accuracy of FFRCT remains consistent. In the ADVANCE trial [6] the investigators found the following:

• 22.9% of the subjects (n = 943) were referred for revascularization after Coronary CTA
• 22.3% of these patients were reclassified to medical therapy alone after FFRCT analysis
  • PCI to medical therapy, 20.9% (n = 198)
  • CABG to medical therapy, 1.4% (n = 12)

In this trial FFRCT demonstrated the ability to dramatically change planned treatment for patients with the most severe disease. Given this clinical utility, access should not be restricted for patients with severe 3 vessel disease and I recommend this restriction’s removal from the final LCD.

I would like to thank you for the opportunity to comment on the proposed LCD for the HeartFlow - FFRCT analysis. I hope the information I have provided on the clinical and health economic benefits of the FFRCT analysis will positively impact the Palmetto local coverage determination. I look forward to continuing this service for Medicare recipients.

References were provided for review.

Thank you for your comments we will address each of your concerns below.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing therefore does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Prosthetic valves: Strict adherence to careful patient selection contributed significantly to demonstrating the net health benefit for FFRCT. No supporting research was submitted with these comments. Pivotal trials for FFRCT consistently excluded patient populations with prosthetic valves and extensive coronary calcification from study. Furthermore, these limitations are reported by the manufacturer in FDA filings. Due to the unavailability of U.S.-based, high-quality, peer-reviewed published clinical research in patient populations with extensive calcification or prosthetic valves, we must conclude that it is still investigational and, therefore, unreasonable to consider for coverage at this time. Upon availability of this evidence, as well as removal of this limitation from the FDA-validated filings, we can reconsider this coverage limitation upon request. The limitation will remain as written.

After reviewing the proposed LCD I have some concerns I would like to comment on.

• BMI restriction on patients >35 kg/m²
• Heart Rate >65 beats/min
• Severe 3-vessel disease
• Renal insufficiency (i.e. estimated glomerular filtration rate < 60 ml/min/1.73m²)

Limiting the use of FFRct to patients with a BMI <35 kg/m² would eliminate a lot of patients that could benefit from coronary CTA and FFRct. Routinely in our hospital we routinely scan patients with BMIs above 40 kg/m². These patients would be getting a significantly higher dose if they receive SPECT tests to assess functional severity of disease, and these are precisely the patients who tend to have false positive non-invasive tests which could result in unnecessary angiography. HeartFlow, the supplier of FFRct, has a comprehensive CT quality implementation process where they come onsite and train our CT techs on best practices to create the best images possible. They also require sites to complete a CT image quality assessment prior to going live with FFRct. New sites must submit 10 CT scans for quality assessment; 7 of these scans must pass must before the site can use FFRct. Furthermore, the ADVANCE registry states that BMI is not a statistically significant predictor of CT quality rejection. Finally, there have been many improvements in CT scanner technology which have enabled high quality images to be obtained, even in obese patients. A study from Macron, et al. showed that physicians were able to successfully analyze 97% of vessels utilizing CCTA in patients with a mean BMI of 36.0 ± 5.9 kg/m². According to our experience, if the CCTA exam passes the CT quality process the FFRct results will be accurate even in patients with a BMI >35. This has been our experience to date, and we do not consider BMI when deciding to send data for FFRct.

Requiring the heart rate to be <65 beats/min is also an unnecessary restriction. Many of the same points can be made here that were covered in the BMI section above. At our site we aggressively use beta-blockade to achieve a target heart rate less than 60. However, in some cases this is not possible despite beta-blockade. In our experience we have gotten high-quality FFRct data in patients with heart rates > 75 beats/min. Heartflow trained our CT techs to get the best image quality possible and required 10 successful submissions to ensure that our CT teams understand how CCTA image quality is related to FFRct accuracy. The Macron study stated above also analyzed heart rates and their effect on CT quality. This study showed that physicians were able to successfully analyze 97% of vessels utilizing CCTA in patients with a mean heart rate of 91.4 ± 48.9 bpm. Other studies such as the NXT trial, and the Pacific trial both demonstrate high accuracy of FFRct regardless of heart rate.

We would also like to address the disqualification of patients with known severe 3-vessel disease (>50% stenosis in all 3 major vessels) detected on a CTA study. As you know, CCTA has extremely high sensitivity, however its specificity is much lower. FFRct can be used to determine if there is really likely flow limiting stenosis in patients with a visual stenosis >50% stenosis in 3-vessels and could suggest a trial of medical therapy in patients not likely to have a mortality benefit from coronary angiography and revascularization. This could reduce the need for unnecessary angiography in a patient that may have only a single vessel with likely ischemia by FFRct. The only way to acquire the functional information needed to determine appropriate treatment is either through an FFRct analysis or an invasive angiography. The ADVANCE study has shown us that among the 22.9% (n=943) of subjects for whom the CCTA based management plan indicated revascularization, 22.3% were reclassified to medical therapy alone after access to an FFRct analysis. Additionally, the SYNTAX III study took this concept a step further and found that the Syntax Score (SS) was able to objectively assess the severity of CAD, aiding physicians in deciding whether CABG or PCI was the best treatment option. The Syntax score becomes important in patients with 3-vessel CAD. For patients with a SS >22 points, CABG is the preferred strategy, whereas in patients with a SS <22, both CABG (Class I) and PCI (Class IIb for the American guideline and Class I for the European guideline) can be recommended. Ultimately Noninvasive functional Syntax score (FSS) using FFRct reclassified 30% of patients from the high and intermediate-SS tertiles to the low risk tertile and potentially an alternative revascularization strategy can be offered.

Finally, we would like to ask that the restriction on renal insufficiency be removed from this LCD. Once the patient has undergone a CCTA no additional imaging, radiation, contrast, or time is required of the patient. The CCTA image data is simply sent to HeartFlow for processing and completion of the FFRct Analysis. Therefore, the patient’s renal status is immaterial to the decision of whether to order FFRct. This requirement, which is too restrictive as is, belongs on an LCD related to the CCTA exam itself. We image patients with contrast with GFR as low as 30 ml/min/1.73m² and would recommend this level as an upper limit for any final LCD language.

Based on the importance of FFRct to decision-making in our patients, I would recommend that the aforementioned restrictions be removed in their entirety from the final LCD. We believe FFRct technology keeps our patients from undergoing a diagnostic angiography, and saves the healthcare system money.

Thank you for your comments we will address each of your concerns below.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate and renal insufficiency: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Advances in CT technology over the last decade have enabled image capture of the coronary arteries, once extremely difficult due to the mobile nature of the heart (Coronary CTA). More recently, these anatomic images of the coronary arteries have been supplemented with functional data made possible by Fractional Flow Reserve (FFR) data computed from CT images, known as FFRCT. This technology is the subject of your proposed LCD: Non-Invasive Fractional Flow Reserve (FFR) for Stable Ischemic Heart Disease (DL38278). This technology offers substantial clinical benefits for patients including a 61% reduction in the number of invasive coronary angiograms performed, leading to lower costs and happier patients. I am appreciative that this technology will now be available for the Medicare patients I evaluate and am happy to offer my comments on the proposed LCD.

FFRCT technology has been shown to be accurate, leading to a change in clinical decision making compared to coronary CTA alone. The two recent trials that have reported this are: 1. The NXT trial which reported a per-vessel accuracy of 86% compared to invasive FFR in 254 patients, and 2. the PACIFIC trial which reported per vessel accuracy of 87% compared to invasive FFR, with higher diagnostic accuracy when compared to SPECT, PET and coronary CTA.

I strongly disagree with the restriction related to “Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000.” The literature shows FFRCT is very helpful with decision making in patients with extensive calcification. FFRCT can be used to assess the functional significance of heavily calcified lesions that could not have been assessed by coronary CTA alone. This has been reported in a study by Norgaard, et al. in which the investigators found that the extent of calcification did not have an impact on per-vessel and per-patient accuracy, sensitivity, and specificity. FFRCT showed a 3-fold reduction in false positives by adding FFRCT to coronary CTA, even for patients with high Agatston calcium scores. The overall conclusion of the paper was that FFRCT provided high per-patient and per-vessel diagnostic performance and discrimination for ischemia over a wide range of coronary calcification severity. Based on this evidence my suggestion is that the restriction on patients with high calcium scores be removed from the final LCD.

Coronary CTA with FFRCT is frequently used to assess obese patients with suspected CAD, and I object to the restriction of use of FFRCT in patients with BMI > 35. I find that other non-invasive testing alternatives like stress tests do not work well in morbidly obese patients. In addition, these patients have high rates of vascular complications in the cath lab, and I hesitate to recommend these patients for invasive procedures. I find that coronary CTA is a reasonable choice in these patients as CT technology has improved over time, and modern CT scanners have powerful x-ray tubes, sensitive detectors and advanced reconstruction methods which allow imaging of morbidly obese patients. This has been reported by Macron, et al. who showed that physicians were able to successfully analyze 97% of vessels utilizing CCTA in patients with a mean BMI of 36.0 ± 5.9 kg/m².

Furthermore, the manufacturer of FFRCT has an acceptance process that evaluates the image quality of each incoming case. Coronary CTA datasets that do not have adequate image quality, and hence are not suitable for coronary segmentation and FFRCT modeling, are not processed by the manufacturer. Notification of this is within 1- 2 hours of case submission, enabling a smooth and quick transition to the next stage in the patient’s care. This is in contrast with other, traditional forms of stress testing in which it sometimes takes up to 2 weeks for results to be available. Finally, for cases for which the manufacturer cannot produce FFRCT results, there is no charge to the healthcare system. For these reasons, I suggest that the BMI restriction be removed from the final LCD.

I also disagree with the exclusion of patients with heart rate > 65 BPM at the time of image acquisition. Recent advances in scanner technologies like dual source, volume detectors, and improved gating software have enabled the use of Coronary CTA in patients with high heart rates with proper protocols. In fact, SCCT guidelines note that these new technologies allow for good imaging success in patients with high and irregular heart rates. This has been shown in a study by Macron, et al. which showed that physicians were able to successfully analyze 97% of vessels utilizing CCTA in patients with a mean heart rate of 91.4 ± 48.9 bpm. The manufacturer’s own studies on the accuracy of FFRCT supports its utilization in patients with higher heart rates. The NXT trial included patients with an average heart rate of 63 ± 10 bpm at the time of coronary CTA acquisition.

As stated earlier, the manufacturer of FFRCT technology has an inspection and acceptance process, approved by the FDA, that excludes cases that are not suitable for coronary segmentation and FFRCT modeling. Multiple trials like NXT and PACIFIC have shown that the accuracy of the technology is high for cases that pass the image inspection process, regardless of patient heart rate. FFRCT results are not provided for the cases that do not have sufficient image quality to pass the acceptance process, and Medicare will not be charged for such cases. Based on this information I believe the heart rate restriction should be removed from the final LCD.

References were provided for review.

Thank you for your comments we will address each of your concerns below.

Restrictions pertaining heart rate: We agree this restriction is more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Coronary CT is now an established modality to evaluate presence or absence of CAD, the severity of disease and its effect on myocardium. Majority of coronary CTA's are done in a patient with clinically suspected CAD with an equivocal or abnormal stress test or symptomatic patients with several risk factors for CAD. A good quality coronary CTA often provides a definitive answer for appropriate management. However, CTA cannot determine the hemodynamic significance of detected atherosclerotic coronary artery disease. Many such patients unnecessarily end up undergoing left heart catheterizations and invasive FFR.

I have researched and read extensively about CT-FFR topic while preparing for the invited lectures for EPIC 2017 and STR 2019 meetings. I understand, the important role CT-FFR can play in select patients and avoid unnecessary invasive catheterization, save money and improve patient's experience. We are currently in the process of implementing CT-FFR use at UAB. I read the proposed LCD by Palmetto and would like to provide my comments in regards to a few of the indications listed under "FFRct testing is not reasonable and therefore not covered for patients with the following conditions:"

Severe Obesity (BMI >35 kg/m²)

Recent advances in CT technology hardware as well as software with availability of dual source and wide detector CT scanners with significantly higher generating power have enabled high quality imaging even in obese patients. A study from Macron, et al. showed that physicians were able to successfully analyze 97% of vessels utilizing CCTA in patients with a mean BMI of 36.0 ± 5.9 kg/m². BMI was not a statistically significant predictor of CT Quality rejection in the ADVANCE registry where BMI characteristics range was 26.3 ± 4.8. In addition, the rate of vascular access-site complications for patients undergoing invasive coronary angiography (ICA) is highest in morbidly obese patients. FFRct offers these patients a chance to avoid undergoing an ICA and therefore potentially avoid a complication.

Heart rate or arrhythmia> 65 beats/min at the time of image acquisition

Again, the recent advances in CT technology now allow us to image patients with higher and irregular heart rate with success. In fact, the SCCT guidelines for the performance of CCTA stated that "heart rates> 80 bpm are used to represent a relative contraindication for CCTA because of high incidence of motion artifacts. The limitations in regard to higher heart rates depend on the equipment used, and scan settings have to be adjusted accordingly." In the study published by Macron, et al. showed that physicians were able to successfully analyze 97% of vessels utilizing CCTA in patients with a mean heart rate of 91.4 ± 48.9 bpm. In the NXT trial and PACIFIC FFRct clinical trials there was high accuracy of FFRct regardless of heart rate (NXT - Heart Rate: 63 ± 10 bpm, PACIFIC FFRct heart rate: 57.8 ± 7.7 bpm).

Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000

Norgaard, et al. in the NXT calcium paper found no statistically significant difference in per patient or per-vessel diagnostic accuracy, sensitivity, and specificity of FFRct across Agatston Score (AS) quartiles. They also noted that FFRct per-patient and per-vessel diagnostic accuracy (74% and 83%, respectively), sensitivity (88% and 82%), and specificity (68% and 84%) in the highest AS quartile were equal to or superior to diagnostic performance characteristics reported for most conventional noninvasive ischemia testing modalities when compared with directly measured FFR.

The SCCT guidelines recommend that "the decision to proceed with CCTA in the presence of high coronary calcium score should be left to the discretion of the referring and attending physician.

Known severe 3-vessel disease (> 50% stenosis is all 3 major vessels)

Management strategy based on non-invasive CT-FFR can be changed due to reclassification of high and intermediate severity to low risk. In the ADVANCE registry among the 22.9% of subjects (n = 943) for whom the CCTA-based management plan indicated revascularization, 22.3% were reclassified to medical therapy alone after FFRct analysis (PCI to medical therapy 20.9%, n=198; CABG to medical therapy 1.4%, n=12).

Collet, et al. in the SYNTAX Ill trial reported that the noninvasive functional Syntax score (FSS) using FFRct reclassified 30% of patients from the high and intermediate-SS tertiles to the low risk tertile in patients with 3 vessel CAD.

In conclusion, due to technological advances in hardware and software of CT scanners, the image quality of coronary CTAs has significantly improved even in patients with higher irregular heart rate, extensive coronary artery calcifications and large BMI. CT-FFR is now feasible in these patients and can provide accurate information which in turn could be very helpful for management decisions in symptomatic patients with CAD. CT-FFR can reclassify intermediate and high-risk category patients to low risk even in patients with 3 vessel disease and avoid further unnecessary, expensive and invasive workup. I strongly believe that expanding the indications for the above discussed categories will be extremely valuable for the Medicare beneficiaries.

References were provided for review.

Thank you for your comments we will address each of your concerns below.

Restrictions pertaining heart rate and arrhythmia: We agree these restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

The American College of Cardiology would like to thank you for the opportunity to comment on Palmetto’s proposed Local Coverage Determination (LCD) on Non-Invasive Fractional Flow Reserve (FFR) for Stable Ischemic Heart Disease (SIHD). The American College of Cardiology (ACC) envisions a world where innovation and knowledge optimize cardiovascular care and outcomes. As the professional home for the entire cardiovascular care team, the mission of the College and its more than 52,000 members is to transform cardiovascular care and to improve heart health. The ACC bestows credentials upon cardiovascular professionals who meet stringent qualifications and leads in the formation of health policy, standards and guidelines. The College also provides professional medical education, disseminates cardiovascular research through its world- renowned JACC Journals, operates national registries to measure and improve care and offers cardiovascular accreditation to hospitals and institutions.

The College has several areas of concern with the LCD as proposed:

• Incomplete indications
• Restrictions on single vessel disease
• Indications which would be most appropriately placed in a different LCD
• Indications related to coronary CT angiography (CCTA) image quality
• Restrictions on use of FFRct in severe 3-vessel disease
• Refinement to documentation requirements.

Incomplete Indications

Several criteria for coverage of FFRct in stable symptomatic patients are listed in the draft LCD. Gaps and odd criteria exist within that list of five criteria. First, one of the major coronary arteries is not even listed. The policy should include the right coronary artery. Second, coverage criteria for the left anterior descending coronary artery and the left circumflex disease are limited to proximal portions of the arteries. This should be changed to include both proximal AND mid-portions of the arteries. Third, the indication for left circumflex disease appears to be limited to the dominant system. The word “dominant” should be removed so that coverage is simply for left circumflex disease.

Single Vessel Disease

The proposed LCD seeks to limit utilization of FFRct in patients with single vessel disease to those with >50% stenosis and insufficient response to optimal medical therapy (OMT). This restriction to patients that have failed OMT is referenced from the Appropriate Use Criteria (AUC) related to revascularization and is not relevant to a diagnostic modality, therefore it should not be utilized to restrict access to a diagnostic test that determines whether a specific lesion is causing ischemia. Depending on the results, a physician may decide they need additional information to determine appropriate treatment or that OMT is the correct course of action. The recent AUC on coronary revascularization in patients with SIHD outlines clinical scenarios for which it is appropriate to revascularize patients. For single vessel disease that does not involve the proximal LAD or Circumflex artery, use of OMT is required in order for revascularization to be considered appropriate. Therapeutic AUC should not be applied to restrict diagnostic decision making. The College requests that Palmetto revise this portion of the LCD and should be modified to remove the OMT requirement.

Lesions in the 40-70% range are termed “intermediate” making a 50% cut off inappropriate. Additionally, a set of lesions even less severe than 40%, in series, can cause ischemia, therefore it is not appropriate to set any cutoff below which FFRct is not indicated.

Indications which would be most appropriately placed in a different LCD

There are several coverage restrictions in the proposed LCD for FFRct that appear to be related to perceived coverage restrictions for CCTA. These include:

• Allergy or intolerance to iodinated contrast material
• Renal insufficiency (i.e., estimated glomerular filtration rate < 60mL/min/1.73m²)
• Uncooperative patient
• Patient unable to sustain breath-hold for 5-10 seconds

These restrictions are not related to the acquisition or performance of the FFRct service. Once the CCTA images are acquired, the decision to order a functional analysis of coronary blood flow via FFRct will be made independent of these four factors. These factors may play an important role in a physician’s decision whether to order a CCTA to evaluate suspected SIHD and should be considered as part of the existing Palmetto LCD on CCTA, not in this LCD.

FFRct in relation to CCTA

There are several coverage restrictions in the proposed LCD for FFRct that are related to image quality of the CCTA and its perceived effect on the accuracy and clinical utility of FFRct. These include:

1. Severe obesity (BMI >35 kg/m²)
2. Bare metal intracoronary stent
3. Prosthetic valves
4. Heart rate or arrhythmia > 65 beats/min at the time of image acquisition
5. Extensive coronary calcification by plain film or high calcium scores with prior Agatston score > 1000

The mentioned criteria do not apply to current generation scanners aimed at image acquisition of the heart (>64 slice scanners). Contemporary scanners contain improved hardware and software components, such as dual-source CT and wide-detector scanners, which allow quality imaging of the most difficult patient anatomies and situations. These proposed restrictions should be removed or broadly expanded in order to ensure adequate access for Medicare beneficiaries. Alternatively, simply requiring a technically adequate CCTA would achieve the spirit of these restrictions in a more flexible way.

There are multiple publications that speak to the performance of CCTA in both the obese and high heart rate populations. A recent study from Mangold et. Al examined that accuracy of CCTA in obese (≥ 30 kg/m²) and non-obese (< 30 kg/m²) patients. The investigators determined that there was no statistically significant difference in overall per-patient sensitivity, specificity, positive predictive value, negative predictive value, and accuracy and concluded that CCTA “provides high diagnostic accuracy in non-obese and obese patients.” Another publication from Zimmerman, et al. evaluated how accuracy was affected by both obesity and heart rate control. Four heart rate groups were included in the trial (< 70, 70-79, 80-89, ≥ 90 bpm) and physicians graded image quality based on their ability to visualize the artery. The authors found that 99% of coronary arteries were of diagnostic quality and concluded that “diagnostic quality [CCTA] examinations can be obtained without premedication regardless of body size.”

The presence of prosthetic valves and extensive coronary calcification only preclude the performance of quality CCTA to the extent that these factors prevent adequate imaging of the lumen of the coronary arteries and prevent the reading physician from interpreting the presence of SIHD. It is the current position of the College that the decision to proceed with CCTA in the presence of high coronary calcium score should be left to the direction of the referring and attending physicians. The presence of other traditional obstacles, including the presence of prosthetic valves should be handled in a similar manner. If the performing and interpreting physicians are confident in their ability to perform and read the CCTA in the presence of these or other factors, then the decision to perform the CCTA should be left up to the attending physician.

Evidence from the recent NXT trial that examined the accuracy of FFRct in patients with a wide range of coronary calcium scores concluded that there was no statistically significant difference in per-patient or per-vessel diagnostic accuracy of FFRct across Agatston Score quartiles. It was also found that FFRct diagnostic performance was equal to or superior to diagnostic performance characteristics reported for most conventional noninvasive ischemia testing modalities.

FFRct in Severe 3-Vessel Disease

The proposed LCD limits FFRct utilization in patients with known severe 3-vessel disease (>50% in all 3 major vessels) referencing a forgone conclusion that coronary artery bypass grafting (CABG) is necessitated by patients whose CCTA finds such disease. There is substantial clinical evidence investigated in the three SYNTAX family of randomized controlled trials, that PCI or stenting is a viable option for patients with 3-vessel disease that is determined to be not functionally significant via FFR information. It is accepted practice that when FFR shows 1 or 2 of the vessels in a 3-vessel disease patient to be not hemodynamically significant, that patient is redefined as a 1 or 2-vessel disease patient.

In these trials, the Syntax Score (SS) and objective assessment of the severity of a patient’s SIHD, aides the heart team in determining the best course for revascularization, either PCI or CABG. Patients with a low risk SS (≤ 22) were shown to have similar outcomes with CABG or PCI while patients with an SS ≥ 23 had better outcomes with CABG after 5-year follow up. Use of FFRct to determine a patient’s functional SS in the SYNTAX III trial reclassified 30% of patients to a low risk SS, potentially enabling a less invasive (PCI) form of treatment.

The College urges Palmetto to remove restrictions on use of FFRct in severe 3-vessel disease.

Documentation Requirements

There are several “Documentation Requirements” in the proposed LCD which would place an undue burden on providers and inadvertently restrict access to the FFRct service for Medicare beneficiaries. As above, heart rate at time of image acquisition and body mass index do not influence study quality and should not be included in the LCD. Thus, the College suggests they be removed from the final LCD in order to ensure parity with other services used to diagnose and treat suspected SIHD.

Conclusion

Several of the restrictions found within the proposed LCD will limit the ordering FFRct and place financial burden on providers on those who do. Access to FFRct can lead to fewer unnecessary diagnostic catheterizations making it a cost-effective diagnostic test. We appreciate the opportunity to comment.

References were provided for review.

Thank you for your comments we will address each of your concerns below.

Documentation requirements: We recognize that the heart rate at the time of image acquisition is best applied to CCTA, rather than FFRCT and is removed. Upon receipt of evidence to support the use of FFRCT in Medicare beneficiaries with BMI greater than 39 kg/m², we can reconsider this coverage limitation upon request. The requirement for documentation of BMI will remain unchanged.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Single vessel disease with OMT: We agree with the removal of the single vessel disease and the reference to optimal medical therapy and have removed these requirements.

Prosthetic valves: Strict adherence to careful patient selection contributed significantly to demonstrating the net health benefit for FFRCT. No supporting research was submitted with these comments. Pivotal trials for FFRCT consistently excluded patient populations with prosthetic valves and extensive coronary calcification from study. Furthermore, these limitations are reported by the manufacturer in FDA filings. Due to the unavailability of U.S.-based, high-quality, peer-reviewed published clinical research in patient populations with extensive calcification or prosthetic valves, we must conclude that it is still investigational and, therefore, unreasonable to consider for coverage at this time. Upon availability of this evidence, as well as removal of this limitation from the FDA-validated filings, we can reconsider this coverage limitation upon request. The limitation will remain as written.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing therefore does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Incomplete indications: We agree with your recommendations and have a bullet for right coronary artery. Secondly, we added proximal and mid portions of the left anterior descending and left circumflex arteries. Lastly, we removed the term “dominant” from the indications.

Restrictions pertaining to image quality and CCTA acquisition, including heart rate, arrhythmias, allergy or intolerance to contrast, renal insufficiency, uncooperative patients, breath-hold inability: These restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

Our cardiovascular service line has adopted widely the use of CTA and FFRct similar to that as proposed by NICE for the NHS due to its impact on diagnostic accuracy and early institution of appropriate medical therapies with resultant impact on cardiovascular mortality. Our diagnostic standard previously had been exercise nuclear scintigraphy in appropriate patients. In a project we have done with Analytics4Life, which I published in PLOS ONE, use of an AI derived tomographic method suggested that real life results of exercise nuclear scintigraphy in patient with chest pain resulted in diagnostic specificity of <30%, accompanied by a high, unnecessary normal catheterization rate. We have subsequently submitted the full trial for publication. These discoveries led to a significant shift in our chest pain approach, with greater use of CTA to achieve the above-mentioned goals in patient care.

I am writing to provide comments on Palmetto’s proposed LCD on the FFRct analysis, titled Non-invasive Fractional Flow Reserve (FFR) for Stable Ischemic Heart Disease (DL 38278). This analysis has become an integral part of the evaluation at our facility. Data supporting CTA use has been widely published, leading to better treatment and improved outcomes. Use of fractional flow reserve, whether in the noninvasive form, or in the invasive laboratory has also demonstrated its utilization in optimal medical care decision making (FAME 1, 2). Hence, a single healthcare visit can lead to optimal decision making, and if revascularization is required, more efficient use of healthcare resources. Use of FFRct frequently alters clinical decision making, creating the right treatment for the right circumstances. Furthermore, datasets that do not meet quality imaging requirements do not get processed.

I believe a generic list that proposes that “FFRct testing is not reasonable and therefore not covered with the following conditions” is much too broad and excludes important groups for which FFRct testing is extremely helpful. Some of these conditions include:

1. Severe obesity - accuracy remains if the CT images pass image quality regardless of BMI. Avoidance of unnecessary procedures is a high priority to prevent unnecessary complications, and invasive testing. Disease severity once known can also guide medical therapy as initial approach if identified. Obesity is a major predictor of access site complications.
2. Heart rate or arrhythmia > 65 beats/min - high quality images can be obtained in the majority of patients, especially as the quality of scanners improve. Moreover, inadequate images lead to FFRct rejection for which the carrier is not charged.
3. Extensive coronary calcification by plain film or high Agaston scores - FFRct provides high per patient and pre vessel diagnostic performance and discrimination for ischemia over a wide range of coronary calcification scores and was superior to coronary CTA stenosis interpretation.
4. Know severe 3-vessel disease - FFRct allows for appropriate treatment planning that allows for incorporation of physiologic information to determine the optimal pathway. Limiting additional invasive testing and overtreatment during PCI results in optimal care for patients.
5. Bare Metal Stents - patients with bare metal stents frequently have other vessels for which this information is helpful.

We appreciate the opportunity to comment on the proposed LCD changes. We hope you take a broader approach for use of this highly helpful technique. The benefit of CTA over prior approaches is known and adding physiologic information can create an optimal approach to patient care.

Thank you for your comments we will address each of your concerns below.

Restrictions pertaining heart rate and arrhythmia: We agree these restrictions are more related to CCTA acquisition, not for FFRCT and will be removed.

BMI restriction: A careful evidentiary review has been performed and based on the published literature available we agree that increasing the BMI to <39 is reasonable but we have not seen adequate research validating an increase to 45.

Extensive coronary calcifications: We agree with the reasoning to remove this as a restriction. Given the SCCT guidelines and the Norgaad study we feel this requirement can be lifted and will add a statement requiring a “technically complete” CCTA to the FFR policy.

• Abbara, S, et al. SCCT guidelines for performance of coronary computed tomographic angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J Cardiovasc Comput Tomogr. 2009;3(3):190-204.
• Norgaard, BL, et al. Influence of Coronary Calcification on the Diagnostic Performance of CT Angiography Derived FFR in Coronary Artery Disease: A Substudy of the NXT Trial. JACC Cardiovasc Imaging. 2015;8(9):1045-1055.

Known severe 3-vessel disease: We agree that this restriction may create confusion and as long as the indications which were clearly outlined in the LCD are met there is no reason to include this statement and it will be removed.

Bare metal stent: We respectfully disagree. No supporting research was submitted with this comment. Instead, the comments confirm the well reported limitations of the technology in the presence of intracoronary stenting. Moreover, the pivotal trials consistently excluded patients with prior percutaneous coronary intervention from participation. Upon review, we find no basis for coverage with the currently available evidence. Incomplete reporting of coronary artery anatomy defeats the purpose of this diagnostic testing therefore does not meet the requirements for payment. Upon availability of evidence to support the efficacy of FFRCT analysis in the presence of intracoronary stents, we can reconsider coverage upon request. This limitation will remain and is expanded more broadly including non-metallic intracoronary stents.