Response to Comments: MolDX: Molecular Biomarkers for Risk Stratification of Indeterminate Pulmonary Nodules Following Bronchoscopy

The following comment was submitted to Palmetto GBA:

I am writing this correspondence to make the following recommendations for the proposed LCD regarding molecular biomarkers for risk stratification of indeterminate pulmonary nodules following bronchoscopy to further improve patient care and improve the efficiency of diagnostic workups through more timely access to validated assays.

The complexity regarding signs and symptoms and the potential for the multiple cancer types that could be the underlying cause of the signs and symptoms is well recognized (Hamilton, BMJ 2015). Consequently, there is a significant unmet need for tests that can support physician management of patients presenting with signs (i.e., abnormal laboratory results or imaging) and symptoms that are consistent with cancer with no direct diagnostic pathways. As more than 60% of cancers are diagnosed following a symptomatic presentation, many patients may undergo undirected and misdirected diagnostic evaluation.

As the genomic landscape is rapidly evolving, the development of tests to facilitate diagnoses will continue to be mature, which underscores the need for MolDX to finalize this proposed LCD as a true foundational LCD that provides coverage for a broad range of tests, including blood-based tests, to expand the availability of options of physicians and patients. In a large prospective trial (Nichols Lancet Oncology 2023), GRAIL's test detected a cancer signal in 323 cases, with 244 cancer diagnoses, yielding a positive predictive value of 75·5% (95% CI 70·5–80·1), negative predictive value of 97·6% (97·1–98·0), sensitivity of 66·3% (61·2–71·1), and specificity of 98·4% (98·1–98·8). For cases in which a cancer signal was detected among patients with cancer, GRAIL's test prediction of the site of origin was accurate in 85·2% (95% CI 79·8–89·3) of cases. Additionally, for many non-specific symptoms studied, the test performance was superior to the physician’s choice of diagnostic clinic (reflecting the intended diagnostic pathway).

Additionally, in the prospective, case-controlled Circulating Cell-free Genome Atlas (ClinicalTrials.gov identifier: NCT02889978; Bryce, JCO Precision Oncology 2023) substudy involving 2,036 cancer and 1,472 noncancer participants, GRAIL's multi-cancer detection test performance in participants with clinically presenting cancers and underlying medical conditions and high suspicion of cancer was analyzed and demonstrated an overall sensitivity of 64.3% (95% CI, 62.2 to 66.4) and an overall accuracy of CSO prediction in true positives of 90.3%. Results from this case-control study indicate this multi-cancer detection test could facilitate efficient workup and stratify cancer risk in symptomatic individuals.

Importantly, individuals at risk of lung cancer may be at increased risk of multiple additional cancer types (e.g., esophagus, liver, stomach, pancreas) based on shared risk factors for lung cancer (i.e., tobacco use). In addition, lung lesions may potentially represent metastatic disease, as opposed to a primary lung malignancy, from a number of different organs outside of the lung.

Based on these data and considerations, I am writing to recommend that this proposed LCD establish coverage guidelines for tests that stratify cancer risk where there is unclear or suspicious clinical, laboratory, or imaging results, not just limited to suspicion of lung cancer. A multi-cancer policy would follow the precedent set by MolDX, including for MRD, and remove the need for multiple single cancer policies (e.g., thyroid).

Thank you in advance for your consideration of this request.

Thank you for your thoughtful comment. A multi-cancer foundational policy for coverage of molecular tests that stratify cancer risk in the setting of unclear or suspicious clinical, laboratory, or imaging results is out of scope of this local coverage determination (LCD), which is focused on risk stratification of indeterminate pulmonary nodules following bronchoscopy, as it is not clear that the evidence that supports use of risk stratification of pulmonary nodules following indeterminate bronchoscopy is broadly applicable to other disease states and clinical decision making processes. We welcome submission of supporting evidence as a new LCD request. MolDX has not reviewed evidence to expand coverage for a multi-cancer foundational policy and will do so when presented with supporting evidence.

The following comment was submitted to Palmetto GBA. We received several comments with the same or similar language and request as Comment 2. These are responded to below collectively:

The Society of Advanced Bronchoscopy was formed in 2018 as a clinically oriented, professional society that seeks to provide experience based, quality education to those training in and performing advanced bronchoscopy. We represent the executive leadership team over a current membership of over 160 members representing a diverse group of community and academic pulmonologists. Bronchoscopic evaluation of pulmonary nodules is a mainstay of our clinical practice. Each year over 1.5 million incidental nodules are found on chest imaging and currently 14.5 million Americans qualify for a low dose screening chest cat scan where nodules may be found. Our ultimate desire is to identify potential lung cancers at their earliest stage when they appear as small nodules on imaging. Determining which nodules need intervention, however, can be challenging. We strongly believe there is a significant need for better tools to risk stratify indeterminate pulmonary nodules prior to biopsy to increase the chances of finding lung cancer in its earliest stages.

It has come to our attention that MolDX, the Medicare contractor responsible for molecular testing coverage decisions, has Proposed a Local Coverage Determination (LCD), Molecular Biomarkers for Risk Stratification of Indeterminate Pulmonary Nodules Following Bronchoscopy, which could provide a path to Medicare coverage for such a tool. However, we understand that this Proposed LCD may limit coverage only after a non-diagnostic bronchoscopy. While we are supportive of this Proposed LCD, we believe there is an opportunity to expand the LCD to provide a path coverage in the indication with the greatest need, which is prior to biopsy.

We understand that this Proposed LCD is intended to be “foundational” in nature so that it provides a path to coverage for this type of testing, but it does not specifically grant Medicare coverage for any particular test. We commend this approach to coverage and applaud MolDX’ s Technical Assessment process to ensure that only high-quality tests with published clinical validity and utility will be granted Medicare coverage.

When assessing pulmonary nodules, risk stratification currently relies on available clinical and radiologic features (size, location, patient history, etc.) to approximate a risk of malignancy (ROM). This risk stratification and classification of pulmonary nodules into low, intermediate, or high risk prior to biopsy has been shown to lack accuracy with both physician-assessed risk and when risk model calculators are used.

While we are bronchoscopists by trade our goal is to only perform invasive procedures when necessary. Unfortunately, many patients with benign lung nodules are subject to potentially unnecessary biopsies each year. Therefore, there is a need for noninvasive options prior to biopsy to classify pulmonary nodules more accurately according to the risk of malignancy.

It is our understanding that there are promising new technologies that will aid physicians in more appropriately risk stratifying pulmonary nodules to reduce unnecessary invasive procedures. By restricting this Proposed LCD to indeterminate lung nodules following bronchoscopy, MolDX is excluding coverage for this class of testing where it is needed most, prior to biopsy. We urge you to remove the post bronchoscopy requirement to this proposed LCD and make this more general to any molecular biomarkers for risk stratification for indeterminant pulmonary nodules.

Thank you for your thoughtful comment. We agree that there is a significant need for better tools to risk stratify indeterminate pulmonary nodules prior to bronchoscopy to increase the likelihood of diagnosing lung cancer in its earliest stages, while also avoiding unnecessary invasive procedures in patients with benign lesions. This local coverage determination (LCD) is focused on risk stratification of indeterminate pulmonary nodules following bronchoscopy and expansion of coverage to pre-bronchoscopy risk stratification is beyond the scope of this particular clinical question and setting. However, given recent expansion of the MolDX program to include proteomics, evidence for all molecular pre-biopsy risk stratification tests in scope of the MolDX program is currently under consideration.

The following comment was submitted to Palmetto GBA:

GRAIL, LLC appreciates the opportunity to submit public comments on MolDX’s proposed local coverage determination (LCD) regarding “Molecular Biomarkers for Risk Stratification of Indeterminate Pulmonary Nodules Following Bronchoscopy" (the “Proposed LCD”). We strongly support the establishment of a foundational LCD for new technologies that assist clinicians in risk stratifying patients, thereby better informing clinical decision-making in the diagnosis of cancer.

The cancer risk stratification market is rapidly growing and evolving with several companies bringing to market single cancer tools, while others, including GRAIL, are bringing multi-cancer tools to this space in the coming years. Specifically, GRAIL intends to bring to market a new blood-based biomarker test with an internal working name of the Diagnostic Aid for Cancer (DAC) test. Based on evidentiary findings to-date from studies that GRAIL conducted in collaboration with respected institutions such as University of Oxford, Mayo Clinic and Cleveland Clinic, we believe the DAC test will significantly improve clinicians' ability to risk-stratify individuals with indeterminate symptoms suspicious for cancer prior to undergoing an invasive biopsy.^1,2 As such, we intend that the DAC test will offer a non-invasive blood-based testing alternative to aid in the diagnosis or exclusion of cancer.

By establishing clear foundational coverage criteria for risk stratification tests, MolDX is continuing its efforts to streamline coverage review of new tests, as evidenced by its approach to Minimal Residual Disease (MRD) testing. Therefore, to ensure specific, clear, and inclusive criteria in the cancer risk stratification space, we respectfully offer the following three recommendations for representation in the Final LCD:

1. The Final LCD should be agnostic with respect to molecular analytes. For example, it should apply to cell-free DNA (cfDNA) tests that otherwise meet the coverage criteria.
2. The Final LCD should be expanded to include tests that risk stratify patients in advance of any invasive diagnostic procedures, by eliminating the requirement for a previous bronchoscopy in the Proposed LCD coverage criteria. We believe that this update will make the LCD more adaptable to newer non-invasive blood-based and biospecimen tests, provided they demonstrate non-inferior analytical and clinical validity compared to existing methods.
3. Finally, we urge MolDX to expand the scope of the Proposed LCD beyond lung cancer and consider multi-cancer coverage guidelines for tests that stratify cancer risk in cases of suspicious clinical, laboratory, or imaging findings.

Background

GRAIL, LLC is a healthcare company with a clear mission: to detect cancer early, when it can be cured. Our focus is on alleviating the global burden of cancer by developing cutting-edge technology that aids in cancer detection and monitoring across multiple cancer types.

We leverage a robust methylation technology platform, next-generation sequencing, population-scale analytical and clinical validation, and state-of-the-art data science to develop blood based tests that have multiple applications such as cancer screening (Galleri^®) and to assess the risk of cancer in patients with cancer suspicion (the DAC test). To date, this technology has been published in over 200 manuscripts and presentations in reputable journals and conferences with more than 300,000 participants in our clinical development program.

Today, there exists a significant unmet need for an efficient evaluation of individuals with symptoms suspicious for cancer such as unexplained weight loss. In many cases, a clear diagnostic pathway is lacking due to lack of focality and specificity for a particular disease process. When a diagnostic pathway exists, it often includes invasive procedures like endoscopy or bronchoscopy, with associated complications. Existing, commonly prescribed screening tests like CEA and CA 19-9 are largely non-specific and challenging to interpret. It is well-documented that delayed cancer diagnosis and treatment can lead to poorer outcomes.^3,4,5 Additionally, most signs and symptoms associated with cancer have a relatively low prevalence. Given these factors, individuals initially suspected of having cancer often undergo various unnecessary and misdirected diagnostic tests and procedures.^6,3,

GRAIL's DAC test is built upon the GRAIL multi-cancer early detection (MCED) technology backbone and has been intentionally clinically validated to accurately stratify patients with symptoms suspicious for cancer, distinguishing between those more or less likely to have the disease. Additionally, the DAC test can accurately predict the origin of the cancer signal, which is particularly important when the presenting symptom is non-specific and could be indicative of cancer from multiple different organs or tissues, e.g., unexplained weight loss. This prediction capability surpasses currently available clinical methods.^1.2

Recommend Clarifications and Modifications to Construct Foundational Cancer Risk Stratification LCD

While GRAIL strongly supports this Proposed LCD, we encourage MolDX to make the following clarifications and modifications in order to better serve the objectives of creating a rigorous coverage policy that can apply to the next generation of tests with minimal modification and adoption in real world clinical practice.

1. Clarify Final LCD is Analyte Agnostic

It is GRAIL’s understanding that the Proposed LCD is not limited to RNA testing, or to any particular type of molecular biomarker. Accordingly, we understand that it would apply to a cfDNA-based risk stratification test that otherwise satisfies the coverage requirements.

cfDNA is a well established analyte that is associated with a shared cancer signal and ultimately cancer diagnosis across the spectrum of cancer biology.⁷ While much of the evidence cited in the Summary and Analysis of Evidence sections references an RNA-based classifier (Veracyte’s Percepta^®), this evidence equally supports the clinical utility of a cfDNA-based molecular classifier that demonstrates clinical validity. Notably, if the cfDNA test provides a clinically valid answer to an analogous clinical question to Percepta^® (i.e. the likelihood that suspicious clinical, laboratory, or imaging results are associated with cancer), a clinician should be able to use the test results similarly. We believe that the Proposed LCD aligns with this understanding, as it applies to all “molecular biomarkers” and there is no limitation to RNA in the proposed coverage criteria. We respectfully request that MolDX clarify and confirm this in the Final LCD.

2. Expand Final LCD to Include Tests that Precede Invasive Diagnostic Procedures

The Final LCD should remove the requirement that “the beneficiary has undergone bronchoscopy for an indeterminate pulmonary nodule (IPN)”. Several clinical commercial laboratories have launched or are developing risk stratification tests for IPNs that are intended to be used based on indeterminate imaging results, in order to guide the determination of whether invasive procedures like bronchoscopy are clinically appropriate. MolDX should ensure that coverage criteria are established for risk stratification tests used prior to any invasive procedure, including bronchoscopy or other biopsy procedures.

For instance, a blood test that can generate accurate, and clinically actionable post-test probability results will further support clinical decision making and [efficiency of] diagnostic workup for individuals with indeterminate symptoms, while potentially avoiding unnecessary invasive procedures. Removing the prior bronchoscopy requirement thus aligns with MolDX’s stated objective that “whenever possible, IPNs in [intermediate risk patients] should be investigated using noninvasive procedures in order to avoid morbidity in patients without cancer.”

Should MolDX determine that currently no new pre-bronchoscopy tests are found to be reasonable and necessary under existing criteria, it behooves MolDX and Medicare beneficiaries to establish coverage criteria for such tests in the Proposed LCD, rather than having to prepare a new LCD at some future point.

3. Expand Final LCD to Include Multi-Cancer Risk Stratification

The circumstance of indeterminate symptoms suspicious for malignancy after imaging or other clinical review is common across different cancers, and is not limited to the clinical challenge of risk-stratifying pulmonary nodules. Cancer prevalence associated with non-specific signs and symptoms, such as unexplained weight loss, anemia or abdominal pain, is well established in NICE NG 12 criteria⁸ which are based on study data from the UK and US patients with signs and symptoms of cancer. NICE criteria threshold for diagnostic evaluation for cancer is a prevalence of 3% (3 - 4x the risk in the average 50 year old). Many non-specific signs and symptoms meet the 3% threshold according to NG12. A majority of patients diagnosed with cancer present with suspicious signs and symptoms, and the ability to intervene early has the potential to greatly improve patient outcomes.

GRAIL collaborated with the University of Oxford in the SYMPLIFY study, where the use of a multi-cancer platform was evaluated in a large-scale, prospective study of symptomatic patients in the UK who were referred from primary care physicians for suspicious signs and symptoms of cancer. GRAIL’s cfDNA multi-cancer platform demonstrated an ability to accurately stratify patients with symptoms suspicious for cancer into those more or less likely to be diagnosed with cancer (positive predictive value 75.5%, negative predictive value 97.6%, specificity 98.4%) and can accurately predict the cancer signal origin (85.2% accuracy among those diagnosed with cancer) beyond currently available clinical and pathologic methods. Additionally, the platform generated clinically actionable pre and post test probability to help guide physician management (Nicholson Fig 2). Aggregate sensitivity across all presenting symptoms and cancer diagnoses was 66.3%. Sensitivity was highest among gastric (95.5%), pancreatic (91.7%), colorectal (70.8%) and lung cancers (67.9%). For cases in which a cancer signal was detected among patients with cancer, the MCED platform prediction of the site of origin was accurate in 85.2% of cases. Additionally, for many non specific symptoms studied, the DAC test performance was superior to the physician’s referral (reflecting the intended diagnostic pathway).¹

Once symptoms become focal, involving a particular organ or organ symptom, cancer is more likely to be advanced. Given the commonality across indeterminate symptoms for cancer, a cancer diagnostic triaging tool such as DAC provides opportunity for the most effective care by increasing efficiency and minimizing medical errors earlier in the diagnostic journey. It is essential to utilize non-invasive tools to determine which symptomatic patients should be evaluated for cancer upstream of focal signs and symptoms to diagnose and treat cancers at earlier stages to improve patient outcomes.

There are significant benefits to evaluating patients with indeterminate symptoms in a multi-cancer approach. For example, pulmonary nodules may also represent metastatic cancers which commonly spread to the lungs like breast, colon, kidney, or less common tumors like carcinoid or extranodal lymphoma, highlighting the importance of multi-cancer testing.⁹ Moreover, many indeterminate lung nodules are identified in smokers who have undergone lung cancer screening, and it is well-recognized that smokers are at elevated risk for over 15 different types of cancer.¹⁰ Other studies have found that following the presentation of symptoms concerning a particular cancer type, a different cancer is diagnosed close to 50% of the time.^1,11 GRAIL’s MCED platform, as used in the DAC test, has demonstrated the ability to predict a cancer signal of origin, supporting a more efficient workup for physicians when a cancer signal is detected.^1,2

To further demonstrate the utility of a multi-cancer platform that stratifies cancer risk for patients with indeterminate symptoms, we developed a model comparing utilization of key diagnostic procedures with the DAC test versus without the DAC test. The model used the Pharmetrics Plus database and test performance from the SYMPLIFY study, and stratified patients by presenting symptoms. Our model demonstrated that by incorporating DAC test to the cancer diagnosis process among symptomatic patients, potential cost offsets can be expected by avoiding some unnecessary procedures. For example, for patients with dyspepsia and unexplained abdominal symptoms like bloating and suspected to have cancer, the DAC test could reduce unnecessary utilization of upper endoscopy; and for patients with anemia, the DAC test could reduce unnecessary utilization of CT scans. All the testing procedures that are potentially avoidable with the DAC test could yield a notable cost offset for CMS.¹²

Given these clinical dilemmas of other indeterminate symptoms beyond indeterminate pulmonary nodules, we encourage MolDX to consider expanding the proposed LCD to an “umbrella LCD” that addresses the cancer risk stratification indication across multiple cancers where such testing is used. We believe this approach would maximize efficiencies in the medium-to-long term by limiting the redundancies involved in developing numerous coverage policies for tests with very similar purposes, but in different cancers. Additionally, a single multi-cancer LCD would allow for a broader category of tests with similar clinical efficacy to be analyzed for coverage through the Technical Assessment process rather than requiring new LCDs, improving both beneficiary access and MolDX’s workflow. This would allow for a more accessible and usable coverage approach for a developing set of pan-cancer stratification tests like GRAIL’s DAC test.

We note that MolDX has recently adopted a very similar approach with respect to coverage of testing for inherited cancer syndromes, minimal residual disease testing for cancer, and molecular syndromic panels for infectious disease pathogen identification testing. Accordingly, we believe that the development of such an LCD would be feasible and consistent with MolDX’s general strategy of establishing broad foundational LCDs.

Conclusion

GRAIL appreciates MolDX’s proposed approach in this forward-looking Proposed LCD to govern indeterminate pulmonary nodule risk stratification testing. While we strongly support the finalization of the Proposed LCD, we urge clarification that the policy be agnostic as to molecular biomarkers, and respectfully request the removal of the post-bronchoscopy requirement for coverage. Additionally, GRAIL encourages MolDX to consider expanding the proposed LCD to an “umbrella LCD” that addresses developing an LCD for risk stratification of indeterminate cancer symptoms.

References were provided for review.

Thank you for your thoughtful comment. Please see below for point-by-point responses to topics identified above:

1. The Final LCD should be agnostic with respect to molecular analytes. For example, it should apply to cell-free DNA (cfDNA) tests that otherwise meet the coverage criteria.

The LCD pertains to molecular biomarkers for risk stratification of indeterminate pulmonary nodules following bronchoscopy that can serve as an adjunct to non-diagnostic bronchoscopy to inform further patient management. This includes molecular analytes within the scope of the MolDX program that form the basis of molecular tests that meet coverage criteria of the LCD. This has been clarified in the policy text.

2. The Final LCD should be expanded to include tests that risk stratify patients in advance of any invasive diagnostic procedures, by eliminating the requirement for a previous bronchoscopy in the Proposed LCD coverage criteria. We believe that this update will make the LCD more adaptable to newer non-invasive blood-based and biospecimen tests, provided they demonstrate non-inferior analytical and clinical validity compared to existing methods.

Thank you. This was addressed in response to a prior comment.

3. Finally, we urge MolDX to expand the scope of the Proposed LCD beyond lung cancer and consider multi-cancer coverage guidelines for tests that stratify cancer risk in cases of suspicious clinical, laboratory, or imaging findings.

Thank you. This was addressed in response to a prior comment.

The following comment was submitted to Palmetto GBA:

On behalf of LungLife AI, I am writing in support of the finalization of proposed LCD to establish coverage criteria for molecular diagnostic testing of indeterminate pulmonary nodules. We agree with MolDX that molecular diagnostic testing for this indication demonstrates clinical utility in guiding appropriate patient management by re-stratifying risk of malignancy.

I. Clinical Background

Differentiating malignant from benign pulmonary nodules represents one of the most urgent clinical problems in early lung cancer detection,¹ particularly as we approach possible widespread adoption of lung cancer screening and increasing use of computed tomography scan generally. The challenges associated with lung nodule evaluation lead to delays in diagnosis and treatment for patients with malignant nodules.² Yankelevitz et al estimate 10-year lung cancer survival rates based on 90-, 180-, and 365-day delays in indeterminate lung nodule diagnosis, demonstrating up to a 62% decrease (absolute) in survival rate due to tumor growth.³ Additionally, reports estimate between 30-40% of all lung nodule biopsies are for benign inflammatory disease.^2,4 Up to 35% of surgeries for lung nodules are on benign disease.⁵ This results in significant unnecessary patient morbidity and utilization of limited health system resources. Together, these data demonstrate the need for both “rule-in” and “rule-out” tests for indeterminate pulmonary nodules.

LungLife AI has developed the LungLB^® test, a four-probe fluorescence in situ hybridization (FISH) (3q29, 3p22.1, 10q22.3, 10cen) assay of whole blood specimen, predictive algorithm-generated evaluation reported as decreased or increased risk for lung cancer. It is intended to be used as an aid in the evaluation of indeterminate lung nodules identified by CT scan. LungLB provides information to physicians considering the best path for nodule work-up, enabling avoidance of unnecessary invasive procedures and reducing delays in treatment. The LungLB^® test is approved by the New York State Department of Health Clinical Laboratory Evaluation Program and is offered in our CLIA-certified and CAP-accredited laboratory.

A multi-site clinical validation study of LungLB in patients with indeterminate lung nodules was published this year in BMC Pulmonary Medicine and demonstrated an AUC of 0.76 in a patient population for which existing diagnostic tools are inadequate, including the Mayo Nodule Risk Model and positron emission tomography (PET), which performed with AUCs of 0.52 and 0.57, respectively.⁶ The majority of patients harbored intermediate risk nodules, which are described as the most challenging to evaluate,⁷ and of those patients whose nodules turned out to be malignant, 68% were Stage I. The AUC of LungLB was measured at 0.80 for these Stage I malignant patients. Notably, these results only include the CGAC biomarker utilized in the LungLB test, and do not include clinical/radiological features that are commonly incorporated into nodule evaluation tools, which together with multivariate analysis suggest that LungLB is an independent predictor of malignancy.⁶ Additional clinical validity for indeterminate lung nodule evaluation performed by a sublicensee of the LungLB test, SanMed Biotech in China, demonstrated similar test performance.^8,9

In addition to LungLB’s clinically-meaningful validation performance, results from Schneider et al show that LungLB is cost effective when integrated into the current clinical diagnostic pathway for indeterminate lung nodules,¹⁰ as described in current guidelines from the American College of Chest Physicians.¹¹ Incremental Cost-Effectiveness Ratio (ICER) is a key metric used in the publication to demonstrate cost effectiveness. Integration of LungLB leads to improvement in outcomes and results in an ICER that was 25% below the willingness to pay (WTP) threshold commonly considered by US commercial payors, suggesting overall savings when LungLB is priced at $2,300 per test. ICERs remain below WTP thresholds at prices up to $3,647 per test.

II. Support Proposed LCD

LungLife supports MolDX’s efforts to establish uniform coverage criteria for Medicare beneficiaries for this important testing category of risk stratification for indeterminate pulmonary nodules. We agree with MolDX that “whenever possible, [indeterminate pulmonary nodules] in [intermediate risk] patients should be investigated using noninvasive procedures in order to avoid morbidity in patients without cancer,” and that “molecular biomarkers can fulfill this need.” Accordingly, LungLife strongly supports Proposed LCD.

Our understanding is that Proposed LCD is not limited to any particular sample type, methodology, or molecular analyte. As MolDX may be aware, multiple tests are in development that use analytes other than RNA and sample types other than bronchial epithelial cells obtained from bronchoscopy brushings. This includes our own LungLB^® test, which employs DNA FISH on rare cells in blood, and Veracyte’s Percepta Nasal Swab test which utilizes RNA isolated from the nasal epithelium. As there do not appear to be any sample type or analyte limitations in the Proposed LCD, our impression is that MolDX intends this coverage policy to set broad rules for molecular biomarker testing for risk stratification of indeterminate pulmonary nodules. LungLife agrees with this approach, which aligns with MolDX’s longstanding position of establishing foundational LCDs that enable categories of molecular tests to undergo the streamlined Technical Assessment (TA) process rather than individual LCDs.

III. Recommended Modifications

While LungLife strongly supports finalization of proposed LCD, we wish to offer certain specific recommendations to make the Final LCD more suitable for pre-biopsy molecular biomarker testing that is increasingly used by clinicians to determine the appropriateness of any biopsy, rather than simply after a bronchoscopy that did not provide a diagnosis.

Criterion 1: Include coverage of molecular tests for patients under consideration for biopsy by removing requirement that patients undergo bronchoscopy.

Most importantly, LungLife requests the removal of the stringent bronchoscopy requirement to allow coverage for testing furnished to beneficiaries for whom indeterminate pulmonary nodules have been identified (e.g. by CT scan) and who are under consideration for biopsy. This modification would allow MolDX to establish clear coverage criteria for the increasing number of blood- or swab-based molecular biomarker tests (like LungLB) that can be performed prior to biopsy and risk-stratify indeterminate nodules without the need for an invasive procedure.

Establishing coverage criteria for these tests would align with MolDX’s view that “whenever possible, IPNs in these patients [who might be subject to biopsy] should be investigated using noninvasive procedures in order to avoid morbidity in patients without cancer.” It is worth noting that MolDX has already covered a proteomic test (Biodesix Nodify XL2) for indeterminate pulmonary nodules for beneficiaries prior to biopsy, indicating that MolDX agrees that risk stratification of indeterminate pulmonary nodules prior to the decision to biopsy demonstrates clinical utility.

Of course, the removal of the bronchoscopy requirement and the establishment of a pathway to coverage for pre-biopsy indeterminate pulmonary nodule risk stratification tests does not require MolDX to establish coverage for any test for these indications without review via the TA process. It simply will enable MolDX to review these tests for coverage based on parameters for analytical and clinical validity without unnecessary beneficiary access delays or the reconsideration process.

If MolDX disagrees with modifying the LCD to establish coverage criteria for pre-biopsy testing generally, we urge MolDX to at least allow coverage for beneficiaries who are ineligible for or decline an invasive diagnostic procedure due to comorbidities or significant danger related to the procedure (e.g. nodule proximity to the heart). Tanner et al reported that 35% of surgeries for lung nodules were for benign disease, which occurred in the absence of biopsy.⁵ This is in part because clinicians are increasingly faced with managing lung nodules in patients in whom diagnostic biopsy is not safe or feasible,¹² so the additional information from a minimally invasive test would provide critical information to guide clinical decision making for these already vulnerable beneficiaries.

At minimum, MolDX should permit coverage of molecular tests for patients who have undergone a type of non-diagnostic biopsy other than bronchoscopy (such as trans- thoracic biopsy). Transthoracic biopsy is non-diagnostic at least as often as bronchoscopy for indeterminate lung nodules. Transthoracic biopsy demonstrates diagnostic yields between 60-80% depending on nodule size and physician experience.^13,14,15 This means that at least 20-40% of the time, patients are receiving a non-diagnostic transthoracic biopsy. Research from the University of Michigan shows that while the positive predictive value of transthoracic core biopsy is high, the negative predictive value varies between 50-75%,¹⁶ which is described by the authors as being “unreliable to exclude malignancy.” The requirement of a bronchoscopy could lead to the exclusion of coverage for patients who have undergone non-diagnostic biopsy other than bronchoscopy. These patients might then have to obtain another invasive biopsy (a bronchoscopy) before molecular biomarker testing is covered, an outcome that makes little sense in light of MolDX’s stated goal of limiting the use of invasive procedures.

Criterion 2.a and 2.b: Expand eligibility to patients with cancer history and certain forms of current cancer.

Previous and (in some cases) current cancer does not impact the risk that an indeterminate pulmonary nodule is malignant, and thus should not prevent coverage of molecular biomarker testing of such nodules. For instance, there is no statistically significant increase in lung cancer risk following primary non-melanoma skin cancer diagnosis.¹⁷ The risk of lung cancer is potentially reduced following prostate cancer diagnosis.¹⁸ Given that the impact of cancer history on lung nodule evaluation may be unclear for certain malignancies, and current cancer may not be relevant to pulmonary nodule malignancy risk for patients diagnosed with a cancer considered to have low metastatic potential (e.g. brain tumor, basal cell carcinoma, or low-grade prostate cancer), MolDX should avoid a blanket limitation on coverage for molecular biomarker testing of indeterminate pulmonary nodules based on this clinical factor.

This modification should also allow coverage for patients with a history of lung cancer. The best therapy for early-stage lung cancer is surgery or radiotherapy with curative intent.¹⁹ However, recurrence occurs in up to 50% of patients and as such CT-based surveillance monitoring is prescribed every 6 months for 2 years to look for new lung nodules.²⁰ However, new nodules in this context are challenging to evaluate: they could be due to fibrosis and scar tissue formation from radiotherapy or surgery, infections acquired during the surgical procedure, or a recurrent lung cancer.^20,21 Risk stratification testing can thus be a useful tool for these patients.

Note that we are not suggesting this policy be expanded to allow coverage of molecular tests for monitoring, as is done with minimum residual disease tests. This modification would only allow for coverage of a risk stratification test for a patient with a history of lung cancer if a new indeterminate pulmonary nodule is found. Furthermore, the narrowing of the coverage exclusion for previous cancer will not undercut MolDX’s ability to evaluate tests on an individual basis through the TA process.

Criterion 2.d: Permit coverage of molecular testing for patients with high diagnostic risk of a nodule being cancerous.

Tanner et al demonstrate high rates of surgery for benign lung nodules, due to the absence of a pathological diagnosis in patients at high-risk for malignancy.⁵ Clinicians are also increasingly faced with managing lung nodules in patients in whom diagnostic biopsy is not safe or feasible. Additionally, the assignment of pre-test risk of malignancy varies by institution and physician, with some using risk calculators (e.g. Mayo Model) and others relying on physician judgement or intuition.⁷ In other words, the determination that a patient is high risk for a pulmonary malignancy is variable in accuracy and should not eliminate or undercut the ordering physician’s discretion to order molecular biomarker testing for an indeterminate pulmonary nodule if they consider such testing to be medically necessary. Once again, any rules governing coverage based on patient risk level can be made on a test-by-test basis in the TA process.

Criterion 3: Provide that a beneficiary who has received testing for the same clinical indication should be eligible for the same or similar testing in certain circumstances.

LungLife believes that the coverage criterion that the “beneficiary has not been tested with the same or similar assay for the same clinical indication” should not apply in instances of a new nodule of concern, or where an existing nodule has a change in appearance such that risk of malignancy has changed. Test developers should have the opportunity to prove test validity for these clinical scenarios through longitudinal patient monitoring situations using the TA process. We note that in CMS’s NCD 90.2, the agency requires that a patient “not been previously tested with the same test using NGS for the same cancer genetic content.” Testing of a new lung nodule or an existing lung nodule with a significant change in appearance can be considered akin to testing for new cancer genetic content.

Criterion 10: Provide that new tests with improved performance (sensitivity/specificity) will not render older tests no longer compliant with this policy.

Clinical validity encompasses important metrics for understanding the performance of a molecular diagnostic test. However, there are other factors that are also important to consider when understanding if a test is reasonable and necessary. This includes compliance with the testing modality. For example, sputum is actively being explored as a sample source for indeterminate nodule evaluation, but it can be very challenging to collect²² and therefore patient compliance may be lower than if the sample came from a blood draw or nasal swab. Analogous to this, a study in Germany for colorectal cancer screening indicates that compliance correlates with less invasive techniques, with more individuals opting for blood tests over stool-based tests or colonoscopy.²³ Better patient adherence, even with lower performance, may lead to better outcomes for Medicare beneficiaries and test developers should have the opportunity to prove this using the TA process.

Redline of Coverage Criteria: In accordance with the foregoing, we respectfully request that MolDX make revisions to the coverage criteria of the Proposed LCD.

IV. Conclusion

LungLife greatly appreciates MolDX’s consideration of the recommendations listed above. We strongly support the finalization of proposed LCD.

Redline of coverage criteria and references were provided for review.

Thank you for your thoughtful comment. Please see below for point-by-point responses to topics identified above:

Criterion 1: Include coverage of molecular tests for patients under consideration for biopsy by removing requirement that patients undergo bronchoscopy. At minimum, MolDX should permit coverage of molecular tests for patients who have undergone a type of non-diagnostic biopsy other than bronchoscopy (such as trans-thoracic biopsy).

The LCD pertains to molecular biomarkers for risk stratification of indeterminate pulmonary nodules following bronchoscopy that can serve as an adjunct to non-diagnostic bronchoscopy to inform further patient management. We agree that there is a significant need for better tools to risk stratify indeterminate pulmonary nodules prior to bronchoscopy (or as an adjunct to other invasive biopsy procedures) to increase the likelihood of diagnosing lung cancer in its earliest stages, while also avoiding unnecessary invasive procedures in patients with benign lesions. While this expansion of coverage to pre-bronchoscopy risk stratification (or as an adjunct to other types of non-diagnostic biopsies) is beyond the scope of this particular local coverage determination (LCD), we welcome submission of supporting evidence for this indication in the form of a new LCD request. Given recent expansion of the MolDX program to include proteomics, evidence for all molecular pre-biopsy risk stratification tests in scope of the MolDX program is currently under consideration.

Criterion 2.a and 2.b: Expand eligibility to patients with cancer history and certain forms of current cancer.

Test eligibility has been expanded to include patients with cancer history other than lung cancer, provided that all other coverage criteria are met. The clinical validity of any analytes (or expression profiles) measured must be established through a study published in the peer-reviewed literature for the intended use of the test in the intended population, i.e. in patients with a history of cancer. In patients with a history of lung cancer, a new nodule most likely represents distant metastasis from the initial lung cancer or a second primary lung cancer¹ and clinical validity and clinical utility of molecular risk stratification in this population has not been established. While data on pulmonary nodules in patients with cancer is scarce, the likelihood of malignancy (whether metastatic or due to a primary lung cancer) is high² and clinical validity and clinical utility of molecular risk stratification in this population has also not been established.

References

1. Araujo-Filho, JAB, Halpenny D, McQuade C, et al. Management of pulmonary nodules in oncologic patients: AJR expert panel narrative review. AJR Am J Roentgenol. 2021;216(6):1423–1431. doi:10.2214/AJR.20.24907
2. Caparica R, Mak MP, Rocha CH, et al. Pulmonary nodules in patients with nonpulmonary cancer: not always metastases. J Glob Oncol. 2016;2(3):138-144. doi: 10.1200/JGO.2015.002089.

Criterion 2.d: Permit coverage of molecular testing for patients with high diagnostic risk of a nodule being cancerous.

Clinical utility of a given molecular risk stratification test lies in its ability to refine patient risk of malignancy to change management (i.e. reduce invasive procedures) within the framework of national consensus guidelines, in an effort to improve patient outcomes. Therefore, patients whose pretest risk of malignancy is too low or too high to result in meaningful changes in management that result in improved patient outcomes are not eligible for testing due to lack of clinical utility.

Criterion 3: Provide that a beneficiary who has received testing for the same clinical indication should be eligible for the same or similar testing in certain circumstances.

Each clinical indication is considered separately and repeat testing with the same or similar test for the same clinical indication is considered duplicative. Serial monitoring of pulmonary nodules over time (i.e. sampling the same nodule in the setting of radiographic changes) is not in scope of this policy and has not at this time demonstrated analytical validity, clinical validity, and clinical utility.

Criterion 10: Provide that new tests with improved performance (sensitivity/specificity) will not render older tests no longer compliant with this policy.

Each test must meet policy criteria and exhibit analytical validity, clinical validity, and clinical utility for a given intended use in accord with the evolving standard of care. Medicare Beneficiaries deserve access to tests with the best performance and this standard is expected to evolve over time due to scientific progress. In order to secure access to the best performing tests for Medicare Beneficiaries, tests with significantly inferior performance will not be considered reasonable and necessary as the standard of care evolves.

The following comment was submitted to Palmetto GBA:

Veracyte, Inc. is writing to comment in support of Palmetto GBA’s Proposed Local Coverage Determination (LCD) on Molecular Biomarkers for Risk Stratification of Indeterminate Pulmonary Nodules Following Bronchoscopy, and to provide the following recommendations and clarifications. As a leading genomic diagnostic company, Veracyte has been a pioneer of multiple advanced molecular diagnostic tests. Two of these include Percepta GSC, which provides a risk stratification of indeterminate pulmonary nodules (IPN) after a non- diagnostic bronchoscopy, and Percepta Nasal Swab, which offers risk stratification of IPN prior to a biopsy procedure such as a bronchoscopy. Percepta GSC has been covered by Medicare since March 2017.

Veracyte strongly supports the Proposed LCD, under which MolDX proposes to maintain existing coverage of Percepta GSC. We appreciate the detailed evidence review undertaken by MolDX in the development of this Proposed LCD, and we urge MolDX to finalize the Proposed LCD with suggested edits. We offer the following four additional comments:

1. Recommend removal of the restriction of post-bronchoscopy testing.
2. Recommend removal of pre-test risk as a basis for exclusion from coverage.
3. Recommend modification to the prior cancer exclusion criteria.
4. Request clarification that the eventual LCD will not restrict the current scope of Percepta GSC coverage.

Recommended Modifications of Proposed LCD

While Veracyte strongly supports the Proposed LCD and urges its finalization, based on our review of the Proposed LCD, we recommend the following modifications to ensure a rigorous and flexible coverage policy.

1. Request removal of coverage restriction to post-bronchoscopy testing to ensure consistent coverage of molecular biomarker tests for risk stratification of indeterminate lung nodules

We commend MolDX’s thorough consideration of the evidence for covering advanced diagnostic tests in the context of risk stratification of IPNs. However, we urge MolDX to amend this Proposed LCD to avoid restricting coverage to testing following bronchoscopy, and instead update the policy to allow for coverage of testing performed prior to any type of invasive biopsy, including bronchoscopy (hereinafter referred to as “pre-biopsy testing”).

We believe this modification is critical to ensuring that a final LCD provides a single “foundational” policy that can apply both to post-bronchoscopy and pre-biopsy testing. The Proposed LCD repeatedly alludes to the desirability of pre-biopsy testing, stating that “additional improvements in risk stratification beyond the available clinical and radiological features is needed to decrease unnecessary biopsy referrals and costs” (emphasis added). MolDX also cites the challenge and often poor performance of taking biopsies of lung nodules through procedures such as bronchoscopy and states that “whenever possible, IPNs in these patients [who might be subject to biopsy] should be investigated using noninvasive procedures in order to avoid morbidity in patients without cancer” (emphasis added).

Clinical data support this position. Tanner et al. published a multi-center observational study of patients with an IPN, who were referred to a pulmonologist for further workup. More than half of patients underwent an invasive procedure despite a prevalence of malignancy of only 25% in this cohort. Specifically, 33% of patients underwent a tissue biopsy and over 20% underwent surgery, and more than a third of patients undergoing surgery had benign disease. This study concluded that “[d]espite advances in imaging and nonsurgical biopsy techniques, invasive sampling of low-risk nodules and surgical resection of benign nodules remain common.”⁹

Veracyte wholeheartedly agrees with the conclusions of the Tanner study and MolDX’s view that there is a clinical need for better risk stratification of IPNs both after a non-diagnostic bronchoscopy and prior to an invasive biopsy of any sort. In accordance with this articulated position and the clinical evidence, MolDX should ensure that its coverage policy for IPN risk stratification establishes coverage criteria for both the pre-biopsy and post-bronchoscopy scenarios.

We note that this modification would not necessitate a categorical expansion of Medicare coverage. Palmetto GBA currently covers IPN risk stratification tests for both pre-biopsy and post-bronchoscopy indications. Specifically, Palmetto GBA has two current test-specific LCDs for risk stratification of IPNs. The first one, “MolDX: Percepta Bronchial Genomic Classifier”, covers our Percepta GSC for its intended use following a non-diagnostic bronchoscopy. The other LCD, “MolDX: BDX-XL2”, applies to the Biodesix Nodify XL2 test, which is used prior to biopsy. In other words, MolDX has already determined that risk stratification tests can be reasonable and necessary for both indications.

MolDX based these two LCDs on evidence of clinical utility for IPN risk stratification both pre- biopsy and post-bronchoscopy. The post-bronchoscopy clinical utility evidence for Percepta is recounted in the Proposed LCD. However, MolDX also conducted a review of the clinical utility evidence for the Biodesix Nodify XL2 test for pre-biopsy patients when establishing that this test was reasonable and necessary. Additionally, new clinical utility evidence has recently been published that adds to the body of evidence since MolDX last conducted a review of these tests. For example, the ORACLE study, which was a prospective, multicenter, observational study, found that the Nodify XL2 test for “patients with a newly discovered PN [pulmonary nodule] has demonstrated valuable clinical utility in a real-world setting. Use of this biomarker can change physicians’ practice and reduce invasive procedures in patients with benign pulmonary nodules.”²

Although we acknowledge that the Nodify XL2 test is protein-based, it is a lung nodule biomarker test that provides a quantitative risk of malignancy (ROM) assessment for IPN prior to any bronchoscopy or other invasive biopsy. Regardless of the underlying technology or platform that provides the ROM, the clinical utility of using an objective risk stratification test prior to biopsy should not vary based on the type of biomarker. MolDX should be able to use the clinical utility evidence for the Nodify XL2 test to establish a foundational LCD that provides a path to coverage IPN risk stratification tests furnished prior to biopsy.

Finally, the establishment of coverage criteria for pre-biopsy tests is important given the increasing number of available IPN risk stratification tests that are principally used for pre- biopsy indications. For instance, there are a number of increasingly well-validated testing technologies that rely on blood-based biomarkers^1,2,3,4 or assess changes in RNA expression from nasal epithelial cells^4,5,6,7 that provide risk stratification for IPNs. Some of these tests are already covered by Medicare. Given the existing evidence for pre-biopsy risk stratification tests like Percepta Nasal Swab and others, and the continuing development of further supporting evidence, as well as the longstanding coverage policy for Biodesix Nodify XL2, it behooves MolDX to avoid beneficiary access delays and unnecessary duplication of work in the future by establishing coverage criteria for pre-biopsy tests in this LCD rather than waiting for a future reconsideration.

For these reasons, Veracyte wishes to make the following modifications to coverage criterion (1) to remove the wholesale exclusion of testing that is not furnished post-bronchoscopy. The inclusion of a CT scan requirement ensures that a gatekeeping function remains, while the new criterion (1)(a) maintains coverage standards for post-bronchoscopy testing. Of course, pre- biopsy and post-bronchoscopy tests would remain subject to the other coverage criteria of the LCD.

This contractor will provide limited coverage for molecular tests to aid in the diagnosis or exclusion of lung cancer in a patient with an indeterminate pulmonary nodule (IPN) when ALL of the following conditions are met:

1. The beneficiary has undergone a CT scan where an indeterminate pulmonary nodule was detected, and test results will be used to meaningfully inform patient management within the framework of nationally recognized consensus guidelines.
   1. In the event of testing used post bronchoscopy, the bronchoscopy must have failed to provide a specific histopathological diagnosis such that further diagnostic procedures are considered necessary to pursue a specific diagnosis (non-diagnostic bronchoscopy)

We note that Veracyte is not requesting that MolDX provide explicit coverage for any particular tests that are not currently covered; rather, that MolDX use this opportunity to issue a robust “foundational” LCD for risk stratification of IPN that would provide a consistent path to coverage of future tests within all the indications that have already been deemed reasonable and necessary under LCDs. Veracyte is very supportive of MolDX’s robust Technical Assessment (TA) process as a means to evaluate test-specific clinical validity and utility and believes the variations of test-specific indications should be evaluated through the TA process, allowing this Proposed LCD to be a more impactful foundational LCD that is not limited by narrow indications.

2. Request removal of pre-test risk as a criterion for coverage

Coverage criteria 2(c)-(d) of the Proposed LCD currently state that a test is not covered if the beneficiary has an “overall low/high risk for pulmonary malignancy such that test results would not meaningfully alter patient management and significantly improve patient outcomes.” Veracyte is concerned that these criteria create potential confusion as to the standards for coverage. For example, it appears from this language that if a test result could meaningfully alter patient care and significantly improve outcomes, it would be covered even in patients with an overall low or high pre-test risk of malignancy. However, the Proposed LCD provides no further information as to how this determination would be made.

More generally, the use of pre-test risk as a coverage criterion presents challenges because of the heterogeneity of how pre-test risk is evaluated and defined. As MolDX aptly referenced in the Proposed LCD, pre-test risk, as defined by risk calculators, has significant variability and should be taken with caution: “Risk calculator performance varies depending on the lung cancer prevalence and clinical characteristics of the original study population. Therefore, risk estimates should be interpreted with caution.”

Further complicating the use of pre-test risk as a coverage criterion is the fact that each test in this space has been developed and validated based on different indications. For example, the Biodesix Nodify XL2 test is validated and indicated for IPNs that have a pre-test risk of malignancy that is less than 50%, whereas a test such as Percepta Nasal Swab does not have a specific indication for pre-test risk and instead is indicated based on clinical factors such as nodule size, smoking history, etc. Given these differences, it becomes difficult to understand how a pre-test risk criterion such as the one articulated in criteria 2(c)-(d) can be applied in an objective and consistent manner.

For the foregoing reasons, we respectfully request the removal of Proposed LCD coverage criteria 2(c)-(d). Veracyte believes that the coverage criteria for pre-test risk would better be addressed on an individual test level during the Technical Assessment process.

3. Recommend modification to the prior cancer exclusion criteria so that it is specific only to prior lung cancers

Coverage criterion 2(a) of the Proposed LCD would exclude coverage for any Medicare beneficiary with a “personal history of cancer.” This exclusion criteria appears to be based on the original validation data of a single test, Percepta BGC (1^st generation), which excluded patients with a prior history of cancer from the validation study. However, subsequent risk stratification biomarker tests, such as Biodesix Nodify XL2 and Percepta Nasal Swab, have been validated in patients with certain cancer histories.

Thus, there is evidence that indicates that certain patients with a prior history of cancer other than lung cancer could still benefit from a biomarker that provides risk stratification for primary lung cancer for IPNs. Non-melanomatous skin cancers rarely metastasize to the lungs. Early stage breast, colon, and prostate cancers treated with curative resection are unlikely to subsequently recur with isolated metastasis to the lungs. In a patient with a new IPN, the treating physician should be able to weigh the risks of a new primary lung cancer against the risk imparted by any prior history of non-lung cancer rather than be restricted in the ability to obtain information from IPN risk stratification testing. As such, consistent with our previous request to remove pre-test risk as a coverage criterion, Veracyte believes the TA process may be a more appropriate means to evaluate whether coverage of individual tests should be limited based on beneficiaries’ prior history of cancer.

We do believe it is still prudent to exclude coverage for patients with a previous history of lung cancer since patients with a previous history of lung cancer that develop a new nodule are at such high risk of the new nodule being malignant that a risk stratification test would have little value. Therefore, we propose the following language for coverage criteria 2(a):

1. The beneficiary does NOT have any of the following:
   1. Personal history of lung cancer

4. Request clarification that the Proposed LCD will not restrict the existing coverage afforded to Percepta GSC

In addition to these general recommendations for the improvement of the coverage criteria, Veracyte wishes to confirm that the coverage language in the Proposed LCD will continue to enable coverage for Percepta GSC equivalent to its current coverage.

First, Percepta GSC is currently covered for physician-assessed low and intermediate risk of malignancy but only if “results will be utilized to determine whether CT surveillance is appropriate in lieu of further invasive biopsies or surgical procedures.” As discussed above, we believe that the most appropriate approach for MolDX would be to remove coverage criteria 2(c)-(d) to avoid potential confusion and inconsistency. Even if these criteria are retained, based on MolDX’s previous evidence review, coverage of Percepta GSC for low- risk nodules should fall within the exception for testing that would “meaningfully alter patient management and significantly improve patient outcomes.”

Specifically, as demonstrated in the evidence review, a finding of “very low-risk” by Percepta has a negative predictive value that exceeds 99%; therefore, patients with this result can safely avoid further unnecessary invasive procedures and confidently be managed with serial CT surveillance.⁸ Accordingly, it is our understanding that existing coverage of Percepta for low- risk nodules would not be disturbed by the finalization.

Second, we respectfully request that coverage criterion 1(c) is removed consistent with current coverage of Percepta GSC, which allows physicians to make the clinical determination as to whether Percepta GSC or another invasive biopsy procedure (e.g. EBUS or FNA) is appropriate after indeterminate bronchoscopy. The current proposed criterion 1(c) could be misinterpreted to imply that EBUS or FNA may be required to be completed prior to coverage for Percepta GSC. While these are biopsy methods that are often completed during the same bronchoscopy as the one in which the Percepta GSC specimen is collected, the treating physician should be allowed to choose the tools that are available at their institutions and appropriate for each individual patient. For example, many pulmonologists do not have access to or are not trained to perform EBUS. Allowing treating physicians to select Percepta GSC rather than a second invasive biopsy would align with MolDX’s objective of limiting invasive procedures where clinically appropriate.

Conclusion

Veracyte strongly supports finalizing this Proposed LCD with the four modifications summarized below:

• Removal of coverage restriction to post-bronchoscopy testing
• Removal of pre-test risk as criterion for coverage
• Modification to the prior cancer exclusion criteria
• Clarification that the finalized LCD will not restrict existing coverage afforded to Percepta GSC

Redline and references were provided for review.

Thank you for your thoughtful comment. Please see below for point-by-point responses to the topics identified above.

1. Recommend removal of the restriction of post-bronchoscopy testing.

Thank you. This has been addressed in response to a prior comment.

2. Recommend removal of pre-test risk as a basis for exclusion from coverage.

ACCP Practice Guidelines recommend that clinicians estimate the pretest probability of malignancy either qualitatively by using their clinical judgement and/or quantitatively by using a validated model. Pretest probability of malignancy guides selection and subsequent interpretation of diagnostic tests and is an integral part of patient management. Results of molecular risk stratification tests are often interpreted in the context of pre-test probability of malignancy and the clinical utility of risk stratification is tied to refinement of the pre-test probability. Management changes based on the post-test probability of malignancy are expected to result in improved patient outcomes. While a specific risk stratification method is not endorsed, pre-test risk stratification cannot be removed as a requirement for coverage.

3. Recommend modification to the prior cancer exclusion criteria.

Test eligibility has been expanded to patients with cancer history other than lung cancer, due to its high risk of recurrence, provided that all other coverage criteria are met.

4. Request clarification that the eventual LCD will not restrict the current scope of Percepta GSC coverage.

All tests must meet LCD coverage criteria and undergo Technical Assessment by MolDX. Molecular biomarkers must demonstrate clear clinical utility of re-stratification beyond currently available clinical and radiologic strategies, such that the post-test risk of malignancy results in impactful changes in patient management with improved outcomes within the framework of national consensus guidelines for patient management. Patients whose pretest risk of malignancy is too low or too high to merit meaningful changes in management that result in improved outcomes are not eligible for testing. The benefit of testing patients with a low pre-test risk of malignancy for whom CT surveillance is the recommended follow-up strategy according to practice guidelines is unclear, as re-stratification to a very low risk category leads to the same recommended intervention (i.e. CT surveillance). The clinical utility of risk re-stratification must be clearly defined and demonstrated in the peer-reviewed published literature.