Local Coverage Article Response to Comments

Response to Comments: MolDX: Guardant360® Plasma-Based Comprehensive Genomic Profiling in Non-Small Cell Lung Cancer (NSCLC)

A56107

Expand All | Collapse All

Contractor Information

Article Information

General Information

Article ID
A56107
Article Title
Response to Comments: MolDX: Guardant360® Plasma-Based Comprehensive Genomic Profiling in Non-Small Cell Lung Cancer (NSCLC)
Article Type
Response to Comments
Original Effective Date
10/16/2018
Retirement Date
N/A
AMA CPT / ADA CDT / AHA NUBC Copyright Statement

CPT codes, descriptions and other data only are copyright 2020 American Medical Association. All Rights Reserved. Applicable FARS/HHSARS apply.

Fee schedules, relative value units, conversion factors and/or related components are not assigned by the AMA, are not part of CPT, and the AMA is not recommending their use. The AMA does not directly or indirectly practice medicine or dispense medical services. The AMA assumes no liability for data contained or not contained herein.

Current Dental Terminology © 2020 American Dental Association. All rights reserved.

Copyright © 2013 - 2021, the American Hospital Association, Chicago, Illinois. Reproduced by CMS with permission. No portion of the American Hospital Association (AHA) copyrighted materials contained within this publication may be copied without the express written consent of the AHA. AHA copyrighted materials including the UB-04 codes and descriptions may not be removed, copied, or utilized within any software, product, service, solution or derivative work without the written consent of the AHA. If an entity wishes to utilize any AHA materials, please contact the AHA at 312-893-6816. Making copies or utilizing the content of the UB-04 Manual, including the codes and/or descriptions, for internal purposes, resale and/or to be used in any product or publication; creating any modified or derivative work of the UB-04 Manual and/or codes and descriptions; and/or making any commercial use of UB-04 Manual or any portion thereof, including the codes and/or descriptions, is only authorized with an express license from the American Hospital Association. To license the electronic data file of UB-04 Data Specifications, contact Tim Carlson at (312) 893-6816. You may also contact us at ub04@aha.org.

Article Guidance

Article Text

The comments in this document reflect all the comments sent to all the MACs in the MolDx project for the Draft Local Determination (LCD) MolDX: Guardant360® Plasma-Based Comprehensive Genomic Profiling in Non-Small Cell Lung Cancer (NSCLC).

Response To Comments

NumberCommentResponse
1

We frequently use Guardant360 in our non-small cell lung cancer (NSCLC) patients at the University of California San Francisco Helen Diller Comprehensive Cancer Center. This liquid biopsy test provides us with more information than other liquid biopsy tests because it includes all four classes of genomic alterations and all seven targetable genes in NSCLC. I use it very frequently both to prevent repeat invasive biopsies in the first line when tissue is insufficient and for complete genotyping or at progression to look for targetable acquired resistance mutations. I have several patients for whom the Guardant 360 test made the difference between starting targeted therapy and significant morbidity and likely death as their tissue biopsies were inadequate for next generation sequencing and they were too sick to wait for an additional biopsy and testing.

Thank you for the comments.

2

I have personally ordered over 100 Guardant360 tests to treat patients with lung cancer, over half of which have an ICD-10 code of C34.00, C34.10, C34.30, C34.80 or C34.90-C34.92. In advanced NSCLC, patients often present with multiple lesions and the exact location of the primary tumor is often not clear. Furthermore, I do not differentially treat lung cancer that was diagnosed in the upper lobe of the right lung vs. the lower lobe of the left lung. I treat lung cancer on the basis of the pathological diagnosis, the somatic mutations that are present in the tumor and various other traits that are not related to the exact location of the tumor in the lung.

It is concerning to me that the draft LCD does not cover the most commonly used and medically appropriate diagnosis codes for advanced cancer. Initial diagnosis of patients is made with the goal of treating them appropriately in a rapid manner. Please update the LCD to cover the entire ICD-10 code set from C33-C34.92 for advanced lung cancer.

Guardant360 has helped me treat dozens of patients with NSCLC and other disease states. I look forward to being able to tell my NSCLC Medicare patients who have insufficient tissue to perform CGP patients that Guardant360 is covered by Medicare.

Thank you for your comments. A number of commenters recommended we add this code set to the acceptable ICD10 codes and we have done so.

3

As a leading patient advocacy group that represents the voice and interest of the national lung cancer survivor community by accelerating research to patients that is meaningful to them, empowering patients to be active participants in their care and care decisions and helping remove barriers to access to high quality care, LUNGevity applauds MolDX for providing a coverage determination for the Guardant360 test and ensuring new testing options are available for lung cancer patients. In this era of unprecedented scientific advancements for the treatment of lung cancer, particularly in the field of biomarker testing, liquid biopsy tests, like Guardant360’s, are a promising new development that identify markers predictive of response to particular treatments for patients in a convenient, low cost, and quickly-responsive manner.

Non-small cell lung cancer (NSCLC) is the more common type of lung cancer, diagnosed in about 85% of people with lung cancer. The complex nature of this disease requires personalized management plans for patients.2 Since the discovery of the first epidermal growth factor receptor (EGFR) mutation in lung cancer in 2004, targeted therapies have become a major component of the treatment arsenal of NSCLC patient. Now at least 10 driver mutations in adenocarcinoma have been identified (EGFR, ALK, ROS, RET, ERB2/HER2 mutations, ERB2/HER2 amplifications, MET amplifications, MET mutations, TRK, BRAF, KRAS). In concert with the identification of an increasing number of targetable mutations is the development of novel, potent, and more specific targeted therapies. For example, at present, third generation EGFR tyrosine kinase inhibitors (TKIs) and second generation anaplastic lymphoma kinase (ALK) TKIs are used in clinical practice. With the increased use of targeted agents has come the problem of acquired resistance, where cancer cells inevitably develop resistance to the targeted agent. The EGFR T790M is an excellent example of a resistance mutation that develops in patients treated with first- and second-generation EGFR TKIs. This mutation can be rapidly detected using a liquid biopsy test such as the cobas EGFR Mutation Test v2. Lung cancer is now leading the field of precision medicine where research is rapidly progressing to (1) develop better targeted therapies that combat mechanisms of resistance, and (2) noninvasive assays – such as liquid biopsies – that can monitor status of the resistance mutations (e.g., cobas EGFR Mutation Test v2), sequentially and in real time.

The utility of liquid biopsies in the clinical management of lung cancer is unquestionable, because 1 out of 4 NCSLC patients may be ineligible for a solid tissue biopsy. In her ASCO 2017 presentation on biomarker testing for lung cancer, LUNGevity Scientific Advisory Board (SAB) member, Dr. Alice Shaw from Massachusetts General Hospital, pointed out that liquid biopsies may help in (1) initial detection of targetable mutations in advanced-stage NSCLC at the time of diagnosis, (2) identification of acquired resistance mutations in patients who have relapsed on targeted therapies, and (3) monitoring response to targeted therapies and predicting outcome in advanced-stage NSCLC patient.

Given the utility of liquid biopsy and monitoring importance, we request that you reconsider the “at progression section” of the coverage guidance to include access for all eligible advanced-stage NSCLC patient at progression rather than limiting it to select mutations. Treatment approaches of lung cancer is rapidly evolving, with third-generation tyrosine kinase inhibitors such as osimertinib, first approved in the post-progression setting, moving to the first-line setting for the treatment of EGFR-positive adenocarcinoma. The use of osimertinib in the first-line setting (FLAURA trial) offers a far superior median progression-free survival of 18.9 months versus 10.2 months median PFS offered by first- and second-generation EGFR TKIs. With this progress has come the need to understand mechanisms of resistance to osimertinib in the first-line setting. In the FLAURA trial, mechanisms of resistance observed in nine patients studied includes a variety of genomic alterations (such as MET amplifications, PIK3CA mutations, or C797S mutations, for example) in the absence of an acquired T790M mutation. Despite the small sample size, this provocative data suggests that detection of resistance mutations such as PIK3CA or EGFR C797S in patients who have progressed on first-line osimertinib, using non-invasive approaches, may help determine second-line treatment options. Currently, drugs targeting MET amplification or PIK3CA are in clinical development and there is evidence suggesting that EGFR C797S is sensitive to first-generation EGFR inhibitors such as erlotinib. Using a non-invasive test at the time of progression would not only be beneficial to the patient but also expedite the selection of second-line treatment options.

Last but not least, we request that you review the CPT/HCPC S Codes section of the proposed LCD and consider inclusion of the new ICD 10 code C34.90 (malignant neoplasm of unspecified part of unspecified bronchus or lung) that was created in 2017 in place of the ICD-9-CM 162.9 code. The ability to use this code will be extremely important for patient access, as often times the specific location of the originating tumor is unknown at time of biopsy.

As a leading patient advocacy group that represents the voice and interest of the national lung cancer survivor community, we are excited about the role of liquid biopsies in clinical management of NSCLC.

Thank you for your comments. We appreciate the suggestion (supported by evidence) to include “access for all eligible advanced-stage NSCLC patient at progression” and we have incorporated that concept into the policy. The cited ICD 10 codes were also added as we received that comment from others and it is appropriate.

4

In routine clinical use I use the Guardant360 in a manner consistent with the proposed LCD. Overall, I feel the indications for use are reasonable and clinically appropriate. At MD Anderson we feel that it is appropriate to test all patients who progress on targeted therapy to determine if there are resistance mechanisms to support moving them to a different therapy. As an example, we recently reported data from EGFR T790M positive patients treated with osimertinib, many of whom were initially identified as T790M positive using Guardant 360; at the time progression, many of these patients were again profiled using Guardant 360, and we identified a number of targetable alterations including MET amplifications and ROS1 fusions (several of which were responsive to the MET inhibitor crizotinib) and EML4-ALK fusion (responsive to alectinib). The initial analysis of this cohort was presented in an oral presentation at AACR this April (Le et al, Proc AACR, 2018). In the research setting, we have used it for eligibilty for our study of a new drug that inhibits EGFR and HER2 exon 20 mutations, which demonstrated >60% objective response rate in initial testing (Robichaux et al, Nat Med 2018).

It appears that the lack of coverage for all ICD codes related to NSCLC is a major gap in the coverage. We estimate that in excess of 1,000 patients from MDACC have been treated with Guardant360 with an appropriate diagnosis code assigned to them who would not be covered per the draft LCD. The ICD codes do not directly align with how we diagnose lung cancer; for example, many patients present with widespread metastases and the initial primary tumor is unclear. It is odd that the C34.90-C34.92 codes aren’t covered for advanced cancers considering that a study published in the New England Journal of Medicine1 demonstrates that more than 50% of patients are diagnosed with NSCLC at stage IV.

I believe that it is a major step forward that Medicare is advancing personalized medicine by beginning to cover tests like Guardant360 but creating coverage policies that partially cover the diagnosis set is puzzling and creates even further confusion about coverage in the already challenging space of molecular diagnostics. I would encourage you to cover the entire ICD code set between C33-C34.92 to ensure patients have the appropriate access to care.

Thank you for your comments. Based on your comments re: all patients progressing on targeted therapy, this has been added to the LCD. The ICD10 code set C33.00 to C34.92 will be added as appropriate diagnosis codes for eligible patients.

5

I and my colleagues in medical oncology at UCSD Moores Cancer Center frequently use Guardant360 comprehensive liquid biopsy in our patients with advanced non-small cell lung cancer (NSCLC) as well as refractory metastatic tumors. This liquid biopsy test provides us with all four classes of genomic alterations and all seven targetable genes in NSCLC. We use it to prevent repeat invasive biopsies in the first line when tissue is insufficient for complete genotyping and at progression to look for targetable acquired resistance mutations (EGFR T790M for example). We thank Noridian for the opportunity to review comments.

Thank your comments.

6

We evaluated several commercially available and inhouse-developed methodologies before bringing the Guardant Health technology into our laboratory. After extensive evaluation of analytic validity, clinical validity and available published and inhouse evidence of clinical utility following the Center for Disease Control and Prevention Office of Public Health Genomics model (ACCE), we brought the assay to our Molecular Testing Evaluation Committee (MTEC) chaired by our Vice President for Clinical Research.

The MTEC is the institutional Committee for vetting molecular tests and is composed of personnel across the spectrum of oncologists and other physicians, Legal Services, Billing Compliance, and Patient Billing Services (PBS) as well as clinical laboratory personnel. The MTEC hears presentations requesting new molecular tests to be added to our clinical labs’ rosters of services for requests for reimbursement and has voting authority based upon the levels of evidence. MTEC maintains a file for each proceeding and provides the information to PBS for discussion of coverage decisions with payers. This process is far from a rubber stamp, as it denies about a third of requests. The molecular coverage table on the website of Novitas, our MAC, is derived in large part from our PBS and MTEC collaboration with the company. The MTEC has voted approval of usage and request for reimbursement of Guardant Health testing for specific indications in our patient population, including NSCLC patients.

We are now actively providing results of our Guardant360 and our derived LBP70 testing in our EHR. We are the only laboratory in the world, beside Guardant Health itself, to provide Guardant Health testing to our patients for our oncologists. As you can imagine, the costs to our institution to build this assay capability were substantial, but we made this expenditure for the following reasons:

  1. Many patients cannot be adequately genotyped for the targetable genes in the NCCN and the newly updated CAP/AMP/IASLC molecular testing guidelines for NSCLC. Tissue is the issue, as the tumor may be in an anatomical location out of reach or of insufficient size for interventional radiology to obtain, biopsy acquisition poses risks of pneumothorax and hemorrhage, when obtained the quantity of tumor may be insufficient, or a repeat biopsy to get additional tissue may be contraindicated. Our thoracic oncologists needed a non-invasive solution for these patients, which may comprise a quarter of the advanced NSCLC patients they see.
  2. At progression, advanced lung cancers (and other tumors) acquire additional mutations that are now targetable, such as EGFR T790M or MET gene amplification, with FDA-approved targeted drugs. Because of inter- and intratumor heterogeneity, biopsy may miss the emerging resistant subclones.
  3. Guardant360 was, and our now in-house derived LBP70 test is, used as standard of care in the above indications by our thoracic oncologists. By assuring that patients can have complete genotyping for all seven targetable genes we:
    1. Get more patients on targetable therapy vs. chemotherapy, with reported 2-3 fold higher response rates.
    2. Avoid use of immunotherapy, which is not usually efficacious in patients with alterations in EGFR, ALK, ROS1, BRAF, and STK11, who comprise a sizeable number of patients who would otherwise be tried empirically on an immune checkpoint inhibitor.
    3. Avoid repeating the risks of invasive biopsies, as described above.
    4. Accrue more patients onto clinical trials, such as the newer targeted therapies that may double survival over the older generation targeted therapies, e.g. osimertinib (for EGFR driver mutations) and alectinib (for ALK fusions).

For a liquid biopsy to effectively prevent repeat invasive biopsies in NSCLC, it must be comprehensive. First of all, a “hotspot” liquid biopsy that looks for only one gene will be negative most of the time and force the oncologist to order a repeat tissue biopsy or additional single-gene liquid biopsies. With “comprehensive” liquid biopsy tests like Guardant360 and our LBP70, all 20 oncogenes that drive lung cancer are included, and if any one of them (even if it is not druggable, like KRAS) is found, a repeat tissue biopsy is obviated. Lung cancers are almost never driven by more than one of these 20 oncogenic driver genes, so finding any one of them ends the diagnostic workup.

Secondly, “comprehensive” also means being able to target with liquid biopsy all four types of genomic alterations in one or more of the seven guideline-recommended genes (point mutations/single nucleotide variants, indels, fusions, and gene copy number amplifications) that are targetable. There are additional genes included in the assay that are useful as markers of the continued presence of or recurrence of tumor, such as TP53. In addition, there are genes that are not in the guidelines yet, but which have shown dramatic improvements in overall survival in phase II/III targeted therapy trials, such as NTRK, that we need to provide to patients so that we can enroll them in the final clinical trials on the path to FDA approval of the remarkable targeted drug.

Please do not hesitate to contact me if you have any questions. As I hope is apparent, coverage of liquid biopsies is greatly needed, as this technology, especially the Guardant Health version, allows patients to avoid unnecessary repeat invasive tissue biopsies, or even more importantly, provides access to efficacious drugs and avoids empirical “next-up” use of expensive immunotherapy and other therapies that are unlikely to work.

Thank you for your comments.

7

This policy provides limited coverage for Guardant360®, a plasma-based comprehensive somatic genomic profiling test for patients with Stage IIIB/IV non-small cell lung cancer (NSCLC) either at diagnosis, when results for EGFR single nucleotide variants (SNV) and insertions and deletions (indels); ALK and ROS1 rearrangements; and PD-L1 expression are not available AND when tissue-based CGP is infeasible, or at progression for patients progressing on, or after, chemotherapy or immunotherapy who have never been tested for EGFR SNVs and indels, and ALK and ROS1 rearrangements, and for whom tissue-based comprehensive genomic profiling is infeasible, or for patients progressing on EGFR tyrosine kinase inhibitors other than Osimertinib.

This policy is written as specifically providing limited coverage for the Guardant360® assay, but only references the use of one CPT code 81479 (unlisted molecular pathology). This code allows each MAC to determine the reimbursement for the test instead of reimbursing at the national limitation amount. It is more appropriate to use a genomic sequencing code such as CPT code 81445 in order to prevent arbitrary reimbursement that varies by MAC.

AMP and CAP also urge you to address the limited indications included in this policy. Testing for acquisition of ALK inhibitor (crizotinib) resistance should also be medically necessary since second line therapies are already present (Costa DB et al).

This policy as written limits coverage to the Guardant360® assay. In its current form, this policy would contradict this long-standing policy in favor of providing coverage for one specific assay and excluding entire categories of testing with a long history of being utilized successfully in CLIA-certified laboratories. The use of these tests are often supported by well-established clinical guidelines that have been developed and endorsed by leading scientists, subject matter experts, and professional societies (NCCN Clinical Practice Guidelines). Medicare must cover any and all tests that direct patient care and are deemed clinically relevant and useful by the team of treating physicians that meet CMS’ medically reasonable and necessary standard.

The policy states that Guardant will collect data on use of its test through independent registries, and that continued coverage will depend on annual review by the contractor of such data, including documentation of new sites of disease and organs with new sites of disease. This process is reminiscent of coverage with evidence development (CED). CED requirements are unnecessarily restrictive, due to the incomplete data on patients available to laboratories that is instead logged by treating oncologists. The same concern regarding data collection and availability for laboratories within CED is present when dealing with the data available to contractors from independent registries. Laboratories may not be involved in the documentation process of new sites and disease, but continued coverage for this test will hinge on this data. Medicare contractors are not authorized to use CED in their determinations and, as the process described above closely mirrors CED, implementing it as written would be at odds with established LCD procedures.

Thank you for the comments. We appreciate the CPT coding comments and we can understand how one could crosswalk the test to a CPT code 81455, “targeted genomic sequence analysis panel, solid organ neoplasm…5-50 genes…interrogation for sequence variants and copy number variants….” However, the test is not a simple crosswalk. The test is not performed on tissue (solid organ) and not a simple targeted assay.
We also appreciate the comments regarding the limitations listed in the policy. We have adjusted those limitations in the final policy and we appreciate the comment.

The comment that the policy limits (liquid biopsy) coverage to this assay (Guardant360) is noted and it reflects our intent to limit coverage to tests that have passed our AV, CV, and CU analysis. This policy does not preclude coverage to other approved test platforms (e.g., Cobas Mutation V2) which we also cover. We do not accept the assertion that “Medicare must cover any and all tests that direct patient care and are deemed clinically relevant and useful.” Our directive from CMS (Decision Memo for Next Generation Sequencing (NGS) for Medicare Beneficiaries with Advanced Cancer (CAG-00450N)) is to perform a vigorous analysis of the AV, CV, and CU of all tests in the NGS oncology space and this test has passed our analysis.

We do agree with the comments on independent registries and the possible confusion with “CED.” We are not requiring the test sponsor to provide us with additional data. Rather, this test, like all approved tests, will be monitored by this MAC to be certain that it performs well for the benefit of our beneficiaries and continuing coverage will depend upon ongoing evidence assessment.

8

As far as this assay itself, I think limited coverage is necessary because there are, perhaps, 10 percent of the biopsies that I'm asked to do mutation gone and there is nothing left in the (block) or there is not enough tumors in order to do the test. So, it is a real problem that needs assistance.
There is an NCD about comprehensive genomic profiling that would not cover any NGS assay that’s not FDA approved for determination of metastatic mutations. Do you interpret that would affect this LCD?

Thank you for the comments. On March 16, 2018, CMS announced the Decision Memo for Next Generation Sequencing (NGS) for Medicare Beneficiaries with Advance Cancer (CAG-00450N). Through this decision, “Medicare Administrative Contractors (MACs) may determine coverage of other Next Generation Sequencing (NGS) as a diagnostic test for patients with cancer only when the test is performed in a CLIA-certified laboratory, ordered by a treating physician and the patient has:

  1. either recurrent, relapsed, metastatic, or advanced staged III of IV cancer; and
  2. either has not been previously tested using the same NGS test for the same primary diagnosis of cancer or repeat testing using the same NGS test only when a new primary cancer diagnosis is made by the treating physician; and
  3. decided to seek further cancer treatment (e.g., therapeutic chemotherapy.”

Further, CMS has outlined the requirements for coverage of these tests to involve an assessment of the AV (analytic validity), CV (clinical validity) and CU (clinical utility) of these tests.
This assay has been rigorously assessed (for AV, CV, and CU) and it fits into the explicit space that CMS has left for MACs to create an LCD.

9

We reviewed this LCD and compared it to the pending central CMS national coverage decision (Next Generation Sequencing (NGS) for Medicare Beneficiaries with Advanced Cancer (CAG-00450)). Both coverage policies are covering essentially the same indications using the same technology – namely looking for alterations in tumor somatic DNA using next generation sequencing (NGS). Yet, there are some major differences between the policies which will create significant concerns if the LCD is implemented as written.
Central CMS may retain the position to uncover all NGS testing that is not listed in the NCD (as was written in the draft NCD), but there is a possibility that the NCD will have the scope limited to tissue based NGS for malignancy and this LCD would be implemented. If so, this LCD would be markedly discordant with the tissue-based NGS testing covered by CMS in the NCD.
General Concerns:

  1. Tissue based NGS has been studied more extensively than liquid biopsy and yet CMS has still required that the test be reviewed by the FDA before it would allow it to be covered. Guardant is in the process of going through the FDA, but until it has, it would create a problem to have a laboratory receive CMS coverage (especially with less evidence and less validation) without FDA approval or clearance when other tissue-based labs have gone to this effort and even have undergone joint review by CMS and FDA.
  2. Tissue based NGS, just as cell free DNA (liquid biopsy) panels contain multiple reported genes that have no or limited clinical utility identified with the panel. Central CMS in its NCD has said that due to these biomarkers not being well studied, that it would only cover the testing under coverage with evidence development (CED). Selected CED requirements from central CMS are compared with the current LCD – and there is marked discordance.
  3. CMS has established that reporting for quality metrics must be done through CMS approved mechanisms and registries. Although the LCD does use the term “independent” it doesn’t specify that the registry must be approved by CMS. This approval is crucial to ensure highest quality of data and complete independence of the registry from the laboratory.
CMS NCD Guardant LCD Concern
Consecutive patients will be enrolled into CED. All patients regardless of mutational status will have the same data collection requirements. Consecutive patients, but with only very limited data collection from the majority of patients (those without a companion diagnostic gene identified), where the only data point is if changed physician’s decision. Paying for testing that reports extended panels where clinical utility is not established beyond the companion diagnostic, does not meet the CMS established scientific rigor for clinical utility. CMS has proposed that this be dealt with CED and data collection for every patient, independent of alteration status. The LCD only asks for data from a limited group of patients. By the CMS NCD, the testing and the reporting and payment are the same for each patient, and so if covering the testing, data collection requirement should be the same independent of the results of that testing
Outcomes required: response rates, time to progression and overall survival AND compared to expected outcomes based on clinical data Outcomes Required: Majority of patients: treatment decision changes of physician. Those with NCCN alterations, 4 months disease control rate, etc. CMS has defined meaningful clinical outcomes as looking at the long-term impact of the testing. Although 4-month disease control rate will help to determine a snapshot of certain patients, it is not enough to know ultimately if the patient has benefited from the treatment. The only way that this can be done is through collecting data on every patient and the comparing both short-term and long-term results. Generally, in oncology studies, meaningful end points are response rates, time to progression and overall survival. Disease control rate at 4 months is not consistent in truly understanding the utility of a test and subsequent treatment. Furthermore, a multi-stakeholder panel of ASCO, AACR, NCCN, EORCT, laboratory industry, pharma, NCI, published a study in Cell in 2017 addressing the “minimum data set” for oncology molecular registries. Central CMS’s data requirements are in line with the Cell consensus paper for minimum data. The LCD does not ask for any mainstream oncology outcomes, making it difficult to compare to other published studies who use RR, TTP, and OS.

The best method to resolve these concerned would be to include the Guardant360 testing in CMS’s NGS NCD, with similar pre-coverage testing validation for the AV/CV and then the same data collection requirements under Coverage with Evidence Development (CED).

 

Thank you for the comments. These comments were received prior to the release of the National Coverage Determination. The decision, announced on March 16, does not require registries:

  1. Coverage
    The Centers for Medicare & Medicaid Services (CMS) has determined that Next Generation Sequencing (NGS) as a diagnostic laboratory test is reasonable and necessary and covered nationally, when performed in a CLIA-certified laboratory, when ordered by a treating physician and when all of the following requirements are met:
    1. Patient has:
      1. either recurrent, relapsed, refractory, metastatic, or advanced stages III or IV cancer; and
      2. either not been previously tested using the same NGS test for the same primary diagnosis of cancer or repeat testing using the same NGS test only when a new primary cancer diagnosis is made by the treating physician; and
      3. decided to seek further cancer treatment (e.g., therapeutic chemotherapy).
    2. The diagnostic laboratory test using NGS must have:
      1. FDA approval or clearance as a companion in vitro diagnostic; and
      2. an FDA approved or cleared indication for use in that patient’s cancer; and
        results provided to the treating physician for management of the patient using a report template to specify treatment options.
      3. results provided to the treating physician for management of the patient using a report template to specify treatment options.
  2. Other
    Medicare Administrative Contractors (MACs) may determine coverage of other Next Generation Sequencing (NGS) as a diagnostic laboratory test for patients with cancer only when the test is performed in a CLIA-certified laboratory, ordered by a treating physician and the patient has:
    1. either recurrent, relapsed, refractory, metastatic, or advanced stages III or IV cancer; and
    2. either not been previously tested using the same NGS test for the same primary diagnosis of cancer or repeat testing using the same NGS test only when a new primary cancer diagnosis is made by the treating physician; and
    3. decided to seek further cancer treatment (e.g., therapeutic chemotherapy).
10

In my clinical practice I follow the NCCN guidelines to perform broad molecular profiling with well validated diagnostic tests and Guardant has demonstrated through 60+ peer reviewed publications that the test is well validated. I encourage you to finalize the LCD to allow patients to access to this important diagnostic tool.

I do have one important potential modification to the LCD. I believe that at the time of progression on targeted therapy, assessment for emerging resistance mutations other than EGFR T790M, is an appropriate use case. There are now, for example, four different ALK inhibitors that can be employed once a patient progresses on first-line therapy. Some of these second and third generation ALK inhibitors are not effective in patients with specific resistance alterations in the ALK, making genomic assessment at the time of progression critical in order to choose the correct therapy. Additionally, with the recent approval of osimertinib in the first line, there is emerging evidence that the specific resistance mechanism can inform next line of targeted therapy. Please consider expansion of the coverage policy to include testing for patients that are progressing on any TKI therapy.

Thank you for your comments. The policy has been expanded to cover patients progressing on any targeted therapy (see previous comments/responses).

11

We use Guardant360 in our non-small cell lung cancer patients at the University of Michigan in accordance with the indications described in the proposed draft LCD. We find that many patients referred to our institution have not been adequately genotyped with the result that they do not have access to highly efficacious and low toxicity targeted therapies. When patients are under-genotyped we have to perform an additional biopsy or get the original biopsy slides sent to our institution. This poses risks and/or delays in the management of this rapidly progressive cancer. Or worse, we have to give chemotherapy or immunotherapy empirically for lack of genomic information. We can optimize outcomes with the Guardant360 blood test in these indications without having to repeat a biopsy and get the patient on targeted therapy quickly, when their blood test results are positive.

Thank you for your comments.

12

We frequently use Guardant360 in our non-small cell lung cancer (NSCLC) patients at the Washington University St. Louis. This liquid biopsy test provides us with more information than other liquid biopsy tests because it includes all four classes of genomic alterations and all seven targetable genes in NSCLC. We use it to prevent repeat invasive biopsies in the first line when tissue is insufficient for complete genotyping or at progression to look for targetable acquired resistance mutations. I have been able to get many advanced lung cancer patients onto highly effective targeted therapy drugs with Guardant360, including patients declining quickly and often too sick to undergo a repeat invasive tissue biopsy.

Thank you for your comments.

13

While tissue genotyping is a powerful tool for physicians, there are limitations with the tissue biopsy paradigm. Tissue may be unavailable for many NSCLC patients and repeat biopsies can cause harm. We feel our Medicare patients must have non-invasive testing as an option to help address the " undergenotyp ing " problem caused by challenges with tissue. For example, data from Gutierrez et al. (Clinical Lung Cancer, 2017) showed that more than 40% of NSCLC patients in a community oncology setting did not receive adequate EGFR and ALK testing, and only 8% were tested for all recommended genomic biomarkers. We are excited that Medicare, through the local coverage process, recognizes the importance of making this test available to every advanced-stage lung cancer patient.

However, as a patient organization, we wish to provide comments and recommendations regarding the draft:

  • We feel that patients progressing on fill therapies should have the opportunity to have plasma testing to determine the mechanism of therapy resistance, and that the LCD should not exclude patients progressing on Osimertinib. Data from Ramalingham et al. (JCO, 2017) suggest that there are a variety of resistance mechanisms beyond T790M relevant for EGFR+ patients.
  • We also believe that the LCD should include the comprehensive set of diagnosis codes used for advanced cancer. All ICD-10 codes in the C34 family should be included in this LCD, similar to how they are included in the tissue LCD. We believe these codes, for lung cancer in an "unspecified part of unspecified bronchus or lung" are essential given the complex nature of diffuse disease in advanced NSCLC. NSCLC patients all need to be tested, regardless of where their tumor resides, and limiting access via the proposed coding paradigm would exclude many patients who could be helped.

Thank you for your comments. Your comments echo others and those recommendations have been included in the final coverage.

14

My clinical experience is similar to findings reported by Thompson et al (Clin Cancer Research, 22/23, Dec 1, 2016) which found actionable mutations available for targeted therapy despite limitations of obtaining tissue for genotyping. Use of approved targeted agents in NSCLC has the significant advantage of improved response rates compared to the use of chemotherapy where mutation status is unknown.
I also use it to prevent repeat invasive biopsies in the first line when tissue is insufficient for complete genotyping and at progression to look for targetable acquired resistance mutations. I have been able to get many advanced lung cancer patients onto highly effective targeted therapy

Thank you for your comments.

15

Since the last open comment on the LCD submission in May 2017, several notable events supporting coverage of the G360 assay have occurred:

  • An updated validation analysis is accepted and in press (1) which includes the COMPLETE study demonstrating extremely low false positive rates at the gene variant level for G360 vs. tissue testing, with positive predictive values (PPVs) for key biomarkers ranging from 92 to 100% compared to tissue-based genotyping in 543 patients.
  • Guidelines from CAP/IASLC/AMP were updated in March 2018 to recommend genotyping for multiple genomic targets in NSCLC, and use of liquid biopsy when tissue genotyping is infeasible, consistent with NCCN NSCLC guidelines (2)
  • G360 is now used at all NCCN member cancer centers.
  • Positive coverage decisions for the use of G360 in NSCLC (but for no other liquid assay) have been made by CIGNA, multiple Blue Cross Blue Shield plans, and other private insurers.
  • 63 peer-reviewed publications utilizing G360 are now in print, 26 of which are outcomes studies (including 16 focused on NSCLC), providing validation comparable to the highest level used by predictive diagnostic tests.
  • G360 recently received an Expedited Access Pathway (EAP) designation from the Food and Drug Administration (FDA).

In this comment on the most recent draft, we would like to emphasize the following:

  1. We believe that the registry section in the draft should no longer be necessary, as the evidence has become even more robust, which now includes 16 peer-reviewed outcomes studies for NSCLC supporting the use of G360 specifically and a guideline from CAP/IASLC/AMP2 supporting the use of liquid biopsy. Over half of the G360 published studies as well as this guideline have been published since the original draft in spring 2017.
  2. It is critical to incorporate comprehensive coding for malignant neoplasms of the bronchus and lung, which includes ICD-10 codes C34.90-C34.92 (the most widely used diagnostic codes for stage III/IV NSCLC) since the location of the tumor in the lung (at original diagnosis) is no longer relevant to systemic treatments such as genomically targeted therapy.
  3. We believe that the indications in the LCD should be clarified, as well as expanded to allow diagnostic testing of all patients who are progressing on any targeted-therapy tyrosine kinase inhibitor (TKI), including osimertinib.4
  4. We address recent critical literature to avoid misinterpretation

These points are discussed below:

    • Registry: Most of the goals of the proposed registry requirements in the draft LCD have been achieved thus removing the need for the registry in the final LCD—half of which was not available at the time of evaluation last year—is sufficient for full coverage.

      In terms of support, the draft LCD referenced 21 publications (5–25) (the last of which has since been updated26). Seven of the original 21 publications were outcomes studies.10,13,14,18,19,23,24 The latest version of the NCCN NSCLC guidelines were updated in February 2018.26 Nine other outcomes studies not referenced in the draft27–35 yield a total of 16 outcomes studies related to the treatment of NSCLC using G360 (summarized in Table 1 at the end of this comment). Of note, 11 publications pertinent to NSCLC have appeared since the last evaluation (summarized in Table 2 at the end of this comment).(4,27,29–31,34–39)

      In addition to the above references specific to G360, (a) multiple citations in the CMS NCD for Next Generation Sequencing for Medicare Beneficiaries with Advanced Cancer (CAG-00450N) (40) and (b) a newer guideline from specialty societies (2) support the use of liquid biopsy in advanced NSCLC. In particular, this guideline advises that “patients with the types of advanced lung cancer in which these targetable molecular alterations typically occur should receive the molecular testing required to identify them, and thereby receive appropriate targeted treatments” and recommends “the use of cell-free DNA to ‘rule in’ targetable mutations when tissue is limited or hard to obtain.”(2) In fact, this guideline directly proposes that “[c]ell-free DNA may be preferred for patients unwilling or unable to undergo a biopsy at the time of progression; moreover, as progression may represent subclonal processes, cfDNA testing may represent a more global sampling of disease as compared with a tissue biopsy.”

      In support of the validity of G360 testing vis-à-vis tissue sampling, an updated validation study reveals extremely low false positive rates at the variant level for G360 vs. tissue testing, with PPVs for key biomarkers ranging from 92 to 100% compared to tissue in a retrospective cohort of 543 patients in which results from both tissue biopsies and G360 assays were simultaneously available.

      Thus, we believe G360 is supported by sufficient evidence to remove the registry section, and hereby supply the local Medicare contractor with an updated list of relevant publications. Any future studies or data currently in review that support the utility of G360 will be provided to the local contractor as they are accepted for publication. Incidentally, Guardant is independently conducting an ongoing registry study (41) the results from which we plan to publish as the data mature; however, we believe that a registry should not be a requirement for coverage given the compelling evidence already in existence.

    • Including a more comprehensive set of codes for neoplasms of the bronchus and lung: It is critical to include a comprehensive set of ICD-10-CM CPT codes used by clinicians to diagnose advanced lung cancer, as suggested in the table below. Currently, codes C34.90-C34.92, which represent unspecified parts of the lung or bronchus, are excluded from the LCD. However, these codes represent roughly two-thirds of the codes enlisted by oncologists when ordering G360 because the assay is used in advanced cancer in which the disease is widely disseminated. With this indication, location is no longer relevant, as treatment is systemic.

      In fact, clinicians need the option of selecting “unspecified” in these patients who often have multiple sites of metastases, some of which may be distant and some even unknown, and therefore cannot be described with a single location. Additionally, codes C34.00, C34.10, C34.30, and C34.80 should be
      covered since the unspecified location in the lung does not change the need to diagnose and treat the disease. In short, a general CMS guideline to code to the maximum specificity of the code set is actually misleading in late-stage cancer patients with widespread metastatic disease. In fact, half of NSCLC is stage IV metastatic at diagnosis, and it is impossible to ascertain which of the many lesions is the primary lesion – so the location is unknown.

      Proposed by MolDX42  Suggested Revision 
       C33  C34.32  C33  C34.80
       C34.01  C34.81  C34.00  C34.81
       C34.02  C34.82  C34.01  C34.82
       C34.11  C38.1  C34.02  C34.90
       C34.12  C38.2  C34.10  C34.91
       C34.2  C38.4  C34.11  C34.92
       C34.31  C38.8  C34.12  C38.1
           C34.2  C38.2
           C34.30  C38.4
           C34.31  C38.8
           C34.32  
    • Clarification and expansion of indications: We believe that although the intent of the LCD is sufficiently clear in encouraging testing for actionable genes, its current language may be confusing, as G360 could potentially be interpreted as not covered if one of the biomarkers has been previously tested. In the table below, we propose a clearer explanation for Medicare Part C insurance plans, clinicians, labs, and auditors. We also believe that the indications themselves should be expanded to allow testing of all patients who are progressing on any TKI.

      Proposed by MolDX42 Suggested Revision
      This policy provides limited coverage for Guardant360® (Guardant Health, Redwood City, CA), a plasma-based comprehensive somatic genomic profiling test (hereafter called CGP) for patients with Stage IIIB/IV non- small cell lung cancer (NSCLC):
      At diagnosis –
      • When results for EGFR single nucleotide variants (SNVs) and (insertions and deletions (indels); rearrangements in ALK and ROS1; and SNVs for BRAF are not available AND when tissue-based CGP is infeasible (i.e., quantity not sufficient for tissue-based CGP or invasive biopsy is medically contraindicated),

      OR
      At progression-

      • For patients progressing on or after chemotherapy or immunotherapy who have never been tested for EGFR SNVs and indels; rearrangements in ALK and ROS1; and SNVs for BRAFs, and for whom tissue-based CGP is infeasible (i.e., quantity not sufficient for tissue-based CGP from original biopsy); OR
      • For patients progressing on EGFR tyrosine kinase inhibitors (TKIs), other than Osimertinib

      This policy provides coverage for Guardant360® (Guardant Health, Redwood City, CA), a plasma-based comprehensive somatic genomic profiling test (hereafter called CGP) in patients with Stage IIIB/IV non- small cell lung cancer (NSCLC), according to the following separate criteria for untreated and treated patients:
         1) If the patient is untreated:
      Results are not available for one or more key driver mutations as defined below, AND tissue- based CGP is infeasible (e.g., quantity not sufficient for tissue-based CGP, or invasive biopsy is medically contraindicated).
         2) If the patient is treated:
      (a) If the patient is progressing on or after chemotherapy or immunotherapy:
      Results are not available for one or more key driver mutations as defined below, AND tissue-based CGP is infeasible (e.g., quantity not sufficient for tissue-based CGP, or invasive biopsy is medically contraindicated).
      (b) If the patient is progressing on any tyrosine kinase inhibitor (TKI)

      Key driver mutations in NSCLC are defined as single nucleotide variants (SNVs) and insertions and deletions (indels) for EGFR, rearrangements for ALK and ROS1, and SNVs for BRAF.

       

      As suggested above, the indications should be expanded to allow testing of patients who are progressing on any TKI, which includes those progressing on osimertinib.

      First, recent evidence warrants expanding testing at progression to patients on any TKI, not just the category of EGFR TKIs. After treatment of an ALK or ROS1 fusion, patients at progression often acquire SNVs in these genes, which are now targetable with FDA-approved next-generation TKIs: for example, ALK fusion patients who acquire ALK F1174C no longer respond to first-generation TKI crizotinib, but do to next-generation alectinib.43,44 This phenomenon has been demonstrated for ALK resistance mutations identified with either tissue-based NGS or plasma-based NGS.35,43 Similarly, ROS1 fusion patients who acquire ROS1 resistance mutations respond by switching crizotinib to FDA-approved next-generation ROS1 inhibitors ceritinib, entrectinib, brigatinib, or cabozantanib.34,43,45 It is important to identify which ALK or ROS1 resistance mutation has been acquired in order to select the appropriate next-generation ALK or ROS1 inhibitor.44 The list of next-generation TKIs is also growing, with recent evidence suggesting that treatment of NTRK1/3 fusions with first-generation TRK inhibitors can now be overcome with next-generation TRK inhibitors.34 Thus, it appears that plasma-based genomic testing is useful at progression in patients on any targeted-therapy TKI. 

      Second, patients progressing on osimertinib definitely benefit from G360 testing. In a study using G360 in patients progressing on first-line osimertinib for EGFR driver mutations, instead of EGFR T790M several different treatable resistance mutations emerged,4 including EGFR C797S, which may be acquired when on first-line osimertinib, and will typically respond upon switching to a first-generation EGFR TKI such as erlotinib, since the latter does not bind to amino acid/codon 797 residue as osimertinib does.46 G360 detects these EGFR C797S mutations, providing in a non-invasive and timely manner genomic information that is critically needed to select optimal treatment for patients with advanced NSCLC.

4. Clarification of publications: Two recent publications contain ambiguities that require clarification.

     a. First, a letter in JAMA Oncology(47) reported poor concordance between G360 and another commercial liquid biopsy in patients with prostate cancer based on what we consider to be a fundamentally flawed analysis that failed to account for differences in

  1. the tests’ reportable ranges (i.e., some variants reported as discordant were not even on both tests),
  2. reporting policies of both tests (i.e., the comparator test reported rare germline variants, whereas G360 does not), and
  3. sensitivities of the tests (i.e., most discordances were below the comparator test’s claimed limit of detection).

Regarding the latter point, Guardant reports very low VAF variants because these are true positives, based on established assay specificity. A comparison of 543 matched plasma-tissue samples found 92-100% PPV for G360, ranging as low as 0.1% VAF.1 A Dana Farber Cancer Institute study of 222 consecutive EGFR/KRAS/BRAF/ESR1 positives reported 99.6% PPA (sensitivity) and 97.8% NPA (specificity), and 97.4% PPV for G360 compared to a well-validated clinical ddPCR assay, also ranging down to 0.1% VAF.1

In addition to the methodological errors in the letter, the authors conducted the study in patients outside of the intended use population of G360. Yet, even despite these flaws, the study still reported 100% concordance for the only clinically actionable alterations (AR resistance alterations).

     b. The second publication needing clarification is a review of ctDNA assays by the American Society of Clinical Oncology and College of American Pathologists.(48)

The authors stated that liquid biopsies have a broad range of potential indications. However, many of those indications are not supported by the published evidence, including molecular residual disease detection, monitoring, and early cancer detection. We agree with that assessment.

The authors also concluded that “[s]ome ctDNA assays have demonstrated clinical validity and utility with certain types of advanced cancer,” although “there is insufficient evidence of clinical validity and utility for the majority of ctDNA assays” (italics ours).

We believe Noridian and MolDX appreciate that the authors are carefully distinguishing between (a) the majority of assays whose validity and utility have not yet been established and (b) those few assays whose validity and utility for certain types of advanced cancer have been demonstrated for specific indications. In this second group, G360 has more supportive publications, and outcomes studies in particular, than any other liquid biopsy test.

G360 has been independently validated in 63 peer-reviewed publications (including multiple prospective clinical trials), has met stringent New York State Department of Health approval, and most importantly, and has yielded the expected positive clinical response rates in 26 published, peer-reviewed outcomes studies.

The authors are rightly concerned that the majority of current liquid assays have an appreciable rate of discordance with tumor-tissue genotyping. However, it is important to realize that analysis of DNA from neither tissue nor blood is perfectly sensitive. Sensitivity may be limited in tissue when samples fail to capture tumor heterogeneity and in blood when tumor DNA is not shed into circulation.

In line with the review, Guardant recommends the use of its validated G360 for the NCCN-endorsed indication of characterizing a tumor for therapy selection in patients with late-stage NSCLC for whom conventional biopsy is infeasible or inadequate at yielding a sufficient quantity or in whom the cancer has progressed despite treatment.

This recommendation to consider a test that has been validated for an appropriate purpose is aligned with the review, which concedes that some assays have demonstrated clinical validity and utility even though the majority have not.

Gutierrez (Goldberg) et al.10 Clin Lung Cancer, 2017

  • 1/3 of 814 late-stage NSCLC patients were not tested for both EGFR and ALK
  • Only 8% were tested for all 7 NCCN-recommended genomic targets across community-based practices in NJ and MD
  • Overall survival with targeted therapy was double that of chemotherapy

Villaflor (Salgia) et al.13 Oncotarget, 2016

  • 1st clinic-based NSCLC cohort; 83% had ctDNA detected; tissue not genotyped in 54%
  • Six (11%) of all ctDNA-positive patients had EGFR mutations (two were T790M), one initial EGFR driver that was missed in tissue
  • 8 responded (75%) and 2 were stable – for 100% disease control rate
  • Median PFS of 11.5 months in the expected range based on published studies using tissue genotyping

Thompson (Carpenter) et al.14 Clin Canc Res, 2016

  • Prospective clinic-based NSCLC cohort (n=102), 53 1st line, 47 at progression (2+ line)
  • Tissue not genotyped in 51% – 7 not biopsy-able and 45 QNS
  • CtDNA positive in 84.3%, including 8 with EGFR T790M for which tissue failed
  • 31% had on-label Rx, 55% off-label and 70% were trial eligible

Rozenblum (Peled) et al.18 J Thorac Oncol, 2016

  • 19 tissue insufficient for tissue NGS, treatment decision was changed in 32% (6 of 19) who had NCCN genomic targets EGFR (2), RET (2), MET (1), and ERBB2 (HER2) (1)
  • Matched therapy resulted in 3 PR and 3 SD out of 5 treated (60% objective response, 100% disease control rate by RECIST)
  • All of the 6 with genomic targets (84%) were missed by local EGFR qPCR & ALK FISH

Kim (Lee) et al.19 JCO Precis Oncol, 2017

  • Of 72 NSCLC patients with insufficient tissue for NGS, 34 (47%) had targetable alterations by G360 (29 EGFR, 2 ALK, 1 RET, 2 ERBB2)
  • Objective response rate was 87% and disease control rate 100%
  • An EML-ALK fusion is picked up by G360 in the 2nd line and an EGFR L858R in the 3rd line, highlighting the undergenotyping problem
  • Of 75 gastric adenocarcinoma, 25 (33%) had targetable alterations (11 ERBB2 amp, 5 MET amp, 3 FGFR2 amp, 6 PIK3CA)
  • Objective response rate was 67% and disease control rate 100%
  • No correlation between variant allele fraction and plasma copy number of genomic targets, but all copy number amplifications were 50th percentile or higher (++ or +++)

Peled (Elkabetz) et al.23 J Thorac Oncol, 2017

  • Local EGFR testing at progression found only the EGFR E19 del, whereas G360 found EGFR E19 del, EGFR T790M, and EGFR G724S
  • Tissue sent for NGS-confirmed EGFR G724S, but missed the T790M
  • Response from osimertinib validated the EGFR T790M finding in plasma, but the EGFR G724S rose in ctDNA concentration
  • Addition of afatinib diminished the EGFR G724S clone, but EGFR C797S emerged in plasma (causing resistance to osimertinib); rapid decline after switch to chemotherapy

Santos (Hunis) et al.24 Clin Oncol, 2016

  • In 81 consecutive NSCLC patients, G360 rescued 10 additional EGFR mutations not identified by tumor biopsy
  • Concordance to tissue NGS for EGFR alterations was 77%
  • In 4 patients in whom tissue was insufficient or unobtainable at progression on EGFR inhibitors, G360 identified EGFR T790M with objective response on osimertinib based on cfDNA alone

Lu (Collisson) et al.27 Cancer Research, 2017

  • 1st in-human finding of multiple acquired resistance mutations after treatment of MET exon 14 skipping with crizotinib: these acquired mutations are treatable with currently FDA-approved type II MET inhibitors
  • At progression on crizotinib, G360 identified the MET exon 14 skipping and amplification, as well as 4 subclonal missense mutations in MET: D1228N, Y1230H, Y1230S, and G1163R – all previously and independently reported to impair crizotinib binding

Schwaederl; (Kurzrock) et al.28 Cancer Res, 2017

  • 670 cancer patients sequenced with G360
  • 48% (320/670) had actionable alterations, of which 36% (241/670) were candidates for on-label or off-label treatment with an FDA-approved drug
  • Case report of patient with metastatic gastric cancer with EGFR amplification detected in plasma and treated with erlotinib and cetuximab with durable 60% reduction in tumor size at 5 months assessment

Schwaederl (Kurzrock) et al.29 Clin Cancer Res, 2017 n = 88

  • 72/88 patients (82%) had ≥ 1 ctDNA alteration(s)
  • 75% carried alteration(s) potentially actionable by FDA-approved (61%) or experimental drug(s) in clinical trials (additional 14%)
  • 25 patients (28%) received therapy matching ≥1 ctDNA alteration(s) and 72% (n=16/22) of the evaluable matched patients achieved stable disease ≥ 6 months (SD) or partial response (PR)
  • All of 5 patients with ctDNA-detected EGFR T790M were treated with a 3rd-generation EGFR inhibitor and all 5 achieved SD ≥ 6 months/PR
  • Patients with ≥ 1 alteration with ≥ 5% variant allele fraction (vs. < 5%) had a significantly shorter median survival (p=0.01)

Engstrom (Christenson) et al.30; Clin Cancer Res, 2017

Two case studies:

  • A non-smoker with stage IV lung adenocarcinoma was refractory to multiple lines of therapy and had no detectable actionable mutations on tissue NGS, but did have a MET exon 14 skipping mutation on G360 – subsequent treatment with a type II MET inhibitor reduced the size of target lesions by 66%, durable for 7 months
  • Another patient with MET ex14 del mutation-positive NSCLC had responded to crizotinib, but eventually progressed. G360 detected 3 MET A-loop resistance mutations D1228N, Y1230C/H, and G1163R – new therapy with glesatinib reduced the size of a MET Y1230H mutation tissue-positive liver metastasis and caused the Y1230 mutations in the blood to disappear (but other lesions grew, consistent with detection by G360 of other resistance mutations not detected in tissue, including D1228N, G1163R, and a new L1195V mutation).

Sarfaty (Peled) et al.31Clin Lung Cancer, 2017 n = 14

  • Case series of 14 patients with RET fusions treated with cabozantanib with 60% ORR and 13.7 months median overall survival
  • None tested for locally, 12 RET fusions found on tissue-based NGS, but 2 had insufficient tissue and were identified/rescued with G360

Helman (Simmons) et al.32 Cancer Res, 2017

  • TIGER X/2 single arm cohort of 77 advanced NSCLC patients tissue-positive for EGFR T790M that were treated with rociletinib, a 3rd generation EGFR inhibitor
  • G360 identified 93% of the initial EGFR activating and 85% of the EGFR T790M resistance mutations in pretreatment samples with detectable tumor DNA
  • Responses and PFS for tissue-detected EGFR T790M treated with rociletinib were equivalent to plasma-detected EGFR T790M
  • Patients with EGFR T790M on Guardant360 as low as 0.1% mutant allele fraction had responses similar to those at ten-fold higher mutant allele fractions, i.e. detection of any amount of EGFR T790M in plasma predicted response

Ahn (Mok) et al.33 J Thorac Oncol, 2017

  • 227 advanced NSCLC patients were studied with Guardant360 in the AURA3 randomized controlled trial of osimertinib for tissue positive EGFR T790M patients
  • Guardant360 was more sensitive for EGFR T790M than the Roche cobas FDA-approved plasma test, and ddPCR conducted by Biodesix
  • Outcomes for Guardant360-detected EGFR T790M in the osimertinib treatment arm vs. the control arm are in a manuscript In Review so cannot be disclosed at this time

Drilon (Hyman) et al.34 New England J Med, 2018

  • 55 pediatric and adult patients with NTRK1/2/3 fusions identified by tissue or G360 in a pan-cancer “basket” trial (4 of which were NSCLC) were targeted with larotrectinib, a selective TRK inhibitor
  • Response rate = 75% (according to independent review) and 80% (according to investigator assessment), with 29% experiencing a complete response

McCoach (Doebele) et al.35 Clin Cancer Res, 2018 n = 88

  • The G360 database was queried to identify 88 patients with 96 plasma-detected ALK fusions
  • 42 plasma ALK positive patients had no history of targeted therapy
  • In G360-positive ALK fusion cases, 5 patients were tissue negative, and 11 had insufficient quantity of sample (QNS) for tissue testing
  • Of the 5 ALK fusion tissue-negative but plasma-positive, 3 received targeted therapy and all 3 responded
  • Of 31 ALK fusions patients retested with G360 at progression, 16 samples (53%) contained 1-3 resistance mutations in the ALK gene that could be used to direct changes to next-generation ALK inhibitors
  • G360 could be used (1) upfront when tissue is infeasible or negative for ALK fusions and clinical suspicion is high (young, non-smoking) and (2) at progression to select the best next-generation ALK inhibitor

Publications Pertinent to NSCLC Since the Last Evaluation

Blakely (Bivona) et al.36 Nature Genetics, 2017 n = 1,222

  • 1,122 EGFR-mutant NSCLC patients utilizing cell-free DNA NGS
  • Critical concurrent oncogenic events present in most advanced-stage EGFR-mutant lung cancers are identified, many of which predict poor response to EGFR inhibitor monotherapy
  • Novel resistance pathways limiting EGFR inhibitor response, including WNT/β-catenin and cell cycle gene (e.g., CDK4, CDK6) alterations, are found
  • Tumor genomic complexity increases with EGFR inhibitor treatment and co-occurring alterations in CTNNB1, and PIK3CA exhibit non-redundant functions that cooperatively promote tumor metastasis or limit EGFR inhibitor response
  • This study challenges the prevailing single-gene driver oncogene view and links clinical outcomes to co-occurring genetic alterations in advanced-stage EGFR-mutant lung cancer patients

McCoach (Doebele) et al.35 Clin Cancer Res, 2018 n = 88

  • The G360 database was queried to identify 88 patients with 96 plasma-detected ALK fusions
  • 42 plasma ALK positive patients had no history of targeted therapy
  • In G360-positive ALK fusion cases, 5 patients were tissue negative, and 11 had insufficient quantity of sample (QNS) for tissue testing
  • Of the 5 ALK fusion tissue-negative but plasma-positive, 3 received targeted therapy and all 3 responded
  • Of 31 ALK fusions patients retested with G360 at progression, 16 samples (53%) contained 1-3 resistance mutations in the ALK gene that could be used to direct changes to next-generation ALK inhibitors
  • G360 could be used (1) upfront when tissue is infeasible or negative for ALK fusions and clinical suspicion is high (young, non-smoking) and (2) at progression to select the best next-generation ALK inhibitor

Schwaederl (Kurzrock) et al.29 Clin Cancer Res, 2017 n = 88

  • 72/88 patients (82%) had ≥ 1 ctDNA alteration(s)
  • 75% carried alteration(s) potentially actionable by FDA-approved (61%) or experimental drug(s) in clinical trials (additional 14%)
  • 25 patients (28%) received therapy matching ≥1 ctDNA alteration(s) and 72% (n=16/22) of the evaluable matched patients achieved stable disease ≥ 6 months (SD) or partial response (PR)
  • All of 5 patients with ctDNA-detected EGFR T790M were treated with a 3rd-generation EGFR inhibitor and all 5 achieved SD ≥ 6 months/PR
  • Patients with ≥ 1 alteration with ≥ 5% variant allele fraction (vs. < 5%) had a significantly shorter median survival (p=0.01)

Ramalingam (Jänne) et al.4 Journal of Clin Onc, 2017 n = 60

  • 3rd-generation EGFR TKI (osimertinib) used in 1st line with 20.5 months PFS, n=60
  • G360 used at progression n=38 samples, of 19 positives 9 had resistance alterations including targetable point mutations in EGFR, MEK1, ERBB2 (HER2), and PIK3CA, indel in ERBB2 (HER2), amplification in MET, and RB1 inactivating mutations
  • EGFR C797S were acquired in 22% (2 of 9), but no EGFR T790M were acquired

Hu (Oxnard) et al.37 Clin Cancer Res, 2017 n = 48

  • G360 accurately distinguished germline and somatic mutations in EGFR T790M in NSCLC
  • Blinded validation in 21 cases confirmed a 100% PPV for predicting germline T790M
  • From Guardant’s advanced pan-cancer databased of 31,414 patients, 911 were EGFR T790M mutations, of which 48 were as germline, with 43 of the 48 (90%) having non-squamous NSCLC (p<0.0001)

Sarfaty (Peled) et al.31 Clin Lung Cancer, 2017 n = 14

  • Case series of 14 patients with RET fusions treated with cabozantanib with 60% ORR and 13.7 months median overall survival
  • None tested for locally, 12 RET fusions found on tissue-based NGS, but 2 had insufficient tissue and were identified/rescued with G360

Mooradian (Sequist) et al.38 JCO Precision Oncology, 2017

  • Lung adenocarcinomas that are EGFR mutation positive can transform to a SCLC phenotype as a resistance mechanism to front-line EGFR inhibitors in 5% of cases
  • RB1 is universally lost in SCLC-transformed cases, and baseline RB1 loss is a strong predictor for subsequent SCLC transformation
  • In this case study, one lesion grew and one responded on chemotherapy
  • Plasma NGS detected a targetable EGFR T790M mutation
  • ctDNA genotyping may be able to paint a more complete picture of the competing clones within a patient

Yang (Zhang) et al.39 J Hem & Oncol, 2017

  • Early- and late-stage NSCLC had similar rates of driver mutation (mainly TP53, EGFR, and KRAS)
  • G360 plasma mutations and tissue driver mutations matched for TP53, ERBB2, KRAS, both ALK fusions, PIK3CA, NF1, and CTNNB1 in 177 consecutive pan-cancer (including lung cancer) patients
  • Actionable ctDNA-detected alterations often became undetectable with targeted therapy and response
  • When patients relapse, driver mutations recur, with new resistance mutations such as EGFR T790M, allowing successful treatment with a 2nd targeted therapy

Lu (Collisson) et al.27 Cancer Research, 2017

  • 1st in-human finding of multiple acquired resistance mutations after treatment of MET exon 14 skipping with crizotinib: these acquired mutations are treatable with currently FDA-approved type II MET inhibitors
  • At progression on crizotinib, G360 identified the MET exon 14 skipping and amplification, as well as 4 subclonal missense mutations in MET: D1228N, Y1230H, Y1230S, and G1163R – all previously and independently reported to impair crizotinib binding

Engstrom (Christenson) et al.30 Clin Cancer Res, 2017

Two case studies:

  • A non-smoker with stage IV lung adenocarcinoma was refractory to multiple lines of therapy and had no detectable actionable mutations on tissue NGS, but did have a MET exon 14 skipping mutation on G360 – subsequent treatment with a type II MET inhibitor reduced the size of target lesions by 66%, durable for 7 months
  • Another patient with MET ex14 del mutation-positive NSCLC had responded to crizotinib, but eventually progressed. G360 detected 3 MET A-loop resistance mutations D1228N, Y1230C/H, and G1163R – new therapy with glesatinib reduced the size of a MET Y1230H mutation tissue-positive liver metastasis and caused the Y1230 mutations in the blood to disappear (but other lesions grew, consistent with detection by G360 of other resistance mutations not detected in tissue, including D1228N, G1163R, and a new L1195V mutation).

Drilon (Hyman) et al.34 New England J Med, 2018

  • 55 pediatric and adult patients with NTRK1/2/3 fusions identified by tissue or G360 in a pan-cancer “basket” trial (4 of which were NSCLC) were targeted with larotrectinib, a selective TRK inhibitor
  • Response rate = 75% (according to independent review) and 80% (according to investigator assessment), with 29% experiencing a complete response

We appreciate the comments. We believe the suggestions made by you and other commenters on the indications for the test, ICD10 codes to add, and deletion of the registry are appropriate and they are included in the final coverage document.

16

In our open comment letter to Noridian dated April 13, 2018, regarding the proposed LCD for plasma-based comprehensive genomic profiling (CGP) in stage IIIB/IV non-small cell lung cancer (NSCLC), we detailed evidence and rationale in support of the (1) removal of the registry requirement, (2) expansion of associated ICD-10 codes, and (3) tightening of the coverage language. Below is a summary of key updates since that letter that reinforce support to approve the LCD and remove the registry component:

  1. New analytical and clinical (AV/CV) publication which demonstrates high concordance with tissue genotyping and orthogonal liquid biopsy testing
    1. Publication of a 543-sample blood-to-tissue concordance (the largest study to date), underscoring the very high positive predictive value (PPV) 98-100%1 of Guardant360 for targetable genomic alterations. This study includes many other elements of AV/CV, including a concordance study of 222 samples in a blinded blood-to-blood comparison of Guardant360 to a well-validated digital droplet PCR (ddPCR) test at Dana Farber Cancer Institute. Both analyses confirm the very high PPV/specificity (>99%) (Odegaard et al., 2018) found in earlier studies of Guardant360 for both blood/tissue and blood/blood comparisons.
  2. New clinical utility data supporting the broader use of Guardant360
    1. Publication status of one of 16 NSCLC outcomes studies cited in the Noridian letter (13 published, 3 in review at that time) of Guardant360 moved from in press to in print, reviewing responses to 1st- and 2nd-line therapy selection for Guardant360-detected ALK fusions and ALK resistance mutations (McCoach et al., 2018).
    2. Publication of a study that utilized Guardant360 to find targetable NTRK resistance mutations in the New England Journal of Medicine (Drilon et al., 2018).
    3. Current version of NSCLC NCCN guidelines (v 4.2018) continues to support “broad molecular profiling” and “plasma-based” biopsy when tissue remains insufficient (Ettinger et al., 2018). This NCCN guideline corresponds with MolDx coverage in the draft LCD.
    4. The most common resistance mutation to osimertinib (as recently approved by the FDA for use in 1st-line NSCLC) is EGFR C797S, as was shown in a study using Guardant360. Because EGFR C797S resistance mutations respond to earlier-generation tyrosine kinase inhibitors (TKIs) like gefitinib or erlotinib, repeat invasive tissue biopsies to find targetable EGFR C797S at progression after 1st-line osimertinib can be obviated with Guardant360 (Ramalingam et al., 2017).

Detailed summaries of the updates:

    1. New AV/CV publication which demonstrates high concordance with tissue genotyping and orthogonal liquid biopsy testing

The most comprehensive analytic and clinical validation of Guardant360 was recently published (Odegaard et al., 2018). This study documents very high PPV for Guardant360-identified targetable driver mutations, in which Guardant-360 results are compared first to tissue genotyping and second to a well-validated orthogonal liquid biopsy method. Summary findings include:

      • This blood-to-tissue clinical validation is the largest ever published, with 543 consecutive lung cancer patients. The scale of this study enabled variant-level (rather than relying on broader gene-level) concordance analysis. Comparison of Guardant360 to tissue genotyping for EGFR exon 19 del (n=291), EGFR L858R (n=181), ALK/ROS1 (n=37), KRAS G12X (n=26), EGFR exon 20 ins (n=16), BRAF V600E (n=5), and MET exon 14 skipping alterations (n=3) yielded PPVs for Guardant360 of 98-100%. Original concordance was 92% for ALK/ROS1, when counting three ALK fusions that were negative on the available tissue biopsy, but positive on G360. All three cases responded to ALK inhibitors, so the positive predictive value for ALK fusions was, in fact, 100%.
      • A second study included in the Odegaard publication was a well-designed, blinded blood-to-blood concordance comparing Guardant360 to a well-validated ddPCR assay in 222 plasma samples at Dana Farber Cancer Institute. This substudy demonstrated Guardant360 clinical sensitivity of 99.6% and specificity of 97.8% for EGFR L858R, EGFR T790M, EGFR exon 19 del, ESR1 Y537/D538, and KRAS G12X down to 0.1% variant allele fraction. (Of note, all four “false positive” G360 calls were at allele fractions below the ddPCR assay’s limit of detection.) Good study design for liquid biopsy concordance studies requires that (i) the studied variants must be reported by all platforms, (ii) alterations must be within the detection range of all platforms, (iii) the concordance must be interpreted based on established detection rates at the corresponding variant allelic fractions, (iv) sample pairs are concurrently obtained, and (v) patients are pre-treatment, or if on treatment, actively progressing (Schwaederlé et al., 2017).
      • We also validated Guardant360 for copy number amplifications. In 119 samples, copy number amplifications in 8 genes compared between Guardant360 and ddPCR revealed 90% sensitivity (63/70) and 100% specificity (49/49) for Guardant360.

This publication also reports data for 10,593 consecutive pan-cancer patients tested in clinical practice:

- The Guardant360 assay technical success rate was >99.6%,
- Report turnaround times were rapid (with a median turnaround of 7 days),
- The NSCLC ctDNA somatic mutation detection rate was 87.7%.

The Odegaard et al. publication builds upon earlier AV/CV studies and consistently demonstrates high PPV/high specificity across all four different genomic alteration types, and across all seven different guideline-recommended gene targets in NSCLC.

This study further demonstrates that mutations found in plasma with Guardant360 are almost always found in tissue; as a result, plasma-based CGP outcomes should be equivalent to tissue-based outcomes. This expectation is confirmed in the NSCLC outcomes studies conducted with Guardant360. Taken together, this new CV study and all Guardant360 outcomes studies support a positive coverage decision.

    1. New clinical utility data supporting the broader use of Guardant360
      1. We summarized prior peer-reviewed outcomes studies evaluating Guardant360-detected genomic targets in NSCLC in our prior Noridian LCD comment letter (Ahn et al.; Drilon et al., 2018; Engstrom et al., 2017; Gutierrez et al., 2017; Helman et al., 2017; Kim et al., 2015; Lu et al., 2017; McCoach et al., 2018; Peled et al., 2017; Rozenblum et al., 2017a; Santos et al., 2016; Sarfaty et al., 2017; Schwaederle et al., 2017; Schwaederlé et al., 2017; Thompson et al., 2016; Villaflor et al., 2016). These 16 NSCLC outcomes studies (now 14 published, 2 in review) do not include single-patient case reports. Please note, Helman et al. and Ahn et al. remain in review at the time of this letter with approval expected imminently.
      2. We summarize and have attached McCoach et al. below (the most recent of the 16 studies now in print). Guardant360 changes treatment decisions and improves outcomes when patients are not tested for ALK fusions because tissue is insufficient or infeasible. Identifying these fusions is crucial because 1st-line treatment can achieve a high 83% response rate with newer ALK inhibitors, 2-4 times better than chemo- or immunotherapy (Peters et al., 2017). McCoach et al. was a multicenter study in which the Guardant360 database was queried to identify 88 consecutive patients with 96 plasma-detected ALK fusions. Forty-two of the plasma ALK-positive patients had no history of targeted therapy, partly because 11 patients (26%, 11/42) had tissue quantity not sufficient (QNS) for genomic testing. It had been previously shown that despite the high response rates with targeted therapy, one-third of NSCLC patients do not get tested for ALK fusions, largely because of tissue infeasibility (Gutierrez et al., 2017). McCoach et al. underscores the undergenotyping problem, and the opportunity for plasma-based CGP to rescue patients not tested for all guideline-recommended genomic targets (see NCCN and CAP/IASLC/AMP guidelines). In addition, tissue biopsy ALK molecular testing was completed, but negative in five Guardant360-positive patients. Three out of the five tissue-negative but Guardant360-positive patients were treated with ALK-targeted drugs and all three responded (two were not treated with an ALK-targeted drug). Thus, plasma-based CGP may be useful when tissue CGP testing fails to detect a mutation because of tumor heterogeneity.

        When patients progress on EGFR TKIs, it is not enough to test solely for the EGFR T790M mutation. For those patients progressing on ALK TKIs, McCoach et al. confirmed the need to test for targetable resistance mutations in the ALK gene. At progression, 31 NSCLC patients on targeted therapy for ALK fusions were retested with Guardant360. Of these, 53% (16 of 31) of patients had acquired 1-3 resistance mutations in the ALK gene that are known to respond by changing to a next-generation ALK inhibitors (Dagogo-Jack et al., 2018; Gainor et al., 2016). The authors concluded that “in the largest cohort of cfDNA ALK fusions reported to date, our data demonstrate that comprehensive cfDNA NGS testing is an additional tool that provides a non-invasive means of detecting targetable alterations in newly diagnosed patients, as well as resistance mutations…in patients progressing on targeted therapy” (McCoach et al., 2018).

      3. The third-generation EGFR TKI osimertinib was recently approved by the FDA for 1st-line use for EGFR driver mutations. This newer TKI produced a leap in progression-free survival to 17.2 months compared to 8.5 months for earlier-generation EGFR TKIs (Soria et al., 2018). As osimertinib use moves to the 1st-line, the most common acquired EGFR mutation driving resistance becomes EGFR C797S, as was shown in a study using Guardant360 (Ramalingam et al., 2017). Because the EGFR C797S resistance mutation responds to earlier-generation TKIs like gefitinib or erlotinib, FDA-approved drugs to treat patients progressing on osimertinib already exist (Ramalingam et al., 2017). Thus, repeat invasive tissue biopsies to find targetable EGFR C797S at progression after 1st-line osimertinib can be obviated with Guardant360.

The new outcome study (McCouch et al.) and the approval of osimertinib as a 1st-line therapy validates the need to test all patients progressing on TKIs.

Guardant remains committed to publishing continued clinical outcome studies for Guardant360. While not part of our core published LCD evidence, several additional studies are under editorial review and have been presented at scientific meetings (Ahn et al.; Geva et al., 2017; Helman et al., 2017). Of note, Geva et al. presented data on 92 consecutive NSCLC patients. 38% (35/92) performed ctDNA analysis before 1st-line therapy and 62% (57/92) on progression. ctDNA analysis yielded lung cancer related actionable mutations in total 39% (36/92) of the patients; 31% (11/35) among upfront testing and 44% (25/57) among patients at progression on matched therapy. Treatment decision was changed to targeted therapy subsequent to Guardant360 analysis in 23% (8/35) and 26% (15/57) respectively (total 25%; 23/92). This larger cohort study confirms earlier reports of meaningful impact on physician decision making in two previous studies (Kim et al., 2017; Rozenblum et al., 2017b).

In conclusion, plasma-based CGP with Guardant360 has demonstrated improved outcomes while reducing the harm arising from both repeat invasive biopsies and empirical administration of chemo- or immunotherapy in undergenotyped patients who, in fact, have targetable mutations. The clinical validity and utility evidence has significantly evolved from the initial draft LCD in May of 2017 and is now well established. We believe the evidence supporting Guardant360 meets Medicare’s definition of reasonable and necessary for the coverage requirements described in the draft LCD without a registry. For your convenience, we have attached a redlined LCD with the proposed edits from this comment letter and the earlier Noridian comment letter.

Thank you for the comments. (These comments were submitted before the later received comments by this party. Many of these recommendations have validity and have resulted in changes to the final LCD.)

17

I am writing to fully support finalizing the draft LCD DL37699 and to request a minor increase to the patient coverage profile. Guardant360 is a validated test with dozens of publications supporting the clinical utility, clinical validity and analytical validity. I treat patients according to the NCCN guidelines; I strive to ensure that the molecular profile of my patients’ tumors is understood at diagnosis and as the tumor mutates over time under treatment pressure.

I have been using Guardant360 since 2015 and I have ordered >60 tests for patients with NSCLC. The test has been a vital tool which I rely on to assist making the right therapeutic selection. I am writing to request that all patients who have progressed on targeted therapy be included in the LCD. As the lead author of the AURA1 study which explored the use of osimertinib in first line treatment of EGFR mutated NSCLC patients, I can attest that therapeutics for NSCLC are evolving at an incredible pace. To ensure that patients have appropriate access to care I would suggest that you alter the progression coverage criteria to include all patients progressing on a TKI, including osimertinib. As osimertinib moves to first line therapy, inevitably other drugs will be determined to be effective to treat in the second line upon progression. Additionally, EGFR TKIs are not the only TKIs that should be included, as patients with ALK-driven NSCLC should be re-assessed at the time of progression to choose the most appropriate second or third generation ALK inhibitor.

Thank you for your comments. The policy has been altered to allow coverage for any patient progressing on targeted therapy (see previous comments/responses).

Associated Documents

Related Local Coverage Documents
LCDs
L37671 - (MCD Archive Site)
Related National Coverage Documents
N/A
Public Versions
Updated On Effective Dates Status
08/21/2018 10/16/2018 - N/A Currently in Effect You are here

Keywords

N/A