Sunday, March 20, 2016

California Q&A on Lung LCD (February 2016)

MolDX - CDD: NSCLC, Comprehensive Genomic Profile Testing Comments and Responses

Comment 1: 
  • Scientific literature demonstrates actionable drivers that may be identified by multiplex/NGS are common in smokers; since finding a target mutation radically alters the treatment for patients, strong consideration should be given to broader coverage;
  • Alteration in different genes by CGP may inform physicians about clinical trial or off-label options that may benefit patients;
  • Since coverage applies only to never- and light-smokers with metastatic disease, under 24% of 26,000 diagnosed NSCLC patients with negative EGFR and/or ELK mutations are eligible for CGP NGS, especially when the rest of lung cancer patients stand to benefit by the application of comprehensive genomic profiling;
  • Allow coverage of other testing platforms if the detect the same classes of alterations with comparable performance as in the policy
Response 1:  Although mutations are common in both smokers and never-smokers, actionable mutations such as EGFR mutations, ALK rearrangements, and ROS1 rearrangements occur at a much lower frequency in smokers than in never- or light-smokers. The probability of finding an undiscovered mutation in smokers given its recorded frequency, and the probability that a mutation will be identified by CGP, is estimated to be less than 1 in 400. This would dilute the potential for this approach to replace alternative testing as the first step.  Furthermore, it has been shown that heavy smokers do not have the same response to EGFR inhibitors in terms of progression free survival (PFS)*. Never- or light-smokers with metastatic disease are a logical first step for coverage because CGP testing has the highest impact on care in the patients who are most likely to have actionable mutations.  This is the premise of the Drilon paper; it is the cornerstone of this policy.
Moreover, this is not a policy about NGS testing.  Rather, this policy is about a specific type of NGS-based testing (i.e., tissue-only somatic comprehensive genomic panels) in patients who are at higher risk of having mutations, despite exhaustive LDT testing that has failed to identify any mutation, and requires the collection of outcome data to inform future patient management decisions.  Although mutations were found by CGP in patients (Drilon), it is unclear that they will respond to treatment.  Thus, CDD in a specific population (never- and light-smokers) allows "proof of concept" before Medicare coverage can be justified for larger patient populations.
*Kim MH, Kim HR, Cho BC, et al. Impact of cigarette smoking on response to epidermal growth factor receptor (EGFR)-tyrosine kinase inhibitors in lung adenocarcinoma with activating EGFR mutations.  Lung Cancer 2014;84 (2):196-202. doi: 10.1016/j.lungcan.2014.01.022. Epub 2014 Feb 3.
Comment 2:  Expand genes in NGS panel
  • Request clarification to assist in triaging patients for such testing;
  • While targeted exon sequencing using a massive parallel sequencing (MPS) platform is not comprehensive genomic profiling, a commenter notes that the MPS to assess EGFR, KRAS and BRAF has identified 8% more mutations than real-time PCR previously used.  By combining MPS on more than one cancer type, they have constructed a single panel with cost and time-savings, as well as tissue savings compared to serial gene testing; request clarification that a small, targeted panel using MPS is an appropriate first-line test.
Response 2:  In addition to SNVs and/or indels in EGFR and KRAS, and rearrangements in ALK, recent NCCN guidelines for NSCLC (v5.2015) indicate that the following alterations have "emerging targeted agents": ROS1 rearrangements, BRAF V600E mutation, MET amplification, HER2 mutations, and RET rearrangements.  The policy will be expanded to include data collection for these additional genes.
Medicare is not specifying technologies or platforms for first-line testing.  This policy specifies that CGP will be covered to identify a small group of patients who have already tested negative for EGFR mutations, ALK rearrangements and ROS1 rearrangements
Comment 3: Need NGS upfront due to tissue availability; also Circulating Tumor DNA (ctDNA) testing; Sequential testing algorithm may harm patients by delaying therapy and exhausting tissue samples - A commenter notes that "As many as 25% of NSCLC patients do not have adequate tissue material for molecular testing. The commenter has developed a nine gene ctDNA test, including alterations in EGFR and ALK for use in these patients that it intends to launch later in 2015. The way this LCD is currently worded patients who lack adequate tissue for EGFR and ALK testing would not be eligible for CGP testing even though this would be possible using these new approaches. In addition, the characteristics of ctDNA testing will be different than tissue based NGS testing.  Because ctDNA levels vary greatly in patients and some patients do not release appreciable amounts of ctDNA, ctDNA testing will not be as sensitive as tissue testing. Additionally, although the critical factors defining ctDNA testing performance are still similar to tissue testing the detectable MAF (or ctDNA level) is much lower with a high-quality ctDNA test.  In our view these MAFs for ctDNA should be between 0.01% and 0.05% for mutations and >0.1 % for translocations. It also should be noted that copy number amplifications while detectable in ctDNA are typically only detectable at MAFs of 1%-2% unless an extremely sequence intensive technique is used (this enables detection down to levels of >0.1%), and recommends that The MolDX Contractor specify that this guidance is specific for tissue-based CGP testing and would not apply when the patient has inadequate material for CGP testing.
Response 3:  Noridian recognizes that availability of tissue is very important.  However, this is a limited coverage policy with specific data development expectations for coverage (CDD) focusing on a specific sub-group of patients who have already been tested and found to be negative by other technologies AND have tissue available for testing.  As stated above, this policy is NOT about NGS testing more generally. The premise of the Drilon paper is that never- and light-smokers have a higher likelihood of having less complex driver mutations than their smoking counterparts.  This premise is well founded with some patients having mutations that were missed by an LDT done at the institution.  However, short-comings of the study were: its small size, predominance of never smokers, potential patient selection bias due to the ability to re-biopsy, and limited follow-up on treatment outcomes.  In addition, because there was no direct comparison to an FDA-approved companion diagnostic, we cannot directly extrapolate from the results of testing to improved outcomes, thus the need to collect data.    Because the outcomes of these patients are still unknown, we must follow the inclusion criteria that were used in the study, thus the limitation to never- and light-smokers.  Similarly, since CGP testing was limited to patients who previously tested negative for EGFR mutations, ALK rearrangements and ROS1 rearrangements, the policy is based on the same testing schema.  Noridian expects that as outcomes data are collected and analyzed, CGP may become first-line testing, and then the issue of specimen availability will be moot. At the current time, CGP is limited to patients with available tissue for testing and does not apply to ctDNA testing, so the policy will remain unchanged. 
Comment 4:  Limit of Detection:  In the proposed LCD, The MolDX Contractor indicates that the studies cited demonstrate test performance of 95%-99% sensitivity across the four alteration classes and >99% positive predictive value (PPV). However, this mention omits the fact that sequence mutation sensitivity and PPV were demonstrated at mutant allele fraction (MAF) levels as low as 5%. MAF is a critical parameter in measuring test performance for the analysis of sequence mutations. It is orders of magnitude easier to have >99% PPV at 50% MAF than at 5% MAF. In addition, a large fraction of mutations occur at lower MAF ranges in the average sample (54% are <20% and 26% are <10%). A test that does not demonstrate these performance characteristics at low MAFs will be significantly inferior to tests that do and miss many mutations that may otherwise be clinically actionable. The commenter recommends that The MolDX Contractor specify that sensitivity and specificity metrics must be achieved at MAFs as low as 5% in order to be considered for coverage.
Response 4:  We agree and have published the MolDX analytical requirements, linked in the policy.  In addition, we note that the clinical significance of mutations detected at very low allele frequencies in adequate biopsy specimens is (to our knowledge) currently unknown.
Comment 5: Differentiating between Somatic and Germ-line mutations - A commenter notes that many CGP tests do not utilize matched patient normal D|NA. The commenter believes the lack of patient normal sample results in sub-optimal result quality because there is no way to be sure that the identified mutations are somatic (tumor-specific) in nature without analysis of the matched normal DNA from the same patient.  They note that "taking action on these mutations not only deprives the patient of the opportunity to receive another more efficacious treatment option but also may do harm to the patient. Note that tests that accurately distinguish between somatic and germline mutations do not need to report any germline mutations."  The commenter recommends that The MolDX Contractor reword the first sentence of the CGP test description to define CGP testing as "testing that accurately distinguishes between somatic and germline mutations."
Response 5:  The LCD has been revised to clarify that it applies to tumor tissue-only testing and somatic-only reporting. 
Comment 6:  The MolDX Contractor asks that reports be submitted according to HIPAA standards; Urge lab reporting requirement to include all actionable mutations
Response 6:  Noridian expects patient line-item specific information and to include all actionable mutations. The policy has been revised accordingly.
Comment 7:  What specific billing code applies? 
Response 7:  An LCD can only contain "reasonable and necessary" criteria.  Coding and billing guidelines will be attached to the final document accordingly to Medicare guidelines.
Comment 8:  Mulitplex/NGS Testing is no longer experimental.
Response 8:  While the testing may not be experimental, the multitude of genes tested does not meet the criteria of "reasonable and necessary". 
Comment 9: Lab Verification by MolDX:  Two commenters strongly disagree with the lab verification requirements imposed under MolDX and assert that CMS is the authority to regulate analytical validity (AV) pursuant to CLIA
Response 9:  CLIA certifies a laboratory and provides an assurance that the lab has Standard Operating Procedures, a Quality Control program, proficiency testing, qualified personnel, etc.  MolDX has the responsibility to assure the public that a given test has been adequately validated (i.e., is safe and effective) as part of the criteria to meet "reasonable and necessary". 
Comment 10: Clinical Outcomes Reporting Requirement:  the proposal outlined by The MolDX Contractor is scientifically, financially, and ethically not feasible for laboratories
Response 10:  We recognize the additional work required of laboratories due to the reporting requirements and have therefore made the reporting requirements as parsimonious as possible to allow the scientific and clinical communities to draw some meaningful conclusions or at least provide some data to support generating testable hypotheses. Most of the genes and variant types that are currently being tested on NGS panels do not have sufficient evidence to be "reasonable and necessary" under Medicare guidelines, hence the reporting requirements in an effort to accelerate our understanding of the clinical utility of testing for alterations in these genes. Such information will benefit all stakeholders, most especially patients and providers.  We believe our reporting requirements have been calibrated appropriately to balance the outcome expectations with the effort required. In addition, it is imperative to document that when comprehensive NGS analysis identifies actionable mutations missed by other methodologies, whether they are FDA-approved companion diagnostics or LDTs, the clinical outcomes are similar.
Comment 11: Support policy as written:  Multiple comments supportive of the policy. 
Response 11: NA
Comment 12:  A commenter requests information regarding how to find the MolDX contractor's published AV criteria and how to send information to the MolDX contractor. 
Response 12:  The MolDX program and contact information can be obtained via: This link will take you to an external website..
Last Updated Feb 25, 2016

No comments:

Post a Comment