Yesterday, two Medicare administrative contractors—National Government Services and Cahaba Government Benefit Administrators—released draft local coverage determinations for next generation sequencing in advanced lung cancer.
The NGS LCD covers the initial diagnosis of lung cancer and Cahaba’s LCD covers patients who have been re-biopsied to test for additional mutations that may have been missed on testing of the initial biopsy by older technology.
These latest decisions confuse the Medicare landscape profoundly. They ignore a well formulated MolDX LCD published earlier this year that covers NGS in lifetime non-smokers who were previously tested negative for certain alterations. Rather than build upon MolDx’s approach, these rogue LCDs create a cacophony.
On the surface, yesterday’s decisions would seem to be a victory for patients and providers. However, on deeper examinations informed by understanding of complex technologies involved, these new LCDs are threatening to create coverage chaos.
If we are going to unlock personalized medicine, we need to move carefully forward in unison, learning from almost every patient. The haphazard patchwork of coverage that is starting to emerge, as highlighted by these draft decisions, reflects misunderstanding of technology and its clinical application, creating the potential for more harm than hope. If our national goal is to truly advance personalized care, we must begin by achieving unification of the differences among MACs.
Far from going down the path to personalized medicine, we are now in danger of spiraling down on a perilous trajectory from which there may not be an easy recovery.
Problems with the LCDs
1) Inconsistency in the Medicare Program: Arguably, the only MAC with the expertise to understand the nuances and application of molecular testing is Palmetto’s MolDX program. They were the first group to introduce a NGS coverage policy in NSCLC.
That policy, although not perfect, attempted to define the parameters of a high quality NGS test. In addition, also required that data would be collected to better understand the testing and its clinical outcomes.
This precedent was a good one: standardize testing, compare to previous knowledge, and collect outcomes to advance science. But, rather than build on and improve what Palmetto had started, both NGS and Cahaba decided to go in completely different directions and not follow the thoughtful precedent of Palmetto. Because testing is paid for where the lab geographically exists and not where the patient resides, a patient could have access to each of these policies as long as there is a lab in the jurisdiction that is covered by the MAC. Rather than consistency, there is utter chaos as to how to apply this new testing.
2) Lack of Support for FDA: Each one of the tests covered by these policies falls under the category of laboratory-developed tests.
FDA has been worried about the lack of consistency of LDTs especially when treatment decisions are based on the results of these tests. In 2014, the FDA announced to congress its intent to strengthen its oversight of LDTs. Margret Hamburg, then commissioner of the FDA said: “Ensuring that doctors and patients have access to safe, accurate and reliable diagnostic tests to help guide treatment decisions is a priority for the FDA. Inaccurate test results could cause patients to seek unnecessary treatment or delay and sometimes forgo treatment altogether. [This] action demonstrates the agency’s commitment to personalized medicine, which depends on accurate and reliable tests to get the right treatment to the right patient.”
By not attempting to approach anything but the most rudimentary standards of testing these new policies continue to perpetuate the Wild West of LDTs that the FDA has wanted to see better verified to ensure patients are receiving the highest quality care.
3) Non-Standardized Testing: NGS testing requires three components: a) pre-analytical preparation, b) sequencing, and c) data interpretation using complex algorithms. Each step in the process can lead to discrepancies depending on decisions made by the lab.
Currently, there has been virtually no comparison between testing from one lab to another. Even if both labs are of high caliber, they may disagree on what is the best method to approach each area.
Published data comparing certain platforms and their associated informatics, have shown a concordance of only 80 percent for the simplest DNA alterations, and less than 20 percent for more complicated alterations. False positives (finding alterations that do not really exist but are reported as being present) on the other hand, can make up as many as 50 percent of the alterations reported. As of such, a substantial portion of patients could receive treatment that they will not respond to due to being falsely told they have an alteration of erroneous reporting. In addition, these patients could be shunted from other valuable treatments that they could have received while chasing this analytical shadow.1–3
4) Increased Sensitivity Does Not Necessarily Mean Improved Outcomes: NGS can pick up alterations in DNA that only exist in a very small portion of the cells (low allele frequency). Depending on the sensitivity of an assay and the heterogeneity of the tumor specimen, only a small percentage of cells may contain the maker (as low as 1 percent) and yet be reported as being positive for an alteration that directly guides treatment.
Given the low percentage of cells, even if there is 100 percent response to the identified cells, this may leave the vast majority of the tumor untreated and the patient with a poor outcome. Yet, it is also possible that these low frequency alterations are driving the entire system. As of such, the increased sensitivity needs to be reviewed with outcomes to determine clinical benefit.
5) Limited Published Information: Most payers require that every new intervention show clinical benefit before being covered as standard of care. The NCCN recommendation for testing for these alterations is listed as “Emerging Targeted Agents.” The evidence for testing for these alterations is based on case reports (vemurafanib, BRAF V600E), ASCO abstracts (dabrafenib, BRAF V600E; crizotinib, MET amplification), and small studies. There are multiple ongoing trials that are waiting for final reporting on these agents.
As of such, even with a NCCN recommendation, there is not sufficient evidence to allow wholesale application. Furthermore, the NCCN has not discussed when to place these agents in the treatment sequence. Whereas EGFR, ALK and ROS1 have published data showing benefit when targeted therapy is used first line, it is unclear where to sequence these other agents. Ongoing clinical trials or outcome registries need to be finished before we will have complete answers.
Yet these emerging markers and other rare markers need to be analyzed, and we need to collect as much information as possible. Collecting data on every patient tested, allows these emerging targets to be verified for clinical benefit. Furthermore, there is no convincing published evidence that reanalysis of a tumor specimen through further biopsies can improve outcomes, and rather than re-biopsy a patient, it may be reasonable to save the morbidity and cost by analyzing the original specimen.4
6) ROS1/ALK Translocations and KRAS Testing: Although there are point mutations and small insertions in these genes that may create driver mutations, the vast majority of clinically significant alterations are translocations, and although these can be identified with specialized, high quality forms of NGS currently done by a few leading commercial and academic groups, most labs feel additional testing must be done to avoid false negatives.
ROS1 and ALK are listed as genes as part of the panel and rationale to allow coverage with NGS. In many cases the most common forms of NGS will miss the translocations if additional testing with fluorescent in situ hybridization (FISH) is not performed. Furthermore, unlike colon cancer, KRAS testing is not listed in the NCCN guidelines for lung cancer as a requirement before starting a tyrosine kinase inhibitor and there is published information that KRAS mutations do not harbor the same negative impact as it does in colon cancer.5-6
By allowing these genes to be tested by any form of NGS, many patients will be put in harm’s way, either missing appropriate testing to identify therapy, or inadvertently missing treatment based on erroneous understanding of KRAS in lung cancer.
7) Coding Issues: Many payers have regulations prohibiting payment for the same service twice, therefore if a payer covers NGS under CPT 81445 (5-50 gene solid tumor based on EGFR, ROS1, ALK, MET, BRAF, KRAS), and ROS1 and ALK have to be tested by FISH to look for translocations, and KRAS is thrown out as not being clinically actionable, the NGS gene number falls to three clinically necessary genes and below the 5-50 gene level and therefore should not be billed or covered. If the NGS can adequately replace the FISH testing, then it is a reasonable advance.
8) Lack of Information Collection: By allowing patients to receive testing and treatments that compete with existing trials and not collect any outcomes on these patients, we potentially place patients in harm’s way by receiving ineffective treatments or missing key toxicities that if captured could be published and lead to better understanding of disease. Off-label use of drug in oncology has been widely practiced, but only when there has been information already published. By allowing off-label use of drug when there is not a body of literature, we run the risk of hampering rather than helping advance science.
9) Expanding Panels of Unclear Merit: By allowing broad panels to be reported, it is possible that the extended information will detract from those things that are truly beneficial for a patient. For example, if an estrogen receptor was identified as a biomarker on a tumor specimen in NSCLC and tamoxifen was given as a therapy rather than standard chemotherapy, it is almost certain that the patient would have a markedly inferior outcome.
10) Inadequate Payment to Guarantee High Quality Testing: Final fee schedules for payers are still unknown or in comment period. Although there are labs that can technically analyze a specimen for a low cost of a few hundred dollars, it is likely that the quality of this testing will be suboptimal. Unless adequate payment to ensure high quality testing, it will likely be a race to the bottom to see who can analyze specimens for the greatest margin, not highest quality.
Solutions to the Problem: Coverage with Requirement for Data Collection through a National Registry or Registries
Although there are concerns with the coverage policies presented, it is clear that NGS is a powerful tool in the war on cancer and needs to be available to patients. Some could argue that we need to hold off on coverage until the industry develops standards and outcomes, but many of the alterations that NGS can identify are rare and can be missed by old technology.
As of such, it is best to introduce high quality testing in settings where information can be collected and analyzed. This requires the use of high quality registries defined by the following characteristics:
1) Require Standardization of High Quality Testing: By making sure that the results of one lab are consistent with another, high quality standards need to be developed and verified. The College of American Pathologists and the Association of Molecular Pathologists are both working on these standards and the FDA is also developing guidance. Until these standards have been developed and vetted, interim standards must be developed and upheld by international leaders in both private and academic settings.
2) Compare Back to Companion Diagnostics Where Available: Clinical progress only advances in small steps rather than giant leaps. As of such, new standards must be compared back to the existing body of literature established by specific testing that was approved in connection with targeted therapies (companion diagnostics). In cases where there is no approved companion diagnostic, the closest standardized testing should be used for comparison. This way we can tie old and new literature together and determine if new testing improves outcomes.
3) Collect Outcomes: High-level clinical outcomes need to be collected and compared back to the new testing standard. These outcomes need to be reported relative to not only the alteration found, but also allele frequency. In this way we can determine threshold response levels that have already been established as being crucial for other disease states like Her2 and estrogen receptor testing in breast cancer. Further, collecting outcomes on each patient greatly expands the understanding of disease especially in alterations that have not been fully catalogued.
4) Attach Testing to Existing and Future Clinical Trials: By creating a high quality standard that can be reproduced, it is possible that this testing can serve as direct inclusion criteria for current and future trials. This way we can markedly increase accrual to trials, especially for new drugs and targets.
5) Aggregate All Data, Nationally and Internationally: Place all the information in a searchable centralized database that can be used to identify trends and improve treatments in an open format. Doing so, everyone can jointly learn from each other and look for new signals and associations that can be used to advance testing and treatment options. Furthermore, rare alterations can be identified and tracked and hopefully patients can be treated on small trials.
1. Boland JF, Chung CC, Roberson D, Mitchell J, Zhang X, Im KM, et al. The new sequencer on the block: comparison of Life Technology’s Proton sequencer to an Illumina HiSeq for whole-exome sequencing. Hum Genet. 2013 Oct;132(10):1153–63.
2. Weiss GJ, Hoff BR, Whitehead RP, Sangal A, Gingrich SA, Penny RJ, et al. Evaluation and comparison of two commercially available targeted next-generation sequencing platforms to assist oncology decision making. OncoTargets Ther. 2015 Apr 24;8:959–67.
3. Jones S, Anagnostou V, Lytle K, Parpart-Li S, Nesselbush M, Riley DR, et al. Personalized genomic analyses for cancer mutation discovery and interpretation. Sci Transl Med. 2015 Apr 15;7(283):283ra53.
4. Drilon A, Wang L, Arcila ME, Balasubramanian S, Greenbowe JR, Ross JS, et al. Broad, Hybrid Capture–Based Next-Generation Sequencing Identifies Actionable Genomic Alterations in Lung Adenocarcinomas Otherwise Negative for Such Alterations by Other Genomic Testing Approaches. Clin Cancer Res [Internet]. 2015 Jan 7 [cited 2015 Jun 18]; Available from: http://clincancerres.aacrjournals.org/content/early/2015/03/09/1078-0432.CCR-14-2683
5. Roberts PJ, Stinchcombe TE. KRAS mutation: should we test for it, and does it matter? J Clin Oncol Off J Am Soc Clin Oncol. 2013 Mar 10;31(8):1112–21.
6. Fiala O, Pesek M, Finek J, Benesova L, Belsanova B, Minarik M. The dominant role of G12C over other KRAS mutation types in the negative prediction of efficacy of epidermal growth factor receptor tyrosine kinase inhibitors in non-small cell lung cancer. Cancer Genet. 2013 Feb;206(1-2):26–31.
The author is the CEO of the non-profit Molecular Evidence Development Consortium, MED-C, which was started to help address these emerging issues. Before resigning from Palmetto, he was involved in formulating that MAC’s local coverage decision for next generation sequencing.