Tuesday, December 26, 2017

NEW 20 PAGE WHITE PAPER ON CMS NEXT GEN SEQUENCING NCD

On November 30, 2017, CMS released a 65-page, 45,000 word NCD proposing an entirely new policy approach to all uses of next generation sequencing in oncology.

I've had the chance to talk to a number of stakeholder groups and read the NCD several times, as well as underlying documents such as the FDA approval review for FMI F1 CDx. 

Here, I post a 20 page (7000 word) white paper on the NCD and its potential implications.

Version: 03, January 2, 2018.  * HERE *  - click arrow upper right to download.


This white paper was downloaded 450 times (December 25-January 5).   More after the break.

Sunday, December 24, 2017

Friday, December 22, 2017

Very Brief Blog: J Mol Dx Article on History of Genomic Reimbursement & Coding

New online at Journal of Molecular Diagnostics (included with Association for Molecular Pathology membership).   Hsiao, Mansukhani, Carter, and Sireci of Columbia University have written an in depth study of actual changes in reimbursement (and coverage) in their molecular laboratory as policy, coding, and fee schedules have rapidly evolved in recent years.

The in press manuscript is online here.  Don't miss the interesting graphics from page 20 forward.

ABSTRACT

Changes in coding and coverage create uncertain reimbursement environment for molecular pathology laboratories. 

We analyzed our experience with two representative molecular oncology tests: a T-cell receptor (TCR) beta rearrangement test and a large (467 gene) cancer next-generation sequencing panel the Columbia Combined Cancer Panel, CCCP. 

Prior to 2013, the TCR beta test was coded using “stacked” current procedural terminology codes and subsequently transitioned to a tier 1 code. CCCP was coded using a combination of tier 1 and 2 codes until 2015, when a new Genomic Sequencing Procedure code was adopted. A decrease in reimbursement of 61% was observed for the TCR beta test upon moving from stacking to tier 1 codes. No initial increase in total rejection rate was observed but a subsequent increase in rejection rates in 2015 and 2016 was noted.

The CCCP test showed a similar decrease (48%) in reimbursement following the adoption of the new genomic sequencing procedure code and was accompanied by a sharp increase in rejection rates both on implementation of the new code and over time.

Changes in coding can result in substantial decreases in reimbursement. This may be a barrier to patient access due to the high cost of molecular diagnostics. Revisions to the molecular code set will continue. These findings help laboratories and manufacturers prepare for the financial impact and advocate appropriately.

Sample graphic:

click to enlarge

Thursday, December 21, 2017

Health Executive Adam Boehler Might Head CMMI Next

According to a tweet from PoliticoPro, health executive Adam Boehler might become the new head of the billion dollar Center for Innovation at Medicare (the CMMI).   Tweet here.


Boehler's Linked In is here.  He's also listed at Bloomberg biographies here, and his profile as CEO of Landmark Health is here.

Boehler comes out of the Wharton School and has had both venture capital and operational roles.  He was previously involved with a 25-lab network with 3000 employees, aLabs.  According to Bloomberg, he's Chair and President at Landmark; Executive Chairman at Avalon Health. 

Landmark Health, where he's currently CEO, is designed to provide superior, next-generation home health care and home based medical care to patients "24/7/365."   Earlier this month, Landmark recently hired Christopher Goldsmith (of Optum and with an INSEAD MBA) as President.

click to enlarge

Previous CMMI director Dr Patrick Conway stepped down in mid-year to join North Carolina BCBS.  CMMI is rebooting itself for new directions under the new administration.

Feds Release Surprisingly Affordable Report on Excessive Drug Prices

On November 30, 2017, the National Academy of Medicine released a free 200 page report on the problem of excessive drug prices. 

Drug costs have been a concern for generations; Senator Kefauver of Tennessee began hearings on drug price-fixing when he was told (in 1955!) that two different antibiotics at his corner pharmacy both cost exactly $3.    (This vendetta eventually transformed and morphed into the Kefauver FDA amendments of 1963). 

The NAM press release is here and the PDF report is available here.  It comes with a one hour panel media presentation as archived video, here.

Coverage at:
This National Academies project is different than a mid November GAO report on drug profits compared to R&D spending; see the 70 page GAO report here and the Endpoints blog on it here.


Bonus reading.

For Bloomberg December as "Diabetes as the Messiest Drug Market," here.   Also from Bloomberg in December, "FDA Targets Insulin" for excessive prices, here.  Fpr a summer article at Bloomberg on "Crazy Math Behind Drug Prices," here.

CMS Extends Comment Period on NGS/Cancer Proposed NCD

CMS has extended the comment period to January 17, 2017, on the currently debated National Coverage Determination for all uses of next generation sequencing in cancer.   The NCD provides coverage only for PMA-approved tests used on-label in the relevant cancers.   The NCD would cover patients who are part of clinical studies (coverage with evidence developments) for other uses of PMA- or 510(k)-approved NGS tests.

The tracking and dates page is here.   The NCD is here.

Comment deadline has been revised (old, Dec 29, new Jan 17)

Tuesday, December 19, 2017

Regulatory Nerd Corner: Medicare Review Board Decision Narrows Scope of ALJ Review of LCDs

A section of law  called BIPA 522, more than a decade old, gives a channel whereby a beneficiary can appeal an entire LCD rather than just his personal claim.   The decision is fast-tracked to an Administrative Law Judge (ALJ).

In past cases, ALJs have ruled that there can be "constructive LCDs" - for example, if a contractor says in an Article that Service XYZ is not reasonable and necessary and is non-covered, that may not be formally called an LCD by its MAC author, but it is a de facto LCD, and can be appealed through the BIPA 522 channel.

A 2017 Medicare review board case reversed an ALJ who had broadly construed his ability to call something a constructive LCD based on a reasonable and necessary decision.  In the case, a DME MAC had ruled that continuous glucose monitors were "not a benefit category" because they were "precautionary"  to check for low glucose levels.  The ALJ ruled this was equivalent to saying they were not reasonable and necessary and found the CGM should be covered.   CMS escalated the case to a panel of review judges, who ruled that the decision of the MAC was a "benefit category decision" and therefore not accessible by the LCD review process, which is only for "reasonable and necessary" decisions.

  • I describe in more detail here and provide links.   
  • DME benefit category decisions  can have close overlap with reasonable and necessary decisions because categorizing something as DME invokes some steps that involve its medical use and purpose.

The potential adverse affect is for devious MACs to couch decisions in articles rather than in LCDs and switch to terms like "experimental and investigational" or as "precautionary" or as "benefit category decisions" with an eye to avoiding the textual keywords that would allow review by ALJs under the LCD review channel.   I never understood the reasoning around "precautionary" - all glucose tests are precautionary for an abnormal glucose finding, INR tests in warfarin patients are "precautionary" for an abnormal INR finding.   An MRI in a unilateral headache patient could be styled as "precautionary" for discovering if he has a brain tumor, and so on.

The particular decision about CGMs is of historic interest now as later CMS decisions allow coverage of some CGMs.

My Article on "DIGITAL GENOMICS" Is UP at Journal of Precision Medicine

Over the past year, I produced a white paper and gave several talks and chaired two panels on "Digital Genomics."   A corresponding article is now online (open access) at Journal of Precision Medicine.

See the journal homepage here.   See the PDF article here.

The article has two parts.  The first part talks about labs that are creating an elaborate digital ecosystem around the traditional molecular lab services - "Lab 3.0."  The second part talks about companies that are being established without any wetlab, and dedicated to providing digital genomic services like state of the art outsourced software, analytics, and bioinformatics pipelines. 




click to enlarge
This continues to be an active area of investment, e.g. a $58M investment in DNANexus was announced on January 2, here.

Monday, December 18, 2017

Very Brief Blog: FDA Releases Two Draft Guidances re Biomarkers for Rare Subsets and IDE IVDs in Clinical Trials

One December 18, 2017, the FDA released two draft guidances (both open for 60 days comment) on precision medicine.
  • The FDA's press release is here.
  • Coverage at Genomeweb here.
  • The FIRST guidance, on biomarkers for rare disorders in clinical trials, is here (8pp).
    • This codifies some of the principles used in the May 2017 approval of Keytruda in all solid cancers with high MSI (here).  The pivotal labeled study involved 90 colorectal cancer patiens and 59 other patients, such as 11 biliary, 9 gastric, and as few as 1 or 2 for a number of cancers.  
    • It was far from conventional "proof" that Keytruda works in "all" solid cancers.
    • This guidance is paired with and timed to an FDA article in Nat Rev Drug Devel (12/2018) by Schuck, Pacanowski, Woodcock, and Zineh, here The FDA guidance and the Schuck et al. article are mutually explanatory.
  • The SECOND guidance, on requirements for Investigational Device Exemption when biomarkers are used in clinical trials, is here (26pp).  For example, an FDA approved KRAS biomarker used off label in a different tissue trial may require IDE.
    • To my eye FDA has already offered this rigorous view of the need for IDE status in clinical trials, but FDA pointedly notes that perhaps not all IRBs and investigators got the point.
    • FDA immediately tags this upgraded regulatory enforcement by adding it is looking to "streamline the review of oncology products and [oncology] IVDs" through additional guidance and in the near future.

Regulatory Nerd Corner: The Regulation for the MSK IMPACT test does not exist in the CFR

(Updated)

For regulatory nerds, the MSK IMPACT test, recently cleared by the FDA as a de novo Class II test, has a product code (PZM) and a regulatory classification, 21 CFR 866.6080.

The Class II decision memorandum is here (DEN170058).   The product code PZM webpage at FDA is here.

  • However, the US Code of Federal Regulations or CFR has no regulation under 21 CFR 866.6080.  Regulations stop at 866.6060.  Here.
  • Yet, the FDA has a webpage saying that 866.6080 was created on April 1, 2017.
    •  (Ha ha?  Not only was that April Fool's Day, but it was a Saturday, when the Fed Reg doesn't publish).  Here.  
  • In short, the regulation as numbered on the FDA CFR webpage at FDA fail to foot to any actual regulation in the real Code of Federal Regulations, which would be here.    
The solution to the puzzle is not entirely satisfying, but here it is.  In order to clear a product as a de novo 510(k) product, there has to be a classification order which means there has to be a 510(k) category to put the product "into" such as 21 CFR 866.nnnn.    (See the FDA guidance on de novo classification processes, October 2017, here.)     So yes, the FDA internally created 21 CFR 866.6080 (sic), but will (eventually) publish an article in the Federal Register creating that regulation, at which point the new category will appear in the national CFR that everyone can access.  In the guidance (UCM080197), FDA says (a) it will issue the order to the sponsor and classify the device in that letter, and (b) "we will then publish a final order in the Federal Register, etc.   So it gives the appearance that the "regulation" exists on the FDA website before it appears in the FR or CFR.  And in this case, the FDA regulatory creation is dated April 2017, the Sloan Kettering first example of a product is November 2017, and the Federal Register publication will be December 2017 or in 2018. 

I suppose a Sherlock Holmes FDA sleuth could have figured out that something big was up if they tracked the sudden mystery appearance of 866.6080 on the FDA website as a Class II gene panel test last spring, but he wouldn't have known it would contain oncology genes that are otherwise PMA level CDx biomarkers.



FDA.gov web page for the regulation, 21 CFR 866.6080 ("Created April 1")

Actual 21 CFR 866.nnnn Section at Federal CFR (Dec 18 2017)
____

FDA has a page on procedures for Direct Final Rulemaking, but it is predicated on publication in the Federal Register, which I've been unable to find for 866.6080.  Here.

FDA Commissioner Gottlieb was sworn in May 11, 2017.



Very Brief Blog: FDA Posts the Approval Documentation for Foundation CDX

UPDATE - MONDAY - DECEMBER 18 2018

FDA HAS POSTED APPROVAL PDF's ON FMI CDX



See also all documents in one cloud zip file (Dec 18):



The Approval Order is 5pp; the Safety & Effectiveness document is 58pp; the Labeling is 43 pp.   




Four conditions of approval include (1) "clinical concordance data to support the performance of your device within the appropriate clinical contexts." [?]  (2) Response for NSCLC patients with EGFR T790M mutation mutant allele frequency (MAF) < 5%.  (3)  "Provide additional results from clinical samples to establish the analytical performance characteristics "for all variant types and genomic signatures." This includes MSI, TMB, CNA, rearrangements.   (4)  Software documentation for: validating and implementing software design changes.

The FDA specifically segregates "test output" or reporting into Category 1 CDx claims, Category 2 output of genes with "evidence of clinical significance," and Category 3 claims of "potential" clinical significance (the latter including e.g. animal studies or other rationales).  See FDA explanation here.  (510K gene panels have only Category 2&3 claims.)

The test runs on an Illumina HiSeq 4000.  It uses paraffin block inputs and must have 55 ng of genomic DNA.   The S&E documentation includes a "Table 5" which lists all FDA approved CDx (in any format, e.g. FISH) that parallels a CDx gene on the FMI label.   Table 7 compares concordance of "F1 LDT" and "F1 CDX" (it's high.)    There are many tables of comparisons to prior PMA tests (therascreen, cobas, etc).   The product was not referred to an advisory panel "because the PMA substantially duplicates information previously reviewed by the panel."   All statistics were based on "non inferiority testing."   

Since the device was classified as a Breakthrough Device, the FDA's review expectations included "a balance of pre and post market data."   The FDA concludes that because F1 CDX was "non inferior" to existing CDx tests, it "does not  introduce additional risks."  (Thus the reporting of Category 2 and 3 results, unavailable from prior single CDx's, were seen as not being germane to additional risks.)   FDA concludes: "The probable benefits outweigh the risks."  [Quotes from S&E document, P170019B.)

_____

FMI's own FMI CDX webpage is here.

Sunday, December 17, 2017

Very Brief Blog: CMS Published CY2018 Full Final CLFS Fee Schedule

It seems anticlimatic after a hectic year of lab policies, unprecedented volumes of crosswalk decisions, PAMA crises and lawsuits...but in 1H December, CMS published the "final CLFS fee schedule" for CY2018.   It's hereDownload the zip file to access the CLFS Excel spreadsheet.

CMS has modified the format from prior years.  There are no longer 57 pricing localities; there is one national price.   All the prices are listed as effective "20180101," so that indicator will be more interesting as the years go by.   CMS now uses the letters "N" and "L" to indicate National or Local pricing, while all the codes with an "L" for local (gapfill) pricing are listed as 0.00 dollars.   Dollar rates are listed in the format 00072.19, for $72.19.   As always, QW stands for Clia Waived (or Qlia Waived?) tests, which have the same price as the standard price.

For additional details, see Transmittal CR10409 (here).  That PDF lists, for example, the final pricing choices behind every price that was crosswalked in summer 2017.   ("New code 81176 is priced at the same rate as code 81218,"etc etc).   New code prices will behave the same as the target code.  Meaning: If New Code is priced to a $200 target code that stays level at $200 each year under PAMA, so does New Code.  But if New Code is priced to a $200 target code that falls by 10% a year under PAMA rules, so does New Code.  By my count 116 codes were crosswalked (many were codes on the CLFS that had no utilization and thus no PAMA pricing). 

18 codes enter the 2018 local gapfill process, in this table:

18 Codes are Class "L for Local" Gapfill in 2018



Thursday, December 14, 2017

Very Brief Blog: At Mid Point of CMS NCD Comment Period, No FDA Documents on FMI CDX Test

UPDATE - MONDAY - DECEMBER 18 2018

FDA HAS POSTED APPROVAL PDF's ON FMI CDX


____________________
____________________

* Original Blog December 14 *

On November 30, 2017, CMS released a wide-ranging National Coverage Determination on next generation sequencing in all Medicare cancer patients, with comment period to December 29 January 17.  (The NCD will be edited and finaled by February 28).

The pivot point of the NCD is the newly approved FDA PMA Foundation Medicine CDx test, FDA PMA P170019. 

Confusingly, as of December 14, 2017, the FDA hasn't released any of its PMA safety effectiveness and validation review for the F1 CDx test.

See the product's FDA webpage here.  These webpages appear first as a placeholder with product name and approval date and high-level label (indication).   Thereafter, multiple FDA lengthy review documents are linked on this same page, as well as supplemental PMA documents (sPMA's).   (For a mature FDA webpage, see the ThermoFisher Oncomine Target Dx Test, PMA P160045, here).

Lack of FDA documentation means only publications on the LDT era FMI CDx test are available.   These LDT tests are non covered under the newly proposed NCD (except in NCI trials.)   This makes it hard to comment on the NCD proposal. 


Wednesday, December 13, 2017

MOLDX: Apparently Cancels Instructions for Modifier 22 Billing for PMA CDx Genetic Tests

For several years, MolDx had special instructions if you billed oncogenes like KRAS or EGFR with FDA approved tests:  Add modifier 22.    While MolDx never published fee schedules for added payments triggered by modifier 22, CMS open payment records by lab for 2015 show that some labs, like Genoptix, were paid about 2X the fee schedule rates for some tests (here).

Policies have changed.   The old MolDx covered test instruction page was "Approved Gene Testing, M00041, V16," and it's gone.  (For a November 2 archive of how it was, here.)

The new equivalent article is, "Approved Molecular Tests for Reimbursement, M00149, V1."  It lists "routine" genetic CPT codes that are considered payable, and it also lists those genes inside of Tier 2 codes that are considered payable.  Special instructions about -22 coding are now gone.

Concurrently, MolDx has published coding articles for FDA EGFR, KRAS, and BRAF tests (here, and screenshot below.)  The articles lack the prior Modifier -22 instruction.

click to enlarge

MolDx Likely to Use Flat Tier 2 Prices in 2018 Under PAMA 

It looks like MolDX has also deleted its quirky schedule of very specific and elastic gene by gene prices for Tier 2 codes (cloud archive here) and presumably will price Tier 2 at flat PAMA prices in 2018.

Multi Gene Discount Article Still In Effect

Still active is the MolDx "panel alert" article that if you bill more than two gene CPT codes they should be replaced by 81479 and a Z code, presumably for discount pricing.

The CMS Proposed NCD for NGS Testing in Cancer: Playing It Out for MSI & Keytruda

On November 30, CMS released a proposed NCD that regulates the use of NGS testing in advanced cancers.   In brief, the NCD provides three tiers of coverage:
  1. PMA Tests Used for On-Label Cancers.
    1. These tests are covered for advanced cancers if the cancer is on label for at least one gene.   Foundation One CdX is covered for paients who have either breast, melanoma, lung, or colon cancer [at least one gene is CDx in each].   ThermoFisher Scientific Oncomine Target Dx is covered patients with lung cancer.
  2. PMA Tests Used in Off-Label Cancers *OR* all uses of 510(k) Cleared Tests (such as MSK IMPACT)
    1. These tests are covered only by concurrently enrolling patient in a detailed, long term clinical registry (CED) including ongoing RECIST objective response and PFS imaging.
  3. All Other Tests (e.g. LDT NGS)
    1. Covered only in NCI trials.
How would the NCD play out for solid tumor patients and Keytruda?

In May 2017, FDA approved immuno-oncology therapy Keytruda for all patients with solid cancers.  While there are several test methods (NGS, PCR, immunohistochemistry), none to my knowledge are FDA cleared.  Nonetheless, use of the cancer drug is on label if used with CLIA MSI testing.  In short, currently, FDA approves a major cancer drug for several hundred thousand US patients without an FDA PMA approved diagnostic.  

Patients can get (1) non-FDA approved immunohistochemistry testing, (2) non-FDA approved PCR, CPT 81301) or (3) non-FDA approved NGS MSI (MSI is found on the Foundation One test but not with CDx status).   For example, the MolDX article for test coverage is here.

Under the NCD, patients with breast, lung, colon, or melanoma have access to MSI testing under the Foundation Medicine test, since the test has one approved gene in each cancer, even though MSI is a supplemental report, not a CDx test.   

Using deaths from cancer as a proxy for advanced cancer, I assumed for sketch purposes that half of cancer deaths were in the Medicare population.  This yields the numbers shown below, assuming the Foundation CDx test is paid at $2900 (81455) and the MSI PCR test (CLIA) at $356.    

64,215 patients with advanced lung, breast, colorectal cancer or melanoma could get the FMI test for MSI, at $186M.   (Had they got the PCR test 81301 for MSI, which CMS also covers, $22.8M).   

More interesting is the other cancers - endometrial, renal, liver, etc.   Here, about 71,470 Medicare patients could get either the FMI or IMPACT tests, at $2900 each, for $207M.   Each patient would cost perhaps $10,000 in CED, for an additional $714.7M, totalling $921.9M.   (Had these 71,470 patients had PCR MSI testing and no CED, costs would be $25.4M.)   

click to enlarge
Before the NCD, total cost of testing of all patients would be $48M under CPT 81301 PCR.   With the NCD, total cost of testing (and CED) is $1.1B, with about $400M being cost of testing for MSI.  

How Much Does CED Cost?  Who pays?

The CED required in the NCD is pretty elaborate, such as RECIST imaging for objective response (OR) and progression free survival (PFS), the latter of which could go on for a year or more.   In addition, I'm assuming that it's not rigorous PFS without monthly imaging.  The imaging would not be required for clinical care, so it would not be covered by the Clinical Trials NCD.   So let's say it is $10,000 per year. 
Lab?
No regular lab could afford that (based on $2900 test payments). 
Wealthy Centers??
Maybe wealthy medical centers could fund the CED, for a limited number of select patients. 
Patient???
Since much of the cost is imaging, it could be coded as imaging and billed to patients under an ABN as "medically unnecessary services," but few patients could pay the cost. 
CMS????
I am not an attorney, but it seems like CMS might be able to cover the imaging costs under CED - the whole point of CED under SSA 1862(1)(e) is to cover non medically necessary costs incurred as expenses in health research.  However, so far CED has only been used when the service under CED is "the" medical therapy, and RECIST imaging is not "the" medical therapy under CED research.  However, it is an expense incurred in health research.  But if the imaging was coded with a modifier and paid by CMS, it would probably still trigger copayments.

Medicare Advantage
Rarely, NCDs trigger an exceptions clause in Medicare Advantage plans, which result in the service being billed to Part B even for Medicare Advantage patients.  It's hard to estimate utilization under the NCD (maximal cases are shown in the Excel above), but this NCD might trigger the exceptional expenses clause.