Saturday, August 18, 2018

Foundation Medicine's Final SEC Report: Medicare Policy as a Chutes and Ladders Game

On August 9, 2018, Foundation Medicine (FMI) filed its SEC 10-Q for the quarter ending June 30. And on August 10, it filed its 15-12B: its termination of SEC registration and filing duties.

So, with FMI's acquisition into Roche, the August SEC document is the last view into its legacy of New England MAC policies, North Carolina/Palmetto policies, and journey through an NCD to ADLT status. 

I've stored a cloud copy of the 99-page 10-Q document here, with some highlighting.  Below, I use paginated page numbers rather than PDF page numbers (which are 1 higher).


What follows is my bullet-point notes harvested from the long document.  Some tidbits of policy you may need to know as background are at bottom [*].   For a Genomeweb discussion of recent conference presentations by CMS, FMI, Sloan-Kettering and others, August 24, see here.

  • Most reimbursement pages are at page 29 or pages 44-46.
  • Page 8 notes that Roche acquired FMI shares at $137.  
    • Historic share price was $36 on 8/14/2017, $65 on January 1, and about $100 on May 30.  
    • Accumulated losses circa $550M.  Market cap at the end was circa $5B.
  • Page 2-3 enumerate about 30 bullet-points of the risks regarding forward-looking statements being made by a genomics company:
    • [Risks include] The evolving treatment of cancer...any perceived advantage of our services...sustainability of our competitive advantages...outcomes of clinical trials...
  • Page 6: R&D for the final quarter was $25M or $100M for the year.  
    • For comparison, GHDX and INVITAE had about $50M each in research in the past year; Myriad around $75M.
  • Page 12 notes, "Historically for certain clinical customers, the company deferred revenue recognition until cash receipt..," but current accounting rules lead to estimations for such customers (e.g. BCBS and other payers).
    • Page 11 notes "Due to our out-of-network status with the majority of payors, estimation of transaction price represents variable consideration...estimate [with] a portfolio approach." [**]
  • I believe that page 14 ff [Section 3] describes transactions or projects that overlap specifically with Roche.   
    • These include an earlier $250M investment by Roche; a $150M payment over five years for "access to its molecular information platform;" and enumerated special projects. The latter include cancer immunotherapy and signatures, an immunotherapy testing platform, liquid-biopsy-based TMB, and a ctDNA project.  
    • At page 17,  FMI notes that it will have access to potential developments of F1 or other test properties as Roche Diagnostics IVDs.
  • Several current litigations noted at page 25.
  • Page 27-28 describes major business events such as: the F1 CDx NCD in March 2018, the FDA granting "Breakthrough Device" for FoundationACT (LBx), planned to have 70 genes, TMB, MSI.  Also discussion of a 3-party collaboration with Roche, FMI, and Dian Diagnostics Group for China.
  • One page 29-30, noted that "most of the commercial third party payors that reimburse us do so based upon CPT codes, or based on other methods such as percentages of charges or other formulas that, to our knowledge, are not specific to us and not made known to us."   
    • Read that again!  A huge, powerful, globally-known services company is paid based on methods and rules "that are not known to us."
    • Some claims are processed via third-party (not the primary issurer) entities. (Believe this refers in part to the laboratory benefit management (LBM) industry.
  • As of June 30, not a participant in any Medicaid programs (page. 30).
  • Regarding the local MAC in New England, FMI reminds us that they deferred submitting claims until November 2013.  
    • An LCD at the NGS MAC in New England provides hotspot coverage for 5-50 genes in lung cancer, but "we do not believe this LCD reflects coverage for our services, which include comprehensive analysis of >50 genes and all classes of alterations."  [***] 
    • Adding, NGS MAC "has either denied the F1 or F1 Heme claims we have submitted using stacked codes, or not processed and reimbursed us" in a manner we believe is consistent with applicable processing guidelines.  
    • However, "In August 2016, we began submitting claims for FoundationACT" from Cambridge to NGS MAC with stacked CPT codes, and we have recognized revenue from many of these claims."
  • In January 2017, FMI began submitting claims for advanced lung cancer to Palmetto [MolDx] MAC from lab in Research Triangle Park, paying $3416 per test.  
    • LCDs for other cancers (melanoma, colon, ovarian" have been pending since 12/2016 and not finaled.
  • There is "a current lack of broad coverage decisions among commercial third party payers."
  • The document returns to reimbursement issues beginning on page 44.
    • General remark: Commercial third-party payors and government payors are increasingly attempting to contain healthcare costs by lowering reimbursement rates, limiting coverage of diagnostic services, and creating conditions of reimbursement, such as requiring participation in clinical evidence development involving research studies and the collection of physician decision impact and patient outcomes data.
  • On page 45, note specifically that F1 CDx has two intended uses, one as a CDx (for a number of genes and drugs) and also, one as providing "tumor mutation profiling" [not specifically or only TMB] for professionals to use with cancer patients.  This is important because the CMS covers NGS tests "for intended uses" (with various other requirements). 
  • From March 16-June 30, 2018, FoundationOne CDx [the FDA-approved version highlighted by the NCD in March 2018] was run and billed from Cambridge MA and paid by NGS MAC at $2919.  
    • As of July 1, the F1 CDx will be an ADLT and therefore paid more, at $3500, the current "list charge."
  • The next paragraph is a little convoluted but states it applies "to NGS tests other than those that have been approved or cleared by the FDA."   
    • F1 [non FDA version without the acryonym CDx] will continue to be provided and paid by MolDx/North Carolina, consistent with an NSCLC LCD.   
  • On page 46, they state, "Although we are performing components of our test service for FoundationOneHeme in our North Carolina facility, Palmetto has provided guidance that [CGP] not covered by an LCD is explicitly non covered including FoundationOneHeme."  
    • Adding, "We are still in the process of determining what types of services we may conduct at this {NC} facility. "  
  • Continuing:  "In parallel, we have been engaged in conversations with Palmetto regarding the potential for coverage of FoundationOne claims ...for tumor types other than NSCLC."  
  • A discussion of 14 day rule and changes on page 48.


_____

From page 52 forward, the SEC report has a very long presentation regarding FMI's supplies & service agreement with Illumina.
___

[*]
   It appears that FMI bills its FMI F1 CDx test from its Cambridge, MA [New England] lab, to one MAC, and its F1 non FDA tests via its North Carolina lab to another MAC.
   The MolDx MAC has a longstanding LCD covering comprehensive genomic profiling (CGP; the F1 test) for lung cancer.  For over 18 months, proposed policies for coverage in ovarian, melanoma, colon, have existed but are not finalized.
   The March 2018 NCD for NGS testing in advanced cancer covers NGS tests if they are FDA approved, if they have at least one combination diagnostic, and if they have an indication in the specific patient's cancer.   To my reading, this covers the F1 CDx test in lung cancer, because there is an on-label indication in lung cancer (ALK testing in lung cancer).  But the NCD text also covers F1 CDx test in any solid cancer, because the FDA provided an additional, more generic FDA indication for use as "genomic profiling" for use by physicians in solid tumors.

[**]
The discussion of an estimated portfolio of value in payer invoices submitted sounds like tranches of commercial debt; some AAA bonds, some junk bonds.
[***] 
Codes 81445 and 81455 differ in number of genes (being 5-50, and 51+), but are identical in other respects (e.g. classes of deletions).

____

While we don't know the distribution of price points paid by those cryptic payers to FMI, we do not the price point distribution for codes 81445 and 81455 in CY2016, from primary PAMA data.  The prices are highly non Guassian.  Bar height shows the "N" of payments to US labs for these codes at various price points from $100 or less, to >$5000.  Data is from 1H2016.