Friday, September 18, 2020

After Five Months, MOLDX Authorizes Use of PGDx ELIO IVD Cancer Assay

Personal Genome Diagnostics (PGDx) got FDA clearance for its multi-hundred gene elio test in late April, 2020 - here.   Styled as clearance K192063, under product code PZM and classification 21 CFR 866.6080 (NGS tumor profiling), the elio test can be run locally on a standard platform.  (Product description clipped at bottom [*]). 

  • The 32-page FDA review is online here.   
  • The full 108 page 510K decision summary is also online at FDA here.   
  • The K192063 home page is here (check for updates).
  • As an IVD, elio should have detailed FDA-authorized "instructions for use" - I haven't seen those yet.  A copy is usually on the FDA website.  
    • Lacking a copy of the IFU, the two-page product brochure is here and a two-page product features flyer is here.
The new-news is that PGDx has released a press release that it is now covered by the MolDx program (here).

Nerd Notes

It's unclear exactly how MolDx will handle usage of the test from this point forward.   

  • Will elio be issued one Z code, that is "turned on" for any hospital client to use?   
  • Or will each hospital using elio have to get its own Z code?   
    • If the latter, with local hospital Z codes, will each hospital using elio have to show its on-site validation to MolDx for review?

  • It's pretty clear coverage falls under the NGS solid tumor LCD, see A57831 here.  So no review for "coverage" or "clinical utility" was involved. 
  • On the other hand, regarding edits, LCD L38045 has been active since February 2020, and in a August 2020 copy of the MolDx Master Edit File, a lot of solid tumor tests were not crosswalked to L38045 for editing (here).


It looks like the MolDx review of the elio IVD took some length of time between the April 24 FDA authorization and the September 9 press release.   

Noting that, several stakeholders have asked me about the necessity of lengthy MolDx review of an FDA IVD for analytical accuracy (the coverage under L38045 was already in place and not under discussion).   For example, CMS Coverage Group endorses and requires coverage of any and all FDA approved or cleared NGS companion diagnostic or any FDA approved or cleared NGS hereditary risk profiling test, two categories which are very similar in FDA review standards to the elio 108-page review released in April.  If the clearance labeling for elio had been worded only a tiny bit differently, the test would have been auto-covered by the somatic and germline CMS NGS NCD without any MAC review of analytical validity of its NGS platform.   



PGDx elio tissue complete, the first FDA-cleared comprehensive genomic profiling kit, is used to identify alterations in the tumor and inform treatment decisions for patients with advanced solid tumors. The kitted system allows molecular laboratories anywhere to perform this advanced genomic testing of cancer in a more efficient, standardized, and accurate manner. By providing tests that can be run locally and automating the data analysis process, PGDx is enabling the adoption of precision medicine in healthcare systems across the country, no matter where a patient seeks treatment.

PGDx elio™ tissue complete

PGDx elio™ tissue complete is an FDA-cleared diagnostic kit and accompanying software for molecular labs that provides comprehensive genomic profiles of all solid tumors. PGDx elio tissue complete detects single nucleotide variants (SNVs) and small insertions and deletions (indels) in 500+ genes, select amplifications and translocations, and genomic signatures including microsatellite instability (MSI), and tumor mutation burden (TMB). Designed to be used locally at any laboratory across the country, PGDx elio testing and automated bioinformatics ensures both consistency and quality of results regardless of location.



Re uncertainly of single, or hospital-specific Z codes, I couldn't find any info yet at MolDx:

Digital Health Applications: Debates Similar to the "LDT" Debate in Labs; Nature Digital Medicine Journal

Part of my work is in digital health, and that community of companies has debates about FDA regulation of digital health, similar to the diagnostics debates about FDA-authorized IVDs versus LDTs.

See a window into that digital regualatory debate in a recent conference summary from "Digital Therapeutics East" - article here.

A current Senate proposal, S. 3532, would allow Medicare coverage of "prescription digital therapeutics" in some circumstances (here).

Meanwhile, I note there is a new NATURE journal, NATURE DIGITAL MEDICINE, now being published.  Home page here.

Thursday, September 17, 2020

Seema Verma Makes Somewhat Vague Remarks on COVID test reimbursement

 As first reported in a subscription article at Inside Health Policy, and as summarized open access by McKnights...  Seema Verma made some remarks on September 15 about potential new reimbursement paradigms for COVID testing.   It seems related to pay-for-performance, but from the quoted remarks, it's difficult to see what any actual change is really being contemplated.

McKnights here open access.   Inside Health Policy here subscription.

The headline "hints at new strategy" is about right.


Medicare Releases Claims Processing Instructions for New 2020 NCD for Germline NGS in Cancer

Early in 2020, CMS released a major update to its NCD for uses of NGS testing in cancer patients.  The update focused on new rules and coverage for germline risk testing (especially for FDA authorized tests) using NGS - here.

On September 11, 2020, CMS released Transmittal 11837, which provides its MACs with official notice that the new NCD exists and claims processing instructions.

  • See CR11837 here.
  • See Medlearn Matters version here.
Notoriously, in 2018, CMS issued a transmittal about the original NGS NCD which included some new text that changed the meaning of the NCD released earlier in the year.  So far, it doesn't look like any weirdness is going on in CR11837.   

The instructions ask MACs to work together for uniform implementation up to such time as CMS may determine shared (mandatory) implementaiton rules are provided. 



Wednesday, September 16, 2020

Very Nerdy Note: CMS Releases State Part B Carrier Data File

 Each summer and fall, CMS releases three forms of Part B data.

(1)  Carrier (State) Data File

CMS releases a data file for each MAC subzone, which usually means state.  While tedious, it is possible to merge and reconstruct data by MAC from these 50-plus excel spreadsheets.

  • On September 10, 2020, CMS releases Part B (State) Carrier Files for CY2019.  Here.

It's a little tedious as the 50-spreadsheets are numbered and you need a PDF mapping of the states (e.g. 03502.xls = Utah Pt B).

Not released yet:

(2) National Data File

At some point, CMS will release a national data file of utilization and national total payment for all CPT and HCPCS codes.   It's not released yet for CY2019.  It will appear here as Part B National Summary Data File.

Not Released yet:
(3) Provider/Lab CPT Payments

Since about 2014, CMS releases Part B payments to physicians and labs by CPT code and utilization and payment for each doc and lab.  (E.g. Dr. Smith was paid $1,500 for 25 uses of CPT code 91234).  This is released at an 18 month lag; CY2018 should appear around mid CY2020 but it's not out yet. Here.   This is a good data source because it's a huge cloud database you can sort and cull and then download in Excel.  (For example, you could sort for all CPT codes in the molecular range and all states involved in MOLDX then download that.)

MolDx Reprices Myriad Genesight Test at Lower Price of $1568

This summer, MolDx retired special coverage of the Myriad Genesight test (a psychiatric drug gene panel with algorithm), and announced the test would fall under a general PGx policy that now covers PGx genes based on whether the patient's gene-drug combination is endorsed by the FDA or by CPIC, the pharmacogenetics consortium.    The policy specifically states that no extra credit/coverage is given for "algorithmic" PGx test.

At its August investor call, Myriad declined to discuss the new price assigned by MolDx. 

A new report by the NEPHRON consultancy releases the price as being $1568.64, under Z-code [Z code redacted by law firm cease and desist order]. (The previous price was $2183.50 under Z-code [Z code redacted by law firm cease & desist order). 

This is still a higher price that for most 10-20 gene panels on the Clinical Lab Fee Schedule. For example, 81432 +81433 (BRCA-related genes, 10 or more, including dup del analysis and including BRCA1-2) pay about $1100 together.    

A CLFS pharmacogenetic 16-gene panel, 0078U, is only $450.
Other MolDx payments to Myriad include the gene set (81292, 81294, 813317, 81319, 81298, 81300, 81295, 91297), which are related to Lynch syndrome, for about $3495 (e.g. these genes paid in 2017 CMS payment by lab data and CMS 2018, 2019 Utah data). Labs that perform a larger gene set will trigger, by definition, the standard Lynch syndrome AMA CPT codes 81435, 81436, including 81292 etc, for which MolDx pays only $1170 altogether.*
One other oddity of the Master Edit File is that the Z code for Genesight would seemingly fall under the MolDx LCD for PGx testing, L38294. Dozens of PGX codes are assigned to this LCD, so MolDx clearly knows how to do that and has had time to do so.  But the Z code for Genesight PGx test isn't assigned to any LCD edits in the most recent Master Edit File from MOLDX. 


Other Master File Oddities

Some large tumor gene panels, which would seemingly fall under MolDx's solid tumor LCD, are also assigned to "no LCD" for edits (e.g. see the OmniSEQ ZBnn7).

In other cases, MolDx appears to ignore PAMA's downward price changes - depending on who the lab is.  

For example, the CARIS BRCA test ZBnnC is priced not at the fee schedule price for 81162 ($1824) but as code 81479 for $2396.   Same for the Sonic BRCA panel ZBnn9, which is described as a BRCA panel (normally 81432+81433 = $1118) but the Sonic BRCA panel assigned code 81479 for $2396.   


* There are also separate Noridian and Palmetto articles that "genes billed together" should "never be stack coded" but should be coded as 81479 for non-stack pricing, eg here.


Noridian's instructions for obtaining information from FOIA is here.  Cloud copy here.  Noridian specifically notes that FOIA info includes: "Contractor priced fees (C-Status) not published on Noridian's website" and "Medicare coverage criteria not addressed in policies or provider bulletins."


Copy of release letter for FOIA for Master Edit File here.

Copy of law firm letter "cease and desist" against describing errors in payment rules here.

Between Fall 2019 and Fall 2020 the Master Edit File expanded from around 12,000 lines to around 20,000 lines, which probably has contributed to the accumulation of errors and a general unmanageability relative to the early vision of the Z code system when there were only a couple hundred codes.   For example, they issue Z-codes for tests paying as little as $5 (Z code ZB6nn), and for some topics like FilmArray 0097U, there are 180 separate NPI's tied to its Z code.  What this accomplishes is unclear (perhaps staff could have been deployed building a house of cards or doing crosswords).   Some 1200 codes have special MolDx prices, but in some cases, as noted earlier, this just seems to shield some tests at some lucky labs from PAMA pricing, by switching them from CPT codes priced under PAMA to then 81479 code and then intentionally assigning that lab a higher than PAMA price.

Friday, September 11, 2020

MEDPAC discusses PAMA, Growth in Molecular Pathology Spending

Just a few weeks ago, the OIG issued its annual report on PAMA and lab test spending at CMS - see report here, see blog here.  While the main goal of the report was to examine the impact of PAMA lab test pricing changes on utilization and sepnding, the OIG report also gave quite a bit of attention to the rapid growth of molecular testing.   MoPath spending rose from circa $500M in CY2017 to circa $1B in CY2018.

Meanwhile, COVID legislation last spring both delayed the implementation of the next three-year PAMA cycle, AND also asked the federal advisory body MEDPAC to study PAMA and report back to Congress.  I think the lab community largely viewed this congressional request to MEDPAC as a good thing, hoping MEDPAC would scold CMS for rules that lowered lab test pricing and for creating a data system that was more burdensome than necessary.   (The MEDPAC report had circulated previously as part of the LAB ACT, a PAMA delay bill last winter; here.)

MEDPAC is well underway with its report on PAMA and lab tests, and it's not all good news.  An interim report was presented last week - written by MEDPAC permanent staff and presented to the dozen experts who are the MEDPAC panelists.  (Note - the panelists have variable acquaintance with the world of Medicare laws and policies).  

The report discussed the basics of PAMA, but also noted that molecular spending rose by a lot, doubling from 2017 to 2018.  Panelists asked a range of questions, such as whether CMS had policies to constrain utilization growth, and whether PAMA should set prices to something lower than the median, e.g. the 25th not the 50th percentile of commercial prices.   Overall, there were a lot of references to "expensive tests."   

One panelists suggest that since genomic pricing falls, CLFS prices should automatically deflate.  

Another panelist suggesting that the pricing of "commodity tests" and genomic tests were so different that different rules should apply.   

Another suggest CMS institute "competitive bidding" for lab tests.  (While CMS does source some DME products by competitive bidding, Congress and CMS tried a demo project for lab competitive bidding around 2007/2008 and it collided with lawsuits.)  

Overall, I suspect lab industry viewers watching online were looking at their watches and just waiting for this discussion period to end.

In another part of the discussion, MEDPAC discusses whether telemedicine visits should allow ordering of lab tests, and if so, only for established patients.  

The deck PowerPoint covers the basics, sort of a PAMA 101, most of the interest is in the discussion in the transcript.  

Panelists noted that MEDPAC is asked not only to opine on the CMS implementation rules and data burdensomeness, but "other topics" MEDPAC finds appropriate.   

MEDPAC may have an interim presentation next spring on its PAMA findings and recommendations; the report to Congress is due next summer.  


I believe one version of the LAB act, PAMA delay, had the Hill requesting a PAMA report from National Academy of Medicine, but the passed legislation gave the report to MEDPAC, an organization that favors frugality in its reports.