CMS Announces MEDCAC on Molecular Prognostic Tests - March 24, 2015

CMS has announced a MEDCAC (advisory panel) public meeting on prognostic molecular tests, focused on breast cancer and colorectal cancer molecular tests.   Such meetings often kick off with a technology assessment subcontracted by AHRQ.

Update [March 25]: Early trade press after MedCAC suggests it went exactly as predicted in this blog many weeks ago.  An advance copy of the transcript is available here.

Details after the break...

The Federal Register announcement is here (80 Fed Reg 3065, 1/23/2015):

Medicare Program; Meeting of the Medicare Evidence Development and Coverage
Advisory Committee—March 24, 2015
AGENCY: Centers for Medicare & Medicaid Services (CMS), HHS.
ACTION: Notice of meeting.

SUMMARY: This notice announces that a public meeting of the Medicare Evidence Development & Coverage Advisory Committee (MEDCAC) ("Committee") will be held on Tuesday, March 24, 2015. The Committee generally provides advice and recommendations concerning the adequacy of scientific evidence needed to determine whether certain medical items and services can be covered under the Medicare statute.
This meeting will focus on selected molecular pathology tests for the estimation of prognosis in common cancers (such as, adenocarcinoma of the colon and rectum, breast cancer - invasive duct and lobular cancers, non-small cell lung cancers). This meeting is open to the public in accordance with the Federal Advisory Committee Act (5 U.S.C. App. 2, section 10(a)).
DATES: Meeting Date: The public meeting will be held on Tuesday, March 24, 2015from 7:30 a.m. until 4:30 p.m., Eastern Daylight Time (EDT).

CMS has also posted the MEDCAC website for this meeting, which provides a slightly different summary, here.   On March 17, 2015, CMS updated the MEDCAC website with a full roster of speakers, agenda timeline, and a ZIP file of presentations.

     The MEDCAC panel will consider the evidence about molecular pathology tests that estimate the prognosis of several frequently encountered types of cancer, including:

• Adenocarcinoma of the colon and rectum;

• Invasive breast cancer; and

• Non-small cell lung cancer.

     Outcomes of interest to CMS include both beneficial and harmful outcomes experienced by patients from anti-tumor treatment decisions based on results of tests estimating cancer prognosis.
     Such outcomes may include, among others: mortality/survival; quality of life; and avoiding harms of anti-cancer treatment. Results of a technology assessment of such tests will be presented.
     By voting on specific questions, and by their discussions, MEDCAC panel members will advise CMS about the extent to which it may wish to use existing evidence as the basis for any future determinations about coverage for tests that estimate cancer prognosis. MEDCAC panels do not make coverage determinations, but CMS often benefits from their advice.

Of more interest, the CMS website includes instructions and voting questions for the MEDCAC panelists.  These are listed in full at the bottom of this article.   Tests for voting including Oncotype DX Breast and Colon tests; the Mammaprint breast cancer test; and individual gene tests such as BRAF, KRAS, and EGFR.  Panelists will vote on the evidence for tests as to analytical validity, clinical validity, whether they impact clinical outcomes [for good or bad], and whether they impact clinical outcomes favorably.

No technology assessment has yet been listed as the basis of this MEDCAC.  However, in 2014, the AHRQ published a technology assessment (as a commission to ECRI) of cancer prognostic tests.  And they do include some major prognostic tests, such as Oncotype DX.  

The assessment protocol as a whole, however, was puzzling (even bizarre) to some experts, as it focused only on the "prognostic value" of tests such as EGFR and KRAS, which have scant prognostic value per se but rather, are used to choose the correct chemotherapy for the patient.   Having receiving blowback on their draft report, in December 2013, in the final 2014 version the ECRI authors publicly confessed that they knew some of the tests they reviewed had diagnostic value, rather than any prognostic value, but that they were compelled by their instructions from AHRQ to evaluate only the "prognostic" value of the tests -- and to exclude and ignore data for diagnostic or chemotherapy predictive value.  This leads to the over-the-top strange conclusion that a test like ALK has no effect on management or outcomes. 

Looking at ALK or EGFR tests but being instructed to ignore their "predictive" value and look only to their "prognostic" value is like asking to look at the clinical benefits of penicillin while excluding any data on its benefits in infectious diseases.  

I am unconvinced that the AHRQ contracted authors couldn't have pointed out initially, on receiving the contract, that in significant part, the instructions were miswritten, crazy and made no sense.   If I was asked to write a report on ALK, with the instruction, "studies that evaluate specific diagnostic/therapy combinations are excluded from this report" I would pick up the phone and ask what in blazes is going on - it makes no sense at all.   The trade press is unable to pick up on this point - see "Medicare Advisors Focus on Cancer Tests" in Congressional Quarterly Roll Call, a publication of The Economist, which ought to know better and bring some content knowledge to its biomedical reporting.

I made similar remarks as an open comment letter to the MedCAC committee (here, 2/23/2015.)

The AHRQ report, "Molecular Pathology Testing for the Estimation of Prognosis for Common Cancers" is available here or here (on the CMS website), dated May 2014.  The AHRQ methods guide for "prognostic tests" is also available online, here or here.

Presentations should be submitted February 23, 2015.  (These can also be in the form of a "letter to the panelists" if you don't want to present live.)   Registration to attend closes on March 17 (but register earlier because seats may fill up.)

CMS has held fewer MEDCACs in recent years.  There was only 1 in 2014, and this is the only MEDCAC scheduled so far in 2015.  There were 4 in 2008, 5 in 2009 and in 2010, 3 in 2011, 4 in 2012, and 2 in 2013.   Of the 22 MEDCACs since 2009, 5 have been on genomics.

The MEDCAC panel is here; each meeting uses about 8-9 selected panelists from the standing committee and often adds 2-3 guest panelists not on the standing roster, such as an NIH subject matter expert.

_____________________________________________

INSTRUCTIONS TO MEDCAC:

On March 24, 2015, the Medicare Evidence Development & Coverage Advisory Committee (MEDCAC) will consider evidence about molecular pathology tests[1] that estimate prognosis for Medicare beneficiaries with several frequently encountered types of cancer:

Cancer Type	Test to Estimate Prognosis	Q1a	Q1b	Q2	Q3
Adenocarcinoma of the colon and rectum	BRAF
	KRAS
	Microsatellite instability
	MLH1 promoter methylation
	Oncotype DX® Colon
Breast cancer (invasive duct and lobular cancers)	MammaPrint®
	Oncotype DX® Breast
Non-small cell lung cancers	ALK
	EGFR
	KRAS

Outcomes of interest for CMS reflect beneficiaries’ benefits or harms due to anti-cancer treatment decisions based on test results. Outcomes of interest include (among others) mortality, avoidance of harms of inappropriate anti-cancer treatments, and quality of life.

Voting Questions

For each voting question regarding each molecular pathology test in the grid above, please use the following scale identifying your level of confidence - with a score of 1 being low or no confidence, and 5 representing high confidence:

1      —      2      —      3      —      4     —      5
Low                   Intermediate                      High
Confidence                                            Confidence

1. 1. For each prognostic test listed above, how confident are you that existing evidence is sufficient to confirm the analytical validity of the molecular pathology test to estimate prognosis for Medicare beneficiaries with that cancer type?
   For each prognostic test, if the average of voting members’ answers for Question 1a is in the ‘Intermediate’ range (i.e., mean score is ≥ 2.5) please vote on Question 1b.
   2. For each prognostic test listed above, how confident are you that existing evidence is sufficient to confirm the clinical validity of the molecular pathology test to estimate prognosis in Medicare beneficiaries with that cancer type?
   [Note: The definitions of analytical and clinical validity, based on the ACCE assessment process, as used by the Center for Disease Control and Prevention (CDC)’s “Evidence of Genomic Applications in Practice and Prevention” (EGAPP) working group, are summarized athttp://www.cdc.gov/genomics/gtesting/EGAPP/recommend/method.htm  from which the following table is adapted.  The terms ‘diagnostic test’ and ‘prognostic test’ are defined in the footnote.[2]]

   Analytic Validity (Technical Performance)	Clinical Validity (Correlation with Prognosis)
   Test’s ability to accurately and reliably measure analyte or genotype of interest. Related concepts: (Instrumental) Sensitivity (Instrumental) Specificity Assay robustness Quality control	Test’s ability to accurately and reliably estimate the prognosis for the condition of interest. Related concepts: (Clinical) Sensitivity (Clinical) Specificity Positive Predictive Value Negative Predictive Value

   For each prognostic test, if the averages of voting members’ answers for Questions 1a and 1b are both in the ‘Intermediate’ range (i.e., mean score is ≥ 2.5) please vote on Question 2.
2. How confident are you that there is sufficient evidence to conclude that using the molecular pathology test to estimate prognosis affects health outcomes (including benefits and harms) for Medicare beneficiaries with cancer whose anti-cancer treatment strategy is guided by the test’s result?
   For each prognostic test for which the average of voting members’ answers for Question 2 is in the ‘Intermediate’ range (i.e., mean score is ≥ 2.5), please vote on Question 3.
3. How confident are you that there is sufficient evidence to conclude that using the molecular pathology test to estimate prognosis has clinical utility (meaning, that it improves health outcomes either due to increased benefits and/or reduced harms) for Medicare beneficiaries with cancer whose anti-cancer treatment strategy is guided by the test’s result?

Discussion Question

4. Please discuss whether each factor below might change the generalizability of evidence about prognostic molecular pathology tests in Medicare beneficiaries with cancer:
   1. Regulatory status of test (e.g., US Food & Drug Administration (FDA) approved/cleared vs. laboratory-developed test)?
   2. Type of performing laboratory (i.e., university medical center laboratories, independent commercial laboratories, or community hospital-based laboratories)?
   3. Subgroups in the Medicare beneficiary population (e.g., by age)?
   4. Genomic variations within cancers (e.g., diversity of cancer genomes)?

---

[1] Molecular pathology test is a type of genomic test.

[2] A diagnostic test is defined as a medical test performed to aid in the diagnosis or detection of disease in a beneficiary with signs or symptoms of an illness or injury.  In contrast, a prognostic test is a test performed for a beneficiary with cancer that measures or assesses one or more biomarkers, including the biological and genetic properties of the cancer cells, which are thought to be associated with future outcomes.