INSIGHTS | POLICY FORUM
ring issuance of final guidance pending further discussion and possibly legislation (8).
The Centers for Medicare and Medicaid
Services (CMS) provide some oversight of
laboratories that perform genetic testing under the Clinical Laboratories Improvement
Amendments of 1988 (CLIA) (9). But CLIA
regulation primarily considers analytical validity of tests (whether the tests accurately
measure markers that they purport to detect)
rather than their clinical validity (whether
there is a reliable association between test
results and phenotype) or clinical utility
(whether use of the test improves clinical
outcomes). There is still debate about what
clinical utility means for genomic testing or
what sort of data is necessary to show it.
How can NGS-based tests be converted
from “investigational” to “standard of care”?
Extensive variation in the molecular profiles
of individual cancers—even cancers of the
same tissue type—means that extensive data
from observational studies will be necessary
to validate the clinical relevance of the less
common variants. Up to
4000 samples of each type
of human cancer could be
required to identify nearly
all functionally significant
mutations that occur at
frequencies as low as 1 to
2% (and at least 50 tumor
types can be ascertained
by pathologists) (10). Such
data sets might never achieve the necessary
scale if coverage by insurers must await prior
proof of clinical utility, because testing is presumably less likely to proceed without coverage. Premarket testing on the scale typically
paid for by pharmaceutical firms seeking
FDA approval of new drugs would not generate enough data to evaluate the importance
of any but the common genetic variants.
The line between clinical care and clinical
investigation is shifting. Cancer treatments
are now often initially approved on an accel-
erated basis, with a commitment to submit
further data from postmarket studies. But
some firms have taken many years to com-
plete such studies, if they complete them at
all (11). Although FDA has statutory enforce-
ment authorities that it can use against re-
calcitrant firms, monitoring progress toward
completion of postmarket study require-
ments has been difficult. Accelerated ap-
proval inevitably comes at the cost of greater
uncertainty about long-term safety and effi-
cacy and a need for continued collection of
relevant data. Meanwhile, as new products
become available for use, caregivers begin to
prescribe them, and patients are exposed to
risks that have not been fully studied. The re-
sult is that insurers are paying for treatment
that is, in effect, still under investigation,
At least one firm, Foundation Medicine,
has submitted applications to both FDA and
CMS seeking FDA approval and nationwide
Medicare coverage of a genomic test to detect
alterations in 324 genes, most of which have
been directly implicated in carcinogenesis
(12). Foundation Medicine already markets
the test as a “universal companion diagnos-
tic assay” for genomic profiling “across a di-
verse range of solid tumors”; FDA and CMS
have accepted the applications for “parallel
review” under a cooperative initiative to co-
ordinate their review processes and thereby
provide Medicare beneficiaries with earlier
access to new products
(13). The outcome could set
an important precedent for
by private insurers.
Panel tests or whole-exome sequencing cost
in the range of $1000 to
(14). Thus, the costs of more extensive testing
would amount to roughly 1% of the total bill
for cancer care. At those prices, the costs of
testing seem justified if the tests bring about
improvements in even a small percentage of
patients. As more data accumulate, formal
methods such as cost-effectiveness analysis
and value-of-information (VOI) studies could
help inform future choices.
We suggest that federal agencies and
those seeking compensation for genomic
tests consider a mechanism that CMS has
used successfully: coverage with evidence
development (CED) (15). When existing data
are inadequate to evaluate whether a new
technology is reasonable and necessary for
treating the Medicare population, CMS has
sometimes used CED to provide coverage
while continuing to evaluate the technology.
Under CED, Medicare covers services for
patients enrolled in clinical trials or regis-
tries in order to obtain additional evidence
for evaluating promising technologies of un-
proven value. An early success for this ap-
proach was the use of CED to monitor the
effects of lung volume reduction surgery for
chronic obstructive lung disease (16). Pri-
vate insurers have also sometimes provided
similar coverage of new technologies for
patients participating in research studies,
supporting, for example, trials to assess the
value of high-dose chemotherapy with au-
tologous bone marrow transplantation for
patients with metastatic breast cancer (17).
CED could encourage patient participation
in registries for studies designed to determine whether genomic testing is reasonable
and necessary for the accurate diagnosis and
appropriate treatment of cancer. The swift
development of a substantial database of
results from clinical sequencing of diverse
cancers (18) suggests that such goals are
achievable, provided that reliable data on
treatment responses can be incorporated. j
REFERENCES AND NOTES
1. D. M. Hyman, B. S. Taylor, J. Baselga,Cell 168, 584 (2017).
2. S. Lernery, P. Keegan, R. Pazdur, N. Engl. J. Med. 377, 1409
3. Green Park Collaborative, Initial Medical Policy and
Model Coverage Guidelines for Clinical Next Generation
Sequencing in Oncology (Center for Medical Technology
4. J.B.Guinto,NCCNTrends:An EvaluationoftheUseof
Companion Diagnostic Testing and Targeted Therapies
(reporting results of 2012 survey); https://www.nccn.org/
5. J.P.Cohen,A.E.Felix, J. Pers. Med. 4,163(2014).
6. S. W. Grayetal., J.Clin.Onc.32, 1317 (2014).
7. U.S. FDA, Draft Guidance for Industry, Food and Drug
Administration Staff, and Clinical Laboratories: Framework
for Oversight of Laboratory Developed Tests (FDA, 2014).
8. U.S. FDA, Discussion Paper on Laboratory Developed
Tests (13 Jan. 2017); https://www.fda.gov/downloads/
9. Centersfor Medicareand MedicaidServices,LDTand
CLIA FAQs (2013), https://www.cms.gov/Regulations-and-Guidance/Legislation/CLIA/Downloads/
10. M. S. Lawrence et al., Nature 505, 495 (2014).
11. U.S.DepartmentofHealthandHuman Services
(DHHS), Office of Inspector General, FDA Is Issuing
More Postmarketing Requirements, but Challenges with
Oversight Persist (2016), https://oig.hhs.gov/oei/reports/
12. Foundation Medicine Inc., Form 10-Q for the quarter ending
March 31, 2017, at 23-24; http://investors.foundationmedi-
13. U.S. DHHS, CMS, and FDA, Program for Parallel Review of
Medical Devices. Fed. Regis. 81, 73113 (24 Oct., 2016).
14. A.B.Mariotto, J.Natl.CancerInst.103,117(2011).
15. CMS,Guidanceforthe Public,Industry,and CMSStaff:
Coverage with Evidence Development (CMS, 20 Nov.
16. CMS, Decision Memo for Lung Volume Reduction Surgery
CAG-00115R (CMS, 2003); http://go.cms.gov/2j0pcWL.
17. S. R. Tunis, S. D. Pearson, Health Aff., 25, 1218 (2006).
18. TheAACRProject GENIEConsortium,Cancer Discov. 7,818
19. F. Nowak etal., Nat.Rev.Clin.Oncol. 9, 479 (2012).
ACKNO WLEDGMEN TS
Supported by funds from the Meyer Cancer Center at Weill
Cornell Medicine (to H.V.) and the William W. Cook Endowment
and the Elkes Fund of the University of Michigan Law School
(to R.S.E.). We acknowledge advice from P. Bach, N. Bagley, J.
Baselga, L. Diaz, M. Hamburg, N. Price, R. Sachs, C. Sawyers, M.
Seiden, T. Simoncelli, and S. Tunis.
1134 1 DECEMBER 2017 • VOL 358 ISSUE 6367
“The line between
clinical care and