Tip:
Highlight text to annotate it
X
Dan Roden: Dr. Faruki from LabCorp will tell us how LabCorp
implements genomic testing now, before, and in the future. Thank you for coming.
Hawazin Faruki: Okay. Thank you very much for inviting me.
It's a privilege for me to address this group and really talk about LabCorp's experience
and perspective on implementation of genomic testing. And I'll start off with a few comments
about, "Who is LabCorp?" or, "What is LabCorp?" This is our headquarters building in Burlington,
North Carolina, which is about 45 minutes from Research Triangle Park. We're a large
diagnostic laboratory company with 30,000 employees, approximately 1,500 patient centers.
These are our draw stations around the country. We serve 220 customers. They are doctors,
physicians, hospitals, medical centers, academic institutions, et cetera. We offer about 4,000
different diagnostic tests, ranging from simple things like CBC and urinalysis all the way
up to the most sophisticated sequencing, SNP arrays, copy number variants, et cetera. And
we have a complicated infrastructure to collect close to 400,000 specimens a day and deliver
over a million results every single day.
So, in order to meet that objective, we employ over 750 M.D.s and pathologists. We have 150
or more genetic counselors who support our operations. There are 10 Centers of Excellence
scattered throughout the country that have specialized expertise, 38 primary testing
locations, and STAT centers, and obviously we need an infrastructure of couriers and
corporate jets to move all those specimens around.
So just looking across a map of the United States, all the light blue are the draw centers,
and they're concentrated primarily in the population-dense areas. The darker blue are
our regional labs, and then the yellow circles are actually the Centers of Excellence that
have unique expertise. And LabCorp operates with a spoke-and-hub sort of, where specimens
are collected regionally. They go through a branch, and then to a regional lab, and
for specialized testing to those Centers of Excellence. That allows us economies of scale
and to offer some very specialized services as well.
We also operate a global clinical trial service and capabilities and have labs scattered throughout
the world. We have a recent acquisition of Clearstone, so we now also have a CAP-accredited
lab in Beijing and a variety of other interactions with various other countries in South Africa,
in Australia, in India, in Abu Dhabi, and so on.
LabCorp also has a companion diagnostics program, and we have been instrumental in the introduction
of many of the companion diagnostics tests that are out there today. And LabCorp serves
in a capacity sometimes in the early development stages, such as with some of the HER2 studies
way back many years ago, but also in a clinical trial capability to support a lot of the validation
and submission for those assays and drugs to the FDA. So whether it's on the science,
on the regulatory or on the commercial aspects, LabCorp has provided a lot of support for
many of the companion diagnostics that are out there today.
We like to maintain scientific leadership, and to do that, we need to continually review
new opportunities, so we review probably over 400 new opportunities for diagnostics a year.
And we introduce every year over 100 -- close to 130 new diagnostic tests. They're not only
genomic. They span the whole range of diagnostic offerings. But they often involve collaborations
with pharma, with universities, with biotech companies, with academic centers, with the
government, et cetera, to make that actually happen.
So where did these new diagnostic opportunities come from? Many of them come from academic
institutions, and LabCorp has developed good relationships with many of the tech transfer
offices at most of the major universities so that we can be in their mind when they
develop new diagnostics and want to -- want assistance in getting it out into the clinic.
The biotech companies often seek LabCorp hoping for access to the market through LabCorp's
infrastructure. We -- through our clinical trials and companion diagnostics, we often
get a first look at diagnostics long before they're useful in the clinic, but in the early
stages of development, drug development or development of a diagnostic. We -- I mentioned
our Ph.D.s and M.D.s, and we have many connections to the scientific literature, to meetings,
presentations, and our peers, who also bring all kinds of new ideas to us. LabCorp's active
in the acquisitions and mergers, and sometimes we acquire technology from companies. And
we also monitor our send-out requests so that, when we do have send-outs and we see it growing
to a certain volume, that necessitates bringing a certain assay in house.
So what do we look for? Scientific and clinical evaluation is key, and that's one of the first
things that we do. And I work with a group of scientists at LabCorp in looking at these
new opportunities. We look at analytical validity and clinical utility. So the basic things,
like sensitivity, specificity, positive predictive value, negative predictive value. But key
also is, "Does it lead to improved outcomes? Does it lead to decreased health care costs?
Is there a therapy that's guided by that diagnostic? Can we avoid a toxicity? If we diagnose something
earlier or if the sensitivity's increased, does that impact the outcome?" We're looking
for an actionable result, something that actually changes management, not just that looks good
in a chart, but that actually changes management.
And, finally, we look for evidence, and that could be publications or guidelines, professional
society endorsements, endorsements from the government, like the Department of Health
and Human Services or CMS. And even if those things are not in place at the time that we
license or look at a opportunity, we are thinking about, "Are these things that we could anticipate
in the future?" or, "How far off might those guidelines be?" Because they're critical to,
as I'll show you in a few minutes, to reimbursement and to growth and acceptance of the test.
So, for financial and other considerations, we also explore, "What's the reimbursement
outlook for that particular opportunity, for that licensing or that diagnostic? How will
we get paid, and how much will we get paid? Is there intellectual property associated
with it, and will we have freedom to operate? And is there a licensing? Is there a need
to have a license? And if so, what's the royalty burden for that license? What will it cost
us to actually bring it to market? And what additional evaluation will be needed? Do we
have to sponsor the study to do that? Will it require -- are the specimens available?
Do we have enough annotated specimens to do that with?" So all of those things are part
of the cost of actually bringing it to market. And then those are weighed against, "What's
the likely reimbursement? Thus, what's the return on the investment?"
Is the assay feasible given our platform? And here, there's a few things that we might
not be able to engage in, such as very -- an analyte stability issue where it's really
not amenable to our sort of infrastructure, or potentially a requirement for the result
within minutes. You know, we do have some STAT facilities, but, again, that's not available
in every location across the U.S. The regulatory landscape, which Debra Leonard touched on
briefly, but practically all the molecular tests, with very few exceptions, are laboratory-developed
tests, and we've been able to react to the changing environment, the developing scientific
information, et cetera, by being innovative and by creating these laboratory-developed
tests. So if we -- if there's movement to restrict our ability to do that, I think that
will have an impact on our ability to react to the changing environment.
And finally, some market dynamics and competition that we also look at as we evaluate new opportunities.
So what is success? And so we've modeled this, because we've looked at -- and I've been at
LabCorp now for over 10 years, and we've seen many test adoptions, and we kind of have a
pretty good feel of what creates success. So we've modeled this, and this particular
modeling was done in the context of HLAB-5701, the association with Abacavir allergic reaction,
but it really is mirrored in a number of other examples, which I'll share. But there's typically
a several-year process early on in which the marker's first identified and early publications
support its use, and people like LabCorp actually develop a lab-developed test and bring it
on and put it on the menu.
And then there's some key studies, and in each one of these examples, I could probably
name the key studies that were drivers. In this case, it was the Predict Study and SHAPE,
two studies sponsored by Glaxo that led to a clinically established test with clear clinical
utility. And what happens with that is -- following that is just a very dramatic test acceptance
period, during which all of these other players fall into place. So the clinical utility's
established. The payer-reimbursement coverage happens shortly thereafter. Professional society
guidelines are happening at this same time. And all of those together drive a huge increase
in volume.
Now, that -- and then that increase in volume goes over a short period of time, a few years,
during which there's logarithmic growth of the particular test. And I can tell you, at
LabCorp we often struggle during those periods trying to meet demand, you know, trying to
have enough people trained and enough people in place to meet the demand for a particular
test.
Okay, so -- and I've mentioned, historically, we see many, many examples of this. *** resistance,
guidelines that followed VIRADAPT and GART [spelled phonetically] studies in 2001, the
ALTS study with HPV testing, the HCV genotyping after the ribavirin interferon trials and
that data, the CF testing after the ACMG guidelines, et cetera. In each one of those cases, that's
the event that drives that huge increase in volume. And we've also looked at it in some
of these oncology markers, and I can say the pattern is pretty much the same. In the Herceptin
-- is this working or not? Yeah, I guess the Herceptin approval here, both the diagnostic
and the test, in 2003, 2004, was quantitative BCR-ABL, and the Glefec [spelled phonetically]
monitoring and NCCN guidelines and ASCO [spelled phonetically] guidelines, et cetera, to support
that. And then the KRAS story, which happened, really, between 2007 and 2008, again, with
publication of guidelines, rapid adoption, coverage policy, et cetera. And then I'll
point out these two here at the bottom. The green and the purple are UGT 1A1 and 2D6 with
Tamoxifin, where coverage policies are not favorable, where guidelines have not included
it, and where the definitive studies we're still waiting for.
So I've already mentioned UGT 1A1 and Cytochrome P450 2D6 in the setting of SSRI's EGAPP has
published on that. In the setting of Tamoxifin, I know UNC is working towards creating the
clinical utility story around that, but it's an emerging area, and we still have very low
volume. And let me say, also, that Warfarin and the Coumadin studies and so on relating
to Warfarin pharmacogenetics, again, very low volume at LabCorp for those applications.
So I can't emphasize enough the clinical utility impact. If the clinical utility's not well-established
-- and that could be because it's a non-actionable result. It could be because there's conflicting
studies of clinical utility. And I'll show you a very good example of conflicting studies.
It could be simply because there's limited availability of well-annotated samples to
conduct the study. The end result is lack of endorsement, coverage and reimbursement
denials, and low utilization, and those three things all happen together.
So this is a really good example of conflicting studies. And I want to show you what's going
on with Plavix 2C19. This is not published. This is just internal data that we have. But
it's an interest of mine, personally, about what constitutes and why is it that certain
tests are adopted and some aren't. So if we look back, again, the early markers that support
the use of this go back many years. And in 2009, early in 2009, there were two papers
published in the New England Journal of Medicine, both of them supportive of the use of 2C19
in the management of patients on Plavix.
And we, shortly thereafter -- actually LabCorp had been doing 2C19 testing for a while with
very low volume, but in light of the New England Journal of Medicine papers, we actually, in
March of 2009, introduced a Clopidogrel-specific test in which the interpretive language was
supportive of that particular application and made certain recommendations based on
what alleles were picked up.
In any case, the FDA alerts happened shortly thereafter, and there was a large increase
in volume as a result of that. It's interesting, though, that shortly thereafter, so in June
-- the FDA alerts happen in November of 2009 and in March of 2010. In June of 2010, the
American Heart Association and the cardiologists put out some guidelines that were very, very
lukewarm about the use of the test and kind of leaving it up to the clinician, and it
may not be necessary, and there may not be enough data, et cetera. And that, followed
by the Canadian study, again, that did not show -- again, had a profound impact on utilization
of that test. And you can see it, and that controversy continues with volume relatively
low compared to where it was at the peak.
Now, there are a variety of reasons for conflicting studies, and many of them are legitimate reasons
about the choice of patient population or the choice of the clinical setting in which
a test is used. But that controversy is kind of lost on the masses, and the masses choose
not to use the test.
So let me move on to payer reimbursement. Payer adoption and reimbursement has been
-- so for the last 10 years or so, we've had these existing molecular CPT codes, procedural
codes. So, provided that the scientific validity, the clinical utility is established, practiced
guidelines were there, coverage policy, we could immediately use those stacking codes
to be reimbursed for the testing we're doing. That model is changing, so we are no longer
going to be able to use those procedural codes. And in 2013 there's a new coding system going
in. And Debra Leonard alluded to this briefly, but the new coding will allow greater transparency,
which is really a positive thing for everybody. In other words, insurers, payers, billers,
hospitals, et cetera, will be able to know what they're paying for in those Tier I codes.
And the Tier I codes will really cover most of the diagnostic -- the molecular diagnostic
tests that we're doing today, virtually 95 percent or more. I think the problem is that
the Tier II -- so those are Tier I, and at least Tier I, our understanding is that those
will be -- fee-setting will be comparable to what's currently being reimbursed for those
tests. Tier II are complexity-based codes, and it's not clear yet whether payers will
be denying those Tier II or will be paying or will choose to deny those. So I think there's
a lot of uncertainty about how we're going to get paid for the Tier II, which are kind
of complexity-based codes.
And finally, historically, getting a new CPT code was a multi-year process. And I was just
talking to someone during the break. I would be applying today for the code that would
go into effect in 2014. So if that's the model we have for introduction of every new genomic
test to get paid for it, I mean, obviously that's a huge problem for us. So it's not
clear yet about how new codes and new genomic testing is going to be handled in this environment.
And I'll just say one other thing. Fee-setting is usually a separate process from CPT coding,
so 2014 is when I would get the code. I would still then have to pursue getting a fee set
for that code.
So -- and finally, licensing and royalty burden, which, again, it's not clear what's going
to happen in that arena. But when we start talking about multiple gene panels, we're
also talking about multiple licensing agreements, multiple royalty, potentially, and that stacking
of licensing and royalty is really not -- there's no CPT code for that, nor is there a CPT code
for another issue that was brought up earlier, which has to do with going back to sequenced
data and doing a reanalysis based on the latest scientific information.
So other factors that might impact market adoption, and these are some analyte stability,
and I mentioned some of these. And for most DNA-based tests, this is not an issue, but
I think for RNA expression, the use of FFPE, formal and fixed tissue, and RNA expression,
that is clearly an issue. Access limitations, that can sometimes slow down test adoption
if it's only available at a single site. That restricts access and slows down. Physician-related
modulating -- and I call these modulating because they don't -- the clinical utility
story that I mentioned is the initial driver, but these can modulate adoption. Economic
conflicts, so if a physician is paid for a certain procedure but the test, perhaps, replaces
a procedure that a physician can be paid for, there's some disincentive there, and despite
everybody's best intentions, sometimes the adoption can be slower in those situations.
The physician specialty group involved, the HLA Abacavir was *** practitioners, who are
really very used to taking molecular information, resistance testing information, et cetera,
and adjusting therapy and treatment. If we go into the psychiatrist realm or something
like that, we're talking about a very different specialty group that's not at all used to
and accustomed to incorporating that sort of information. So we need to kind of take
that into consideration with how rapidly the adoption will occur.
And then, finally, physician education. Although I think this area, there is improvement with
greater number of physicians being connected to the Internet, and dissemination of information
is clearly happen a lot faster than it used to, I think this is an area where there's
still a lot of room for improvement.
So, in conclusion, I want to just summarize the points that I made. Well-controlled and
adequately powered studies demonstrating analytical validity and clinical utility are key. We
see these areas where there's clear actionable results, and I can tell you that LabCorp is
currently engaged in multi-gene panels that address some of these areas -- in other words,
areas that prevent drug toxicity in the drug metabolism area, multi-gene panels that identify
a treatment path or select for drug. This is in tumor profiling, and the diagnosis of
rare heritable disorders or carrier testings. I mean, these are three areas that we think
are kind of ripe for prime time and where LabCorp is currently engaged in developing
multi-gene panels using next-generation sequencing to offer clinical services in this area. And
then path to fair reimbursement is -- and freedom to operate, be it from a perspective
of intellectual property and/or from the regulatory world.
Thank you very much. I'll take questions.
Dan Roden: Thank you. Marc, and then Teri has a question.
Marc Williams: I really was interested in the adoption curves
that you presented. I think those are fantastic. And, again, these are the sorts of data that
I think groups like this need. In my view, as I think about the takeaways for the genomic
medicine group, it seems to me that this, if we could have access to data like this,
it would be a really nice metric to look at some of the things that either come out of
this group or other NHGRI-related activities, to say, you know, when we have a publication
that comes from a funded study or something else, what impact does that actually have
on the adoption of a given test or technology? I think that would be a very interesting thing
to do.
The other thing I'll just notice about those types of curves is that Clay Christianson
[spelled phonetically] has noted in his study of innovation that that's the typical, of
course, innovation adoption curve of the early adopters, and then the rapid, and then the
hold-outs at the end, the luddites at the end. You can actually take that curve and
integrate it so that you can actually say, "Here is the point at which 50 percent of
people will consider this standard of care," and plot that on a timeline. So it would also
be very useful to be able to use that prospectively in addressing what Sean had talked about for
prioritization, saying, "Where do we need to put our efforts?" Because these are going
to be emerging sooner than those, if we had regular access to those types of data.
Hawazin Faruki: So we can talk about that further. I think
in the past LabCorp has not disclosed publicly their actual volumes, although we can reconsider.
Teri Manolio: That would be great. Maybe for those of us
who aren't familiar with CPT, could you maybe give use two or three minutes on what they
are, why they're important, and how one gets them changed?
Hawazin Faruki: Yeah. Okay. So the laboratory industry -- CPT
coding is kind of the basis for how laboratories get paid for their testing. And CPT stands
for procedural terminology or --
Male Speaker: current procedural --
Hawazin Faruki: -- current procedural terminology. And these
cover all of the medical procedures that physicians do. And, initially, in the molecular diagnostics
space, initially what happened is, as molecular infectious disease testing got started with
viral loads 10 or 15 years ago, people like LabCorp -- in fact, I wrote some of those
initials ones for resistance testing and so on -- you would apply to the AMA, and you
would fill out an application that had to do with how the test was being used and what
procedures it included, and et cetera, and request a unique CPT code, and then that application
would go to a committee, and I think it's run by AMA, if I'm not mistaken, that then
evaluated it and decided either yes or no that they would give it a code. And within
a couple of years, you'd have a code, and then you would follow that up with efforts
with CMS and et cetera to get a fee set for that particular code. And so that's why I'm
saying it's -- it used to be a several-year process to do that, and for the infectious
disease applications 10 years ago, people applied and they added each organism as a
code. So *** viral load got a code, and HPV got a code, and HCV got a code, and a quant
had a different code, then a quall, and a genotype had a different code that a quant,
et cetera.
Now, for molecular oncology and molecular genetics, although I have to say 10 ten years
ago I lobbied against this, but I was overruled, the group decided that they didn't want to
have specific codes for specific genes, that they wanted to use procedural codes, things
like an extraction or a hybridization or a sequencing reaction, et cetera. So each one
of those procedures got a code, and then molecular testing labs were left to stack the codes
as they saw fit, based on their particular assay format.
So what ended up happening is payers -- and maybe the total dollars weren't significant
enough for them to care at the time. But now, as the volume of genomic testing is growing,
there's an increasing need to be able to say, "Well, that hybridization and sequencing and
probe and interpretation and et cetera, what was that for? Was that a KRES [spelled phonetically]?
Was that a oncotype? Was that a -- you know, which test does it relate to?" So this new
effort is really to give unique names to a KRES test, to a CF test, to et cetera, so
that payers and everybody else has greater transparency.
Eric Green: So is there -- it does sound like a '90s construct
for, you know, now a much more contemporary, you know, pursuits, so that -- but, I mean,
is there a single CPT code for sequencing a genome?
Hawazin Faruki: There is a -- no. There is a code for sequencing,
and it's, I think, based on one intron or multiple introns or --
Eric Green: That just seems absurd, right? Has anybody
taken a step back and just said that this is like fitting, you know, a round peg in
a square hole or something? This is absolutely not going to be generalizable to the circumstance
we're going to find ourselves in.
Hawazin Faruki: I think there are many people on the committee
who are familiar that this is not going to work for the future, and I think that's been
brought up several times on those calls.
Dan Roden: Let's -- Debra, can you add your comment?
And then we're going to move on, I think.
Debra Leonard: So one of the issues here is that these stacking
codes, while they usually don't pay very well for single-gene tests, if you actually stack
them together -- you talk about being able to do a genome for $1,000 or $3,000 -- you
actually can get paid $1 million for doing a genome.
Male Speaker: I knew there was a reason I [inaudible]
[laughter]
Debra Leonard: So --
Male Speaker: [inaudible]
Debra Leonard: I'm sorry to tell the payers that, but --
Marc Williams: [inaudible]
Debra Leonard: That's right, so --
Female Speaker: The payers aren't aware. From a practical
--
Debra Leonard: It's got to be rationalized somehow for genomics.
Hawazin Faruki: Okay, well, and I want to add one other thing.
Really, I doubt anybody's paying $1 million, because most payers have limits on how many
times a recurrent CPT code will be paid.
Dan Roden: So Paul Ridker, Kate, and then we're going
to move on to Howard.
Paul Ridker: So two very quick comments, and then a question
for you. The comments are, to go back to those curves, this is really important, I think,
for Eric to sort of hear, as someone who lives in this very translational end of the spectrum,
and I'll have the opportunity to talk tomorrow more about this in length, but the X axis
is typically --
Male Speaker: You'll have 20 minutes tomorrow, not "at length."
Paul Ridker: Sorry. The X axis is typically 15 to 20 years,
and you need to understand that. And there's reasons for that that are very important.
The second is, your very last slide appropriately showed not just laboratory utility but the
clinical utility piece, and it's really hard. And I'll come back to that.
My question to you, though, is something that LabCorp -- I don't know if they do it or not
-- but your view of it. How do you feel about the marketing of this? I mean, what's the
role of your industry for selling these tests and the promotion of them? Because sometimes
what we've seen, at least in my field, are things being promoted that probably aren't
the right ones to promote, and that doesn't help us, versus things that we actually think
work not getting much promotion, because perhaps they're not particularly cost-effective or
set for someone, and just your perspective on if that happens, what we do about it.
Hawazin Faruki: [inaudible]
Female Speaker: Turn your microphone on, please, and speak
really into it.
Hawazin Faruki: Yeah, I don't know if I can speak for marketing,
but what I can tell you is that there's no marketing efforts that can achieve what the
clinical utility can achieve, and there's really no comparison -- we're talking about
orders of magnitude -- with any marketing effort versus what that clinical utility,
professional society endorsement, guideline, et cetera, does. There's just no comparison.
Female Speaker: So I want to point out that we're talking
a lot about CPT codes but we're not addressing the issue of ICD-9 codes, and so I can tell
you, as a clinical geneticist, the dearth of ICD-9 codes, once you make a diagnosis,
is incredible impediment to getting to care for patients. And I think that, while it's
not totally germane to this conversation, that I think, as you make additional genomic
diagnoses, the lack of having ICD-9 codes, even from known genetic disease, really actually
makes it very difficult to do -- even follow NCCN guidelines for screening, because we
don’t' have those codes. I just want to point that out.
Marc Williams: If I could just quickly respond to that. Of
course, we'll be moving to ICD-10 next year, which is one of the reasons that the molecular
codes got pushed back a year. That has better specificity, but certainly not perfect specificity,
to address the question that was brought up. But this whole issue of the coding problems
has been addressed in a number of publications, including secretary's advisory committee report
on oversight. It's a huge issue that the ability to actually impact the codes is limited because
of who actually maintains the code. The AMA running CPT, you know, groups like WHO and
CDC are involving in the ICD-9 and 10. And actually having tried to interact in those
systems to make these types of points that are being raised here is extraordinarily difficult.
Dan Roden: So that was a really great morning. We look
forward to more discussion this afternoon. We're going to change gears a little bit and
sort of intersperse. During the meeting will be reports from the current five or six -- I
can't even keep track -- working groups --
[end of transcript]