SINGAPORE: Genomic Predictors of Clinical Outcome in Gastric Cancer – Patrick Tan

SINGAPORE: Genomic Predictors of Clinical Outcome in Gastric Cancer – Patrick Tan


Patrick Tan:
Thanks very much, and it’s a great honor for me to present. My name is Patrick, and I represent
a number of different institutes in Singapore. And I think, in the next 15 minutes, what
I’d like to do is to give you an oversight of what we’re doing in Singapore and how we’re
trying in a early way to try to move some of the research that we’ve been doing into
the clinical practice. And I do like Dr. Manolio’s point that a lot of this must be pitched as
an evolution of current practice rather than as a revolution. And I think you’ll see some
of that here. Speaking from an even smaller country — [laughter] — than Belgium, where we have population
size of about 5 million, I think, you’ll see that this presents some opportunities but
also challenges as well. So how do I move this? So just one slide on
the overview. So this is be a very pragmatic, practical talk, I think, compared to maybe
some of the high-level stuff that you’ve been hearing before. Just one slide. So the — our
foray into biomedical science really began about 10 years ago. And it is over the past
five years it’s seen steadily increasing support from the Singapore government, and the last
[unintelligible] was about $2.6 billion in biomedical funding for the past five years. And this, the funding, arises from three major
ministries: trade industry, education, and health. And while that’s good in the sense
that there are many different grant agencies, it can be a challenge trying to align the
different types of needs of these different ministries when we are trying to talk about
something like genomic medicine. One positive aspect is that this has resulted
in a lot of research institutes. Some of you may be familiar with the Singapore Biopolis,
and also a number of academic medical centers, and some of my colleagues in Singapore are
here representing those. So I think that there is this receptacle, I think, for the use of
genomic medicine. And I’m just going to use stomach cancer as a basis for that in terms
of some of the experiences. So the area that I think Singapore has tried
to positioned itself is as a receptacle for understanding Asian-specific conditions. And
one of these is obviously Asian cancers. This is the main area that we’ve been working in,
and this is stomach cancer and the incidence of stomach cancer in different parts of the
world, and you can see that it’s primarily present in many parts of Asia. I think my
colleagues from Japan and South Korea will agree. But globally you will see that stomach cancers
are actually the second highest cause of global cancer deaths. So what this means is that
this is a disease that we need figure how to treat better. And I think the intersection
of this with genomics represent some very interesting opportunities in being sort of
like poster children for how we can use genomic medicine. Just two studies that I think highlight this,
and more, how do we use this in a way that builds on current clinical practice. So this
is just one study that is — highlights what is — that basically this paper published
in the Lancet highlights the use of transducer map targeting gastric cancers that are amplified
in the HER2 oncogene, similar to breast cancer. And this is the first targeted therapy in
the gastric cancer. And this, actually, is a positive for these patients, sort of amplification
of HER2, but for other, 90 percent, of cases, there are no targeted opportunities available.
So this is where genomics can come in, where we can do a landscape survey, find new opportunities.
And in this particular work, what we find, it was quite interesting from a kettle [spelled
phonetically] point of view is that besides the HER2 population, there are distinct segments
of gastric cancers that actually amplify different components of the RTK ras map kinase pathway.
And this obviously is a pathway where there’s significant type of therapies available. So,
in one sense, this is a very simple way of genomic medicine where we can stratify patients. Validation is very important to have all of
these studies, given the set of data. This is a one-off finding. And this is just a comparison
of the data from the Chinese gastric cancer patients to the KCG cohort, primarily patients
from Europe and Caucasia. And for the most part, you can actually see many, if not all,
of the similar amplifications being seen in both. So, in a sense, I think that for many of these
conditions, we have the validation and prevalence. But how, then, do we move this into the clinic
is, I think, another challenge that we need to think about. Another study that I think highlights, you
know, how do we use this as a way of evolving our current knowledge rather than as a revolution
that comes from a transcriptome profiling study where I think it really highlights the
challenge of how can we position this sort of work as an improvement on current pathology
rather than a “let’s keep the pathologists out of a job.” And so this is a study whereby using contents
clustering, we can fine treat distinct subgroups of gastric cancers. And by doing the standard
pathway analysis, we can find distinct pathways associated with each cancer type, and also
these different subtypes actually have different preclinical, at least, drug sensitivities.
This has yet to be validated. What was interesting from this is that this
particular subtype over here that I think represents how we can use this as an evolution.
So it’s been known since the 1960s that gastric cancers — and I apologize for those of you
who don’t work on stomach cancer, but this is the only thing that I know — [laughter] — is that a gastric cancers can be divided
into intestinal and diffuse. And this is the standard pathology practice. However, what’s interesting is that a number
of years ago, it turns out that this — that what pathologists have known for a while but
it’s sort of embedded in the pathology world is that intestinal subtype — actually there
are two particular variants: one type that has markers of normal gastric epithelial and
one not so. So this actually there’s full heterogeneity here, and when they map it to
the gene expression subtypes, it’s possible in that this actually may correspond to one
of those molecular subtypes and one of a different one. So I think that by talking to pathologists
and asking them how does this reflect your own personal experience, we can begin to build
bridges between the genomic sector and standard pathology. And so this is currently the consensus being
done. This is the pathway that most of us will learn in medical school about how stomach
cancer develops. But by intersecting pathology information, there may be a separate subtype
over here that represents the new genomic subtype we are seeing, and we’re working with
these pathologists to see if we can improve that. So this is a way of getting them used
to understanding genomic data. This is just one sort of approach. And, you
know, this is not just stomach cancer. I think that, over the past number of years, Singapore
has done pretty well in identifying genes, gene polymorphisms associated with different
types of Asian cancers. I just highlight, this is a very nice one,
where this is a deletion and 2.1 kb indels, so very hard to see by standard NGS sequencing
in them and only present in Asian populations. That seems to predict inferior responses to
tyrosine kinase inhibitors. And also this paper from the Cancer Science Institute looking
at prognostic markers in liver cancer and other Asian-specific cancer. So I think from a discovery biology point
of view, we’ve been doing a pretty good job at finding things to translate. So the next question is that how then do we
move that to the next step? And this way it comes — intersects with the medical system.
And this is where, you know, we have challenges but opportunities. So one of the things that we tried to do — and
this is not — this is — it’s not the solution. It’s a pilot thing — is to establish a program
called POLARIS, which is the North Star, this is where we want to get to in future, where
we can try to see, can we have a national structure to look at how we can implement
genomic medicine. It’s funded for three years, about $20 million, from the Agency for Science,
Technology and Research — one of those three agencies. And I just pilot the use of — clinical
use of genomic testing, starting with cancers and genetic disease. And in terms of operationally,
what it really means is a nationwide network of College Medical Pathologist-certified genomic
laboratories running the same tests, running common informatics systems, sample preparation
systems, so that we can standardize across the whole island. I mean, it’s a tall order,
so we’re starting in a very pragmatic, small way. Now show you what I mean, and we brought
one year all into this program. Some operating principles that we feel is
important — and I’d like to raise these for discussions — is something we felt that would
be important to — for genomic medicine to work. The first operating principle that we
thought would be important is that genomic medicine labs running your HiSeqs, and your
MiSeqs should be deployed within an existing clinical framework. So as opposed to setting
up your center as a stand-alone center, that — and the main — one reason for that is
just purely financial. If you’re going to be trying to reimburse, if you have a stand-alone,
that center’s going to be bleeding red right from Day 1. But if it’s part of a pathology
department, pathology departments can cross cover through parts of lab medicine and so
on. So I think that’s one thing that — and this allows them to see it not as a threat
but as an evolution of their current practice. The second thing is that, you know, in single
gene tests, probably have had your heyday. But I would say that for people that want
to start genomic medicine, you know, you need to have the framework to do basic genetic
testing, single gene testing before you can even start to think of doing a whole exome.
And I think that is — you know, so we need to learn by doing it, finding out where the
pressure points are, and moving on from there. The third point is that genomic tests should
leverage on existing competencies. And I think this way the genome England study is that,
you know, when you want to set up a new center, it is already a lab that’s doing really good
Sanger sequencing. You can use that lab for validation rather than having to set up your
own system. But that requires trust. And how do we get the funding to flow across different
centers to reimburse that lab for that? And so I think that this is a more cost-effective
way, but it does take more teamwork and more effort in play. So a lot of my time is spent
meeting people and taking them out for lunch. And I think last year, I think it’s one is
that starting with tests as initial proof of concepts that provide true clinical utility
will lead to clinician buy-in. So what we try to do is to highlight disease champions.
These are people working in a particular domain. They are very influential, and we ask them,
“What would you like to measure that would make you change the way you treat your patient?”
And we start there and we build a test around that, and then we use that as a basis to learn.
And they, themselves, participate in how the clinical report is being generated. So this is the current status of the one year.
Our first POLARIS test — I’ll spend the next two slides talking about — is a very simple
single test. But I think it has identified some of the issues that we need to fix. And
it’s going to be launched end of the month. The genomic labs themselves are going for
national certification in mid 2044. It’s going to be based on illumine. I’d like to talk
to people here about what do they think is the necessity for reflex validation, when
you see a new variant caught by, let’s say, on a MiSeq. How important do you think that
is? The test revenue — so these patients actually
being charged for these tests. And what happens is that the revenue that is incurred from
this test is equally split among the different centers contributing to the test on a cost
recovery basis. And so this is important for sustainability. And finally, the next POLARIS test that we’re
going to do, and by end of this year, is in MiD [spelled phonetically] gene GI, gastrointestinal
cancer test, that has all of your favorite players on it. So let me talk about this test over here because
it’s a Sanger test. It’s a very simple test. But I think, you know, there are certain — it
led us to identify certain things that we need to fix in the system. The tests is targeting
a condition called stromal corneal dystrophies, and most of the patients with this condition
have a mutation in this gene. It’s one gene, so it’s very easy to do by Sanger. But I think
it hits the flavors of the things that we can use to prove the use of genetic and genomic
testing. For instance, by testing these patients, we can clinch the disease diagnosis; the location
of the mutation in the gene directs different therapeutics that — therapies that the eye
surgeon has to use. And it’s important for screening of unaffected family members because
these patients that are — have the test, even though if there are no symptoms, should
not go for Lasik. So I think it’s all the flavors of something
that has true clinical utility. And the fact that we have a national center for eye diseases
allows us to fund our patients through this particular test. More importantly, it’s allowed us to try to
integrate a number of the different academic medical centers. So, one, the National Eye
Center provides the patients on consultation, the ordering. They do the blood collections.
So, all of this is intersecting with the standard hospital practice that we’re doing here. The
national hospital system provides the DNA sequencing and the mutation report and the
GIS bioinformatics. And so — and the test revenues are actually
split among these three places. So in this case, it’s actually this particular place
gets the bulk of the revenue from the test. And so at front [spelled phonetically] of
this, with how — so then challenges I’d like to close now, just three or four that I’d
like to — that I’d like to get your feedback on. The first one is that because of different
ministries trying to get legal and licensing agreements across the institutions, it’s quite
complex. So that’s the first one. Second one, Singapore has a very interesting
system that the moment a patient’s test crosses an academic medical center, that test is charged
at what we call a full rate. It can be. And so this is something that we need to fix in
the system because the goal of this is to get all of the people from different centers
working together. There’s a general lack of expertise in genetic and genomic counselors.
So one of — a lot of doctors have told to me, “We would like the test, we would like
to prescribe it, but I don’t feel comfortable with it without a qualified counselor sitting
in my clinic at the same time.” So this is something that we are trying to address by
sending more people for training. And finally, I think — and one of the things
I’d like to get back from this one, is that we lack official policies on how do you write
the — an informed patient consent form? How do you deal with these incidental findings?
And how do you aggregate your data from the patients that you’re running a test on, such
that you can have a database that itself is very valuable for future discovery? And what’s
the –the fact that it’s clinical service, how does that feed into research? So I’d like to close there. I’d like to thank
my colleagues in the audience: Professor John Wong, Professor Chng Wee Joo, who are here.
And thanks. And any questions? Thank you. [applause] Marc Williams:
Marc Williams, Geisinger. Could you talk a little bit about why TGF B1 was chosen as
your exemplar for a single? Is this a condition that’s more prevalent in the population, because
it’s not one that has — at least that I’m aware of in this country as being tremendously
prevalent. Patrick Tan:Yes. So, again, I think it’s very
pragmatic. We needed to have a poster child of a test that hit all of the bells of clinic
utility, so, firstly. And we need it to have a — what we call a disease champion, a clinician
that was willing to work with us over a six-month basis to polish and build a test. The third
thing was then there was actually research be ongoing in the center on this condition.
So they have all of the patients, the wild type and the mutant patients that we can assemble
to rapidly validate the test. So I think that — so we wanted something
that we could — at least from the technical standpoint, all of the ingredients were there.
But then we could — and we could use it as a test to see what the other challenges — the
framework for how do you do the reimbursement because I think that’s the important part.
The technology, we can swap in and swap out. But how do we do those other bits? I needed
something that we could stress test the system and find out where the weak points are. So we won’t make a million dollars from it,
but I think we’ve done about five to 10 patients already. And, you know, the — I think that’s
— and that’s been clinical benefit from these patients already. So I think that’s a — and
we’re going to do more. Once we get the framework set up, we can then use the same basis to
do other ones along these similar frameworks. So that was the reason. Teri Manolio:
Great. Teri Manolio. Very neat project. And I’m wondering the data that you generate from
this are going to be useful to others who might want to interrogate the TGF B1 gene
as well. Have you thought about how both your interrogating other databases to see when
you have a whole host of variants that may have no meaning at all, how you’re doing that,
and then how you can contribute to those databases, such as through things like the ClinVar database
at NCBI or others? Patrick Tan:So, speaking as a researcher,
I firmly believe the 100 percent about the importance, the value of open sharing. And
I think — however, I think that in order to reach that vision, we have to do it in
steps. And there are significant concerns, some of them emotional, some of them, you
know, from a patient privacy point of view as to what is the right version of the data
that we can open up? Another tension is that a lot of the funding
in our — was from the Ministry of Trade ministry. So they’re obviously interested in this from
a commercial standpoint. So you have that question of open access versus not. These
are all, I think, issues that can be solved with further engagement of the different — because
I think all of us have the same vision. But the getting there, I think just takes step
by step. So the vision, I think, is then to share all of that data that we have. We just
need to figure out the best way that won’t led us into trouble. Teri Manolio:
Sure. Well, and I might just suggest David Ledbetter and Heidi Rehm are here and are
quite familiar with data sharing in ways that probably don’t affect intellectual property
and that sort of thing, but might want to talk with you a little bit about getting those
data. Patrick Tan:Definitely. I’m here to learn
on this. Thank you. George Patrinos:
George Patrinos. This is actually a question I wanted to ask to Tim, to lecture before.
One important argument for policymakers to incorporate genomics into health care is that
they save money, the national healthcare expenditure. So have you considered doing economic evaluation
using this test so that it can become eventually cheaper for the patients? Patrick Tan:So, yes and no. And you may not
like what I’m going to say. But, again, it’s, I think it’s a very pragmatic. What we do
— are going to do is that we’re going to compare the cost of developing this test in
Singapore and doing it versus the cost of this same center sending the test out overseas
where — and so the price — we’re still working on the final pricing of the test, but we can
basically do it in-house at about one-third of the cost, particularly if you factor in
— so that’s basically the argument that we’re going to say for cost-effectiveness. Now, we could talk more about this later on
because a lot of this is patient pace. So but at the very least we can say that, you
know, we’re actually bring that test to Singapore, leveraging on our research, and to bring — at
least have that cost for the individual patient down. I think we’ll start there. Male Speaker:
The last speaker this morning is —

Leave a Reply

Your email address will not be published. Required fields are marked *