Tip:
Highlight text to annotate it
X
Robert Nussbaum: Let's see. Right there. So this is the report
with the authors, and I think I can point out, you can see that Les Biesecker and Robert
Green were really the conveners, and did the real heavy lifting in the discussions that
we had about this topic. I'm glad Les is here, and he's going to be making comments. Also
point out that Amy McGuire, who is here on the council, was also co-author was well as
a first author of an interesting paper that was published in Science recently on some
of the ethical aspects of this set of recommendations. And I'm very hopeful that Amy would be willing
to make some comments as well.
So, the ACMG recommendation is that all laboratories conducting clinical sequencing should seek
and report expected pathogenic mutations for a short list of carefully chosen genes and
conditions. And the reason that I put this up is because, as you'll see later, this has
been a report which has been met with a lot of, I'd say, some misunderstanding, and perhaps
some miscommunication. And I think it's important to remember what it is that this recommendation
was about. The idea was to generate a minimal list. It's -- the importance was to establish
a precedent for the idea that incidental findings needed to be reported. It is a recommendation
and not a requirement.
So why did ACMG feel that there was need for guidance? Dr. Mike Watson is here in the audience
as well, I saw Mike, and maybe he would like to also comment on this. But I think it's
because there was a felt to be a lack of consistency and consensus on what to report in the face
of growing clinical use of whole exome and genome sequencing for specific indications.
So these would be children who had an unsolved developmental delay problem, or you may have
a family in which there is an unexplained familial cancer syndrome that's not -- that
-- for which no genetic basis had been found, and a variety of other sorts of applications
where you're taking a whole genome or a whole exome approach to try to understand a disease,
and recognizing that in addition to, perhaps, finding the underlying basis for the problem,
you're going to find other things as well. And it just didn't seem to be a general consensus
as to what the laboratory's responsibility was. So the idea was to provide guidance for
the laboratories.
So what's in the ACMG report? It defines obligations of laboratories to report incidental findings
that meet a high threshold of clinical utility. And that high threshold included high penetrance,
presymptomatic intervention was beneficial. It acknowledges that this is a work in progress,
and it recommends a mechanism for ongoing review and revision. So the report is not
without controversy. Some people say the report asks for too much. That's -- sorry -- T-O-O
much. I know those things you don't want to hear, but I'm enjoying the hell out of saying
them.
So the idea was that we were forcing information down people's throats. The report does not
say that patients should be forced to learn information they may or may not want to know.
What the report does say is that these recommendations for laboratories who have little, if any,
relationship with the patient, and it's not recommendations for the physician, whose professional
relationship lead to ordering the test. It does put a lot of extra burden on the physician.
On the other hand, the physician has ordered this test, and it's -- and I think this emphasizes
the importance of appropriate clinical consent when a patient and the patient's family and
the physician are deciding to undertake such a kind of broad genetic sweep in trying to
make a diagnosis.
The other thing is that the report contradicts previous testing policies for children. This
is something else that has been said. And I have to say the report does not recommend
overriding the autonomy of children deciding whether to be tested when families know the
children are at risk for adult onset disease. So in a classic situation where a parent might
have, let's say, a mutation for BRCA1, and they have a child who's two years old, that
the recommendations are, and I think that many people accept, that that child be given
appropriate autonomy, and for a decision as to whether that child wishes to have testing
done. And so the child can do that when he or she grows up to the age where they can
make an informed decision.
But the report does -- so the report does not recommend overriding the autonomy of children
in that situation, but it does weigh the beneficence of alerting families of risk versus respecting
the child's decision making autonomy when the families don't even know they are at risk.
I think it is just a completely different situation. Where instead of saying, yes, we
-- you know you're at risk, and we're going to force you to have that information, versus
it's sort of difficult to respect the autonomy of somebody to make a decision about something
that they know nothing about. They don't even know they're at risk. And so it's still -- it's
a very different situation. And so I disagree with the claim that this report contradicts
previous testing policies for children.
Another complaint has been that the report asks for too little, that we left out pharmacogenetic
variants, carrier state for recessive disorders. And we also left out variants that might be
of some personal utility, but with some question as to their actionability. The idea was to
try to have a limited list, one that most people would agree on, constituted a set of
information that people would really need to know, and would be unhappy if they suffered
a clinical event later on which could have been anticipated, and intervened on, and prevented
if only they had known that it was present in their whole exome or whole genome sequencing.
Another complaint was that the report was developed in an ivory tower with no outside
input. The authors spent over a year discussing the issues. The authors reached out to a panel
of domain-specific experts to comment on the recommendations. The authors held public fora.
So I think as one person said, we all sat around and drank the Kool-Aid. I don't think
that that is really an accurate description of the process by which these recommendations
were made.
Finally, I would just like to point out, there's a very nice paper that came out in Science
Express on Ethics and Genomic Incidental Findings. Robert is a co-author, Amy McGuire is a co-author,
and although they both, as it points out, the disclosure here, although they were both
on the original AGMC Working Group that wrote the recommendations, this paper does not represent
the official views of the Working Group. But I think it is a very careful and thoughtful
contribution to the whole area, and I would recommend that people having a look at it.
And I think with that, I'm going to stop and take questions. But more importantly, be very
good to get comments from the other people in the room who really have a lot to say,
I think, on this topic.
Eric Green: Can I make a -- Les is here. Les, why don't
you at least come to a microphone because we have free chairs in our system -- just
come sit here. Because I suspect if we're going to have a discussion about this, you'll
be brought in. I suspect you'll be brought in; you might as well be close to a mike.
So we'll open this up for -- Jim.
Male Speaker: So I support the recommendations. And one
of the things that I think is critical that hasn't been extensively discussed is really
what is the nature of incidentals? So one of the criticisms is you have said that labs
need to go looking for these mutations in these genes. And I think what that fails to
recognize is that incidental findings in medicine are almost always arrived at after a methodical
analysis of the data that is before you. So if you think about a radiologist who's reading
a CT scan of the chest, she engages in a very methodical analysis of the data that's sitting
there in front -- incidental findings doesn't just leap out. I'm going to follow the contour
of the ribs, I'm going to look at the plural surface, I'm going to look behind the heart,
et cetera.
So I think the thing to remember is that when genome-scale sequencing is done, the data
has already been generated, and you're going to parse those data through a whole series
of informatics filters. And what the ACMG recommendations say is you should also parse
it with a filter that looks at these.
And the other thing I would just comment on as a physician is that all of this controversy,
it's understandable, and it's all the more reason that when you order -- that you should
be hesitate to order a broad test, right? You should be -- you should be seeking in
medicine, usually, to order narrowly-focused tests that answer a specific question. Sometimes
we don't get to do that. Sometimes we need to do a CT of the chest because there's something
wrong in the chest, and we don't know what it is. Likewise, I think much of the problems
that people envision with this can be gotten around if physicians do not just willy-nilly
order whole-genome sequencing. That's probably a bad idea, and most of us would agree. But
sometimes it's going to be the right thing to do because we really have no knowledge
of where to start.
Robert Nussbaum: Yeah. Tony.
Male Speaker: In all the grants that we're looking at this
time on newborn sequencing and others, there's always a component of that which says that
they're going take this information and try to understand from the patient or family point
of view, how they're internalizing it, what does it mean to them. So, you know, you say
that the criticism was done in an ivory tower, but in retrospect, do you think these types
of recommendations should have come out without having public fora with the stakeholders to
understand their prospective, in retrospect?
Robert Nussbaum: Well, I think there certainly were public
for a, but the public fora may have been too restricted in that they were predominantly
medical professional fora. And that's -- therefore, perhaps some more involvement with consumer
groups of some kind, I think, would have probably been a good idea. And perhaps we had hoped
that there -- that there would have been more of that at the fora that we held. Yeah.
Male Speaker: So some of the controversy seemed from disagreeing
-- to just from disagreement with the actual recommendations, but some seemed to be misunderstanding
of what those recommendations were, according to you. Is that all rectified and clarified
now?
Robert Nussbaum: I would say that -- I would say that most
of the controversy that comes from a misreading or misunderstanding -- that the fact that
there's been a misreading or a misunderstanding has been pointed out. And so I think that
part, I think, is better. I think there's still a fundamental feeling among some individuals
that it's just the wrong idea, just from the get-go. The whole thing is just a bad idea.
And I think, to some extent, that comes from a lack of personal experience with what it
means to be a clinician, and what it means to take care of a patient. This puts a lot
of responsibility on the physician-patient relationship.
I mean, so, for example, we had a meeting last week, the ICCG meeting, in which one
of the people in the audience got up, and said, you know, "I come from the orthodox
Jewish community, and these recommendations show a profound misunderstanding of some of
the cultural contexts. We would certainly not want certain information passed on to
the patients." And I think the issue here is that it's the laboratory that's being asked
to report these to the clinician. The clinician should have the -- I mean, the laboratory
has no way knowing what the cultural background is of the patient on which they have sent
the sample. It's the clinician who should understand that, know their patients very
well, and know what -- based on the basic consent that has gone on between the clinician
and the patient, what needs to be communicated and what should not be communicated.
So these are recommendations for laboratories. I think that's kind of a fundamental problem
that people don't understand, and misunderstand the patient-physician relationship. Les.
Les Biesecker: I think that's a really key observation, and
I think the answer to your questions is "no." And I think the reason why it's "no" is I've
come to think of these recommendations as being kind of a Rorschach test for how people
think about our field. And I think what we have realized in the process of doing this,
you know, and as Bob said, we were working on this for 14 months, and wrestled with it,
and struggled a lot with a lot of these questions, and then launched them a bit abruptly, I think,
on people. And I think what the reactions reflect is as much a consequence of how people
view the field, because people have very differing views about what medical and clinical genetics
is or should be, how exceptional it should be compared to the rest of medicine. And I
think what I underestimated is how high and strong the walls are that we, as a field,
have built between ourselves and the rest of medicine. They're really impressive. And
I think people are shocked and dismayed that we may have actually taken down a couple of
those walls, and people are upset about that.
And we struggled for a long time, and no one in our group could come up with good reasons,
no coherent reasons that I can recall, why a number of these situations we're talking
about returning are any different from imaging, or routine clinical chemistry, or the physical
examination even. There isn't a good distinction, and no one has come up with a coherent reason
why these kinds of conditions should be handled completely differently than all those other
things in medicine. There's just -- there aren't -- I haven't heard one. And so what
I think what it reflects is people have notions about what medical genetics is and is not.
And it's a challenge for us to change some of those perceptions because this is a microcosm
of what we're really talking about when we say genomics needs to move into the mainstream
of medicine. That's the challenge that we are now facing, and I think it is a bigger
challenge than I initially anticipated that it would be.
Male Speaker: [inaudible]
Male Speaker: So I think this is a wonderful problem. The
response is fantastic. Because too often, when we put these things together, no one
cares. No one hears it. No one, there's nobody even bother with that. I remember when interventional
radiology started out, where we're talking about the radiologist -- the idea that radiologists
would actually do something with a patient -- I mean, these aren't real doctors, they're
people that read films. These are photomap people, not doctors. And yet they're going
to stick a needle where?
And so it's very similar now that your geneticists are people go do some things somewhere, and
now they're actually interacting with patients. So I -- the fact that there was a response,
and the fact that so many different types of people cared, I think reflects that the
field is starting to move into, maybe not mainstream [laughs], maybe not even close,
but at least towards something that is going to be more widely perceived. And hopefully
now the question is can we maintain that, and push it to the point where, you know,
people see the usefulness, and get the past just the laboratory, but also the doctor part
of it.
Amy McGuire: Yeah. So I think there is -- I think, Jim,
this has already been touched on, but I think there are two fundamental issues that are
kind of at the core of the controversy around this. And one is, what is the nature of the
test, and this issue of incidental finding. So I think that some people view this as dictating
the scope of analysis, which is typically an expert type of judgment that you make in
terms of I'm analyzing a test, how broadly do I analyze it, how deeply do I analyze it,
what do I look for, those sorts of things, which is not something we usually asked patients
about. It's a pretty technical question. Versus whether you're actually doing something outside
of the scope of the original test, and this is something different that you then need
to get permission for. So I think that's one fundamental piece of technical, I guess, controversy
that really informs the debate.
And I think the second piece is the nature of these results. And so there's developed
from a -- the ethical standpoint there's developed a really long history and discussion about
sort of the right not to know, and the role of patient preferences, and pre-test counseling
and stuff in genetics. And most of that discussion revolves around the types of results that
are late onset, you can't do anything about. A lot of people have expressed that they don't
want them. The uptake from genetic testing is very low. And the recommendations really
address things that are supposed to be, you know, actionable, highly clinically beneficial.
And so I think that there is debate about whether there is evidence to support whether,
in fact, we have that level of clinical benefit to justify what went on the list, right? But
that's the question of is what's on that list, what's the right thing, or is it not the right
thing, or do we have enough evidence yet, or have we not built enough evidence yet,
et cetera. But it seems to -- but in terms of the ethical issues that have really been
at the, I think, the crux of the debate around this. The question about how much we should
be asking patients their preferences for what we look for when we do the analysis, I think,
turns on those two things. And I think that the presumption of the group, if I -- I don't
want to speak on behalf of the group, but my understanding of our conversations was
that just like any time a patient goes in for a comprehensive evaluation, you can presume
that a reasonable person is going in there for an evaluation would -- and would expect
to be informed about things that are clinically relevant to them that they can do something
about. And we, all the time, do diagnostic work-ups on people with a general consent
to treat under that presumption that we will give them back information that that we find,
that the clinician finds to be relevant to them.
So I agree these are recommendations for laboratories. I wouldn't -- I would say that I think it
is little bit strong to say that the group didn't say anything about [laughs] the physician-patient
relationship. I think they did leave room. Just like when anybody goes in for any other
type of diagnostic test, patients always have the right to refuse information, and that
gets negotiated, whether the physician feels comfortable continuing a relationship with
that patient if they're not going to give them information that they think is important
that they now have possession of, and what kind of position that would put them in. So
there's always room for that, and that's -- there's always room for negotiation. But I don't think
that they were completely silent on or antagonistic about whether this goes from the laboratory
to the physician to the patient, or just from laboratory to the physician, so.
Robert Nussbaum: I think what I was trying to say was that
I thought that the -- that these recommendations actually put a greater emphasis on the patient-clinician
relationship. You know, and, so, for example in the last year I -- if you want to think
about a whole genome test that's being done fairly routinely, and I would say under the
presumption that Amy just described, which is that you're treating the patient, is Array
CGH. We're doing lots of Array CGH for children who have autism or children who have a developmental
delay. In the last year, I've had two patients, one with a deletion of Von Hippel-Lindau,
and the other with deletion in P53. So now I made a Li-Fraumeni diagnosis and a Von Hippel-Lindau
diagnosis. And, you know, you'd better be sure when you order a test like that, that
you warn the patients and the parents in advance. We may not find what the cause is. We may
find the cause. We may find something you didn't anticipate, and we may find things
that you may not even want to know about that we may find -- we may find out. So you and
I, in ordering this test, we have to be sure that we're on the same page as to what the
test is being done and what I'm going to look at, what I'm going to look for in this test,
and which I'm going to feel I need to communicate to you. You may have brought your child in
for developmental delay, but now we have to worry about Von Hippel-Lindau.
Male Speaker: So I want to echo that; I think that it's
incredibly important that this report came out, and that you guys clarify that it really
relates to sort of standards for labs versus the patient-doctor relationship.
One question I had was in your internal discussions, were there concerns about false negatives
and how those would be interpreted? So is this a question about, you know, in order
for you to be able to return exome results, you must meet the criteria that you could
find all known or expected pathogenic variants in these tests because when the doctor gets
it back, they're testing for X but they see that the patient is cleared for Y, that they
may return and say, "Well, we didn't see any BRCA1 or 2 mutation." But, in fact, since
that wasn't the primary aim of the test, it may not have met that kind of criteria. How
would those exceptions come about?
Robert Nussbaum: Do you want to comment on that? Go ahead.
Male Speaker: Yeah, it's a big concern. And it's a very
-- and Robert Green has unearthed the term, it's called "opportunistic screening," because
the data present an opportunity to detect a subset of things that are highly actionable.
And they are not -- they're obviously not a complete set of highly actionable things
that you might want to do for a patient, nor are they even nearly complete as to the list
of variants in those 57 genes that you might be concerned about. But we were very concerned
that we wanted to keep the positive predictive value high because the prior probability of
having the disorder is low. And so we focused on raising the threshold quite high so that
we were confident that if a secondary incidental finding was returned, it was highly likely
to be correct, acknowledging that we've sacrificed sensitivity in doing that.
And that's an issue -- and, you know, we deal with that in the clinical setting all the
time, clinical research setting, in that people who've undergone genome or exome sequencing
in my protocols call me up and say, "Oh, gosh, you know, I had nodule in my colon, or a polyp
in my colon; did you find anything in those colon genes?" And I say "Whoa, we are not
screening for those genes." If you have a clinical indication for that single gene test,
you have to get that single gene test. This is not a substitute for that. And so -- and
just as newborn screening is not a substitute for working up a kid who has a metabolic collapse.
If a kid has a metabolic collapse, you use the newborn screening results if they're positive;
if they're negative, that does not dissuade you from doing a complete workup of that baby
for that phenotype. So it's exactly the same. But it's a great question, because there's
a lot of presumptions about that, and we have to make sure that patients understand that
that's what we're doing.
Male Speaker: So did you come up with any language that
might, you know, company reports or something that could be communicated to that degree?
You know, even -- sort of what's the confidence in the genes that had been screened, and that
sort of label that might come...
Robert Nussbaum: Yeah. We didn't come up with any specific
language in the report, but, of course, all laboratories do report their level of sensitively
and specificity. So I think it's going to be implicit in that, but we didn't come up
with any specific language. Maybe that would be good thing to do. But we would do that
with the laboratories, [unintelligible] laboratories. Mike.
Eric Green: Well, let's go to the panel up first.
Robert Nussbaum: Oh, sorry. Okay.
Female Speaker: That's all right. So I just wanted to follow
up on something that Les said, because I think that it's part of the why there was misunderstandings.
And so, on the one hand, there's the argument that although many people have tried to find
some essential difference between what kind of findings you produce from a test like this
versus the kind of findings you might produce from radiology or from physical exam, and
so if you can't find an essential difference, then why are you treating it essentially differently?
It's an argument that makes sense. On the other hand, I think that one of the issues
is that there's also a contradictory sentiment that overstates the importance of genetics
and genomics, and that there's a lot of enthusiasm for these tests and a lot of interest in these
tests, and although, presumably, other technologies, when they were new, also had too much enthusiasm
and too much interest. But I think that it's the fact that that is as true, is recognizable
as the fact that there is no essential difference, right, that the people still treat them differently.
And so what needs to happen, which people have said, but I don't think was emphasized
in the report is that, yes, this speaks to the need for a very strong relationship between
the doctor and the patient. It also speaks to -- and to create that requires an incredible
amount of education because that is not typically available right now. And so you can say we're
going to put the burden here, but you have to be very aware that it might crash because
people can't necessarily pick up that burden. So, you know, like the issue that Carlos raises
is, is there language to talk about how much confidence you have in -- is there language
to talk about false negatives? I think that's the feeling, that there was a sort of prematurity
to the announcement because there are all these steps that still remain to be gone through.
Male Speaker: Just to respond. I think that those are all
great points. And I think, all that being true, I think that we've been in cycle for
a fairly long time of -- my perception is -- of trying to convince the wider medical
field that genomics and genetics is relevant to them. And they're saying "Go away. Come
back when you have something useful for us." And so I think there's a catch-22 there. And
I think this -- part of my reasoning for pushing this pretty hard is that I think we've just
somehow have to break out of that cycle, and it's going to awkward and difficult wherever
we break out of the cycle, but I think one of ways to do that is to put something in
front people, and, you know, they're using the test. Those results are there. And they're
sitting in front of their patients, and the labs are looking at these variants, and the
patients are sitting in the offices. There it is. And whether we like it or not, whether
we're all completely ready or not, we're doing it. And so we just have to -- I think our
feet are in the fire now. We have to figure out how to solve it.
Male Speaker: Yeah, I think one of the sticking points in
all of this has been the issue of patient choice. And it's very seductive to say, well,
patients should have choice. And, you know, before they have the test they should be able
to decide what gets reported and what doesn't. I think the real problem with that is -- I'm
not -- I think that's a kind of elucidatory choice. And I think that holding somebody
to a decision they make, you know, even if there was a half an hour of counseling, which
is probably unrealistic for every one of these test about that particular subject, and holding
them to a hypothetical decision that I don't want to be informed about something that we,
as a community, think is probably, you know, could be life threatening and preventable,
it is hard for me to do.
I think that the other thing that I would just bring up is, I worry more, in some ways,
about false positives than false negatives. I sat down with a patient just last week who
is in our CSER study, and I told him, as are the terms of our study, that we found MSH2
mutation in you that was utterly unexpected. We had sequenced him because of a visual deterioration.
And this was in our class of genes where we had said, in the consent, that we're going
to tell you if we find a mutation. Now the reality, as I think we all know, as we begin
to ascertain people through sequencing as opposed to family history, that the penetrance
for most of the thing we're dealing with is going to drop, right? And that's a real concern,
and those are data we have to collect.
But nevertheless, as Bob said, we have to make decisions now -- and as Les said -- about
how to integrate these things into medicine, and the best evidence we have now is that
a clearly -- what appear to be a clearly deleterious mutation in MSH2 is going to carry with it
a very high risk of a serious, but preventable, disease. And I think we owe it to patients
to act on that, while continuing to collect evidence to see, are we so wildly wrong about
penetrance estimates that would it obviate that responsibility? I doubt we are, but I
think we need to entertain that as a possibility.
Male Speaker: Mike.
Mike Watson: So just to address a couple of things that
Male Speaker: Turn on the mike.
--
Mike Watson: Hello. Hello.
Male Speaker Yep, there it is.
Mike Watson: Okay. So the -- just to answer a couple of
the questions. This report was -- this group did not operate in complete isolation on this
topic. There's another group in the college that's working on technical standards and
guidelines for the laboratories that include sample reports. I think it was approved by
the board. It's just delayed in publication. So a number of issues related to the clinical
laboratory have been addressed and should be out pretty soon.
I think that we acknowledge that there's a lot of system issues in place that are going
to make this difficult for many labs. You can't get paid for all that extra work right
now. So there's a lot of changes that would have to take place for this to be fully realized.
And we often wonder if we should have been a little bit more emphatic about that need
for system change for a policy like this to really be enabled.
And then I would say, echoing what Les said, there is a lot of this clinical testing already
going on, and the bar has gotten pretty low on the kind of results being returned. And
I think it was important to set some parameters around what's going to be acceptable in a
health care system. Otherwise, everybody is going to get another follow-up test that's
going to cost the system a ton of money. So that -- you know, those two extremes aren't
going to work, and finding that midpoint is what we're going to be looking for.
And I think there's a lot of things we need to downstream. And Jim alluded to one already,
which is, we have to figure how we're going to capture this data. This is the ultimate
unbiased ascertainment population that came in for one indication. It is not what this
incidental finding is. It's unrelated to what they had. And if we don't find ways of capturing
that unbiased ascertainment data to really understand the risks associated with these
people in their particular medical and family histories, then we're missing a huge opportunity
because it is really the pilot for what's already starting to appear on the horizon
which is primary genetic screening. You know, none of us think that the cost equation fits
that at all, nor does the data right now. But this is the kind of incidental finding
data, it's the kind that can really inform how that whole part of genetics develops over
the long haul.
So I think that this could very important, as a research institute, that we being to
think about how are we going to capture this data to inform future indications for how
this kind of information is used. We are putting together the system by which we want to be
able maintain this list over time. I doubt it will be done before July, when we will
be ready to accept people's recommendations of what they think we should take off the
list or should be added to the list. I think those are my main comments.
Robert Nussbaum: Thanks.