1
DEPARTMENT OF HEALTH AND HUMAN SERVICES
FOOD AND DRUG ADMINISTRATION
CENTER FOR DRUG EVALUATION AND RESEARCH
PEDIATRIC ADVISORY SUBCOMMITTEE OF THE
ANTI-INFECTIVE DRUGS ADVISORY COMMITTEE
Wednesday, February 4, 2004
8:00 a.m.
Advisors and Consultants Staff Conference Room
5630 Fishers Lane
Rockville, Maryland
2
PARTICIPANTS
P. Joan Chesney,
M.D., Chair
Thomas H. Perez,
MPH, Executive Secretary
SGE CONSULTANTS
(VOTING):
Mark Hudak, M.D.
David Danford, M.D.
Richard Gorman, M.D., FAAP
Robert Nelson, M.D., Ph.D.
Susan Fuchs, M.D.
Robert Fink, M.D.
Victor Santana, M.D.
Norman Fost, M.D., MPH
Judith O'Fallon, Ph.D.
Ralph D'Agostino, Ph.D.
Mark Fogel, M.D.
Tal Geva, M.D.
Craig Sable, M.D.
Vasken Dilsizian, M.D.
Marilyn Siegel, M.D.
Phillip Moore, M.D.
MEMBERS
(VOTING):
Mary Glode, M.D.
Steven Ebert, Pharm.D. (Consumer Representative)
FEDERAL EMPLOYEE
(VOTING):
Mario Stylianou, Ph.D.
INDUSTRY
REPRESENTATIVE:
Samuel Maldonado, M.D.
FDA:
Julie Beitz, M.D.
Sally Loewke, M.D.
Susan Cummins, M.D.
Diane Murphy, M.D.
3
C O N T E N T S
Call to Order,
P. Joan Chesney, M.D.
4
Recap of Day 1,
David Danford, M.D.
4
Discussion of
Questions to Committee
11
4
1
P R O C E E D I N G S
2
Call to Order
3
DR. CHESNEY: Welcome back,
everybody.
4 Dr. Loewke is going to give us an
overview of the
5 questions again but, first, the
folk at the FDA had
6 asked for one of us to do a recap
of what we
7 covered yesterday and Dr. David
Danford, our
8 resident cardiologist, has
offered, under duress,
9 to do that.
10
[Laughter]
11
So, why don't you go ahead?
12
Recap of Day 1
13
DR. DANFORD: Thank you, Dr.
Chesney. The
14 subcommittee had quite a full day
yesterday of
15 excellent, very informative
presentations on
16 pediatric cardiac imaging and the
agents currently
17 in use to enhance that
imaging.
18
FDA began by identifying four classes of
19 these injectables, including
gadolinium agents for
20 cardiac MRI, radiopharmaceuticals
for evaluation of
21 myocardial function and perfusion,
microsphere
22 contrast for echocardiographic
image enhancement,
23 and iodinated contrast for
angiography and CT.
24
With the exception of the iodinated
25 contrast, which is labeled for
angiographic use in
5
1 children as young as one year of
age, all pediatric
2 cardiac use of these agents is
currently off-label.
3 Desiring to obtain information
about these agents
4 that would allow labeling for
pediatric use, FDA
5 assembled speakers to address,
one, what pediatric
6 subpopulations receive these
agents; two, what
7 diagnostic purposes are being
served; three, how
8 the imaging data affects patient
management; and,
9 four, what additional labeling is
needed.
10
From this, it was hoped that we could
11 determine what, if any, pediatric
labeling
12 information could be extrapolated
from the adult
13 experience and what research
studies should be
14 designed to obtain the data
required for
15 responsible pediatric
labeling.
16
Dr. Geva introduced the concept that there
17 are huge numbers of patients
living with congenital
18 heart disease in the United States
that are
19 surviving longer, and frequently
they have residual
20 anatomic and functional
cardiovascular impairments
21 and may have a lifelong need for
medical
22 surveillance that includes cardiac
imaging.
23
We repeatedly heard from multiple
24 presenters that unenhanced
standard, regular old
25 echocardiography was the imaging
modality of first
6
1 choice for most of these
patients. It accounts
for
2 more procedures than MRI, cath, CT
and nuclear
3 studies combined. Its shortcomings are
poor
4 diagnostic quality in certain
subgroups of
5 patients, like older and bigger
patients, those
6 with chest wall deformities or
prior cardiac
7 surgeries, those with pulmonary
disease and those
8 in whom the primary focus of
diagnostic interest is
9 outside the heart, for example,
aortic arch,
10 pulmonary artery branches,
systemic or pulmonary
11 veins. Unenhanced echo is also suboptimal when
the
12 diagnostic question is one of
coronary perfusion.
13
So, when standard echo fails to provide
14 the diagnostic information
required for management
15 of heart disease one of the other
imaging
16 modalities is selected. MRI is one of
those
17 modalities, and we heard from Dr.
Fogel that
18 gadolinium contrast is injected in
the large
19 majority of pediatric cardiac MRI
examinations.
20 MRI often provides images superior
to echo for
21 aortic arch and its branches,
pulmonary arteries
22 and veins and the systemic veins,
and it can also
23 provide information on myocardial
perfusion and
24 tissue
characterization.
25
The anatomic MRI
data is suitable for
7
1 processing into 3D reconstructions
that are
2 aesthetically impressive and
highly clinically
3 relevant for the guidance of
surgeons and
4 interventional cardiologists as
they plan
5 treatment. MRIs applications are limited
by
6 artifact when objects made of
certain metals are in
7 the field of
interest.
8
There was support in our discussions for
9 investigations to define the
appropriate pediatric
10 gadolinium dose, its safety in
children with heart
11 disease and the diagnostic
accuracy in pediatric
12 cardiac
applications.
13
Like MRI, cardiac CT is also superior to
14 standard echo for imaging of
extracardiac large
15 vessel abnormalities like aortic
aneurysm, double
16 aortic arch and other vascular
rings, pulmonary
17 artery sling, pulmonary branch
stenosis, aortic
18 coarctation and pulmonary systemic
venous
19 anomalies.
20
Nonionic iodinated contrast is used in
21 essentially all pediatric cardiac
CT exams. It
has
22 a long record of safe use in
children and is
23 approved for angiography in
patients as young as
24 one year old. The subcommittee heard
concern,
25 however, about radiation exposure
from CT imaging
8
1 and struggled with the issue of
separating the
2 risks of the contrast agent from
the risks of X-ray
3 exposure.
4
Like CT, cardiac catheterization with
5 angiography utilizes nonionic
iodinated contrast
6 material and X-rays. It has broad
diagnostic
7 applicability in a wide range of
conditions,
8 including both intracardiac and
extracardiac
9 anomalies, and is increasingly
performed as a means
10 to treat the condition by means of
balloon
11 valvuloplasty or angioplasty,
stent placement, the
12 creation of holes where they are
physiologically
13 advantageous and the closing of
holes where they
14 are not.
15
The diagnostic information obtained with
16 angiocardiography is, therefore,
often with
17 immediate application to
therapeutic intervention.
18 Even in the shrinking minority of
such procedures
19 now done for purely diagnostic
purposes the
20 anatomic details provided
angiographically often
21 guide surgical
treatment.
22
One speaker suggested that there were few,
23 if any, pediatric labeling issues
remaining about
24 the use of iodinated contrast
material in the
25 cardiac cath lab, but another
suggested that we
9
1 actually lack information about
the true maximum
2 safe dose and in some complex
cases in the cath lab
3 we enforce an artificial maximum,
resulting in
4 deferred angiography and return to
the cath lab for
5 second procedures that might not
be in the
6 patient's interest if it were
established that
7 greater volumes of contrast could
safely be
8 administered in a single
sitting.
9
We heard that nuclear cardiac imaging
10 differs from the modalities we
have discussed so
11 far, and its focus is not on
anatomy but on
12 function, blood flow and
myocardial perfusion.
Not
13 surprisingly, its applications in
the
14 cardiomyopathic processes and
abnormalities of
15 pulmonary blood flow and coronary
arterial
16 perfusion were emphasized. The use of
radioactive
17 pharmaceuticals to obtain this
information is
18 associated with radiation exposure
to the patient.
19
While there was support for studies to
20 determine the pediatric safety and
appropriate
21 pediatric dosing, concerns were
raised that NIH
22 guidelines for radiation exposure
in pediatric
23 research subjects may be an
impediment.
24
Contrast echocardiography employs
25 encapsulated air or other gas
bubbles to enhance
10
1 endocardial edge detection by
harmonic ultrasound
2 imaging. This adds information on
myocardial
3 function and perfusion to the
cardiac anatomic and
4 Doppler blood flow diagnostic
information that is
5 generally available on standard
echo.
6
The potential for pediatric application or
7 contrast echocardiography is
currently largely
8 unrealized as most pediatric
centers do not provide
9 routine contrast echo
services.
Nevertheless,
10 there was interest in obtaining
pediatric safety
11 and efficacy data for these
contrast agents as some
12 experts would estimate that as
many as five percent
13 of all patients having pediatric
echocardiography
14 would benefit from the clinical
information about
15 myocardial perfusion and
ventricular function that
16 contrast
provides.
17
Finally, representatives of a number of
18 national professional
organizations, including the
19 American Academy of Pediatrics,
the American
20 Society of Echocardiography,
American Society of
21 Nuclear Cardiology, Society of
Nuclear Medicine and
22 a representative of one
pharmaceutical company all
23 spoke in strong support of FDA's
initiatives to
24 promote responsible pediatric use
of these agents
25 through labeling.
11
1
DR. CHESNEY: Thank you very
much. That
2 was excellent. We should have asked you for
copies
3 last night.
4
DR. DANFORD: I wasn't ready
last night.
5
DR. CHESNEY: Thank you. I did have one
6 announcement to make. Dr. Hari Sachs had asked
me
7 to tell you that on March 29-30
the FDA and NIH are
8 jointly sponsoring a neonatal
workshop, in
9 Baltimore, which will cover pain,
pulmonary,
10 neurologic and cardiac issues, and
there is more
11 information available on the web
site for anybody
12 who is interested--and
ethics.
13
Dr. Loewke, would you like to get us
14 started on the job at
hand?
15
Discussion of Questions to the Committee
16
DR. LOEWKE: Good
morning. I just
wanted
17 to clarify a couple of points and
maybe run through
18 an example that might help the
discussion for
19 later, so bear with me here. Let me find my
20 slides.
21 [Slide]
22
I wanted to talk a little bit about
23 extrapolation. The agency has commented that
when
24 there is potential to use adult
efficacy data and
25 extrapolate that to the pediatric
population--I
12
1 just wanted to clarify that we
would fully intend
2 to do PK parameters, PK studies
and safety studies
3 in the pediatric population. So, the question
that
4 is posed to the panel today is
whether or not there
5 is any case in which we can
extrapolate efficacy
6 data to children so we wouldn't
have to do large
7 efficacy trials in the pediatric
population.
8
[Slide]
9
I don't want to beat a horse to death but
10 I wanted to just throw these back
up so we can see
11 what really is approved in the
pediatric
12 population, and reiterate that
everything that
13 really was talked about yesterday,
most of it is
14 being used off-label in the
pediatric population.
15
What I really wanted to focus our
16 discussion on is what products are
currently being
17 used in a large enough population
that additional
18 drug labeling would make a
considerable health
19 benefit and make efficacy trials
feasible.
20
I just wanted to walk through an example.
21 I hope it might help. Dr. Fogel talked
about
22 gadolinium yesterday and he had
identified that MR
23 angiography is performed in
patients with
24 congenital heart disease to look
at vascular
25 anatomy. So, I am thinking that obviously there
is
13
1 benefit in this particular area to
study the
2 gadolinium product. The question is that first
we
3 need to identify which patient
populations.
I
4 assume you are looking at things
such as anomalous
5 vessels, aneurysms, coarctations,
etc. that you
6 pointed out. So, what are the
relevant
7 populations? We just need to identify
what
8 specific groups of patients we
want to look at
9 anatomy for.
10
Within that population, are all the
11 abnormalities considered
equal? I don't mean
from
12 a clinical standpoint but I mean
from an imaging
13 standpoint. Could you do a general
angiography
14 exam with gadolinium and see all
of these different
15 types of abnormalities, or would
you have to change
16 your procedure or modify your
procedure for any
17 particular one? If that is the case, we would
tend
18 to probably exclude that. We want to try to get
a
19 homogeneous group of patients in
which all those
20 types of anomalies or
abnormalities of vasculature
21 you want to look at would be
captured in a standard
22 MRI angiography. I don't know if it is possible;
I
23 am just throwing it out
there.
24
Then, we would need to identify whether
25 just knowing vascular
abnormality--do you find that
14
1 clinically useful? Do we have to prove that
that
2 is clinically useful? If we have to prove
that,
3 how would we go about proving that
as part of the
4 clinical
trial?
5
Then, how do we validate the findings on
6 the MR angiography? I was thinking last
night
7 maybe many of these patients go on
to
8 interventional angiography. Maybe we could use
the
9 findings of that procedure to
confirm the
10 abnormality seen on MR. Maybe many of
these
11 patients go on to surgery and we
could use surgical
12 findings to confirm the
abnormalities picked up by
13 MR.
14
That is just sort of an example of how we
15 are trying to work through these
and the types of
16 information we are trying to get
so we can try to
17 figure out where to
go.
18
DR. CHESNEY: Just before you
sit down,
19 could I ask the committee and our
consultants
20 whether you have questions for Dr.
Loewke as to
21 exactly what they are looking
for? Yes, Dr.
22 D'Agostino and then Dr.
Fogel.
23
DR. D'AGOSTINO: If you break
down every
24 possibility we could go on
forever. Are
you
25 looking for some sort of general
type of indication
15
1 so that then there is sort of a
guideline or
2 response to these questions that
sort of gives some
3 input on how one would go about
putting the trial
4 together? I am just concerned that
the
5 specifications, you know, can get
very, very
6 detailed. When you were saying can you lump
these
7 together, there were some people
on this side of
8 the table shaking their heads, no,
you can't.
So,
9 does that mean that for each
possible condition
10 there is another trial, or are you
just looking for
11 some sort of generalities in terms
of how you could
12 give guidance to industry and the
FDA and some
13 sense from the advisory
committee?
14
DR. LOEWKE: We are trying to
capture what
15 information and what populations
are large enough
16 that we can pursue efficacy trials
that would give
17 benefit to the pediatric
population. If we
can't
18 lump patients, then we have a
problem.
19
DR. D'AGOSTINO:
Right.
20
DR. LOEWKE: But I needed to
hear from the
21 panel what they think we can or
cannot do, given
22 their
experience.
23
DR. D'AGOSTINO: I went
through this with
24 the FDA in terms of pain models
and we ended up, in
25 terms of analgesics, laying out a
tremendous number
16
1 of pain models, and
what-have-you. In the end
we
2 just said we can't fill page upon
page upon page
3 but there are some general
principles, and I am
4 gathering that that is where you
are heading, that
5 there are specifics but there are
still general
6 principles that would lead from
one condition to
7 another so that we could give you
decent input to
8 putting trials
together.
9
DR. LOEWKE: From the talks
yesterday,
10 generally a role that I was seeing
is that we are
11 largely doing these studies to do
anatomy. CT
does
12 that; MR does that. There are reasons to
do
13 perfusion studies in kids. So, knowing those
14 global areas, now I just want to
get a little more
15 to what specific populations shall
we be looking at
16 because why are you doing these
studies? Then,
how
17 we should at least design the
endpoints that would
18 have clinical value to the
community? Then we
can
19 go from
there.
20
DR. D'AGOSTINO: We do a lot
these--the
21 Framingham study, and we find
oftentimes that you
22 find calcium or something like
that and you get all
23 upset about it. To produce a better image of
that
24 that we don't know what to do with, or
anything
25 like that, isn't very
helpful. So, when you say is
17
1 it clinically useful, then does
that mean that we
2 have to be able to identify in the
protocol that it
3 actually has a clinically
meaningful condition that
4 is tied into it? Or, is it just an enhancement
of
5 the image that we may not know
anything about?
6
DR. LOEWKE: Well, that is
what we are
7 trying to get
at.
8
DR. D'AGOSTINO:
Right.
9
DR. LOEWKE: We don't just
want
10 enhancement. If it doesn't mean anything and
isn't
11 useful to the practicing
community, then that is
12 not the
endpoint--
13
DR. D'AGOSTINO: So, we need
a standard
14 beyond just the
image.
15
DR. LOEWKE:
Right.
16
DR. D'AGOSTINO: Thank
you.
17
DR. CHESNEY: I have Dr.
Fogel, Dr. Geva,
18 Dr. Fink and Dr.
Fost.
19
DR. FOGEL: I wanted to step
back for just
20 a minute. Yesterday we got a lot of
speakers
21 together and we heard all sorts of
wonderful talks
22 about how we use contrast agents
and how we might
23 eventually design efficacy and
safety trials in
24 children to prove that that,
indeed, is efficacious
25 and safe in children and adds
clinical benefit and
18
1 value to their medical
care.
2 I
guess I want to step back for one
3 second. I guess this was bothering me last
night
4 and this morning, that is, a
number of people
5 brought up the notion that a
number of these agents
6 are off-patent. We could sit here all morning
and
7 talk about what wonderful trials
we would design
8 and how we would do it but, from a
practical
9 standpoint, how will the FDA
approach, once we do
10 give recommendations--how will the
FDA approach
11 getting the trials done? I mean, is there
some
12 sort of carrot that you guys think
you are going to
13 stick in front of the
pharmaceutical industry, as
14 you did with the pediatric
exclusivity rule, that
15 would be able to accomplish
this? Or, is
this
16 really more an academic
discussion?
17
DR. CUMMINS: This is not
just an academic
18 discussion. I want to reassure all of you of
that.
19 I spent most of my time yesterday
talking about the
20 on-patent process. There is an off-patent
process
21 as well. That off-patent process is done
in
22 collaboration with the National
Institutes of
23 Health. It is specified in the
Best
24 Pharmaceuticals for Children
Act.
25 Annually--actually it has been a
couple of times a
19
1 year, the NIH lists in the Federal
Register drugs
2 that are high priority for study
and the FDA
3 develops a written request for
those high priority
4 drugs and issues them to
industry. If
industry
5 does not want to conduct the
studies, then those
6 off-patent written requests are
referred to the NIH
7 and the NIH then translates them
into request for
8 proposals and they are awarded for
study.
9
We have been doing this now for about 18
10 months. We have a process in place. A couple of
11 contracts have actually been
awarded. There is
a
12 coordinating center that is
coordinating all these
13 studies and there is definitely a
mechanism for
14 translating the recommendations
that we get from
15 you all into studies for
off-patent products.
16
DR. FOGEL: That is
great. I must
have
17 missed that yesterday. Thank you.
18
DR. CUMMINS: Well, I don't
think I
19 explained it enough so rest
assured.
20
DR. CHESNEY: Could we ask
you--we were
21 discussing this in the van this
morning--does the
22 NIH have money to do these
studies, or do they go
23 in a list and maybe the top one is
funded and the
24 other hundred
aren't?
25
DR. CUMMINS: Without being
responsible
20
1 for spelling out the NIH process
in detail, yes,
2 they have funding to do these
studies.
These
3 studies are not funded through the
NIH Foundation;
4 they are funded through the NIH
budget.
5
DR. D. MURPHY: They have
some funding,
6 yes. You know, Congress suggested that they
have
7 200 million dollars for this and
then appropriated
8 none. So, the issue is that the institutes are
now
9 each having to find this money,
and they have.
10 But, you are right, there clearly
are limitations
11 and you will have to march through
the priority
12 list.
13
DR. CHESNEY: Thank you, Dr.
Murphy. Dr.
14 Geva, Dr. Fink, Dr. Fost and Dr.
Nelson.
15
DR. GEVA: I wanted to ask a
question.
16 What are the advantages or what is
the incentive
17 to, let's say, get gadolinium for
pediatric
18 cardiology applications approved
by the FDA?
19
DR. CUMMINS: The approval
would allow for
20 labeling of a product in the
pediatric population
21 so it would give us data on
efficacy, on dosing and
22 on safety that we could then put
into the label.
23 Currently, as you all
acknowledged, we don't have
24 that information in the
label. The products
are
25 all being used off-label.
21
1
DR. GEVA: Perhaps if I may,
to just
2 answer some of the questions that
were asked
3 earlier about the patient
populations and specific
4
diagnoses--to answer your
specific question, I
5 would think that what you are
looking at is
6 essentially the entire population
of patients with
7 congenital heart disease as far as
designing these
8 studies. I don't think that it makes a lot
of
9 sense, at least early in the
process, to break it
10 down to very specific
diagnoses.
11
I can tell you that of the patients who
12 come for an MRI examination, as
Dr. Fogel mentioned
13 yesterday and that is the
experience in our center
14 as well, the majority get
gadolinium MRI studies.
15 So, to start breaking it down into
specific
16 diagnoses you would probably be
doing yourself a
17 disservice. Now, there are a number of ways
by
18 which you can address the issue of
efficacy and
19 that would be a fascinating
academic discussion.
20
DR. CHESNEY: Thank you. Before we get
21 down to that issue, I think Dr.
Fink had a question
22 and Dr. Nelson had a
question.
23
DR. FINK: I guess mine
actually coincides
24 with what Dr. Geva was
saying. I don't
really
25 think that you can measure
efficacy for these
22
1 agents and what we really need is
safety and PK/PD
2 data for their usage in children
because what I
3 heard yesterday is that there are
differences in
4 technology, and which agent is
most effective may
5 depend on how many tesla your MRI
scanner has; the
6 experience of the operator;
whether the
7 cardiologist prefers MRI or CT
after they have done
8 the echo. And, the effectiveness we are
really
9 looking at may depend on the
thoracic surgeon or
10 the skill of the interventional
cardiologist and
11 these are really agents for
helping to just do the
12 image. So, I think efficacy for these
agents
13 really is the imaging and what we
primarily want in
14 pediatrics is safety and PK/PD
data.
15
DR. LOEWKE: We need to give the
user
16 information about the performance
of these agents,
17 and we look at the image and we
compare it to a
18 standard of truth to show them how
it performs to
19 what may be currently used. It gives them a
sense
20 of how this performs; do they want
to use this in
21 place of something else. So, there is value and
it
22 is very important to look at the
efficacy of these
23 products.
24
DR. FINK: I didn't hear a
whole lot of
25 discussion yesterday about
different agents that
23
1 are used or what we are comparing
it to. It
seemed
2 like for MR it was pretty much
gadolinium and that
3 there wasn't a lot of variation in
agents.
4
DR. LOEWKE: I am not sure I
understand
5 your
question.
6
DR. FINK: For efficacy are
you looking at
7 the various efficacies of the
different gadolinium
8 agents or variation between, let's
say MR versus
9 CT?
10
DR. LOEWKE: No, you are
looking at the
11 particular MR agent and you are
comparing it to
12 what--depending on what we define
as a standard of
13 truth, whether it be a comparator
agent that is
14 already approved for the
indication or whether it
15 is a standard of truth such as
conventional
16 angiography that is currently a
gold standard for
17 use in the cardiac population
today. So, it
gives
18 the user a sense of where this
falls into their
19 arsenal, and how to use it, and
how to rely on the
20 information that they get from
it.
21
DR. FINK: But it would seem
that that is
22 primarily machine driven. A 3 tesla coil is
better
23 than a 1.5 tesla coil, or a 16
detector CT is
24 better than a 4 detector CT. It would seem like
25 the technology available has a far
greater impact
24
1 than potentially these
agents.
2
DR. LOEWKE: I think we heard
yesterday
3 that it is a combination of both
the agent, the
4 drug and the user too. It is a combination. We
5 try to put factors in place to
accommodate for some
6 of those issues when we design a
trial.
7
DR. CHESNEY: Dr.
Nelson?
8
DR. NELSON: I guess I have a
question and
9 a comment. I just want to make sure I know
which
10 are on-patent and off-patent so as
we are
11 discussing the issue I have a
sense of the public
12 health impact for the feasibility
of doing the
13 trials. I guess since Bristol-Myers Squibb
likely
14 is the one who put in those two, I
am inferring
15 from the fact that they presented
on nuclear
16 imaging that, in fact, the nuclear
imaging products
17 are on-patent. I am assuming most of
the
18 gadolinium products, unless there
is some fancy one
19 in the wings, are off-patent. The nonionic
20 contrast is probably off-patent
but the fancy
21 echocardiography bubbles are
likely on-patent
22 because that is new. Have I gotten that right? I
23 am just trying to understand which
are on- and
24 off-patent as we are discussing
these different
25 modalities.
25
1
DR. CUMMINS: Whether or not
a drug is on
2 patent is actually more
complicated than one might
3 imagine, and I would encourage you
to put aside the
4 patent status of any product and
focus on the
5 product. We really need your scientific
advice.
6 Then we can think about how that
fits into the
7 whole on-patent/off-patent
process.
8
DR. NELSON: So be it. Let me then
9 continue. I am still unclear about the issue
of
10 extrapolation and trying to
separate out in terms
11 of properties of the agent and the
resolution of
12 the imaging versus application to
the population.
13 Part of this, in my mind, then
translates to how
14 you would try to design a
trial. For example,
when
15 I listened to the echocardiography
presentation and
16 looked at the slides, it sounds
like that one of
17 the issues is the ability to
differentiate a
18 tissue-liquid interface and the
relationship
19 between the resonance of the
bubbles and the
20 harmonics of the machine in
relationship to the
21 harmonics of the tissue and the
harmonics of the
22 liquid--so, a very complex
interaction. If
you
23 said that what you need to see a
good image is a
24 resolution better than--I think
you mentioned 1 mm
25 or 2 mm, 1-2 mm, the question then
is under what
26
1 circumstances can you demonstrate
that you have a
2 product that gives you that 1-2 mm
resolution
3 assuming all other factors remain
constant as far
4 as tissue harmonics and liquid
harmonics.
5
So, if I was looking at a protocol and
6 asking do you need to do that in a
neonate to
7 answer that question, you know,
can you demonstrate
8 a 1 mm resolution, my question
would then be are
9 all other factors equal apart from
the properties
10 of the agent itself? That would be the
question
11 and then you would have to tell me
what are the
12 harmonics of the adult tissue of
the heart. I
13 would then say, well okay, if it
is the same use
14 adults; if it is not use
kids. So, that is
the
15 kind of technical question I would
ask in
16 evaluating a
protocol.
17
If then you said, okay, we have an agent
18 that has demonstrated 1-2 mm
resolution, do we then
19 take it to a pediatric population
and try and show
20 that we can find clinically useful
information?
21 That then goes to the next step
and I think it goes
22 to trying to sort out what is the
nature of the
23 kinds of questions that we need to
ask.
24
I must confess, you know, I understand an
25 image. If anybody in the audience is in art
27
1 history or art appreciation, I
mean imaging is--but
2 it is unclear to me--if you took
the gadolinium MRI
3 scans, if you see a double arch
you are going to
4 see a double arch and I am not
sure you need to
5 know 1 mm resolution to see a
double arch.
So,
6 some of the questions are going to
vary depending
7 upon the modality. It sounds to me, from what
I
8 saw, I confess it looked like most
of the complex
9 questions are in the nuclear and
the echo where
10 there are still a lot of
unanswered questions,
11 whereas in the CT and the MRI it
was more a
12 question of extracardiac use of it
for imaging of
13 vessels that you can't see in the
other modalities,
14 putting function
aside.
15
So, I think, you know, as we go through
16 these--I mean, all the questions
break down each
17 particular imaging modality so it
is not only
18 population but it is what do we
really need to see
19 via CT, via MRI versus imaging and
the like. So,
I
20 think separating out those
questions is important
21 and that is why when I think about
PK and safety
22 data I also consider the basic
properties of the
23 agent, independent of whether or
not you are using
24 it to find clinically useful
information.
25
I don't know if that helps.
You know, can
28
1 you extrapolate? If all you need to do is see a
1
2 mm vessel, find a 1 mm vessel in
an adult
3 basically, if that is all you need
to see. You
4 don't use kids until you have to
use kids.
5
DR. CHESNEY: Lots of hands
and lots of
6 lists here but, Dr. Siegel, do you
have a specific
7 response to his
comment?
8
DR. SIEGEL: To three of the
comments.
9 First on the patient population
for each
10 examination, I think the patient
population is
11 really more than complex heart
disease. There
are
12 several things that have been
brought up here.
13 One, the extracardiac or the
vascular lesions; two,
14 valvular lesions; three, simple
septal lesions;
15 and, four, complex heart
disease. So, if we
are
16 looking at that, and that does
bring up the point
17 of, as you said, if we see a
vascular lesions how
18 sophisticated do we need to
get? We have seen
it
19 and that is the end of the
imaging. So, looking
at
20 the patient population we need to
sort of deal with
21 those areas, where each of these
exams fits in and
22 do we need more to confirm it or
do we stop at a
23 certain
point.
24
I think the other confusion in my
25 mind--there are two things. There is endpoint and
29
1 gold standard. I think maybe we are
overlapping
2 them. The endpoint would be a clinical
outcome.
3 If we were doing antibiotics, you
know, does the
4 patient get better? An endpoint here is a
little
5 bit more difficult to define. But if you look at
6 it clinically, there are I think
at least two
7 endpoints. If somebody has a widened
mediastinum,
8 whether it is on a CT or MR to
look at the arch,
9 and we find something and we can
say what is the
10 clinical usefulness? The clinical usefulness
there
11 terminates further imaging
studies. We are
not
12 going to have a correlate on
that. So, in
some
13 cases the clinical usefulness is
that it terminates
14 additional imaging studies or
diagnostic workup.
15 In other instances it is going to
lead to further
16 evaluation or treatment. So, if we do a study
and
17 you see a septal lesion, it is
perhaps going to
18 lead to echo or
catheterization.
19
So, to me, the two clinical endpoints, at
20 least in a simplistic world, would
be termination
21 of additional studies and end of
the workup or if
22 we need further workup. That is a clinical
23 outcome.
24
Gold standard then is if we wanted to
25 confirm a lesion that needed
further workup, how do
30
1 we do it? Again, it is going to depend on
the
2 modality. If I am doing CT, you know,
probably
3 echo is going to have to be my god
standard if I
4 wanted to do a study. If this is an
incidental
5 pickup, it probably would go to
echo but somebody
6 might say do an MR. I mean, that is going to be
a
7 little tougher. If you design a study you
could
8 probably get very specific on what
you wanted to do
9 clinically after you find a lesion
and it is up to
10 us perhaps to suggest or up to a
clinician to
11 decide what they want. But I think we are
dealing
12 with, you know, gold standard and
endpoint here,
13 and gold standard might vary for
each study.
14
DR. CHESNEY: Thank you. I have Dr.
15 Dilsizian, Dr. Gorman, Dr. Geva
and Dr. Fogel.
16
DR. DILSIZIAN: I guess I
wanted to
17 respond to Dr. Loewke's
request. There are
two
18 questions you asked. One is extrapolation
from
19 adults to kids. I think yesterday we all
said
20 adults and kids are
different. But, at the
same
21 time, I would like to emphasize
that a lot of the
22 things we use, let's say, in
nuclear medicine
23 perfusion imaging and
function--the concept of flow
24 and function can be extrapolated
from adults to
25 kids but what we need to do then
is that at the end
31
1 it would be wise to test these in
adults because
2 the dosimetry is much more
favorable. Once
you
3 have shown your efficacy and the
accuracy in
4 adults, now you can, in essence,
apply this in kids
5 but with the caveat that it has to
be retested
6 because their vessels may be
small; the organs are
7 smaller; the radiation exposure is
now different.
8 But I think that it is perfectly
safe or wise, at
9 least in my mind, to have it
approved in adults at
10 first and then accept it in kids
but then repeat it
11 in kids to see whether a
difference exists or not.
12
The reason I say that is because, for
13 example in nuclear, as you know,
perfusion and
14 function has been approved by the
FDA. It is
one
15 of the few indications that has
been done. But
we
16 now are extrapolating use in kids
but we haven't
17 really tested in kids. So, I think it would
be
18 wise, again, for echo bubbles or
DTPA to do the
19 same thing. I think we have to first show
efficacy
20 in adults and then apply it in
kids. I know
they
21 are different but, given the
safety issues, I think
22 it is always wiser to test it in
adults first.
23
DR. CHESNEY: Can I ask you
how do you
24 define
efficacy?
25
DR. DILSIZIAN: There are two
approaches.
32
1 For perfusion imaging one would
say, in adults for
2 example, I would like to detect
coronary artery
3 lesions so I can angioplasty that
lesion or send
4 the patient to surgery. Therefore, we use
5 traditionally coronary angiography
as the gold
6 standard and say can I
non-invasively predict a
7 perfusion defect which will then
guide cath and
8 angioplasty or
surgery?
9
We have learned, however, since then that
10 there could be perfusion defects
that are not
11 necessarily anatomical. There could be
12 vasoconstriction or other
physiological parameters
13 of hypertrophic cardiomyopathy
where we have no
14 coronaries but the demand is
different.
So,
15 physiological information is not
necessarily
16 equivalent to anatomical
information.
17
So, the next question is how do I judge
18 those patients? I hope I made the case that
in
19 those patients you look at
outcome--syncope, sudden
20 cardiac arrest--to see whether
identifying those
21 patients and treating them or not
treating them
22 changes the outcome of that
patient's symptoms.
23 So, those are the two
endpoints. One is
an
24 anatomical correlate as a gold
standard and the
25 other one would be
outcomes--syncope, sudden
33
1 cardiac arrest or some other
adverse events.
2
DR. CHESNEY: I think Dr.
Siegel mentioned
3 yesterday that accuracy is
efficacy here, efficacy
4 of diagnosis. Next, Dr. Gorman and then Dr.
Geva
5 and Dr.
Fogel.
6
DR. GORMAN: The question of
gold standard
7 is one that we didn't discuss much
yesterday but
8 the clinical definition I think
has already started
9 to be expressed and expounded
here, which is that
10 the clinical definition of a gold
standard is if
11 you stop intervening at that time
and your clinical
12 predictions come true, then you
made a clinical
13 diagnosis that was
appropriate. If you
continued
14 to intervene after you do a
procedure of any of
15 these sorts and the next procedure
confirms your
16 diagnosis, then it was again an
efficacious
17 procedure. So, you then begin to have a
moving
18 gold standard target which is that
for each of
19 these many lesions that we could
discuss there is a
20 series of modalities that would
help diagnose them.
21
Clinically, there is a pediatric
22 population that is enriched for
cardiac lesions and
23 everyone undergoes cardiac imaging
and I would
24 suggest them as a potential first
place to start
25 study design. Every single one of those undergoes
34
1 a cardiac imaging procedure I
think in the United
2 States for whether they have
clinical findings or
3 not. They also have potential
benefit.
4
DR. CHESNEY: Dr. Geva, Dr.
Fogel and Dr.
5 Sable.
6
DR. GEVA: To go back to the
efficacy
7 issue, I would like to expand on
Dr. Siegel's
8 comment and that is that there are
at least two
9 ways of defining efficacy in the
context of this
10 discussion. One is diagnostic accuracy and
it
11 depends on the specific trial and
the specific
12 lesion or group of lesions that
are being
13 investigated. One can choose an appropriate
"gold
14 standard" and that may be
something that has been
15 around for decades, such as
angiography which is
16 commonly accepted as the best that
is currently
17 available; surgical observations;
a compilation of
18 all available imaging tests--there
are several ways
19 of going about putting together a
reference
20 standard. Not all of these are true gold
standards
21 but they have been around long
enough and that is
22 what is being used most commonly
so if a new agent
23 or a new technique is being
proposed it is a common
24 thing to test it against
those.
25
Then, a different approach to efficacy is
35
1 to look at an outcome, clinical
outcome with the
2 use of a new diagnostic
technique. To give
a
3 specific example, currently all
patients, let's
4 say, who are candidates for a
certain surgical
5 procedure, let's say the Fontan
operation,
6 routinely undergo cardiac
catheterization and one
7 can design a study whereby instead
of routine
8 cardiac catheterization selected
patients undergo
9 non-invasive preoperative testing
and that is an
10 arm in a clinical trial. Patients are
randomized
11 to standard invasive testing
versus non-invasive
12 testing. Then one can look at set
clinical
13 outcomes--freedom from
intervention, length of stay
14 and so on and so
forth.
15
So, studies like that can certainly be
16 designed. Although you don't directly test
the
17 diagnostic accuracy of, let's say,
gadolinium MRI,
18 you are testing whether the use of
gadolinium MRI
19 can be used in order to achieve
equivalent clinical
20 outcome but with less cost, less
risk for the
21 patients, less radiation and so
on.
22
DR. CHESNEY: Dr. Fogel, Dr.
Sable and Dr.
23 D'Agostino.
24
DR. FOGEL: Yes, I have a
number of
25 comments. First going to the efficacy issue, I
36
1 just wanted to say that I strongly
support the
2 notion that efficacy is a very
important part of
3 this entire discussion. It goes towards the
whole
4 notion of, because we have seen a
potpourri of
5 diagnostic imaging modalities,
obviously, if you
6 have efficacy on the various
imaging modalities in
7 a given patient population or a
given category of a
8 patient population you can then
compare the various
9 diagnostic imaging modalities and
say, well,
10 imaging modality X is more
efficacious than imaging
11 modality Y in this particular
instance and that
12 would actually improve patient
management and
13 patient care in the sense that you
would then have
14 some real data to say, well, if
this patient comes
15 along with a certain likelihood
the best clinical
16 pathway for one to follow would be
to get imaging
17 modality X and then Y and then go
on to
18 intervention Z because we have
shown that X is more
19 efficacious than Y in this patient
population.
20
So, I think that that would be very
21 useful. It would improve patient safety
because
22 you wouldn't have to do sedation
for an
23 echocardiogram and then do
sedation for an MRI; you
24 could just do it once and then
move on to the next
25 diagnostic imaging modality or
therapy. So, I
37
1 think that would be very important
to do and I
2 voice strong support for
efficacy.
3
In terms of efficacy being a clinical
4 outcome, which I have heard a
number of speakers
5 talk about, we all have to
recognize that imaging
6 in an of itself, to use the
clinical trial
7 terminology, is really a
surrogate, and it is a
8 surrogate for something that is
really true, which
9 would be holding the heart in your
hand and being
10 able to see the whole heart, being
able to
11 miniaturize yourself down to a
teeny little person
12 and see that little coronary
artery and walk
13 through it. But apart from that, it really is
a
14 surrogate.
15
As such, with clinical outcome there is so
16 much that--let me step back for
one second.
The
17 imaging itself is just one
component of a
18 multi-faceted thing that is going
to happen to the
19 patient. There are all sorts of other
imaging
20 modalities that might occur, as
well as
21 interventions and postoperative
care.
22
So, although I guess you could design
23 trials that would have imaging
modalities and look
24 at the clinical outcome, I would
imagine you would
25 need a lot of patients and it
would be very noisy
38
1 because there are so many other
factors that go
2 into a patient's clinical outcome
other than the
3 diagnostic imaging modality. I think it would
be
4 very, very difficult in terms of
being able to show
5 efficacy in that particular
way. Now, if you
want
6 to do it against a gold standard,
that would be
7 surgical observation,
unfortunately, sometimes
8 pathologic observation. That is totally
different.
9 But clinical outcome sounds like
it would be pretty
10 noisy data.
11
Finally, the last thing I wanted to
12 mention is the extrapolation issue
of Dr. Nelson.
13 I have to say that I don't really
think you can
14 extrapolate from adults to kids,
as we all
15 mentioned yesterday. I don't think that if
you
16 have a 3 mm or 4 mm aorta in a
child you can then
17 say, well, can I see a 3 mm
coronary artery in an
18 adult? Well, if I can see a 3 mm coronary
artery
19 in an adult, then I can certainly
see a 3 mm aorta
20 in a child. That doesn't really work. There are a
21 lot of technical issues that go on
in there--tissue
22 attenuation, the size of the
patient, how big a
23 field of view you need to see the
various
24 structures--a lot of technical
things go into the
25 fact that I don't think you can
really do a good
39
1 extrapolation from adults into
children and I would
2 be very wary of doing
that.
3
DR. CHESNEY: Dr. Sable and
Dr.
4 D'Agostino, and then I would be
very interested in
5 polling all our experts to see if
they agree with
6 you. Let's do that right now, if you don't
mind.
7 Would you all agree that you can't
extrapolate from
8 adult data to children? I think that was one
of
9 the big
issues.
10
DR. MOORE: I would not
agree. I
would
11 say, just to focus on what I think
the issue here
12 of this subcommittee, whether
additional labeling
13 is required for some of these
agents and labeling
14 specific for pediatrics to make
sure that these
15 agents are safe and effective, I
would argue a
16 little bit along Dr. Nelson's
lines that gadolinium
17 and certainly iodinated contrast
have a lot of data
18 that is available both in adults
and children in
19 terms of their safety and efficacy
in other areas
20 in the body and in other
modalities which can be
21 translated over to cardiac
imaging. I would
argue
22 that the focus really needs to be
on some of the
23 newer agents and perhaps some of
the
24 radiopharmaceuticals and some of
the echo contrast
25 agents in terms of the specific
issues with safety
40
1 and
efficacy.
2
Just
to speak to that point, you know the
3 gold standard in many institutions
nowadays for
4 some of these cardiac lesions is
no longer
5 angiography; it is already
considered gadolinium
6 MRI or iodinated contrast CT. So, to then go
back
7 and say we are going to evaluate
efficacy in these
8 agents that are already clinically
being used in
9 many areas of the country as the
gold standard in
10 these applications doesn't make a
whole lot of
11 sense to me, and I think we can
extrapolate from a
12 lot of the data that is already
out there for some
13 of these very experienced
agents.
14
DR. SIEGEL: Well, I am going
to go the
15 opposite
way.
16
DR. CHESNEY: Dr.
Siegel?
17
DR. SIEGEL: I don't think we
can
18 extrapolate because of the various
varying factors
19 in children, which would be the
smaller size; the
20 faster heart rate; the inability
to hold their
21 breath; the motion. I think that is going to
make
22 it harder to see or more difficult
to see these
23 smaller
lesions.
24
As far as just following up on another
25 comment, I do agree that safety
issues have been
41
1 proven in the iodinated contrast
media, but I am
2 not sure about the efficacy
because that has really
3 not been shown in children. I think we still
have
4 to prove
that.
5
DR. DILSIZIAN: I actually go
somewhere in
6 between.
7
[Laughter]
8
And the answer is, as I said before, yes,
9 you can extrapolate but do test
again in the kids.
10 The reason I disagree with the
comments is that
11 everything we have talked about,
whether it is
12 gadolinium, micro bubbles or
perfusion, we tested
13 in adults first and then we are
testing it in kids.
14 The knowledge base came from
adults. We
15 extrapolate to the kids but we
haven't really
16 checked the efficacy in the kids,
which has to be
17 tested. Yes, there is extrapolation but test
again
18 in the kids.
19
DR. SABLE: I think, as
everyone seems to
20 be agreeing, it is not a simple
answer. First
of
21 all, we can't come up with a
blanket answer for our
22 different modalities. Just to use echo as
an
23 example, I think if you divide
patients by weight
24 or size above a certain age and
weight there is
25 probably reasonable utility to
extrapolating for a
42
1 given patient population. For example, a
14-year
2 old who had Kawasaki disease with
a structurally
3 normal heart would be a very
reasonable population
4 to study, very much based on
extrapolating from
5 adult data, although I think it
should be done in
6 children also. Conversely, a 3-year old who had
a
7 transposition repair in whom we
want to try to
8 assess regional wall motion I
think has a lot more
9 unanswered
questions.
10
Just to kind of cover one other thing
11 about gold standards versus other
ways to design
12 tests, I think a lot of us feel
that MRI or
13 contrast-enhanced CT may be a gold
standard for
14 some things, but the reality is
that in most adult
15 studies that I would pattern my
pediatric studies
16 after they are not using gold
standards because it
17 is much more difficult to design
tests using a very
18 subjective standard which is
widely accepted as
19 having a physician or a group of
physicians look at
20 different segments of the heart
and saying I can
21 see it well; a little bit; not at
all, and asking
22 the question does this modality
improve my ability
23 to see what I am trying to
see. Most tests
are
24 much more easily designed but
clearly not as
25 elegant as having a gold standard
such as MRI or CT
43
1 or the ultimate gold standard
which would be
2 surgical or pathology which we
rarely have.
3
DR. CHESNEY: Dr. Geva, can
we extrapolate
4 from adults to
children?
5
DR. GEVA: I agree with Craig
that this is
6 complex. There is no blanket answer. I would say
7 with regard to the gadolinium MRI
that it is age
8 related and you can extrapolate a
little bit to the
9 adolescent and adult with
congenital heart disease
10 perhaps. But when it comes to young children
with
11 small body size the answer is
no.
12
DR. CHESNEY: Dr. Loewke,
does that help
13 with your question about whether
we can extrapolate
14 adult to pediatric
data?
15
DR. LOEWKE: Yes, it
does. Thank
you.
16
DR. CHESNEY: Yes, Dr.
Fogel?
17
DR. FOGEL: Listening to all
my colleagues
18 talk, you know, I do agree that
for children who
19 are in the adolescent age group
that are getting
20 close to adulthood you could
potentially
21 extrapolate from adults to
children. But I
guess,
22 again using the terminology of
surrogate, when you
23 are talking about this you are
really talking about
24 using adult studies as surrogates
for looking at
25 childhood efficacy in these
patients. You know,
44
1 using surrogates has all sorts of
issues and
2 problems. I mean, the Fleming and Demetz
article
3 basically states that a whole lot,
and I would
4 still be very, very wary about
doing that.
5
But using gadolinium-enhanced MRI or CT as
6 a gold standard, if you do it
already why do more
7 clinical trials? I think what we are missing
in
8 the literature is rigorous,
large-scale trials that
9 look at this. We have numerous reports with
small
10 numbers of patients that add up to
a certain
11 number--maybe add up to a mildly
large number of
12 patients but we don't have
large-scale, rigorous
13 clinical trials that look at
it. Then, there
is
14 anecdotal evidence but I think if
we are going to
15 serve our patients properly we
need to have the
16 data to then show
them.
17 DR.
CHESNEY: Dr. D'Agostino, did you
have
18 a comment?
19
DR. D'AGOSTINO: I wanted to
comment on
20 the trial design. I am not sure, given what I
have
21 heard and what I know about these
procedures, that
22 clinical outcomes are necessarily
a useful way,
23 just to endorse what Mark was
saying, because there
24 are so many other things that go
along with the
25 actual decisions in terms of what
medical practice
45
1 is going to do beyond the
imaging.
2
The other comment is that I would have
3 thought, again from what I know
and what I have
4 read, that a simple trial that you
can do here is
5 basically to have the individual
go through this
6 procedure with and without the
imaging agent, or
7 different levels of the imaging
agent, and then ask
8 the question does the higher level
of the imaging
9 agent somehow or other add more
information to
10 improve the clinical decision on
that individual.
11 It is a simple trial and the point
is how do you
12 decide on the clinical
information. You know,
the
13 sort of subjective way of having a
panel do it, and
14 so forth, blinded or unblinded, is
a matter for
15 discussion but I don't think we
want to run to the
16 notion of clinical outcomes, and I
do think that
17 the trial design doesn't have to
be very
18 complicated and we should try to
avoid that.
But
19 the outcome being clinically
meaningful is a real
20 trick, be it a gold standard or
something else.
21
DR. CHESNEY: Dr. Sable, and
then I think
22 we will go on to question number
two.
23
DR. SABLE: I want to add one
more comment
24 about extrapolation. I think that it is
25 important--and I am kind of
biased--to
46
1 differentiate what I do from what
all of my
2 colleagues do. All of my colleagues are
already
3 using contrast in some percentage
of the studies
4 and that is probably the rule
throughout the
5 country. Conversely, there are almost no
pediatric
6 echocardiographers using contrast
and the idea of
7 us using contrast, although I am
obviously an
8 advocate for it, is a much bigger
leap. For us
to
9 even think about using it in our
clinical practice
10 needs an incredible amount of push
and support.
11 So, even if you could extrapolate,
if I have a
12 17-year old who comes into my lab
who has the exact
13 same criteria as an adult and I
want to do a
14 contrast study, it is going to be
a much bigger
15 issue for me to do it. But we do have
patients
16 that we would like to do in our
lab. So, the
17 practicality of the issue is that
even if you could
18 extrapolate, the pediatric cardiac
community needs
19 additional enhancement to
undertake contrast.
20
I will just kind of end by using the
21 example from Dr. Gardiner's talk
yesterday. A
22 company that makes Definity and a
nuclear medicine
23 agent was very adamant that we
think about using
24 his agent for a population of
maybe 4,000 studies a
25 year but didn't even mention using
one of his other
47
1 agent for a population that has a
million studies a
2 year. So, I think that just kind of brings
home
3 the point that there is just a
huge gap between
4 using contrast echo in the
practical setting and
5 using the other
agents.
6
DR. CHESNEY: Dr. Siegel and
then Dr.
7 Santana.
8
DR. SIEGEL: Just one comment
about the
9 research possibilities, I think
designing these
10 trials in children is going to be
difficult because
11 you can't really use different
concentration doses
12 of drugs. It would be very difficult to get
it
13 through an IRB and you certainly
can't do it in the
14 same patients. You would have a very mixed
patient
15 population.
16
One of the issues we haven't addressed is,
17 you know, do we need to get down
to the level of
18 doing animal research and really
getting back to
19 basics? It is the only way I think we will be
able
20 to look at different doses versus
enhancement and
21 different flow rates, if that is
important to you,
22 versus enhancement, and I don't
think we will be
23 able to do that on a pediatric
population.
Adults,
24 yes, probably but not in
children.
25
DR. CHESNEY: Thank you. Dr. Santana and
48
1 then we will see if we can
start--
2
DR. D'AGOSTINO: Can I make a
comment?
3
DR. CHESNEY:
Yes.
4
DR. D'AGOSTINO: When I was
talking about
5 the trial I was saying a simple
trial but I didn't
6 say it would be simple to
do.
7 [Laughter]
8
It is a different matter altogether in
9 terms of can you operate it. But the design of
10 running to a clinical outcome and
so forth I think
11 is a much harder to thing to do
and probably has
12 tremendously difficult
interpretation problems.
13
DR. SIEGEL: I think we are
proving this
14 whole thing is going to be
difficult to do.
15
DR. CHESNEY: Dr. Santana
first and then
16 Dr. Loewke.
17
DR. SANTANA: Having
experienced sitting
18 through pediatric oncology
committee meetings at
19 two separate meetings where we
discussed the issue
20 of extrapolation of adult oncology
data to
21 pediatrics, I have learned two
lessons that I think
22 may be relevant to this
discussion. The first
is
23 that although I think in general
we agree that it
24 is not wise to extrapolate adult
data directly into
25 pediatrics because there may be
different disease
49
1 processes; there may be different
issues of
2 tolerance; and ultimately there
are differences in
3 functionality, PK, organ maturity,
when forced to
4 think about this issue, the
pediatric oncology
5 committee did come up with a few
examples in which
6 we were able to fulfill the
criteria that the
7 disease process was similar enough
that it was not
8 ethical to do efficacy trials in
children, and we
9 should put our resources in doing
the type of PK
10 safety studies that are more
relevant.
11
So, the challenge I think for my
12 colleagues--although we all like
to say that in
13 general terms we should not
extrapolate, the
14 challenge is to come up with
examples in which you
15 can extrapolate and that will save
us time, effort
16 and safety for our patients so
that then we can do
17 those studies more wisely and
capture that data
18 quickly and get more information
out to consumers
19 and
practitioners.
20
So, that was just a word of wisdom by
21 extrapolation. We all like to say, no, let's
not
22 extrapolate; they are
different. But
force
23 yourself to think that there may
be scenarios in
24 which you will be able to
extrapolate and those are
25 the ones that I think we need to
bring forward to
50
1 resolve some of these
issues.
2
DR. CHESNEY: Dr.
Loewke?
3
DR. LOEWKE: I just wanted to
make a
4 comment that seeing more doesn't
necessarily mean a
5 benefit. These drugs are not without risk. So,
6 obviously, the utility of the
information you are
7 getting is very important and that
is, again, a
8 risk-benefit
assessment.
9
DR. CHESNEY: Dr. Glode and
Dr. Fink, and
10 then I think we need to push on to
begin question
11 two.
12
DR. GLODE: I just wanted to
clarify a
13 question and I think reemphasize
the comment that
14 Dr. Siegel just made. It seemed to me, or at
least
15 I wanted to confirm that for some
of these agents
16 not only dose but infusion rate
are issues to be
17 potentially
studied.
18 The
comment I wanted to make is just a
19 comment very similar to what Dr.
Siegel just
20 commented on in terms of if your
goal was to find
21 the lowest effective dose--again,
a presumption
22 that a lower dose translates to a
safer dose--I
23 don't know how you are going to do
that in
24 children. In animals, yes, and hope that
that
25 translates or something. But it does seem very
51
1 problematic to say here is our
standard dose X and
2 we are randomizing people to half
X, and the
3 endpoint is that we couldn't read
your study and it
4 gave us no valuable
information. So, now we
need
5 to sedate your child again and do
another study.
6 So, the study design is pretty
problematic in
7 trying to get to the lowest dose
that gives you an
8 interpretable
image.
9
DR. CHESNEY: Dr.
Fink?
10
DR. FINK: It strikes me that
we are
11 spending all this time talking
about these agents.
12 It is wonderful. It would also be interesting
to
13 see if equal time was spent
looking at the
14 equipment. How much of the equipment we
are
15 talking about is actually licensed
for use in
16 neonates? There are huge improvements
in
17 resolution at least with MR and CT
that could be
18 done with better design of the
equipment or
19 attachments that optimize it for
the infant where
20 you get the collectors and the
collimators much
21 closer to the
patient.
22
My guess is that there would potentially
23 be more to gain by equipment
redesign and algorithm
24 specifically designed for the
neonate than by the
25 dyes, and you might be able to cut
dosages far more
52
1 dramatically by getting
manufacturers of the
2 equipment interested in looking at
the problem.
3
Just out of curiosity, are any of these
4 devices actually licensed for use
in premature
5 infants or neonates? Because it seems like
they
6 come on the market for adults and
they get used in
7 kids because that is what is
available.
8
DR. LOEWKE: I don't think
that CDRH is
9 here--they were here yesterday--to
answer that
10 question.
11
DR. CHESNEY: Dr.
Maldonado?
12
DR. MALDONADO: Just about
that, actually
13 I approached Dr. Feigel, who is
the Center Director
14
of Devices, recently because
I was curious about
15 how we will go to approve a device
for children.
16 He told me that the Center for
Devices doesn't
17 approve those devices for
particular populations.
18 You are right, Dr. Fink, they are
approved for a
19 participant image in this case but
there is no
20 reference to where these devices
could be used.
21
DR. CHESNEY: Approved for
human use and
22 neonates are human. So. I keep putting off
23 question two but let's have two
more, Dr. Fogel and
24 Dr. Danford.
25
DR. FOGEL: Yes, I just
wanted to respond
53
1 to the question about dosing. At least for MRI
for
2 example, as I mentioned yesterday,
gadolinium is an
3 adjunct to the rest of the study
and not a study in
4 and of itself for the vast
majority of the studies,
5 not all but for the vast majority
of the studies.
6 So, if you have an MRI scan that
has half a dose of
7 gadolinium versus a full dose of
gadolinium versus
8 a dose and a half of gadolinium,
you wouldn't
9 necessarily get uninterpretable
information from
10 the entire study because you would
have done the
11 non-contrast part as well and maybe
gotten the
12 information but you certainly
would be able to make
13 a diagnosis. Now, would it change the
clinical
14 outcome? Would the surgeon not like it as much
as
15 if we had done the 3D and had them
take a look at
16 the 3D? Probably not but you certainly would
get
17 that
information.
18
If you address it along the same lines as
19 you would in a blood pressure
clinical trial, it is
20 the same thing versus getting a
placebo. I
mean,
21 you know, you have to accept that
when you enter
22 into a clinical trial there are
some people who
23 will benefit and some people who
won't benefit.
24
DR. CHESNEY: Dr.
Danford?
25
DR. DANFORD: I am going to
quibble for
54
1 just a minute with Dr. Loewke's
remark that we
2 really need to prove that better
imaging translates
3 into better outcomes. In an ideal world,
of
4 course, we would prove that but,
as a practitioner
5 in pediatric cardiology, I think
that the better
6 you see this stuff the better job
your surgeon and
7 your interventional cardiologist
is going to be
8 able to do for the patient. We haven't yet
reached
9 the plateau where we have such
high quality imaging
10 that we absolutely know
stuff. It is still
shades
11 of grey and degrees of confidence
and we are still
12 surprised sometimes by what our
surgeons find that
13 we were not
expecting.
14
And, I think the proliferation of all of
15 these imaging modalities that we
have heard about
16 speaks to that. You wonder why are we
developing
17 all of these things. Don't we
already have either
18 an accurate diagnosis or not? I think it is more
a
19 shades of grade phenomenon and the
better imaging
20 we get, I think the better
outcomes we are going to
21 have. I have no data to support that but I
think
22 that is
true.
23
DR. CHESNEY: Thank you, Dr.
Danford. I
24 know Dr. Siegel has to leave a
little bit early
25 this morning--oh, that is
different than my
55
1 question two. My question two says please
discuss
2 each of the following questions
for cardiac CT.
I
3 must have the wrong set of
questions.
Sorry.
4
DR. SANTANA: Dr. Chesney,
may I make a
5 comment?
6
DR. CHESNEY: Dr. Santana?
7
DR. SANTANA: As I have heard
all the
8 discussions yesterday and today, I
am still a
9 little bit like Skip was
yesterday, disoriented,
10 because we are talking in certain
scenarios about
11 anatomy, in certain scenarios
about perfusion, in
12 other scenarios about the tools,
the machines, the
13 operators, in other scenarios
about the agents.
14 So, one thing that would be very
helpful to me, as
15 we go through each of the
modalities, is if the
16 panel of experts, one or many,
could specifically
17 tell us what is the question that
is most
18 clinically relevant to them. If they were given
19 one choice to do a study with this
modality and
20 this patient population, what is
the burning
21 question that they want
answered. Rather than,
you
22 know, trying to design fifty
trials, it may be
23 better if they would help us or
the FDA by saying
24 this is the question that is most
relevant right
25 now. Let's put our money into it; let's put
our
56
1 effort into it; let's move
forward.
2
DR. CHESNEY: Thank you. That was maybe
3 your idea yesterday. Somebody raised that as
a
4 potential way of addressing
this.
5
DR. D'AGOSTINO: That is what
I raised
6 yesterday but was 24 hours too
early I guess.
7
DR. CHESNEY: Well, you
phrased it
8 differently in the van. It came out very
clearly,
9 what is the burning issue for each
one of our
10 experts. The FDA has put a lot of thought
into
11 these questions so we want to be
sure to address
12 them as well, but maybe each of
you could start by
13 saying in the best of all possible
worlds, this is
14 the question that I would like
addressed and then I
15 will address (a) through (f). Dr. Siegel, you
are
16 starting.
17
DR. SIEGEL: Okay, we will
start with
18 cardiac CT. I think there were sort of three
basic
19 elements discussed yesterday and
it is really
20 safety, dose and efficacy. If I look at that
for
21 CT, the safety has been
proven. My issue is
dosing
22 and actually other elements of
technique.
23
I don't know the dose that will work best
24 for CT. We use doses that are based on
information
25 dating back to the '60s and '70s
and that is the
57
1 standard dose we use now. My feeling is that
for
2 CT we can get away with a lower
dose. I have
used
3 it but we have no large series on
that. So, my
4 question is what is the minimum
dose that we can
5 use that will provide an effective
or diagnostic
6 image?
7
The other issue for CT is what is the flow
8 rate that will also provide an
effective and
9 diagnostic image? So, those are the issues I
need,
10 the more technical factors to
optimize a study for
11 children.
12
DR. CHESNEY: That is very
valuable.
13 Maybe we could go (a) through (f)
now and you can
14 just give us one-word answers and
then we will move
15 on.
16
DR. SIEGEL: Okay, imaging
agents further
17 study? No, I think it is a mature population
and
18 the safety of these agents has
been proven.
19
What population should be studied?
I
20 think we addressed that
before. We could divide
it
21 into four populations, the
vascular lesions,
22 valvular lesions, septal lesions
and complex heart
23 disease.
24
I will step back for a second and say if
25 we look at the vascular lesions
such as the aortic
58
1 lesions, the arch lesions and some
of the pulmonary
2 slings we may be able to
extrapolate on that.
3 There are series both in the MR
literature,
4 primarily in the MR literature and
some in the CT
5 literature and certainly in the
adult literature
6 that CT is efficacious for the
diagnosis.
Those
7 are large structures; it is going
to be valuable.
8
But I think the other three categories,
9 valvular lesions, septal lesions
and complex heart
10 disease are patient populations
that need to be
11 studied. You can further say patient population
by
12 age, and I think the age we really
need to look at
13 is the younger patients. For CT, those are
14
patients who are six years
of age and younger, the
15 ones who are more likely not to
cooperate or hold
16 their breath and are smaller in
size.
17
Moving on, what disease states should be
18 studied? To me, that is the same as sort of
the
19 patient population unless you have
another
20 definition.
21
What endpoints? Again,
endpoints, to me,
22 are going to be different from
gold standard, and
23 that would be clinical outcome
either leading to
24 further studies to validate the
finding on CT or
25 termination of imaging
studies. We could, of
59
1 course, talk about research but I
think that will
2 come a little bit
later.
3
How should a trial be designed?
If I
4 think the burning concern is dose
and flow rates,
5 as I mentioned, it is going to
have to be animal
6 studies. We cannot do that on children. It just
7 will not be approved. I can't imagine any
IRB
8 approving that. So, that would have to be
an
9 animal study with varying
doses. I have
the
10 numbers but I don't think we have
to say the exact
11 numbers. Varying flow rates and then looking
at
12 enhancement, standardizing the
study by automated
13 means and looking at various
structures in the
14 heart and even outside the
heart. That I think
is
15 the type of trial that I would be
designing.
16
By designing that type of trial you would
17 also be able to look at whether
there is diagnostic
18 information, whether we can see
these structures.
19 Hopefully, at that point we would
be able to
20 translate some of this use in
children.
Perhaps
21 these studies could also be done
in adults; they
22 are being done and we might want
to look at that
23 information when those trials are
completed to see
24 if we can extrapolate that
information and where
25 our starting point would be.
60
1
How should the standard for comparison be
2 defined? Is there a gold standard? I think if we
3 were to do those studies the gold
standard would be
4 cardiac cath. I think that has been the
gold
5 standard for a while. That is probably what
I
6 would suggest for the animal
studies.
7
I think if we do a pediatric population it
8 is going to be more different
because of the
9 radiation issue and we would not
really be able to
10 say let's do a cardiac
catheterization on
11 everybody; the risk is going to be
too great.
You
12 would have to redefine your gold
standard and then
13 I might say let's go for
echocardiography,
14 hopefully with some contrast agent
by that time, to
15 minimize radiation risk. That is always going
to
16 be the concern when we design any
study for CT.
17
DR. CHESNEY: Thank you. Comments? Dr.
18 Nelson and Dr. D'Agostino and then
maybe Dr. Loewke
19 could tell us if we have answered
everything for
20 question number two or
three.
21
DR. NELSON: I agree with
your
22 observations about the risk and
how it would be
23 hard to design a trial like this,
but let me see if
24 I can ask you a question that
might give a little
25 bit of an opening. Would there be a population
61
1 that might be going to surgery
anyway where the
2 surgeon would say if you don't see
this as well as
3 you would because you have done
half a dose of
4 contrast I can check it
operatively and it won't
5 put the patient in any different
risk relative to
6 having been exposed to the risk of
a lower quality
7 study because you have done a
lower dose of the CT?
8 If it is possible that the gold
standard would
9 still be done, in whatever
instance this might be,
10 and the person doing it would not
have lost
11 information that they wouldn't be able to
verify
12 at that time that you might be
able to make an
13 argument for putting the child at
that risk.
You
14 might, but it is a
reach.
15
DR. SIEGEL: Right now you
couldn't vary
16 the dosage. I think if I go and start
saying
17 instead of using 2 mL let's do our
studies with 1
18 mL I am experimenting without
approval.
19
DR. NELSON: I am assuming
you would
20 design a protocol that way. I am just thinking
21 that the point at which, from a
risk perspective,
22 an IRB might say it is justified
is if the gold
23 standard would still be done, and
at the time that
24 that gold standard would be done,
such as surgery,
25 the operator would not have lost
information that
62
1 they couldn't otherwise verify,
and there might be
2 a chance that they would let you
take the risk of a
3 lower quality
study.
4
DR. SIEGEL: In some places
it might, but
5 you are absolutely right, you
might say that I want
6 to do, you know, 1 mL/kg based on
the adult
7 work--it has to be based on
something, and that
8 would be a possibility. Then the patient--to
be
9 part of the study the clinician
would either have
10 to agree to do a cardiac cath
because he is going
11 to do it anyhow or the patient is
going to surgery.
12 I mean, I have that type of study
now, a rather
13 limited study, so I think that is
doable. But,
14 again, it is going to be a little
more difficult to
15 get through a number of
IRBs.
16
Just as a quick comment, a few years ago I
17 tried to get a similar study
through by saying I
18 would like to do patients with
reduced
19 milliamperage or current. We were using 200 and
I
20 said let me drop it to 150, 100
and then 50, and I
21 couldn't get it approved because
they were
22 concerned it wouldn't be a
diagnostic study and I
23 would be repeating it. So, by dropping
the
24 contrast, I think there may be the
same concern
25 about that. I think we can design a study. It is
63
1 going to be a little bit more
difficult to do given
2 the radiation. That is why I suggested the
animal
3 model. But I agree with you, there would be
some
4 possibility to do
that.
5
DR. CHESNEY: Dr.
D'Agostino?
6
DR. D'AGOSTINO: I have three
comments.
7 The answer to part (a) where you
said nothing needs
8 further study, I thought that was
the whole purpose
9 of the question, to sort of
identify which agents
10 do need further
study.
11
The second and third questions I have is
12 that if we had the design that I
was calling simple
13 before and one was the echo and
the other was the
14 imaging agent, that gives you the
two measurements
15 on the individual to make those
comparisons and try
16 to get the clinical benefit, and
so forth, so it
17 fits in very much I think with
what I was
18 suggesting
earlier.
19
The third question in terms of the dose,
20 couldn't one do some animal
studies, maybe some
21 sort of Phase II type of studies
getting some idea
22 of the dose, and then move on to
the Phase III
23 study where you have the dose
fixed and also the
24 injection rate, and so forth? I mean, a little
25 mixture of the animal studies to
get some
64
1 information and move to something
like a
2 dose-ranging study with a small
number of subjects
3 to give you an idea. The study for the efficacy
is
4 the fixed dose, fixed infusion
echo versus the CT.
5
DR. SIEGEL: Going backwards,
I think I
6 agree with you on the last
point. I think
I
7 mentioned that we start with an
animal study with
8 the varying doses and then
translate it to
9 pediatric patients using echo as
the comparison of
10 the
standard.
11
The contrast agents that are being used
12 for this have been studied in
detail. There is
a
13 lot of information out there. They are approved.
14 Their safety is known. I don't think we are
going
15 to see new contrast agents. It is not the
contrast
16 agents; it is really the dose and
flow rate that we
17 are dealing with. These are safe. They work. We
18 don't need to develop new
ones. What was
your
19 second
question?
20
DR. D'AGOSTINO: What I was
calling a
21 simple design before, that you
need two
22 measurements and you could do an
echo on an
23 individual and then the imaging
agent at a fixed
24 dose.
25
DR. SIEGEL: I agree. You start with the
65
1 CT and then we do the echo to
confirm it.
2 DR.
CHESNEY: Dr. Stylianou, you had
a
3 comment?
4
DR. STYLIANOU: I have a
comment also.
As
5 far as the animal studies are
concerned, even if
6 you do the animal studies you
still have to test in
7 humans eventually. And, my guess is that
a
8 clinical trial is probably
unrealistic because of
9 the toxicity involved. One possibility would be
a
10 case-control type of study. You could have two
11 groups and match them by some
characteristic like
12 age, body mass index or some kind
of
13 characteristic, and you can have a
study doing it
14 that way.
15
DR. CHESNEY: A prospective
study?
16
DR. STYLIANOU: A prospective
study.
17
DR. D'AGOSTINO: What are you
matching?
18
DR. STYLIANOU: At this time
I am not sure
19 how to match but at least it would
be a way--
20
DR. D'AGOSTINO: But what is
it? People
21 with two different
procedures?
22
DR. O'FALLON:
Stratifying.
23
DR. STYLIANOU:
Right.
24
DR. SIEGEL: But I don't see
how this gets
25 us to dose or flow rate issues.
66
1
DR. STYLIANOU: You test
it. You said
the
2 doses are already safe but is it
tested?
3
DR. SIEGEL: The current dose
is tested
4 but we don't know how low we can
go on the dose--
5
DR. STYLIANOU:
Right.
6
DR. SIEGEL: --and get a diagnostic
image.
7
DR. STYLIANOU: So, basically
you have to
8 test a lower dose to see if it is
effective.
9
DR. SIEGEL: Correct, and
again it can be
10 different in a pediatric
population because if you
11 get a non-diagnostic study you
have irradiated a
12 patient for no reason and then you
have to either
13 repeat that study or do another
study. That is
the
14 dilemma we are in with CT because
of the ionizing
15 radiation.
16
DR. CHESNEY: Dr.
Sable?
17
DR. SABLE: I think we need
to be a little
18 bit careful when we are designing
studies. If
we
19 are going to use echo as a gold
standard, which is
20 safe, simple, low cost and
portable, then why do we
21 need to do another study that may
be more risky?
I
22 think CT has a lot of wonderful
potential for many
23 things that are much better than
echo--
24
DR. D'AGOSTINO: Couldn't you
ask do you
25 get more information out of the CT
than the echo?
67
1
DR. SABLE: Well, if we are
asking that,
2 then we shouldn't be using echo as
a gold standard.
3
DR. D'AGOSTINO: It is not a
gold
4 standard, it is a
comparison.
5
DR. SABLE: I think when we
design our
6 studies we just need to be
careful--
7
DR. D'AGOSTINO: But you can
use a gold
8 standard if you have a gold
standard or you can use
9 a comparison. The question is do you get
some
10 information from the
CT.
11
DR. SIEGEL: Right. I mean, we are not
12 saying that CT becomes the first
imaging study.
13 Echo is still the first imaging
study. But
let's
14 say the echo is equivocal, then we
are going on to
15 CT, and I am basing this on our
adult population,
16 as I said, with congenital heart
disease which is
17 1,200 patients and we have done a
number--at least
18 300. We are doing them because of equivocal
study
19 or sometimes there is a murmur and
it is the first
20 diagnostic study we are
doing. So, the
question
21 is, you know, is it efficacious
and can we use it
22 if there is an indication for it
because of an
23 equivocal echo or because it is an
incidental
24 pickup. If it is an incidental pickup, do we
need
25 to go further? But I don't think this is a
68
1 first-line imaging
study.
2
DR. SABLE: Sure, and I
certainly agree
3 with all that. This is not a question that
CT
4 doesn't add a lot to equivocal
echoes; the question
5 is when we are designing studies,
if we are putting
6 echo as part of your study design
to validate CT,
7 then I think an IRB could look at
that and say,
8 well, why are you even doing the
study? I
think
9 that is a different question than
whether or not CT
10 adds to equivocal echoes. I think we need to
be
11 careful about using circular
logic.
12
DR. SIEGEL: Yes, I am
agreeing with you.
13 I think if we do the echo and it
is diagnostic we
14 don't go further. But we would have to
identify
15 the population that would have an
equivocal echo,
16 or perhaps postoperative if it is
Mustard or
17 Senning procedure and there is a
question of a leak
18 and you need a better
definition. That is a
large
19 population and perhaps the
postoperative patients
20 might be another population. But we are not
here
21 to really design the study in
detail right now.
22
DR. CHESNEY: If I could ask
FDA about a
23 procedural issue here, if we are
going to have to
24 discuss (e), trial design, on each
one of these we
25 are going to be here for several
days. I am
69
1 wondering if we can't just omit
(e) and--
2
DR. NELSON: And the ethics
disappear--
3
DR. CHESNEY: And I am not
making the
4 ethics disappear; just to get
through each one of
5 the questions for everything but
(e), and then we
6 address issues of trial
design. Can I get a
show
7 of hands from the pediatric
committee? Does
that
8 sound like a reasonable
approach?
9
DR. O'FALLON: I think we
could talk about
10 design in about three minutes and
get that off the
11 board. All right? May I do that?
12
DR. CHESNEY: Wait just a
minute, I have
13 to absorb
that.
14
[Laughter]
15
DR. SANTANA: Joan, I agree
with that. I
16 think if we frame the question
that Dr. D'Agostino
17 and I have been trying to push,
which is tell us
18 what is the question that is more
relevant in your
19 disease and what you want to do,
then we could have
20 a brief discussion about how that
trial should be
21 designed rather than discussing
every single
22 permutation of every possible
trial to be done.
I
23 think if we look at it that way we
should be able
24 to help the
discussion.
25
DR. CHESNEY: All right. So, before we
70
1 address any of these questions we
will just address
2 the most important thing for you
and how you would
3 like to set up the study, and then
we will come
4 back to these questions. Is that what I am
5 hearing? That is not what Dr. O'Fallon
is
6 suggesting, Dr. Santana. You are suggesting
that
7 we ask each person to tell us the
most burning
8 question and how they would design
the trial.
9
DR. SANTANA: Right, like Dr.
Siegel did.
10
She did that I think very
appropriately. She
told
11 us what her issues were if she
wanted to answer
12 this question. She wanted to look at dose. She
13 wanted to look at infusion
rate. She wanted
to
14 look at animal models and then she
was thinking how
15 she would take that into a
clinical trial,
16 comparing it to another
modality. If we have
that
17 kind of discussion, we may be able
to get some
18 comments like Skip was making
about whether it was
19 ethical or whether there would be
issues that would
20 have to be approached in a
different way.
21
DR. CHESNEY: I just have a
feeling that
22 we are going to be going on and on
if we get into
23 that. All right, Dr. O'Fallon, if you can
solve it
24 in three minutes we are wide
open.
25
DR. O'FALLON: I have been
sitting here
71
1 quietly, letting you guys have
your say, but I
2 think that we can cut through on
the issue of
3 design. I think that there is a basic
strategy
4 that applies to all of them. Not all of them
will
5 use every piece but there is a
basic procedure that
6 has to be used to go through this
and we don't need
7 to deal with it for every single
modality.
8
I think that basically you have to define
9 your study goals. Are you looking at
movement?
10 Are you looking at anatomy or are
you looking at
11 what? Disease identification, whatever? But you
12 decide the goal. Then you have to rank in
your
13 particular disease the contrast
agents that are of
14 most importance to you. Then you have to
define
15 what initial dose levels you want
to study based on
16 adult levels and/or animal models,
but, you know,
17 you have to decide what you want
to do. Then
you
18 do your pharmacokinetics and dose
levels and
19 flow--in this case flow levels,
but that would have
20 to be well defined before you went
into the
21 children. But then when you had realistic
levels
22 you would go ahead and perform the
PK and dose
23 level which could include flow
level studies.
24
Now you have to define your age groups.
25 Are you going to do it in
adolescents? Are you
72
1 interested in neonates? What are we dealing
with?
2 But you have to define that and
you would have to
3 do them I think in each age group
in order to
4 characterize the adverse
events. You
know,
5 everybody is making the assumption
that they know
6 what they are, but they have to be
defined to at
7 least get some preliminary data on
adverse events
8 in each of these age groups that
you choose to use
9 it for.
10
Of course, you have to define your success
11 endpoint which would be in terms
of image quality
12 or diagnostic utility. That you would have
to
13 design for each one of your
things. That is
what
14 you would be talking about up
here.
15
I mean, there is a basic strategy for
16 doing the design in these studies
and, like Dr.
17 D'Agostino was saying, it is
pretty much a simple
18 deal because they really do have
PK and dose level
19 information in order to provide
the kind of
20 information that will be needed
for labeling.
21
DR. CHESNEY: Dr.
Beitz?
22
DR. BEITZ: I would say that
what Dr.
23 Siegel responded with was really
excellent and is
24 the kind of thing we are trying to
get from the
25 panel. So, if we could go through the different
73
1 modalities in turn and just get
some brief answers
2 and then let the panelists and
other members have a
3 discussion for maybe five, ten
minutes afterwards
4 and then go on to the next, that
would be I think
5 plenty.
6
DR. CHESNEY: Thank you. So, we will
7 proceed to cardiac MRI. Dr. Maldonado?
8
DR. MALDONADO: I just have a
quick
9 question for Dr. Siegel. I think that you seem
to
10 be comfortable with the safety of
these contrasts,
11 as I heard, but I still don't
understand why you
12 want to go down in the doses if
you feel that the
13 safety is not a problem. The reason I ask this
is
14 because when we are trying to go
to small molecules
15 in my field, go down in the doses,
we are trying to
16 optimize safety without losing
much in efficacy.
17 Since you seem to be comfortable
with the safety,
18 are you trying to optimize the
efficacy with going
19 down with the
doses?
20
DR. SIEGEL: Well, i think as
we discussed
21 yesterday, we think that less is
better so it would
22 be nice to be able to use
less. The safety
is
23 proven. The other thing in CT is if we can
lower
24 the dose and give less volume we
may be able to
25 inject it faster and get better
enhancement because
74
1 if we can increase the flow rate,
then we can
2 increase our enhancement so we
will get better
3 images. They will be better diagnostically; I
am
4 not sure, you know, that they will
be better images
5 from our quality standpoint. So, lowering the
6 volume makes it easier to get the
total amount of
7 contrast in somebody who is
small.
8
DR. CHESNEY: Thank you. Dr. Fogel, if
9 you would first tell us what is
the most burning
10 issue for you if you had a
wish-list, and then
11 address (a) through (f) and very
briefly (e)?
12
DR. FOGEL: Sure. Well, in my mind,
I
13 have to say there are two most
burning issues.
One
14 is anatomy and being able to get
efficacy data and
15 safety data on anatomy with
relation to dose.
The
16 second burning issue, real
quickly, would be
17 perfusion and viability, which I
think is very
18 under-utilized in congenital heart
disease in our
19 patient populations and I think
gadolinium-enhanced
20 MRI could add greatly to
that. So, those are
in
21 general the two burning categories
which I would
22 like to see
addressed.
23
What imaging agents need further study?
24 Well, in MRI it is fairly
easy. The vast,
vast,
25 vast majority is gadolinium and
nobody is using the
75
1 manganese or the superoxide iron
particles,
2 although there are some studies
being done but I
3 don't know if they are being done
in cardiac very
4 much. So, for me gadolinium would be the
only
5 agent.
6
What patient population should be studied?
7 Again addressing the anatomy and
perfusion, for
8 anatomy I think you could probably
lump all the
9 extracardiac vasculature into one
patient
10 population. The key would be the size of
the
11 patient whether they be neonates,
infants,
12 toddlers, children, and then
adolescents.
There
13 was a good case made that you
could probably
14 extrapolate adolescents from adult
data so I am not
15 as strongly married to that as I
am to neonates,
16 infants, toddlers and
children. So, I think
those
17 would be the patient
populations.
18
In terms of the types of disease processes
19 and the patient population, it
would be those
20 patients who have extracardiac
anomalies like
21 coarctation, postoperative
tetralogy, postoperative
22 transposition. Those would be the
patient
23 populations--the postoperative
Fontan patients.
24 Those would be the patient
populations that I would
25 target.
76
1
In terms of perfusion and viability, again
2 I would say that we would have to
address both the
3 size issue--neonates, infants,
toddlers, children,
4 and I would put in as a patient
population the
5 people who are at most risk for
myocardial
6 perfusion defects and scarring of
the myocardium.
7 Those, for example, are patients
after a Ross
8 procedure where they get coronary
manipulation;
9 patients after transposition of
the great arteries;
10 after arterial switch procedures
who also get
11 coronary manipulation; and those
patients, although
12 rare, who have native coronary
artery anomalies,
13 like anomalous left coronary, and
come to medical
14 attention. All those patients would have
the
15 opportunity to benefit from
myocardial perfusion.
16
What endpoints should be used?
As Skip
17 was alluding to, I think the gold
standard would
18 probably be surgery, and for
perfusion I think
19 nuclear medicine would probably be
the gold
20 standard that I would use for the
perfusion defects
21 because that is the most widely
accepted, although
22 it still has issues with radiation
and things of
23 that nature. But this is a wish-list; this
isn't
24 how we would actually do it in
practice.
25
In terms of dosing, presently for
77
1 extracardiac anomalies, for
example, we usually use
2 a double dose of gadolinium so I
would advocate
3 maybe just four categories; double
dose, one and a
4 half, one and then half a dose of
gadolinium, just
5 thinking off the top of my head
how one would do
6 that and randomize people to those
four dosing
7 levels.
8
I would just like to point out that with
9 MRI gadolinium is an adjunct and
we will get other
10 information from the study which
will help the
11 surgeon. I would also have a gold standard
which
12 would be surgical
observation. You know, in
any
13 trial in the high risk procedures
that we do,
14 unfortunately, sometimes we will
have pathologic
15 observations but in either case we
will have direct
16 human observation which would be
the gold standard.
17
I also want to point out that when we are
18 looking at these dosages we know
what the upper
19 dose is and we have gotten a lot
of safety data
20 from anecdotal evidence from
various studies,
21 numerous studies in the literature
on upper dose of
22 gadolinium. It is the lower dose and the risk
of
23 not getting a diagnostic
gadolinium study rather
24 than giving too much and causing
toxicity. So,
I
25 think that is an important point
for us to
78
1 remember.
2
Then, how should the standard for
3 comparison be defined? Is there a gold
standard?
4 In answering the other questions,
you have to
5 necessarily answer that. So.
6
DR. CHESNEY: Thank you very
much. I am
7 hoping that when we get through we
can come back to
8 issues of study design. Dr. Nelson?
9
DR. NELSON: Mark, for
perfusion issues
10 now what tests are being
done? I mean, would
you
11 do a nuclear scan?
DR. FOGEL: Normally
what
12 we will be doing will be nuclear
scans and/or
13 cardiac catheterization to see if
there was any
14 coronary artery stenosis or some
microcirculation
15 perfusion abnormality. So, these would have
been
16 done clinically anyway and the
question would be
17 whether or not MRI--because of its
greater tissue
18 characterization, no ionizing
radiation, being
19 non-invasive--would have a benefit
so that in the
20 future you would be able to
obviate the need for
21 cath and/or nuclear studies to a
great degree and
22 just be able to use MRI
instead.
23
DR. CHESNEY: Go
ahead.
24
DR. NELSON: Just as a
follow-up, I think
25 there can be some general
principles outlined in
79
1
terms of trial design that
if, in fact, the gold
2 standard would be performed
anyway--I mean, I think
3 that is an important one, then you
want to avoid a
4 repeat procedure. So if, in fact, the
gold
5 standard would be done anyway and
the risk of a
6 repeat procedure is not there
because you would
7 then proceed to that gold standard
without
8 repeating your MRI, I mean, I
think that is the
9 general principle. So, I think you can
outline
10 some general principles of a trial
design that
11 would allow you to generalize
across all of these
12 possible
scenarios.
13
DR. CHESNEY: Dr. Ebert had
his hand up,
14 then Dr. Fost and Dr.
Santana.
15
DR. EBERT: Just a follow-up
question, Dr.
16 Fogel, I think earlier in your
comments you
17 mentioned that you could also
design this is a way
18 where you would not use contrast
in an MRI.
That
19 could also serve as a control in
some of these
20 studies.
21
DR.
FOGEL: Yes, well, what we
basically
22 do is we basically do the
non-contrast studies
23 first, if nothing else, as a
localizer to how we
24 are going to do the contrast
studies. So,
the
25 contrast is more of an adjunct to
it rather than
80
1 standing on its own merit,
although there are some
2 times when it does stand on its
own merit but as a
3 general rule we do the
non-contrast enhanced first,
4 get some information that way and
add more
5 information by doing the
gadolinium.
6
DR. CHESNEY: Dr. Geva, you
have some
7 expertise in this area. I wondered if you
wanted
8 to comment. Then I have Dr. Fost, Dr. Santana
and
9 Dr.
D'Agostino.
10
DR. GEVA: I just wanted to
comment about
11 the endpoint and reference
standard for a potential
12 study design. I would have some concern
about
13 relying on surgical observations
alone. Number
14 one, it does have its own
limitations. Although
it
15 appears on the face of it as if
the surgeon opens
16 the chest and sees everything,
that is far from
17 being the case. I would propose for
consideration
18 as a blanket reference standard
for studies on
19 diagnostic accuracy you might want
to look at
20 something like summation of all
available
21 diagnostic information on a
patient. Some of
these
22 patients will have clinically
indicated cardiac
23
catheterization with extra
angiography.
Actually,
24 some will also have CT. Some will have
surgical
25 observation. Some will have autopsy findings.
81
1 That information can be combined
together.
2
DR. FOGEL: I just want to
say that I
3 understand that surgery is not a
be-all and end-all
4 in and of itself, but any gold
standard has a
5 false-positive, false-negative and
sensitivity and
6 specificity rate. And, I think that for most
gross
7 anatomical manipulations that the
surgeon is going
8 to do for a diagnosis or for an
extracardiac
9 structure that they are going to
be sewing
10 together, they are going to be
deeply involved in
11 manipulating the tissue itself,
and the success or
12 failure of the surgery depends
upon how well they
13 manipulate the tissue we are
trying to image
14 non-invasively and that is the
best gold standard
15 that we have. I don't pretend to say that that
is
16 the be-all and end-all by any
means, but at the
17 moment I think it is the best we
have.
Comparing
18 it to echo and angiography, I
don't think that they
19 are gold standards in the sense
that for the things
20 we are talking about, that
patients have to go
21 surgery for, it is ultimately
going to be up to the
22 surgeon to be able to manipulate
the tissue in such
23 a way to have a good clinical
outcome for the
24 patient, and that seems like to
would be the gold
25 standard we want to shoot
for. Again, surgeons can
82
1 be wrong, heaven forbid, and it
certainly is not
2 100 percent of a gold
standard.
3
DR. CHESNEY: Thank you. Dr. Fost?
4
DR. FOST: A couple of
questions, Dr.
5 Fogel. So, you are proposing doing children
who
6 are already scheduled for a cath
to look for
7 perfusion problems, and you are
suggesting doing an
8 MRI before you go to
cath?
9
DR. FOGEL: Well, this would
be patients
10 who you would be considering who
might have some
11 coronary artery issues and some
coronary artery
12 problems. I mean, there would be
clinical
13 justification in all patients who
have coronary
14 artery manipulation that you would
want to see
15 whether or not coronary artery
manipulations that
16 were preformed by the surgeon, for
example after a
17
Ross procedure or after an
arterial switch
18 procedure, whether or not that put
any of the
19 myocardium at risk. We do have some
individuals
20 after those surgeries who then get
coronary
21 ischemia. We see this on the EKG and other
things.
22 Or, decreased myocardial
performance that might
23 suggest that there may be some
coronary perfusion
24 issues that we would need to
address. Now,
the
25 knee-jerk reaction and the first
thing you would go
83
1 for would be a nuclear medicine
study, and other
2 individuals would go for cardiac
cath.
Therefore,
3 as people are saying, you would
have done those
4 things anyway. These would be patients who are
at
5 risk who you would have done those
things anyway
6 for, and now you would add on the
MRI as an
7 additional
test.
8
DR. FOST: So, this would be
a
9 non-therapeutic MRI for this
child.
10
DR. FOGEL: Correct.
11
DR. FOST: And that meets
minimal risk
12 criteria.
13
DR. FOGEL: I would believe
so, yes.
14
DR. FOST: Would they need
separate
15 sedation for
that?
16
DR. FOGEL: Well, depending
on the age
17 group, they could potentially need
extra sedation.
18 That is
correct.
19
DR. FOST: And are there data
on that
20 question in adults? That is, does MRI predict
or
21 correlate with cath data for
perfusion problems?
22
DR. FOGEL: There is a number
of papers
23 that have been done in adults,
looking at ischemic
24 heart disease and comparing it
against PET, that
25 have shown that MRI was very good
in that sense, in
84
1 actually advocating the use of MRI
for that patient
2 population. Can you then say that the
coronary
3 artery disease that we see in
kids--can you then
4 extrapolate that from ischemic
heart disease to
5 congenital heart disease coronary
artery issues is
6 another question. I don't think you can but if
you
7 have information in adults saying
that it could
8 potentially be useful, then I
think that would be a
9 good basis for you to then go
ahead and move along
10 into kids.
11
DR. FOST: Might it be
different in
12 infants than
adults?
13
DR. FOGEL: Well, most of the
time the
14 microcirculation and the actual
obstruction that
15 you might find in the major
coronary arteries are
16 atherosclerotic in nature, as
opposed to patients
17 who have undergone
cardiac-pulmonary bypass and
18 actually taking the coronary
arteries and moving
19 them, and flipping them, and
putting them in all
20 sorts of other geometric ways you
might not
21 necessarily think that it may be
as efficacious in
22 kids as it might be in
adults. Plus, with kids
you
23 have smaller children and you need
a greater
24 resolution to tell differences in
myocardial
25 perfusion. In children you might need a 1 mL or
85
1 sub milliliter pixel size to be
able to tell issues
2 of hypoperfusion whereas in an
adult it may be 1.5
3 mL, 2 mL limit of resolution with
which you might
4 be able to tell perfusion
defects. So, you may
not
5 necessarily think that you could
do it in adults
6 and not doing it in
kids.
7
DR. FOST: Thank
you.
8
DR. CHESNEY: Dr. Santana,
D'Agostino and
9 Nelson, and we will let you go
first and then we
10 will go on to the next
question.
11
DR. LOEWKE: What I am
hearing is you are
12 looking at probably two types of
clinical trials,
13 one to get an anatomic delineation
type of a claim,
14 and one for a functional perfusion
type of claim.
15 I know you said this was a
wish-list but I have to
16 go back to your perfusion gold
standard, just to
17 throw it out there. Nuclear medicine is
not
18 approved. The radiopharmaceuticals are
not
19 approved for perfusion in
kids. So, do you
have
20 any other
suggestions?
21
DR. FOGEL: Yes, but it is
actually, in
22 fact, in clinical practice used
all the time in
23 children. I don't know the numbers
specifically
24 but the numbers were shown
yesterday. It was
a
25 considerable number of patients in
the childhood
86
1 population in whom it is
used. I guess outside
the
2 regulatory arena it is considered
the gold
3 standard. Cardiac catheterization
doesn't
4 necessarily address the
microcirculatory issues
5 that would be addressed with
perfusion defects that
6 are shown by MRI as well as by
nuclear studies.
7
So, I think if you were just going to use
8 cardiac cath alone it would be a
suboptimal trial
9 and less accepted by the general
community of
10 physicians than if you use
the
11 radiopharmaceuticals. I know that that
might
12 present a regulatory issue from
your standpoint but
13 I think you might have--and I
don't know if that is
14 a total brick wall that can't be
broken down or if
15 it is something that can be
finessed and
16 side-stepped, but I think it would
be better for
17 general acceptance among the
entire medical
18 community if something like
radionuclide
19 pharmaceuticals were used. And, I am not a big
fan
20 of radiopharmaceuticals but it is
a gold standard.
21 So, that is what I would
use.
22
DR. CHESNEY: That is why we
are meeting.
23 Dr. Santana, Dr. D'Agostino, Dr.
Nelson, and then
24 we are going on to the next
question. We have
just
25 had another question added so we
need to get
87
1 moving.
2
DR. SANTANA: Can you clarify
for me--I
3 should have asked this yesterday
but it didn't come
4 up until today when I realized
what you were
5 talking about in terms of your
potential trial
6 designs--how many times within a
given MRI can you
7 administer gadolinium or, because
it has such a
8 half-life time, is it that you can
only do it once
9 and you are over with
it?
10
DR. FOGEL: Well, we can give
it a couple
11 of times as long as the dose
during that entire
12 session does not exceed the
maximum dose which is
13 40 cc or a double dose up to 40
cc, depending on
14 the kilo body weight. We do that a number of
times
15 for the perfusion abnormalities
so, for example, we
16 will inject half a dose of
gadolinium, get three or
17 four slices, and then wait a few
minutes, inject
18 another half dose, get three or
four at different
19 orientations and then do that a
couple of times;
20 then wait five minutes and then do
the viability
21 portion. So, you get basically two for the
price
22 of one.
23
DR. SANTANA: So, you could
do a study in
24 which there was an intra-patient
escalation of
25 dosing once you defined what the
target lesion was
88
1 that you were after. So, to address some of
your
2 issues of dosing of gadolinium the
patient could
3 have an escalation--I was thinking
about anatomy
4 actually, not perfusion. Once you identified
what
5 the target lesion was that you
were looking at with
6 X dose, you could administer that
patient a
7 different dose and see if you
improved your
8 efficacy of defining that target
lesion within the
9 same patient. So, the incremental risk would
be
10 the risk of giving another dose
certainly, and the
11 incremental risk of more time
under the machine.
12
DR. FOGEL: Right, you could
do that with
13 half a dose and one and a half
doses, which would
14 actually add up to two doses. You can't do it
with
15 the double dose because that is
the maximum you can
16 give. And, you couldn't do it with one
dose
17 because you couldn't give one dose
and then do
18 another dose because they are the
same dose.
But
19 you could potentially do that with
half a dose and
20 one and a half dose so you could
simplify the trial
21 to a certain extent that way. That is a very
good
22 point.
23
DR. CHESNEY: Dr.
D'Agostino?
24
DR. D'AGOSTINO: Fortunately,
Victor just
25 asked half of my question. The other half is to
89
1 the FDA. If you did a design that had no
contrast
2 versus contrast at some fixed
level, would that be
3 an acceptable design if you could
show clinical
4 benefit, gold standard and so
forth with none
5 versus some and get more
information standard some?
6
DR. LOEWKE: You would have
to be able to
7 identify that the added
information had clinical
8 value.
9
DR. D'AGOSTINO: Exactly, you
would have
10 to show that you do get clinical
benefit but, you
11 know, could you do the MRI without
any gad in it
12 and then do it at a particular
level and show that
13 that particular level does, in
fact, add
14 information? Because you automatically do it at
no
15 level,
right?
16
DR. LOEWKE: I mean, we have
approached
17 things before in that fashion. That was before we
18 have moved forward with clinical
utility. So,
it
19 would be very important that the
added information
20 really had value that you could
clearly identify.
21
DR. D'AGOSTINO: Right, but
it is not an
22 unacceptable
design?
23
DR. LOEWKE: It is something
that would
24 need further
discussion.
25
DR. D'AGOSTINO: Yes, thank
you.
90
1
DR. CHESNEY: Dr.
Nelson?
2
DR. NELSON: I have just two
comments to
3 follow-up on some of Norm's
questions. There
is
4 precedent both for local protocols
as well as for
5 NIH-funded studies for limited
procedural sedation
6 to be considered a minor increase
for
7 non-therapeutic procedures. There is also
8 precedent for trying to minimize
the risk of
9 sedation by combining the MRI
being performed when
10 there is an anesthetic being
provided for other
11 reasons, either operatively or
that real fancy
12 picture you showed us yesterday,
Phil, of the UCSF
13 slide in and out between MRI and
catheterization
14 which, sounds to me like the
perfect venue for this
15 kind of MRI/catheterization
because you are just
16 sliding the patient back and forth
a few feet, it
17 would seem.
18
DR. CHESNEY: We have been
asked to
19 include Dr. Moore in all of these
modalities.
So,
20 that is going to be an additional
question and I
21 would like, unless there is some
strong feeling, to
22 move on to the cardiac ultrasound,
hoping to take a
23 break at the end of that and then
we can address
24 the nuclear imaging and
angiography. But
Dr.
25 Siegel looks insistent.
91
1
DR. SIEGEL: One quick
comment as we go
2 through the rest of it, I just
thought about one
3 other way to do research and to
maybe complicate it
4 more, we forgot about simulation
models. With
all
5 the computer designs out there
now, we would be
6 able to look at certain facets at
least in CT and
7 maybe in MR using computer
models. That is just
a
8 thought. It just dawned at me that if I am
looking
9 at dose and I am measuring density
I could probably
10 do this with a computer phantom,
setting up the
11 appropriate computer example. So, it is just
12 another thing to put on the table
if anybody thinks
13 that is appropriate as we discuss
other modalities.
14
DR. CHESNEY: Thank you. I am glad you
15 were insistent. Dr. Sable, are you
ready?
16
DR. SABLE:
Sure.
17
DR. CHESNEY: Please tell us
what is your
18 most burning issue in the best of
all possible
19 worlds, and then if you could
address (a) through
20 (f), please.
21
DR. SABLE: Well, I think to
me the most
22 burning issue is to try to
incorporate contrast
23 echo into evaluating left
ventricular function and
24 wall motion in complex patients in
whom we can't
25 get good pictures with routine
echo. I think that
92
1 is probably the most important
thing and would be a
2 starting point as a basis to do
other things with
3 contrast
echo.
4
I think in terms of what agents need to be
5 studied in pediatrics, there is
the most experience
6 in adults with Optison and
Definity so I would
7 clearly focus on those two drugs,
both looking at
8 the necessary dosing and safety in
anything that we
9 do.
10
In terms of which populations should be
11 studied, I think there are a
couple of groups that
12 I would divide them into. If you think about
13 patients with poor windows, they
can be patients
14 like after cancer where you just
need functional
15 studies, or patients with complex
heart disease
16 that have unusually shaped
ventricles or single
17 ventricles, right ventricles
acting as systemic
18 ventricles such as in Mustard or
stenting repairs.
19 I think another group of patients
would be those
20 who would need stress echo
evaluation, including
21 patients with Kawasaki disease,
heart transplant
22 and patients who have undergone
operations that
23 involve the left coronary artery
or the arterial
24 switch procedure. I think later on we could
move
25 towards doing perfusion studies,
but I think the
93
1 first step would be to look at
left ventricular
2 opacification both at rest and
exercise.
3
In terms of using endpoints, I think that
4 the ideal is using a gold standard
and in this case
5 probably MRI or nuclear medicine
could be a gold
6 standard, but I think the more
practical approach
7 to using an endpoint, and it is
probably easier
8 with echo than other modalities
because we can vary
9 without increasing risk sedation
or time, we can
10 get pre- and post-injection images
to see if there
11 is an improvement using the
standard American
12 Society of Echo wall motion
score. We have
22
13 segments for every patient so the
power of the
14 study could be achieved relatively
easily with not
15 a huge number of
patients.
16
In terms of trial design, I would start
17 with a group of patients such as
those with poor
18 windows that I had mentioned, and
pick a drug like
19 Definity or Optison and use
incremental dosing
20 based on weight to get a sense of
do we get
21 improved images; how long does
that image last for;
22 and comparing it to the pre- and
postop and pre-
23 and post-injection
state.
24
A second, probably softer endpoint would
25 be does adding contrast obviate
the need to do more
94
1 invasive studies? We could do a
randomized,
2 controlled study, although the
ethics of that may
3 be challenged based on adult
literature and we may
4 need to consider using historical
controls.
5
That probably gets to trial
design. I
6 think I would start with using an
older population
7 and then gradually work my way
down to smaller
8 patients. I think I will stop there and
answer
9 questions.
10
DR. CHESNEY: Dr.
Loewke?
11
DR. LOEWKE: I was wondering,
you had
12 mentioned stress testing, are you
talking about
13 exercise only? Pharm stress only? If you do
14 exercise, how low can you go
age-wise and actually
15 get patients to cooperate? I will throw a
monkey
16 wrench in here as well. I believe--and there
is
17 somebody here I believe from
Cardiorenal--that the
18 pharm stress agents aren't
approved in kids.
19
DR. SABLE: Certainly, that
is true and
20 there is maybe one more paper on
stress echo in
21 children that is on contrast echo
in children.
22 That being said, a large number of
us do dobutamine
23 stress echo. It is a little bit cumbersome, as
one
24 can imagine, to have somebody run
and then throw
25 them onto the bed to image
them. You can do a
95
1 little better with some of the
odometers where you
2 are lying flat. Tal can probably speak to
this.
3 They probably do more than we do
in their lab.
But
4 it is much easier to do dobutamine
stress echoes
5 and that would probably be the way
we go. If you
6 look at Dr. Kimball's study, I
think they did 19
7 dobutamines and 2 ergometer stress
echoes in their
8 study. Clearly, we could be having the same
panel
9 meeting about dobutamine stress
echo and come up
10 with all the same issues that I
just mentioned for
11 contrast echoes. We are a little bit further
along
12 in that
realm.
13
In terms of age, we have done them down to
14 a year, a year and a half. Then you get into
the
15 whole issue of sedation and that
clearly needs to
16 be a part of any equation with
echo if you are
17 doing very small children. With older
children,
18 probably beyond four or five, it
is not as much of
19 an issue.
20
DR. CHESNEY: I have Dr.
Ebert, Dr.
21 Gorman, Dr. Moore and Dr.
Geva.
22
DR. EBERT: I have a question
for Dr.
23 Loewke. As we go through some of
these
24 specifically, does the agency feel
pretty
25 comfortable about measures of
safety, either short
96
1 term or long term, that you are
going to
2 incorporate into these, with MRI
or with any of the
3 other
modalities?
4
DR. LOEWKE: Obviously, we
would welcome
5 any information that you can
provide on how you
6 feel safety should be incorporated
into trial
7 design.
8
DR. SABLE: I think with any
type of
9 stress echo or contrast echo we
would need to
10 monitor vital signs and pulse
oximetry very
11 frequently, probably similar to
some of the
12 conscious sedation protocols, for
a set period of
13 time after the study is done,
probably at least an
14 hour. Obviously, we would be watching for
more
15 severe adverse events and
reporting those.
16
DR. CHESNEY: Drs. Gorman,
Moore, Geva and
17 Nelson.
18
DR. GORMAN: I was hoping not
to have to
19 ask this question because I was
hoping it would
20 come out, but what has been the
barrier that has
21 prevented these contrast agents
from being used in
22 echocardiography? Clearly, it is not fear of
using
23 things off-label because
pediatricians do that all
24 the time. Clearly, it is not that it hasn't
been
25 used in adults. So, what has been the barrier that
97
1 has prevented this modality from
leaping, as all
2 the other technologies have leapt,
to pediatrics?
3
DR. SABLE: There are two
answers to that.
4 The first one is that every
patient that comes to
5 one of my colleague's labs is
probably the only
6 patient they are dealing with for
at least a few
7 minutes, and they are all going to
have an IV and
8 they are all going to be prepared
to get contrast.
9 So, it is kind of the mind
set.
10
Whereas, in a busy echocardiography
11 laboratory--we do about 50 studies
a day in our
12 laboratory and we usually have
three or four rooms
13 going at once--we have very
limited nursing.
We
14 have maybe one nurse that is there
to cover a
15 sedated echo which we still use
oral sedation for.
16 So, putting an IV in, in the midst
of a very busy
17 echo lab is much different than
for some of the
18 other
modalities.
19
The economics of echocardiography is that
20 in many cases we are supporting
other programs.
We
21 have a huge volume, a huge
money-maker and it is
22 hard for me to convince my
administrators who are
23 looking at the practicality of
doing this that
24 instead of doing seven echoes
using one sonographer
25 I want to do one echo using one
sonographer, one
98
1 doctor and at least one or two
nurses. So, I
think
2 that is a huge barrier. It is inappropriate but
it
3 is the
reality.
4
The second barrier is what I alluded to
5 earlier, that is, the drug
companies which are
6 making these agents--I have talked
to them at
7 several meetings--don't seem to
have much interest
8 and they seem to be scared of
getting into
9 pediatrics. Clearly, the talk we heard
yesterday
10 from Bristol-Myers was much more
focused on the
11 nuclear agent even though the
discussion included
12 both.
13
DR. GORMAN: I can understand
the economic
14 argument inside a hospital, but
not much
15 echocardiography actually goes on
in hospitals.
16 So, why isn't some entrepreneurial
private practice
17 group that does echoes doing
this? I mean,
if
18 there is really a need out there
for this--if there
19 is really a diagnostic utility to
this that
20 clinicians will use, then
generally what happens is
21 people use it and either show it
works or doesn't
22 work and reimbursement follows
after that.
23
DR. SABLE: Yes, I think that
even though
24 a large proportion of pediatric
echo is done in
25 community hospitals or in smaller
clinics, contrast
99
1 echo is going to start in tertiary
care large echo
2 labs. So, I think it would primarily be
hospital
3 based early on. It has been something that I
have
4 been trying to push in my
institution and I think
5 it is somewhat resources and just
a different way
6 of thinking. A lot of the people who
refer
7 patients--a lot of my colleagues
who refer patients
8 for echoes are used to getting an
answer in five or
9 seven minutes so it is kind of
changing the mind
10 set. Hopefully, Dr. Kimball's paper will
circulate
11 through the pediatric cardiology
community and the
12 American Society of Echo that you
heard yesterday
13 will change the mind set. I think in general,
for
14 lack of a better term, it may be
ignorance.
Very
15 few of us even think about
it.
16
DR. GORMAN: One more, is
there something
17 uniquely different about these
agents? Are
these
18 really something that has traveled
the border of
19 drugs and devices? Are these bubbles
really
20 bubbles or are they particles that
don't break
21 down?
22
DR. SABLE: I think they
break down. I
23 think another issue is that the
way they interact
24 with the echo machine--before I
prepared for this
25 talk I read a couple of contrast
echo books and the
100
1 principles behind them, and the
way you use them is
2 complex and intimidating and it is
really a whole
3 new science to learn. So, I think it is more
that
4 than actually concern about safety
or particles
5 breaking down. I mean, the potential, as
I
6 presented yesterday, is so great
and if it could be
7 done in a routine echo setting--I
have talked to a
8 number of adult echocardiographers
who do this on a
9 daily basis and there is a huge
learning curve to
10 get started and a lot of us aren't
willing to take
11 this learning curve. Clearly, I am here as
a
12 representative of an academic echo
lab that feels
13 the learning curve is certainly
worth it.
14
DR. CHESNEY: Drs. Moore,
Geva, Nelson and
15 then I think we will take a
ten-minute break.
16
DR. MOORE: Well, I will just
follow-up on
17 that. Dr. Gorman played devil's advocate for
me so
18 I appreciate that. But our experience,
interacting
19 quite a bit with our own echo lab
which has had an
20 interest in contrast echo for
years and interacts
21 very closely with the adult echo
lab because we are
22 not a free-standing children's
hospital, has really
23 found a relatively limited utility
in the smaller
24 patients with regards to the
current echo contrast
25 agents. I wouldn't say that is to say there
aren't
101
1 applications and there may not be
huge future
2 applications for
it.
3
I would say in our own experience the
4 limitation has primarily been
added value in terms
5 of the younger patients. We even started
using
6 some of the Optison type contrast
in the cath lab
7 looking at different shunts in
very specific
8 indications and found over time
that it really
9 wasn't giving us a lot of added
value. Because
of
10 that, we limited its
use.