Latest / EHA Unplugged / Mapping Rare Anemia Patient Data: RADeep (with Dr. María del Mar Mañú Pereira)

Transcript

Mapping Rare Anemia Patient Data: RADeep (with Dr. María del Mar Mañú Pereira) · EHA Unplugged

• 0:03
  Greetings and welcome to EHA Unplug, the official podcast
• 0:06
  channel of the European Hematology Association EHA.
• 0:10
  Hello, I am Isabella Rivera, a medical writer for EHA.
• 0:14
  Welcome to today's episode on the reuse of clinical and
• 0:18
  multiomics data, specifically focusing on rare disease
• 0:22
  patients. Our guest today is Maria Del Mar
• 0:26
  Mano Pereira, an expert in the field of rare anemia disorders.
• 0:31
  She's here to share her insights on how every deep registry is
• 0:35
  changing the way we use patient data to promote research and
• 0:40
  improve healthcare outcomes. Maria is also actively involved
• 0:44
  in the enroll platform from EARN Aeroblue Blood Net, Which aims
• 0:52
  to standardize and systematically gather clinical
• 0:56
  data for reusing research. Thank you very much for being
• 0:59
  with us today. So you are involved in all these
• 1:04
  initiatives to gather, standardize and reuse clinical
• 1:08
  data. Why is this important?
• 1:12
  Well, we started with this with the registry in 2016 and when
• 1:19
  why we start with this initiative?
• 1:22
  Because we realized that it was impossible to have any robust
• 1:27
  evidence and and clinical outcomes that are reliable if we
• 1:33
  don't pull the data from different countries because as
• 1:36
  you said rare anemia are rare diseases and then we have a few
• 1:41
  patients in each country. So we really need to have a
• 1:44
  European approach and to be able to pull all this data in order
• 1:49
  to have the minimal numbers to prove hypothesis or to advance
• 1:55
  in research, so. What is exactly area deep?
• 2:01
  There are DEEP is the acronym for Rare Anemia Disorders
• 2:04
  European Epidemiological Platform.
• 2:07
  So is a registry for a patient registry and it is an umbrella
• 2:12
  for for both already existing registries, national registries,
• 2:18
  but also new registries. The the aim of of RADIP is to
• 2:23
  standardize the clinical data and in a way that is aligned
• 2:28
  with the new developments for anemia disorders and that we can
• 2:33
  reuse this data for advancing research, both translational
• 2:37
  research and clinical research. The problem is that it's very
• 2:42
  difficult if you think about pulling data from different
• 2:46
  countries, first barrier is language.
• 2:49
  So in each country we have a different language, but not only
• 2:53
  this also the understanding of the OR the definitions for many
• 2:59
  of the parameters are are also different.
• 3:02
  So we need to understand each other first as physicians,
• 3:09
  researchers that we are working with this type of disorders and
• 3:13
  we have we need to have a common understanding or what we mean by
• 3:18
  hemoglobin concentration, something that is simple, but
• 3:22
  you have different units. You can collect a hemoglobin
• 3:26
  concentration as ranges or you can have a number.
• 3:32
  So you need to standardize the way that you are collecting the
• 3:35
  different information. And then also the second step is
• 3:40
  to have some interoperability for machines.
• 3:43
  So it's not only between us, but also we need that all this
• 3:48
  information is can be understand by them computers.
• 3:54
  So this is the the part of the interoperability and we do this
• 3:58
  using as ontologies and codifications, international
• 4:02
  codifications. So we are we are sure that they
• 4:07
  when we pull the data we are putting together data with the
• 4:12
  same meaning and this and then we can have comparable results.
• 4:17
  So how does a very deep compared to enroll Euro blood net?
• 4:21
  There are many differences. So for the the ENROLL is the
• 4:26
  central registry of the year end Euro blood net.
• 4:30
  So this is on the strategy of the European Commission and the
• 4:33
  Member States and it's part of the European reference networks.
• 4:37
  There are 24 European reference networks.
• 4:40
  They are in Euro blood net, is on hematology, both oncological
• 4:43
  and non oncological and ENROLL is the central registry.
• 4:48
  The idea of ENROLL is not really exhaustive registry, but it's
• 4:54
  more a platform to promote the interoperability standards and
• 5:00
  also the FAIR principles. So the third principles is the
• 5:06
  the, the data needs to be findable, interoperable,
• 5:11
  accessible and reusable. And these are this is what we
• 5:14
  call the third principles. And in the strategy of the
• 5:18
  European Commission, the idea is to cope all the fragmentation of
• 5:24
  data for rare diseases and then work on these principles in
• 5:29
  order to be able to pull the data, to understand the data to
• 5:35
  to share the data and reuse it. But Enroll is more on this
• 5:40
  implementation and promoting all these standards.
• 5:45
  And also it collects A minimal data, but it's really a minimal
• 5:50
  data for understanding more epidemiology, but no more than
• 5:54
  this. Now to the use of these
• 5:58
  platforms. How has this systematic
• 6:00
  gathering of information and standardization of data impacted
• 6:04
  now research in rare anemia disorders?
• 6:09
  Well, I think that we are in the, in the beginning, let's
• 6:13
  say, because for rare anemia disorders, there was really a
• 6:16
  big need of, of research and data-driven research.
• 6:22
  So for instance, when we started with the RADIP in 2016, there
• 6:28
  were only three registries in Europe for sickle cell disease,
• 6:33
  the one in Germany, Belgium and Spain.
• 6:38
  There were no more registries. There were some databases, but
• 6:42
  not really registries with our national coverage.
• 6:47
  What this means that they at the end you cannot prove the
• 6:52
  hypothesis. You are working always with the
• 6:55
  of course with your expertise in the clinical settings, but you
• 6:59
  cannot really prove the hypothesis.
• 7:01
  We need more data to really prove all these hypothesis or
• 7:10
  outcomes or how the different drugs work in the in the
• 7:14
  patients. So anemias are complex diseases
• 7:20
  are chronic. So you have at the end a lot of
• 7:24
  complications, the disease progression, the you need to
• 7:28
  have reliable information for the for identifying reliable
• 7:36
  indicators for the outcome for the endpoint.
• 7:40
  And if you don't have all this, it's very difficult to do any
• 7:44
  type of research and and drug development.
• 7:48
  So now in the in the RADIP, we have for the first time is that
• 7:55
  the first initiative in Europe for collecting all this data.
• 8:00
  And we have a standardized clinical data for almost 5000
• 8:04
  patients in Europe from 13 countries.
• 8:08
  This means that we are supporting the countries for
• 8:12
  collecting all this data. We established the the different
• 8:18
  expert groups, we were discussing together among all of
• 8:22
  us what was important for these diseases, what, what were the
• 8:27
  priorities, the challenges. And I think it's really very
• 8:34
  important that we work in this way in a European approach.
• 8:39
  So for on one hand we established A robust basis for
• 8:46
  all these a standardized clinical data collection and on
• 8:52
  the other hand we have started reducing all this data with
• 8:56
  research projects. So we are contributing to to EU
• 9:01
  funded projects that are Genomide for all and Syntema.
• 9:06
  And in the two projects, what we are aiming is to develop
• 9:12
  artificial intelligence models for risk scores in sickle cell
• 9:17
  disease and for basocclusive crisis, for acute chest
• 9:21
  syndrome, for chronic organomet. And in syntema, we aim to
• 9:28
  develop synthetic data generation.
• 9:33
  So this is important because there are many unmet needs in
• 9:39
  research for rare anemia. They have been neglected for
• 9:42
  many years from health planning. So I think it's the first time
• 9:48
  that they are really overrun. Exactly.
• 9:51
  And I, I'm very optimistic and I think that there are more and
• 9:56
  more clinicians and researchers than that are ENCODE right now
• 10:01
  with this initiative to, to pull the data standardized.
• 10:05
  They do more research because for many years we say that
• 10:09
  personalized medicine was very difficult in sickle cell disease
• 10:13
  and it is. It's difficult because it's a
• 10:15
  very complex disease. But also we realize that the
• 10:19
  lack of a standardized clinical data hamper really a lot this
• 10:24
  research. How is AI helping?
• 10:29
  So you said for developing risk course or yes?
• 10:33
  And so you're also you mentioned synthetic data.
• 10:39
  Yes, there are different aims for synthetic data.
• 10:42
  One of them is to enlarge the data sets.
• 10:46
  You can also use synthetic data for imputation.
• 10:49
  So if you have some missing data in your data set for some
• 10:53
  patients, you can also predict some of the parameters.
• 10:57
  So then you have a complete data set for the patients.
• 11:01
  This is this is imputation and then to enlarge the data sets is
• 11:06
  known as augmentation. So there are different
• 11:08
  strategies. There is another point that is
• 11:11
  as you know in in the European Union we have the GDPR, that is
• 11:17
  the the General Data Protection Regulation and and also the
• 11:23
  generation of synthetic data will help for all, all these
• 11:31
  barriers that we face for sharing the data across
• 11:34
  countries. So the way that we are working
• 11:38
  is also through the development of Federated platforms that then
• 11:44
  the the data for of the patient doesn't leave the hospital.
• 11:49
  So all the nodes are connected to a central node, but the
• 11:53
  patient's data is not leaving the country, it's not even
• 11:57
  leaving the hospital. So this is what we are working.
• 12:01
  The synthetic data also can be used not now, but if we do
• 12:07
  properly, everything could be used as external control arm for
• 12:13
  clinical trials. This is also a problem, let's
• 12:18
  say that for many patients with a rare anemia disorder, they
• 12:24
  have really very few options for treatment.
• 12:27
  And then when there is some drug in clinical trials is like a bit
• 12:33
  also from an ethical point of view, having patients in
• 12:37
  controlled arms is reducing their possibilities to have a
• 12:43
  treatment. And and then this could be used
• 12:47
  the synthetic data as this external control arm.
• 12:51
  So you you don't need to put patients in a controlled arm and
• 12:55
  not receiving the treatment. Are you already doing this?
• 12:59
  We are doing this in one project is I think it's the first time
• 13:04
  for sickle cell disease. We are doing the models for
• 13:08
  generate for the generation of the synthetic data, but is not
• 13:12
  used for as yet known. Because it will be, it will have
• 13:15
  also regulatory of. Course, of course, of course.
• 13:20
  So one step back. So the first thing that we are
• 13:24
  doing with the RADIP that we have the real data from the
• 13:28
  patients we are now in, in while in contact with AMA, with the
• 13:33
  European Medicine Agency. And we already had one meeting
• 13:37
  in Amsterdam with the different coordinators of the different
• 13:41
  registries. And I, I was there as
• 13:44
  coordinator of the RADIP registry.
• 13:46
  And, and what the first thing that we want to do is to have
• 13:50
  our real data sets from the registries qualify for
• 13:55
  regulatory purposes. So first step in my opinion is
• 14:00
  to have the real data set qualify.
• 14:03
  Then if for this I have to say that all the recommendations
• 14:08
  from the EMA are really very good for us.
• 14:13
  So we are in the Radip, we are developing all our data quality
• 14:18
  framework based on the recommendations of EMA.
• 14:23
  And first thing is to have good quality of data, real data.
• 14:28
  And why? Because also the synthetic data
• 14:32
  is developed through a model that has been trained in the
• 14:36
  real data. If the real data is bad quality,
• 14:40
  the model will be bad quality and the synthetic data as well.
• 14:45
  So first point is really to be sure that we have good quality
• 14:50
  data from the patients. We are doing this in the Radip
• 14:54
  and then in the in Syntema. We are using a subset of this
• 14:59
  data set that we already prove all the quality and did all the
• 15:03
  cleaning of the data. And then we are now about to
• 15:07
  start with the models for generating the synthetic data.
• 15:12
  Then once that you generate the model, you need also to do a
• 15:15
  clinical validation of the data. And for this, we work with the
• 15:20
  experts of course. And from a statistical point of
• 15:24
  view view, we are developing the clinical utility frame also to
• 15:30
  prove that this data is good, announced the synthetic data.
• 15:35
  And you will use this to get also regulatory approval.
• 15:38
  To use this, yes this is the idea, but step by a step.
• 15:42
  So first you need to start with. But this is the idea, yes.
• 15:46
  So you have implicated how many countries?
• 15:50
  1313 countries, 12 member States and Norway.
• 15:54
  How has been this experience of collaborating with all these
• 15:58
  different systems, these different cultures, obviously
• 16:02
  languages, but these different systems to get all this data
• 16:06
  gathered? So it was really very, very
• 16:09
  difficult to understand or to support our national
• 16:15
  coordinators for all the legal and ethical part of of setting
• 16:21
  up our registry. So you have to, at the end you
• 16:25
  have to learn about many things about the legal, the GDPR, the
• 16:30
  agreements, ethics, the informed consent.
• 16:34
  So the regulation is different from 1 country to another
• 16:39
  country. So what is validity in Spain is
• 16:42
  not valid in France or in or in Germany.
• 16:46
  So it's it's not easy. Yes.
• 16:48
  How long has it taken you to set up then this registry?
• 16:51
  We started with this idea in 2016 and we started with ultra
• 16:56
  heart disorder, with pyruvate kinase deficiency.
• 17:04
  And then we started with PKD and for 2-3 years, we were working
• 17:14
  on setting all the, all the legal, the ethics also to
• 17:20
  develop the electronic case report form.
• 17:24
  So we were mapping all the existing registries, asking for
• 17:29
  access for the metadata or for the registries, because we
• 17:33
  wanted to build something that was based on the existing
• 17:37
  registries. No, not to start from a scratch,
• 17:41
  but we then realized that that actually there were very few
• 17:46
  registries in place. And then in 2020, we had the two
• 17:52
  years of the pandemic. So this also when we were about
• 17:56
  to really the pandemic also has slowed down, no, the
• 18:02
  development. But I have to say that after the
• 18:05
  pandemic, now we are in an exponential phase and we were
• 18:11
  able to fix all these legal and ethics issues and also to
• 18:18
  develop a modern platform and, and implement all these quality
• 18:23
  standards. And now the, the, the progress
• 18:28
  is really very good. We have almost 5000 patients
• 18:32
  already registered. So this is really very good.
• 18:36
  And also I want to thank all our contributors because it's not,
• 18:41
  it's really not easy. This, this is done out of the
• 18:46
  the clinics work no so, but I think that now we have a very a
• 18:52
  very good group and the support of the of the EMA of the
• 19:00
  European Commission. We also collaborate with pharma
• 19:04
  industry patients. Patients are involved in in the
• 19:09
  RADIP from the onset. So we have two representatives
• 19:14
  for patients and one for a sickle cell disease and
• 19:19
  thalassemia and another one for the ultra rare disorders.
• 19:23
  And we would like to have more but it's also difficult to to
• 19:27
  have patients with enough time to to be spent in the registry.
• 19:32
  But, but we have them and they are part of our data access
• 19:36
  committee, in which we also have two legal experts and one expert
• 19:41
  on ethical issues. I think this is very important
• 19:45
  to have all these different expertise.
• 19:48
  Perspectives, yes. And perspectives, exactly.
• 19:51
  So now that the data has been standardized, can research teams
• 20:00
  or people who want to make use of this data access it and use
• 20:04
  it for their own projects or is still not ready?
• 20:09
  Now we are about to to finalize the first RADIP manuscript and
• 20:19
  then we will publish everything that we did so far and also we
• 20:23
  will include some epidemiological data.
• 20:27
  And then for next year also we plan to have the the second
• 20:32
  publication on all the data that we have collected.
• 20:36
  Then once it's published, we can open all these data set for a
• 20:42
  request from researchers and perhaps even a bit before.
• 20:48
  So the answer is yes. So the idea and the and the
• 20:54
  division of Doradip is to open up the the data sets.
• 20:59
  So we are already reducing this data as I told you before with a
• 21:03
  group of experts in two European projects and but then we will
• 21:08
  open also the the, the data sets to to the research and community
• 21:15
  also patients or industry or any legal, any legal entity in
• 21:22
  Europe could request data not only in Europe, but it's a bit
• 21:26
  complex. But we have already a policy for
• 21:31
  data request and there is a submission form.
• 21:36
  So you have to complete the like all why you are asking for the
• 21:41
  data? What is your institution?
• 21:43
  If you have an ethics approvals and then the data access
• 21:47
  committee will review the the form and then give the green
• 21:52
  light or some modifications. But of course first we need to
• 21:56
  publish the results and then we open the the data sets.
• 22:00
  And are you going to be constantly gathering more
• 22:04
  information? Are you?
• 22:07
  I imagine this is also a lot of organization to keep this
• 22:11
  running. Yes, the idea is the Radip is a
• 22:14
  prospective longitudinal registry.
• 22:18
  So the idea is to have a long term registry that can be
• 22:23
  upgrade and it's a modular structure.
• 22:26
  So we can have different modules for different purposes and and
• 22:32
  we collect the baseline information from the patients
• 22:36
  first time when they are registered and then all the
• 22:39
  longitudinal data is collected once per year.
• 22:44
  We have also standardized procedure to review every year
• 22:50
  the type of data that we are collecting, if we need to do any
• 22:56
  change or modification or for adapting the data that we are
• 23:00
  collecting to the new developments on drugs or the new
• 23:05
  knowledge on the disease. And so, and we are funded by
• 23:13
  public and private resources. So first thing we have our
• 23:18
  institutions behind that all the institutions are supporting the
• 23:23
  the initiative. And this also means a number of
• 23:28
  human and resources that the institutions put in place for,
• 23:33
  for the maintenance of of the registry.
• 23:37
  And also we have this EU calls and funded projects and we have
• 23:42
  also agreements with pharma industry.
• 23:46
  But the right, the RADIP is an academic registry.
• 23:49
  So all the ownership belongs to the to the RADIP members.
• 23:56
  But we have agreements, research agreements with pharma as well.
• 24:01
  What role is EHA playing or will play in this?
• 24:07
  We, we have, we had a lot of discussions with EHA and I'm
• 24:13
  very happy to say that we were officially endorsed by EHA just
• 24:19
  one month ago. And yes, we collaborate with EHA
• 24:25
  also at the level of the ERN of the ERN Euroblood Net that I am
• 24:30
  in the coordination of the ERN Euroblood Net as scientific
• 24:33
  Director and recording. We collaborate with EHA for, for
• 24:40
  education for patients engagement for guidelines.
• 24:45
  So and this will be another field of collaboration.
• 24:50
  I think that the role of VHA is really very important to engage
• 24:56
  the community to also to discuss the unmet needs of rare anemia
• 25:03
  disorders to increase visibility.
• 25:07
  So we need to start the discussion on on how we are
• 25:11
  going to collaborate in concrete actions.
• 25:14
  But for sure, I'm I'm confident that will be really.
• 25:18
  Very good. Now there is gene therapy for
• 25:21
  sickle cell disease and thalassemia.
• 25:23
  So these patients are cured somehow or at least have a very
• 25:29
  long time of remission. But you have to keep track of
• 25:33
  what's happening to them. So I imagine they are going to
• 25:35
  be in this registry. We will have it in our registry
• 25:41
  and I think this is also an important point because the
• 25:44
  indication for gene therapy are those patients that are suitable
• 25:50
  for a bone marrow transplantation, but they don't
• 25:54
  have a donor, a compatible donor.
• 25:58
  So I think that the the RADIP registry will be very important
• 26:04
  to understand where are the patients and how are the
• 26:09
  different. What we include in the RADIP
• 26:11
  registry are also some parameters that are for
• 26:15
  understand the clinical practice at the national level.
• 26:19
  Because this is also very important.
• 26:21
  It's not only the variability of the patient or but also the
• 26:25
  National Health system and how the different care pathways are
• 26:31
  implemented at the national level.
• 26:33
  So we will be able to see differences on when the patients
• 26:40
  are offered with a bone marrow transplant for instance.
• 26:45
  It's not even if it is a standard of care, it is not
• 26:49
  indicated exactly the same way in all the countries or at the
• 26:55
  same age. And there are many patients,
• 26:59
  adults that they are not screened for a compatible donor
• 27:02
  perhaps. So this is the one part that the
• 27:07
  RADIP registry can contribute a lot for a better understanding
• 27:11
  on how to find and how to improve access to gene therapy.
• 27:18
  Then after the gene therapy, there is the European Bone
• 27:24
  Marrow Transplant Registry, the EBM team that they are focusing
• 27:29
  on this innovative therapies. So of course we don't want to
• 27:35
  duplicate what is already in place.
• 27:37
  And then the idea is to have a discussion with the EBMT and
• 27:42
  then to see if we have a complementary module for the
• 27:46
  follow up of the patients or if they are transferred completely
• 27:51
  to the EBMT. This is the same registry where
• 27:55
  you register car T therapy. Yes.
• 28:00
  So it would be all in the same place.
• 28:01
  The at the moment ABMT is gathering one registry for all
• 28:08
  patients that have some kind of genetic engineered therapy.
• 28:14
  And also the, the conventional bone marrow transplant.
• 28:16
  So that's clear that we don't want to duplicate, yes, so, but
• 28:22
  have some synergies because we have the information of the
• 28:26
  patient before yes, the, the gene therapy.
• 28:30
  So the idea is to to to have a collaboration with ABMT.
• 28:36
  And mostly you will from the start help to identify which
• 28:40
  patients should could be candidates for yes.
• 28:44
  So this is something that we are working a lot, that is we are
• 28:47
  reviewing the Radip, CRF, all the data that we are collecting
• 28:52
  in order to be aligned with the inclusion and exclusion criteria
• 28:57
  of clinical trials. So in this way we are able to
• 29:02
  have a European referral system so we can reach the patients.
• 29:08
  This is also very important for ultra rare diseases, Yes.
• 29:12
  So thank you very much, Doctor Manu Pereira, for sharing your
• 29:17
  expertise with us and thank you to all of you for listening.
• 29:21
  If you enjoyed this episode, don't forget to subscribe and
• 29:24
  share it with your friends and see you at the next.
• 29:27
  Episode, thank you very much.