Latest / European Society for Vascular Surgery / TOGETHER Registry - with K.K. Yeung and T. Resch

Transcript

TOGETHER Registry - with K.K. Yeung and T. Resch · European Society for Vascular Surgery

• 0:07
  Welcome to the ESBS podcast. My name is Agla
• 0:10
  Kovidunete. Today, we'll be talking about real
• 0:13
  -world aortic data, in particular, the Gore -Together
• 0:17
  Registry, and what large, long -term registries
• 0:20
  can actually tell vascular surgeons about durability,
• 0:23
  anatomy, and re -intervention. Our first guest
• 0:26
  is Professor Kaki Young, vascular surgeon and
• 0:29
  professor of personalized therapy in vascular
• 0:32
  surgery at Amsterdam UMC in the Netherlands.
• 0:36
  where she leads major programs on aortic diseases,
• 0:39
  advanced imaging, and data -driven innovation.
• 0:42
  Professor Young has also been recently elected
• 0:44
  to become the ESVS president for 2026 -27. Congratulations.
• 0:50
  Our second guest is Professor Timothee Resch,
• 0:52
  Professor of Vascular Surgery at Copenhagen University
• 0:55
  Hospital. Internationally recognized for his
• 0:58
  work in complex endovascular aortic repairs,
• 1:01
  including fenestrated and branched techniques.
• 1:04
  Welcome to our podcast. Thank you. Great. Let's
• 1:09
  get started. We've been doing EVAR for years.
• 1:12
  What do we still not know well enough about long
• 1:14
  -term outcomes that this registry is trying to
• 1:17
  answer? The nice thing about the together registry
• 1:20
  is that it's not only doing capturing the real
• 1:26
  life data of one. excluder but it's the newer
• 1:32
  version of the excluder with the conformability
• 1:34
  and secondly it's an umbrella of other other
• 1:38
  ones as well so tim you have of course i think
• 1:42
  the tbe right the branch devices are also included
• 1:48
  in this registry so and then if there is another
• 1:52
  device coming out from the Goal platform, then
• 1:56
  it also goes into this really large umbrella
• 2:00
  that actually captures the real -world data.
• 2:03
  And what is also new is including an imaging
• 2:06
  platform. And that is really different than the
• 2:10
  Great Registry, which is also a long -term follow
• 2:14
  -up registry. But currently, this is also including
• 2:18
  imaging. And I think that is something we do
• 2:21
  need. Yeah, I agree. I think it's a good add
• 2:25
  -on. I think, you know, kudos to Gore, who have
• 2:29
  always been good on doing vascular registries.
• 2:32
  They're a great registry, and now they've expanded
• 2:34
  that. And we're actually – another feature, I
• 2:38
  think, of it, at least in a lot of the modules,
• 2:39
  is that we're actually looking at up to 10 -year
• 2:41
  follow -up, which is unique for the registry.
• 2:43
  I mean, we've been – in the old registry, very
• 2:46
  happy with the five -year follow -up scheme,
• 2:48
  but – Things have changed, and now we're offering
• 2:52
  T -VAR, E -VAR, ILLIAC branch devices more across
• 2:56
  the board, so we need longer follow -up. So I
• 2:59
  think that's very good, and of course, the imaging
• 3:01
  data. And I think also, and Kaki can answer this
• 3:03
  much better than I, because she's the AI expert,
• 3:06
  but I think also now we're living in a day and
• 3:08
  age where we have all this data accessible to
• 3:10
  us in a completely different fashion than we
• 3:13
  had 10, 15 years ago. So I think our hope is
• 3:16
  that we'll have longer follow -up. more complete
• 3:19
  follow -up, and also more extensive follow -up
• 3:22
  with regards to granularity, which hopefully
• 3:26
  will give us some answers. Talking about the
• 3:29
  general knowledge in current data, what are the
• 3:33
  blind spots we currently have? Is it durability,
• 3:36
  re -intervention, SAC behavior, iliac branches?
• 3:40
  What makes this registry necessary? There are
• 3:44
  several components, actually, that you already
• 3:47
  mentioned. But I think why this is necessary
• 3:52
  is that we are looking at new devices, but also,
• 3:57
  you know, actually add -ons on already existing
• 4:00
  devices, like the TBS, a branch extra, the IBE
• 4:04
  device which has never been captured long -term
• 4:07
  data. And this is an executable form world, so
• 4:10
  it has like with an angulation in the neck, which
• 4:14
  is possible. So there are a little bit of add
• 4:16
  -ons that actually includes more challenging
• 4:19
  aortic aneurysms. So not the standard ones only,
• 4:23
  but also challenging ones. I think this is a
• 4:26
  real add -on. And indeed, what Tim already said,
• 4:31
  we are living now in an AI area, and I guess...
• 4:36
  We can very nicely look at all these images with
• 4:41
  artificial intelligence techniques and also make
• 4:43
  predictions. So we can study it even better.
• 4:46
  What is necessary is a standardization of, of
• 4:50
  course, all the data collection. And that is
• 4:54
  also something that this registry does really
• 4:58
  well, is that everyone is doing exactly the same.
• 5:01
  We're going to collect exactly the same. And
• 5:04
  we're also collecting quality of life, which
• 5:08
  is also new in this registry. Now, I was just
• 5:12
  saying, just like Kaki, just to reinforce it,
• 5:14
  I think what's unique about this registry is
• 5:16
  a couple of things. And I think the objectivity
• 5:19
  of the data is very, very important. I think
• 5:21
  the imaging data that Kaki is referring to in
• 5:24
  the old registries was all center -based. So
• 5:26
  together, great registry. Kaki and myself and
• 5:30
  the equivalents pushing in the data. It was non
• 5:33
  -validated. It was nothing. And the really nice
• 5:35
  thing about the imaging data here is that it's
• 5:37
  objective because we have the imaging itself
• 5:39
  can run it through AI when that becomes readily
• 5:42
  available and then analyze everything. So that's
• 5:44
  really nice. And again, I think a lot of the
• 5:49
  data that we have on endovascular repair is still
• 5:52
  from pivotal trials of the highly, highly selected
• 5:55
  patients, whereas the registries, of course,
• 5:57
  provide something else they provide real world
• 6:00
  data we can use this it doesn't have to necessarily
• 6:02
  be on ifu or off ifu we can use it as we clinically
• 6:06
  find correct and then get answers to that great
• 6:12
  walk us through the design is this randomized
• 6:15
  perspective retrospective are there any other
• 6:18
  design details we need to know about No, it's
• 6:23
  simple. I think it's a prospective database.
• 6:26
  So you get informed consent from the patients
• 6:28
  and you enroll them as they come along. For some
• 6:31
  of the half acute cases, you can get consent
• 6:34
  up to 24, 36 hours after you've included the
• 6:38
  patients or treated the patients, but it's all
• 6:41
  prospective data. So it's not retrospective.
• 6:45
  It's also not randomized. It's just a longitudinal
• 6:48
  registry covering all these patients. So that
• 6:51
  has strengths and weaknesses. Again, it's real
• 6:54
  -world data. Hopefully, we can capture most of
• 6:57
  the patients. How broad is the inclusion? Are
• 7:01
  you only enrolling the perfect neck anatomy patients,
• 7:04
  or are you deliberately trying to capture real
• 7:08
  -world hostile anatomy? It's really real -world
• 7:12
  hostile anatomy or not hostile. So everyone that
• 7:16
  you see in your clinic, it's a real -life data
• 7:20
  collection. Yeah, so that is, I think, nice.
• 7:26
  For instance, it's a large registry. I think
• 7:29
  it's around 3 ,500 patients that will be collected
• 7:35
  worldwide. So it's a really large registry. Of
• 7:39
  course, Real data is not standardized data. So
• 7:45
  that means that you have all kinds of anatomy
• 7:47
  that you will collect. However, it's also nice
• 7:50
  to see how this really works then, how the prosthesis
• 7:54
  is behaving in all these years. And if you have
• 7:58
  only standardized inside -out view patients,
• 8:01
  then it's not real data, of course. And you would,
• 8:05
  of course, miss a lot of information. Well, this
• 8:11
  leads perfectly to our next question, talking
• 8:13
  about anatomy. Is there anything specifically
• 8:16
  that you're looking at challenging necks, angulated
• 8:19
  necks, short necks, the things that historically
• 8:23
  made people nervous about durability? It's a
• 8:28
  good question. I think because the gore exclure
• 8:33
  conformable is intended to place in highly angulated
• 8:38
  necks. and also in relative short extra, like
• 8:42
  10 millimeters. So beforehand, the normal extra
• 8:44
  was 50 millimeters. I think that is something
• 8:48
  that has broadened the eye of you. I guess a
• 8:53
  lot of physicians will place it even in a more
• 8:56
  challenging anatomy. We will see that. With the
• 9:00
  imaging, we can capture how the device is behaving.
• 9:03
  And of course, right now, the discussion is...
• 9:05
  also on sec regression type two and the leaks
• 9:08
  and and these things and our analytics in itself
• 9:11
  type one is highly angulated next i guess that
• 9:15
  is uh nice of this registry we will know um after
• 9:20
  a few years how this will behave in mid and long
• 9:23
  term yeah what's really really interesting also
• 9:27
  is that there's a tbe module of course to look
• 9:30
  at the side branch for the for the lsa which
• 9:34
  is where we mostly use it. I think you can get
• 9:36
  some really interesting data there because we
• 9:37
  really have very little knowledge about what
• 9:41
  it means to revascularize. We have historical
• 9:43
  data that suggests that revascularization of
• 9:49
  the LSA is beneficial in a lot of situations,
• 9:54
  but there's very little proof. So I actually
• 9:55
  think that, and that's part of the registry data
• 9:58
  that will come out reasonably fast because it's
• 10:02
  It's only a two -year inclusion process for that.
• 10:05
  We've almost already included all these TBE patients.
• 10:07
  So I think quite quickly we'll get some perioperative
• 10:11
  and midterm outcomes for that device in the real
• 10:14
  -world setting. People, you know, you go around
• 10:17
  the world and you hear that, oh, this is a great
• 10:19
  procedure, no complications, and it prevents
• 10:22
  a lot of complications with standard TVAR. But
• 10:24
  I think now is actually the first time we'll
• 10:26
  get fairly large. data set of patients that have
• 10:30
  actually done this procedure for aneurysms, PALs,
• 10:34
  dissections and everything, and can say something
• 10:36
  about stroke risks, spinal cord ischemia risk,
• 10:40
  and so on and so forth. And at least, even though
• 10:42
  it's not a comparative study, compare it to historical
• 10:44
  data with covering the LSA or doing bypasses
• 10:49
  on the neck and such. So that short term is really
• 10:52
  interesting, I think. some of the main endpoints
• 11:00
  you care about as vascular surgeons? Is it sac
• 11:03
  regression, type 1 -3 endoleaks, need for secondary
• 11:07
  re -intervention? Are there any specific endpoints
• 11:11
  that you are specifically interested in? I think
• 11:14
  I'm really interested in the quality of life
• 11:18
  data because I think for me, there's been a shift.
• 11:22
  I think that's really interesting. I think the
• 11:25
  fact that we're concerned scientifically about
• 11:30
  endo leaks or sac regression or like i don't
• 11:34
  want to say minor details but these small finer
• 11:37
  points of endovascular repair is indicative of
• 11:41
  where that we're now at a point where this really
• 11:44
  works really really well in most patients and
• 11:48
  our our expectations for the treatment are very,
• 11:52
  very high. We expect everything to work and it's
• 11:55
  supposed to be good. And so we're focusing, we're
• 11:57
  refocusing on what does this mean to the patient?
• 12:01
  What does all this follow -up mean? What is the
• 12:03
  quality of life for the patients? Re -interventions
• 12:06
  that we often say are insignificant and we do
• 12:10
  them in local anesthesia, but it's also a constant
• 12:12
  worry to the patient when we say you have to
• 12:15
  come back, we have to do things. So I think that
• 12:18
  part of the data set is very, very unique and
• 12:22
  something new. So I think for me and as a clinician,
• 12:25
  that's becoming more and more important. Not
• 12:28
  if they have a small type 2 endoleak somewhere,
• 12:30
  but how do the patients actually perceive this
• 12:32
  treatment after five, six years? I think that
• 12:38
  is very important. And we clinicians have made
• 12:42
  a lot of... endpoints up for ourselves like set
• 12:45
  regression which is not a thing like five years
• 12:48
  ago but right now it is and all kinds of tools
• 12:53
  or actually to assess our success rate after
• 12:58
  surgery. We as physicians are making up all kinds
• 13:03
  of endpoints that we thought it is important
• 13:06
  and one of those are set regressions. Five years
• 13:10
  ago this term didn't exist and now we just made
• 13:13
  it up and all kinds of things that we make up
• 13:18
  to assess our success rates after the surgery.
• 13:22
  But what does it really mean for the patient
• 13:24
  and how are these complications that we make
• 13:29
  up for ourselves as a physician related to the
• 13:32
  quality of life of these patients? So I don't
• 13:35
  mean to say we don't need to assess all these
• 13:38
  complications or the sex regression. But we do
• 13:41
  also need to, you know, give it into relationship
• 13:45
  and what the patient thinks of it. On the other
• 13:49
  hand, I think that type 2 aneurysm and sacrogression
• 13:52
  is still a mystery to be solved. I think it's
• 13:57
  ongoing growth of the aneurysm without we know
• 14:01
  why it's an aneurysm that is growing. And sacrogression
• 14:05
  is actually funny because it's... The second
• 14:09
  is regressing, which is the opposite of growing.
• 14:13
  Biologically, it's super interesting because
• 14:16
  that means the order is remodeling. I still think
• 14:20
  we can learn a lot from this. That means that
• 14:23
  a disease is reversible. I guess that is just
• 14:27
  funny to think about. And I'm also looking forward
• 14:31
  to this data to analyze it. I think also, again,
• 14:35
  like I was mentioning, how endovascular repair
• 14:39
  has matured over the past 25 years. And I also
• 14:44
  think that has allowed us now to actually get
• 14:49
  closer to the type of follow -up we're interested
• 14:52
  after aortic repair full stop, even open repair.
• 14:56
  I mean, it used to be, Kaki knows this, and we
• 14:58
  used to like, we did open repair and then we
• 15:01
  would give the patient a tap on the shoulder
• 15:03
  and send them home and say, everything is great.
• 15:06
  Don't ever come back. You don't need to come
• 15:09
  back because we did a great repair. And then
• 15:11
  actually, believe it or not, worked really well.
• 15:14
  And then EVAR came and there was all this intensified
• 15:17
  follow -up, which happened to coincide with the
• 15:20
  development of CT imaging. So we could see all
• 15:22
  these things. And then we went down this rabbit
• 15:25
  hole of looking at small decals. And now I think
• 15:28
  we're almost returning back to say, what's important
• 15:31
  to us and to the patients? Because you have to
• 15:34
  realize that for a lot of patients that we treat
• 15:36
  with endovascular repair, you know, 80, 75, 80
• 15:41
  % live less than 10 years after the repair. And
• 15:45
  we're looking at minute things that doesn't really
• 15:47
  change their course of life. slightly turning
• 15:51
  back to say, maybe we can relax, follow up. Maybe
• 15:53
  we can look at the quality of life. Maybe we
• 15:56
  can say, well, it's fine if you have an endoleak
• 15:57
  after 10 years, you're now 92. And maybe it doesn't
• 16:01
  matter. So yeah, it's interesting to see this
• 16:05
  data for sure. Great. This perfectly leads us
• 16:08
  to our next question. How do you expect this
• 16:11
  data to influence how we counsel our patients
• 16:14
  about surveillance? probably tighten follow -up
• 16:18
  or potentially let us safely relax follow -up
• 16:21
  in certain anatomies if durability looks solid
• 16:25
  yeah i think we hope to answer this question
• 16:31
  by looking at this data and also by capturing
• 16:35
  more registries more standardization using maybe
• 16:39
  artificial intelligent techniques to have a prediction
• 16:42
  on which groups will have uh you know you know,
• 16:46
  we can loosen up the follow -up. However, I do
• 16:52
  think that it's not only about that. I mean,
• 16:56
  it's not only about follow -up, but it's also
• 16:57
  about what do the patients want. That's one.
• 17:01
  And secondly, later with all kinds of the mobile
• 17:04
  devices, it could be also that we can have follow
• 17:07
  -up on a distance. And we can have different
• 17:12
  kinds of follow -ups as well. So it might also
• 17:15
  be that we don't want to know how it looks like
• 17:19
  on an image, but we just want to know how the
• 17:21
  patient is living. Is he living a good life?
• 17:24
  Yes or no? And then we have a follow -up to the
• 17:27
  questionnaire. I'm not sure. It could also be
• 17:30
  that we have a marker, a biomarker, or one image
• 17:34
  that will lead to this. So I'm not sure if only
• 17:38
  the together registry will answer all these kinds
• 17:40
  of questions, but I do think that we are, as
• 17:43
  vascular surgeons, we are looking for this together
• 17:45
  with our patients. And because a lot of patients
• 17:48
  get unnecessary follow -up, unnecessary stent
• 17:51
  grafts, and unnecessary treatment. Actually,
• 17:56
  yeah. So I hope we all can, you know, have a
• 18:00
  better healthcare for our vascular surgery patients.
• 18:04
  I'm not sure if only the Together Registry will
• 18:07
  answer all of these questions that I have, but
• 18:09
  it's a beginning. It's a start. No, but I think
• 18:13
  like Kaki was saying, I think obviously the Together
• 18:17
  Registry is not the full answer. I think it provides
• 18:20
  a piece of the pie. I think a lot of Khaki's
• 18:24
  research with AI prediction models is one piece
• 18:27
  of the pie. What the patient wants and needs
• 18:29
  is another piece of the pie. I also like to add
• 18:32
  that, I mean, we hear this discussion that we
• 18:34
  want less follow -up for these patients. I think
• 18:36
  what we're talking about is we want less imaging
• 18:38
  follow -up. But as Khaki just very succinctly
• 18:41
  pointed out, there are lots of types of follow
• 18:43
  -up. And it has other important aspects, not
• 18:46
  the least to make sure that the patient is still
• 18:50
  not smoking. taking their medicine with their
• 18:52
  blood pressure as well, controlled. I mean, there
• 18:54
  are a gazillion things that happens around the
• 18:57
  patient that I think, where I actually think
• 18:59
  our follow -up helps, that we don't think of.
• 19:02
  We take it for granted. But if you drop the patient
• 19:05
  somewhere and say, bye -bye, we don't have to
• 19:08
  see you again, then we lose that as well. And
• 19:10
  we drop that to primary health care, but some
• 19:13
  sort of system where the patient feels that we're
• 19:16
  holding their hand along this pathway, which
• 19:18
  is aortic care. Plus, it's not only, you know,
• 19:22
  and I mentioned it's funny with sac regression.
• 19:24
  I think we don't have the image, but I think,
• 19:28
  you know, we're treating a focal segment of the
• 19:30
  aorta with an endocrine when we know that the
• 19:34
  whole organ is sick in most cases. So that's
• 19:36
  one of the reasons to do subsurface. Yeah, I
• 19:41
  think it's also nice that... we have looked at
• 19:46
  the data with the vasculator consortium is that
• 19:50
  the patients actually do want to have follow
• 19:54
  -up and maybe it's not imaging, you know, but
• 19:57
  they just want to have like a connection. You
• 19:59
  want to have a good communication with the physician.
• 20:03
  And that is something they find really important.
• 20:06
  So it might be that, indeed, it's not a form
• 20:09
  of imaging that we need in these four or five
• 20:12
  years, because that's what the guidelines say
• 20:14
  of ESVS, mostly on imaging, and like three years
• 20:20
  or five years imaging if you have sac regression.
• 20:24
  But I guess you can also, you know, have other
• 20:28
  type of follow -up. Yeah, definitely. Great.
• 20:32
  and last question what are the limitations here
• 20:35
  what will together not answer well it's still
• 20:42
  a registry so it will capture some data but not
• 20:45
  all data the biggest problem with registries
• 20:50
  sometimes is to complete the follow -up with
• 20:54
  patients but you only include them after five
• 20:57
  years or ten years so That might be a limitation.
• 21:00
  The other limitation is obvious for the Together
• 21:03
  registry. It's only one brand of devices. Both
• 21:06
  Kaki and I, we know that today you have a vast
• 21:09
  array of treatments, both open repairs, endovascular
• 21:14
  repairs of these patients. So we're just seeing
• 21:17
  a very small segment of it, of course. Yeah.
• 21:23
  Yeah, I totally agree with that. So, and of course,
• 21:27
  this is, just a small set of data that we collect.
• 21:31
  I mean it's a registry with standardized data
• 21:35
  that we found important beforehand and with some
• 21:38
  quality of life but what we have done in previous
• 21:42
  research is to look at all data that has been
• 21:46
  reported as an outcome so we have made a data
• 21:51
  dictionary. with the vasculate project and there
• 21:55
  were actually 342 parameters that we wanted to
• 21:59
  capture which is not all of those in this together
• 22:03
  registry so actually we are missing out on a
• 22:06
  lot of other data uh that might be not important
• 22:11
  but they have been reported before as being related
• 22:14
  to an aneurysm which we got not going to capture
• 22:17
  in the together registry it's always about a
• 22:20
  registry to get people to Just actually sign
• 22:23
  up and participate and do it. It's a lot of work.
• 22:26
  It says you want everything. We want all the
• 22:30
  information. That means that nobody is going
• 22:31
  to enter anything. So you have to find a strike
• 22:35
  of balance. Yes. We don't have genomics in this.
• 22:39
  We don't have omics or plasma markers because
• 22:41
  we do not collect bloods. We do have imaging
• 22:44
  and clinical parameters, but not all clinical
• 22:47
  parameters. So still probably something will...
• 22:52
  will be missing because 10 years is a long time
• 22:55
  and maybe something comes up which we don't do
• 22:59
  not have been you know thought of before and
• 23:04
  we don't have it in the generation because we
• 23:06
  didn't collect it yeah thank you very much for
• 23:11
  our discussion today well have a good day everyone
• 23:16
  Thank you and happy holidays. Happy holidays.
• 23:18
  Bye. Bye. Thank you. You can find more published
• 23:23
  episodes on Spotify, SoundCloud, the Apple Podcast,
• 23:27
  or the ESVS library. Bye for now.