EU Orphan Devices: Understanding the Regulatory Pathway and Clinical Evidence Requirements

Speakers: Matthias Fink, Gavin Quigley, and Nebojsa Serafimovic 

[Download the EU Orphan Device Clinical and Regulatory Playbook]

Index

[Download Slides]

Introduction and Overview (02:20 - 06:24)

• Introduction by Matthias Fink
• Overview of orphan medical devices in Europe
• Focus on the latest MDCG document published in 2023
• Panel discussion and audience Q&A setup

Current Challenges in Orphan Device Market (06:25 - 15:43)

• Shortage of pediatric cardiology devices
• Manufacturers removing low-sales devices
• Comparison with US and Japanese orphan device programs
• Regulatory context and importance of the MDCG document

MDCG Document Structure and Key Criteria (15:44 - 21:30)

• Three parts: Clinical evaluation, procedural aspects, and critical appendices
• Definition of orphan device: Less than 12,000 patients/year in the EU
• Criteria for orphan status: Unmet treatment needs or significant improvement over existing options
• Use of data from similar devices and extrapolation methods

Regulatory and Clinical Data Considerations (21:31 - 33:11)

• Flexibility in clinical data requirements for orphan devices
• Use of in silico data, non-clinical data, and registries
• Importance of early engagement with notified bodies and expert panels
• Ethical and scientific standards for data quality

Panel Discussion: Perspectives from Notified Bodies and Competent Authorities (33:12 - 45:47)

• Transparency in orphan device decisions
• Role of expert panels and structured dialogue
• Importance of aligning expectations across notified bodies
• Advice for manufacturers on state-of-the-art justifications

Audience Q&A and Closing Remarks (45:48 - 58:29)

• Questions on in silico testing and clinical study population sizes
• Discussions on post-market data collection for orphan devices
• Future outlook on regulatory flexibility and guidance updates
• Final summary: Importance of justification and alignment in the orphan device pathway

Full Transcript

Matthias Fink - 02:20
We have a very interesting topic for this one-hour webinar today, as Sean just announced, for two very highly competent and interesting co-panelists later on.

I'm going to start with a brief, let's say, maybe max 15-minute quick introduction on today's subject, which are often medical devices, how they're going to be reviewed regulatory and clinically in Europe, especially focusing on the Latest MDCG document that was published last year, and afterward, we have a probably 35-40 minutes time for a panel discussion, taking questions from the audience. So, as you already mentioned, in the Q&A, you can put in any questions, and we will try to address as many questions as possible. And without further ado, I'm going to start right away.

As an aside, the usual disclaimer, but that's why we have Nebojsa and Gavin here from the, let's say, regulatory side, from the authority and from the notified body BSI.

So they're probably able to address the question more definitely than I can as a consultant. So why were you having this? As I was saying, this discussion about the MDR is in place now for four years. So why, in 2024, do they now decide, okay, we need an MDCG document on that? And if you look at these, this is an ongoing discussion now, and this is really something that is apparently a problem that wasn't really there at the end of the MDD.

Matthias Fink - 03:56
I'm a physician by training. I cannot remember during my 7 years of a clinical career really discussing any shortage of devices and whatnot. But it looks like it's becoming a big, big problem, and is also in the media. This comes from a webinar from the end of 2023. And they said, especially in pediatric cardiology if you remember the previous slides, it was also coming from the European Cardological Society.

Apparently, there's a big issue right now with these types of catheters and other devices specifically used in pediatric cardiology, and this is driving a lot of this discussion that at the end, leads to this device definition.

So it looks like we have a shortage of devices, a lot of manufacturers have removed their devices with low sales numbers because they're used for, let's say, niche indications, and for that reason, that currently, clearly, is a problem. Certain patients with very rare indications cannot be treated sufficiently in Europe.

Hence, we have this whole discussion. So, looking at other countries, as I said, under the MDD, it didn't look like it was a real problem in Europe. The US, for example, the FDA has this humanitarian device extension program, which is similar to what we have now with this orphan program here in Europe.

They have a similar definition, and this program has run since the beginning of the 90s.
But if you look at this data from 2021, they only approved 78 devices over 30 years in this program. So it's not really something that we're probably going to see very often. It's for rare diseases, rare indications.

Australia, for example, allows the use of data which is not approved for certain indications, so basically a lot of off-label use. Japan has a similar program to the US with the FDA, but they only approved 22 devices in over 30 years, so they have less than one device per year in Japan that falls under this definition. And what does the European Union do?

I said we got this MDCG document published last year. In my opinion, it's one of the best MDCGs we have seen so far. It's quite clear, but it still leaves some room for questions and interpretations. And we're trying to dissect this a little bit.

Matthias Fink - 06:24
I'm not going to go into all the details. It's really like a high-level introduction, but this MDCG It's quite detailed; it has three parts:

One talks about the clinical evaluation, because, as you can see, as you know, clinical data is always the relevant topic under the MDR.

It talks about procedural aspects that probably more address notified bodies and even expert panels if they got involved.

And it has three, some critical appendices that provide some examples, some templates that manufacturers can use if they are considering an orphan device submission.

So what defines an orphan device?

There's a clear definition in this MDCG document, and this is that it should be less than 12,000 individuals per year in the European Union that get affected by this disease. So it's not like, okay, the manufacturer says we only plan to sell 11,500 devices per year, so we have an orphan status. It's how many patients were diagnosed or treated with this specific rare condition per year in the EU. And this number is extrapolated from the FDA definition of rare diseases and then basically up-calculated to the European population. And this is a criteria number one that has always to be met.

And there are two additional criterias: either, one, there is right now insufficient available treatment options for these patients; or the new orphan device, and this is also, of course, targeting, let's say, innovative devices for the European market, that the expected benefit for the new device is higher or is better than the benefit of the currently available treatment option.  So it's not just a number of devices. They clearly have to show, okay, there's a real benefit regarding the outcome for treatment, either better performance or lower side effects.

So, big question: how do I determine if I have 12,000 patients per year?

So, when I prepared for this webinar, I said, okay, I do what everyone does these days.
I asked ChatGPT, talking about these pediatric intervention procedures, and ChatGPT said, okay, it's between 10,000 to 15,000 catheter-based interventions for pediatrics per year. Great, 10,000 would be fine; 15,000 would be too high for the off-device definition. So, we have to look elsewhere, of course, and I'm pretty sure Gavin will agree that a Notified Body is not accepting GPT as the substantiation.

Matthias Fink - 09:12
So we have to look a little bit more in detail, look at registries.

In the epidemiologic databases, there are a lot of publications that talk about incidents with maybe in certain countries and certain regions that can be extrapolated. I mean, you do not need to have a hard proof number of this 12,000. There are a couple of different data sources you can use. To get this as long as you need to justify that, and it's a comprehensive justification why you’d come to this. And if it's 12,050, I'm pretty sure they're not going to deny this. It's not possible. There are no databases in Europe that give the exact number of cases per year per disease.

So we always have to be a little bit flexible, but there are a lot of data, so then it takes a little bit of homework, of course, to come to this justification. 

And then, of course, it's highly recommended, also in the guidance document, get this clarification upfront. So do not put everything together and then submit to the notified body and say you have an orphan device. Get that clarification, either get in dialogue with your notified body. The MDCG also clearly recommends, okay, talk with the expert panel. The role of the expert panel could be that they could basically confirm the often device status upfront. 

Of course, what of clinical data? And this is the big, big question.

I mean, this MDCG document clearly allows for clinical data of lower quality and quantity compared to what we would expect for a normal EU MDR submission. But then the question is, what is lower quality and quantity?

It's a little bit like this: when the EU MDR was published, it said, okay, you have to have sufficient clinical data. No one really knew what is sufficient clinical data. It's always a little bit of a case-by-case decision, probably the same over here.

Matthias Fink - 11:09
So you have to look at the data. Are there other even non-clinical data you're able to use? And these are some examples. These are coming from the MDCG. I already said it's quite a useful guidance document, so certainly some laboratory and animal testing. I mean, you can never replace clinical data fully with animal data, but it might have a higher weight compared to, let's say, non-orphan devices.

Similar to testing in silico, this is a field that's becoming more and more important. You can replace certain animal and even maybe a limited number of patient studies with these models. Look at data from similar devices.

So if they are similar devices and also similar devices could mean, okay, maybe for similar conditions of use, but maybe if the device is already approved for an indication, you might use the data.

Then the MDCG explicitly allows the use of data from similar devices, even if they are not equivalent. Look at the state of the art, especially for the technology. As I said, for the determination of the status, look at the data that's out there, look at any possible data sets that you might be able to use, and also say, okay, any other relevant data, including humans, which does not qualify as MDR clinical data.

So it is quite flexible, but of course, as I said in the previous slide, get in contact with your notified body quite early and probably try to understand and get an outfit from them, what they would consider as qualified data, that you are able to use, because it could be that sometimes not all notified bodies might have the same expectation on that.

Of course, there are always limitations in clinical data, but you still need to ensure you have to follow ethical, scientific standards. You cannot just say, oh, we found something here, in, I don't know, whatever a journal, and they didn't even say it's a study, they didn't even say anything was followed. Make sure it still has valid ethical and current scientific standards. It depends, of course. The less you produce right now, and clinical data in the pre- or for the submission.

Matthias Fink - 13:36
You always need some. As you can see, the last point here, you need to always have a PMCF, and probably a lower quality and quantity of the data as you're able to submit initially, probably the more you need to put up in the PMCF. But I mean, it's still possible. A lot of these rare cases or rare diseases are treated in specific centers and not at various little hospitals.

Or it might be that if you go to a country like Germany or Austria or whatnot, that only a handful of specialized hospitals might treat this very rare indication. So, you might be able to get some data out of there. Some of them even have registries because they monitor these patients because usually, for a lot of reasons, there are not that many treatment options available. You might be able to extrapolate data you're having already.

As I said, it might be feasible to use data from another, like a common indication, to transfer them to this pediatric indication, or it might also be possible to use adult data and maybe extracurricular pediatric data, which normally is never feasible. Also, that's the only time that the MDR basically allows with this MDCG the use of off-label data, as we all know, but the MDR is certainly not allowed.

But for legacy orphan devices, meaning devices which already were CE marked under the MDD or AI MDD, now probably are removed from the market or will be removed from the market.

And the manufacturer is aware, okay, well, in the past, our device, yes, even was approved for indications A and B. We know from publications, case reports that it was used quite often for rare indications, which is not on label, and the manufacturer probably decided, okay, the number of patients per year is too low. It doesn't make sense to get through the whole process of study.

This can be used, but it's very important to mention that this off-label data can only be used for legacy devices.

Matthias Fink - 15:43
If you have a new innovative off-label device, you cannot use any kind of off-label data.

And again, always the full MDR requirements, you need a key revelation, you need a PMCF strategy, you would always need some form of PMCF, but as for the clinical data, probably the expectation on the notified body is a little bit lower for the PMCF.

In my opinion, it would obviously lead to some form of specific PMCF, but again it has to be adapted and reflected the type of rare indications of this often device status.

So just to quickly summarize, I think this MDCG is very important because it allows the continuation of state-of-the-art treatment for patients with the rare indications in Europe. The early exchange with a notified body, if you have your notified body, contact them very early, discuss with them, try to get their opinion, especially when it comes to the determination of this orphan device status. Get this upfront before you even start your whole clinical evaluation process. Saves your time, saves your money.

Also, if you want to get a clarification about the expert panel, get in contact with them early. You need a robust justification. I said, look at out what databases, epidemiological studies, and publications for a lot of indications. I did this for three or four just because I was interested. For a lot of indications, it's possible to get this type of data from the internet. Again, it doesn't have to be the exact number 10,855, but at least you can get an overview that it roughly would be within this 12,000 patients per year.

Again, there's a lot of, let's say, flexibility when it comes to quality and quantity of the data, but you still have to make sure that you justify.

Like with everything else, you submit to a notified body. If you come to the conclusion that, depending on the often device status, it is sufficient data at this point to get the CE mark, make sure you justify this appropriately. As I said, you always need some specific PMCF, in most cases, because you usually do not have significant clinical data. And, of course, you need to collect more data at the post-market setting. And also we haven't talked about this - this was really more the MDCG - but of course, all the notified body treaters, if you look at the MDCG, it talks about structured dialogue and how you have to be a little bit more flexible, it mentions certificates issued with a condition, so for example, and I'm happy to hear Gavin’s statement on that later on, a certificate with a condition saying, okay, you need to include a certain number of patients in the PMCF, but there will also be a possibility for a notified body to issue certificate with a little bit of limited data.

This was a really quick overview, and now we jump into the panel discussion, which is the main focus here.

And I'm really happy to have two such excellent speakers. First, I ask Gavin quickly, before we jump into the discussion, maybe you and then Nebojsa Serafimovic, afterward, give a quick summary of your thoughts. What do you think about this current device status, and what do you think is really the added value here for manufacturers? Okay, Gavin, you may want to go first.

Gavin Quigley - 19:16
Thanks, Matthias. Thank you all for coming.

So we don't expect to see huge numbers of these. I think that's probably fair from what everybody's said. I think what we have got is a framework where we can work together with the Commission, with the expert panels, with manufacturers, to make sure these devices are available, because that's really what we want. I think what we've got here is a framework that will allow that. You know, it's certainly something that we've been looking for. I think it gives us confidence where we can be involved in discussions with the expert panel with manufacturers and agree on a way forward.

And I think that's what we're trying to do for these rare and orphan devices. We understand there is a need and we have no wish to slow them down. I'll hand it over to Nebojsa.

Nebojsa Serafimovic - 20:15
Thank you very much.

As a representative of a competent authority and one of 27, I can say that the reason for this guidance and this work came firstly from the industry and, of course, healthcare professionals who saw the risks coming up in the future also with the MDR leaving legacy devices that will be revoked from the market.

Therefore, the MDCG recognized this issue in 2022 in its position paper, and we're very happy also in the CAE Task Force Working Group that this guidance was published last June, and it's already ongoing. We have four applications selected for orphan device advice and this is ongoing.

We, as competent authorities, hope that this guidance will lead to a fruitful and good solution, and we should keep in mind that this is a guidance that will also need to be developed in the next years when we all gather more experience on that.

Matthias Fink - 21:30
Okay, thank you, Nebojsa, and thank you, Gavin, for these initial thoughts. Yeah, I think that it's a very helpful guidance because it at least provides some clarification. As Gavin said, I don't think we're going to see a lot of these per year because, looking at the experience that the US and Japan have similar programs over 30 years now, we're talking about maybe two or three devices per year that really fall under that.

It's not really happy manufacturers who probably just have limited data and probably sell to not that many patients per year. But looking at that, the MDG was published, I think, in June last year.

Maybe Gavin, did BSI already get some information already? Did you get requests from clients? Did you already go through a process? I'm not saying to the certification, but maybe you at least got some initial discussion, or maybe already have you have some insights on how BSI is probably going to confirm that off-the-device status at the end?

Gavin Quigley - 22:32
I think we've not seen it, I've been involved in one discussion with the manufacturer, but in the end, I think we felt they didn't perhaps need to use that, and obviously, we can't advise them on their route to conformity, but we felt they probably wouldn't qualify for this particular designation. I'm aware of one other potential client who's discussing it with us, but we really haven't seen huge amounts. 

I think there will inevitably be a number of devices that perhaps feed in early in the process, but as you've said, per year, it's likely to be a relatively small number, but I think we hope there are maybe five or ten devices out there amongst all the notified bodies who will meet this designation and we will be able to ensure that they're available, but I think the important thing is the transparency of discussion.

We're very much viewing it as part of structured dialogue. We will discuss it, record it, and make those discussions available to our competent authorities, you know, if they come and want to see it.

We want to have that transparency and say this is what we're discussing, and I think if manufacturers can lead in and lead in with the expert panel early, as you've said, approach us early, approach the panel early, then we can, you know, we can work out what the best way forward is and make sure that they're not, ultimately for the manufacturers, wasting time and money to get somewhere where they don't need to be, or we can listen to what advice has been given from the expert panel and come a way forward and a pathway for the manufacturers if we're all agreed.

So I think that's really what it's about. And I think, for me, the guidance is about that need that was identified, but also the transparency. We're happy to discuss this with manufacturers; we're happy to discuss it with the expert panels; we want people to see how we make our decisions and our justifications.

Matthias Fink - 24:38
Okay, thank you. Nebojsa, any experience you, as a representative from the Commonwealth Authority in Austria, have? Did you ever get, I don't know, any feedback on that yet? Or is it still, as I said, probably too new? And since it's basically happening between the notified body and the manufacturer, maybe the expert panel, probably the authorities aren't really involved at this early stage.

Nebojsa Serafimovic - 24:58
Exactly, so if we look at the numbers from 1990 to 2021 the FDA approved only 78 such orphan devices in this time period and extrapolating from that we also don't expect a big number. On orphan device applications or requests in Europe, the MDCG did recognize the need for this issue because it's a very important field, and the patients, if collected together, are quite high numbers if we look at all the patients in Europe. And therefore, I also asked a couple of my colleagues. We did not have contact with orphan device applications yet, whereas the expert panel has. 

So in June, the guidance was published. There was the project from July running until the end of 2025, but we already picked four applications for orphan device advice. And I believe this work is ongoing, so we will see the outcome from this first try of expert panel advice on orphan devices. When looking at the single member state working with orphan devices, we see a lot of derogations where clinicians ask for derogations for devices that possibly have been on the market, have been legacy devices, and now are not on the CE market anymore. And they often have orphan device status.

Matthias Fink - 26:37
And I think this last comment is really important because, as I said, we've learned from a past webinar at the end of 2023 that a lot of, especially pediatric cardiac devices, have already been removed from the market because the manufacturers started to say the business revenue is not high enough and there are always the costs.

Since, of course, this is very important, you still have to go through the whole conformity assessment procedure. That's not like an abbreviated version. There, the MDCG provides you some shortcut, so to say, getting the data you need to provide, but it does not shorten the whole process.

So, talking about costs and maybe there's a question for you, Gavin, since it's still the whole confirm the attendant procedure is required, but of course, manufacturers probably don't have, let's say, it's not a big field where they can sell a lot of these devices because it's just such a rare number of patients. And even if, even if there are, let's say, 7,000 patients per year, it doesn't necessarily mean that all 7,000 would need that specific procedure of that device. So, as a notified body, a limited number of devices that I want to get used for this indication. So somehow reflected in all the costs for the notified body review because we all know, especially, and we talked about a lot of these are class three devices, high-risk patients.

And it could be quite expensive. Is there some, I don't know, did BSI or do you discuss other knowledge about this? Is there some leniency saying, okay, we have a reduced key or whatnot to reflect this orphan device status?

Gavin Quigley - 28:14
We don't at the moment, because we have to be completely transparent and be fair to everyone. So we also have to play by the, we publish our fees, BSI, I think, you know, do publish our fees, but there is work going on to look at what we can do, you know, for some of these either orphan type devices or supporting, you know, smaller innovative manufacturers.

So that is a piece of work that's ongoing, certainly within BSI. And I think within the notified body community in general, you know, I think we're seeing that, but at the minute, you know, we have our set fees and our set rates, and you're right, you know, these by and large are class three devices. They're not on quality certificates. They require that extra period of consultation as well. So, I think the other thing I would say is look to manufacturers, get in there now, and get started.

We can, you know, start the process; we can engage with the expert panels. There's currently, I think, no fee from the expert panels. Nebojsa, is that right?

I think they're not charging a fee at the moment. So, that may change as things go on. I would encourage people where they think there is a need and a case to try and engage with the notified bodies. And see what the most efficient way forward is for the device and what is likely. Obviously, there are commercial decisions to be made and we can't get involved in that.

Matthias Fink - 29:48
Okay, thanks for clarifying that, and I think probably it's really something that, of course, manufacturers are interested in saying. We still have to, let's say, they are businesses, they are for-profit organizations that at least have to say, okay, we cannot pay more for the certification and the re-certification every five years than the revenue for this type of devices.

Another interesting question when we're just talking about this more as logistics here, since we are all, I think we can all agree, not expecting to see that many devices looking at the other countries. And it's quite, let's say, I'm not saying, a delicate process, but of course, it probably requires a lot of a learning curve on the notified body. Reviewers like to understand, okay, what limitations; do I able to accept the clinical data? The MDCG provides some examples, but not in detail.

Question for both of you: Maybe Nebojsa should first come from the authority side. Do you think it makes sense to say, okay, we only dedicate certain notified bodies in Europe to do this? Let's say, give them a little more experience because if you only talk about 10-15 devices per year with right now 15 notified bodies. Then, basically, every four years or so, a notified body would see one of these often device applications. Do you think it makes sense to say, okay, we decide on a handful, let's say 10 notified bodies who are going to get this designation of orphan devices, or is it basically nothing that was discussed?

Nebojsa Serafimovic - 31:08
Not an idea that I've heard of. We have to keep in mind that the guidance on orphan devices is not extending what is written in the MDR, but just maybe trying to push some borders that we have here. We implemented a definition of orphan devices that there is no easier pathway for CE certification for orphan devices in the MDR. But the guidance has only the possibility to clarify a lot of questions, and a lot of issues for both manufacturers and possibly also notified bodies to just give clarity and definitions here.

So, giving just a few notified bodies a designation is not really something I could decide, but I think it's also not necessary because, for every orphan device, it will be a case-by-case decision depending on the clinical data available. And the guidance here gives the opportunity to rely on a lot of data that is not as strict for the conformity assessment as for not orphan devices, where we ask for the highest quality of RCTs, of studies, of clinical investigations inside the European Union.

That gives a lot of explanation on the kind of ideas to gather data that is, as you mentioned in your presentation, not non-clinical data, data from outside the European Union, extrapolated data, even off-label use. And, therefore, I believe we have to grow with future requests from manufacturers and also from the expert panel advisors.

Matthias Fink - 33:11
Okay, thank you. And I just saw that Sean posted something in the chat that the orphan device expert panel pilot program is free and runs until the end of 2025. So, for anyone who is interested, you might have something you can contact right now for free and get advice or feedback from the expert panel.

So apparently, every notified body can do that. And as we all know, there's always this criticism from manufacturers: notified bodies are not that well-aligned sometimes between the different notified bodies. When I was still working as a clinical reviewer for a notified body, we initiated the clinical evaluation at a Team NB-level quarterly meeting, so trying to get a better alignment. Are you aware, is there currently any discussion on Team NB-level trying to align a little bit better on these expectations to avoid that? There are too many different interpretations because it's something that everyone has to go through in a certain learning curve right now.

Gavin Quigley - 34:09
Yeah, I think, you know, Rich Holborrow, who unfortunately couldn't be here today, I know many of you will have turned up to hear Rich, and you're stuck with me. And despite my regional accent, it's not going to help, is it? The audience, I suspect, wants to hear the lovely Rich, and I can never replace him, but he sends his apologies. But I know Rich and the rest of Team NB have been looking at this and trying to make sure we are consistent.

I think Nebojsa's right in that the commission is unlikely to restrict notified bodies to specific things. There is a bit of restriction in scope, so not all notified bodies will have the full scope for implantable cardiac devices that perhaps some of the notified bodies have.

So inevitably, there will be some restriction, and we've looked at it very much as the clinical team is going to be the focus of our initial discussion, so we have a lot of experienced clinicians, certainly in our notified body, 13, 14 now and growing, who will be part of the central discussions and offer that degree of pragmatism, I hope, to manufacturers to say, look, this feels like a device that may meet the criteria.

And I think what I would say Nebojsa has come to it as well. Nothing has really changed in terms of what the guidance says for the requirements for MDR. We still have to meet those requirements. The really crucial bit for us is that state of the art, which we say to everybody, that state of the art search. Where does your device sit in this state of the art?

Show us where it sits and why your device may be the best option or a particular option.
And if there aren't comparable devices, what are the other treatments? What are the other options? Because if there aren't good options, then that's a good justification for a device. So we want to see that, even if you say that we don't have comparable devices, there aren't equivalent devices, that's okay, but make sure it's brought out in those discussions and really emphasize where your device sits in this overall treatment pathway, and perhaps why there are limitations.

We know Matthias was talking about the septostomy balloons; they're used in an emergency. There is no long-term follow-up because the patients go on and have another procedure by and large, but we know they're widely used. So registry data, Matthias touched on it as well, rest of the world data. Why is it applicable? Why can we look at it? Why do you think, if you're going to start talking about things like adult data, why can it be scaled down in this case of pediatrics? Why is it still, you know, appropriate? Have you got supporting evidence with in silico models? Can you show us these models that say, actually, if you change the stent size or the, you know, the weave or whatever it is of your stent, that it still works? 

There are lots of really good work going on in Europe. In the US, the FDA has in silico modeling centers. Let's look at that sort of thing, and we want to reassure ourselves that there are models that support these devices, and that we can use that in a sensible way to support these devices going on the market.

Matthias Fink - 37:34
Thank you, and I think this is a very, very important what you said at the beginning of your comment here, is make sure you look at the state of the art and make sure you look in where in the whole realm of the state of the art for these indications your device fits in, because quite often manufacturers are, they’re fully aware, our device is useful to these patients, but sometimes they are not having, let's say, the full overview look in the state of the art, and quite often, you can come up with a justification probably why you can use this lower quality of data. Looking at the poll you just had, I didn't want to ignore the poll results.

So it looks like the majority is here because they want to a little bit better understand, okay, how is the determination? How do you determine if I want to offer the advice? And also there was the same question coming up in the Q&A here. So I will probably hand this over to you, Gavin, first because you are also a physician like myself.

How does BSI, what are BSI's expectations? How should the manufacturer present the case? And what is, let's say, I have one quick slide on that. What could be data sources? How do you justify why a device would have PD orphan status?

Gavin Quigley - 38:44
So I think, again, start from your state of the art. Where does your device sit? Why does it only apply to this particular? Why is it likely to benefit this group specifically?

Go to the guidance document and bring your argument from there, and then construct it and say, look, this is what we have. This is where our gaps are. If you can identify your gaps as well, it's much easier to have that transparent discussion and say, well, look, that's an understandable gap.

I understand why there's that gap because there are problems with follow-up or it's not entered in registries very well. Look at where your device really sits in the treatment pathway for these patients and which patients are going to benefit.

We're talking about a relatively small number here, and I think that's the first thing is to really look at - the definitions in the MDCG guidance and say, where would our device fit? Because it may be a device that only applies to a very small number of patients, even though it's a device that can be used in perhaps other conditions. And it has a very specific indication where there are limitations.

So, you know, there are ways of interpreting, I think. We would try and approach it with our pragmatic heads on and say, okay, there is some evidence, we know this has been widely used for years, we know what group of patients, what we need is a sensible, logical argument constructed around that basic tenant of where it's used and why it's used.

And if you can bring in that clinical, and if you've been to the expert panel and they have shared your view, And you've got an idea this is where our gaps are, we know what we're going to do, we know what our PMCF is going to look like. It's much easier for us to say, okay, that is a device we can hang our hats on.

You know, this is about a bit of risk sharing as well. You know, we get reassurance from the expert panel from manufacturers that what they're doing is sensible and in the interest of patients. You know, and that's really what we're, from the notified body as well, that we need to be convinced that we're doing the right thing for everyone, for the patients, for the risk, for the whole sort of ecosystem and for the lifetime of your device.

Matthias Fink - 41:04
Regarding the number of 12,000 patients, are there any, let's say, data sources that basically BSI considers, let's say, as I have thoughts on examples, their registry if there are maybe publications, there might be in some countries even publicly available, let's say, epidemiologic databases and any tips for the audience what they probably could look at?

Gavin Quigley - 41:26
Yeah, I think that wherever your device sits, whether it's paediatric cardiology or neurosurgery or rare disease processes in the GI system, look at the professional societies that are involved in treating those patients. I think Matthias touched on it already. There tend to be very small groups of clinicians who treat these patients. If you think about hyperplastic left heart, there are only three or four centers in the UK for 60-odd million people who actually perform that and children.

So that will inevitably apply across most of the European population where there's a subspecialty. Find out from the special interest groups, from the patient groups as well. They have really good data by and large. So there are professional patient groups who recognize these conditions who will be able to tell you what sort of numbers you'll expect for your given population, or if you're given genetic makeup if that is relevant. So those are the places I would always start.

The WHO is always a good bet, as well. You know, FDA data, Australian data, the UK has a ton of NHS data which hopefully will become available over the years, and we're hoping that those sorts of things will be accessible.

Matthias Fink - 42:48
Okay, thank you, and interesting comment saying look at other countries, other regions also of Europe. So, is this something also that the authorities would agree on if, let's say, there's nothing because we all know about the high data protection regulations in Europe. So, it might be a bit more challenging to get this data. And we know that other countries, US and Australia, for example, might have a little bit more accessible data. So would this something that you would think as an authority is also be acceptable if the manufacturers actually couldn't find any real European database, but we found that, say, an epidemiologic database in Australia, and they ex-populated certain incidents for a certain indication of the European populations to something that could be considered as a relevant data source, or this is something also of Europe is not acceptable?

Nebojsa Serafimovic - 43:33
Well, the guidance does outline the possibility of using data outside of Europe. This, of course, is focused on the clinical evidence that can be provided based on possibly the use of devices that haven't been used in the European Union. On the other hand, we did extrapolate the number of 12,000 incidents per year from the US definition.

Therefore, I personally believe that this is a possible source of showing how many incidents are to be expected also in the European Union, but that really depends on two entities on the expert panels and the notified body reviewing the application. So if we look at the definition, we have to also consider that there might be not only an orphan device, but also just an orphan indication, which is more than one indication for one device. And therefore, this might be a solution if no literature research limited to Europe leads to a result that is satisfying.

Matthias Fink - 44:46
And also one, I think, interesting question we just got from the audience is, will the decisions on orphan devices be made publicly available? Like a little bit similar, like what we see with the legal relation and consultation opinions. Is there any plan to make this publicly available? Probably not, because it still follows the law formal MDR conformity assessment procedures. So I don't think there's going to be something planned, but maybe two of you could quickly address this question. Maybe Gavin, you go first?

Gavin Quigley - 45:21
I think, as far as I understand it at the moment, it won't be publicly available. But as I say, we're very happy to share that with competent authorities. I would hope that the panels will produce some sort of summary document, but I think Nebojsa may be able to come to that in a bit more detail.

Matthias Fink - 45:44
Nebojsa, any thoughts on that?

Nebojsa Serafimovic - 45:47
Yeah, maybe I start with referencing a very great website, both of the European Commission and of the EMA for orphan devices, that shows the pathway of applying for a request for orphan device status. And also the process that is very helpful besides the guidance to reach this goal. As I understood it, but I'm not sure, I thought it is intended to make something public, possibly the opinion of the expert panels.

As far as I have read it on the website but just please refer to the website in this regard. That is very helpful and has a lot of information on this. And also one more aspect that I would like to mention here regarding the expert panels is that we differentiate between early advice and advice if a clinical assessment is already advanced. So, as a manufacturer, you have the possibility for an early request to the expert panels, as mentioned today, the project is ongoing until the end of 2025, but I think the expert panels are picking limited numbers of requests and also have some criteria which requests they will pick.

So we're looking forward to hopefully gathering more information and also experience on that, and hopefully, we will get something published when the expert panel issues.

Gavin Quigley - 47:21
Our conformity assessment bit remains between us and the manufacturer, but yes, I think there is hope that particularly where there is advice given that will be publicly available from the panels and, again, hopefully a kind of summary every year to give us a flavor of what's been happening, because the notified bodies, as you said, there's what 40, 50 nearly If it's split between lots of them, we would like to get that consistency.

Matthias Fink - 47:53
Yeah, because especially, I think, probably if it's two manufacturers, let's say, contact two different modified bodies with a similar device for the same indication. And you do not want, because manufacturers, there's MedTech Europe, there's other trade associations, they're talking to each other. And of course, the worst case would be that one manufacturer would get the orphan device designation, and they could use, let's say, limited clinical data might get the market advantage because they might access the market earlier than the other one who has to go through the full, let's say, clinical investigation. That's certainly, I think, something no one would like, would like to see that, and there's a long, long discussion to be more aligned.

Another question from the panel, probably quite easy. I could answer it right away, but I would probably give this to Gavin since he might address the notified body. Is the company's certification ISO 13485 or 9001 required for the off-device pathway?

Gavin Quigley - 48:47
Well, I mean, you must have a quality management cert.

The MDR, as you know, doesn't actually specify what it says. I think the wording is a “recognized quality management system” is the phrase that it uses. But I think inevitably, that means, yes, you have to have one of those by default.

Matthias Fink - 49:10
And yeah, I think Nebojsa, you're just nodding. I think you can fully agree. And I mean, looking at the MDR and just looking at my 10 years of experience as a clinical reviewer and now as a consultant, I would say everyone has an ISO 13485. I think if you look at the overlap between the ISO 13485 and what's required for the UMDIs, the QMS, I think, it's a mandatory requirement that you need to have an ISO 13485. 

Another question from the audience is, okay, can you show an example of in silico testing regarding orphan devices? I cannot. I mean, in silico testing is something that's really picking up now. There's a lot of research into that, but it's still kind of new. I remember when I was at the notified body, we got some questions from manufacturers saying, can we use that? And I know there was a lot of discussion within my notified body because this was really something new the MDR doesn't mention in silico testing. We see in silico at the moment in Europe as supportive data.

Gavin Quigley - 50:44
We don't see it replacing human data, because when you look at clinical evidence definition in MDR, it talks about human subjects. That aside, we certainly think it has value. The FDA is a bit further ahead, and myself and one of your former colleagues have been involved in discussions with the FDA and MDIC about in silico modeling and what that will look like.

I can share an example, there's been a couple of good examples, one on fatigue fracture testing for pedicle screws in the orthopedic indication that the FDA and researchers have published in the last year. And there was a recent webinar looking at stents and testing nitinol stent failure and predicting when it would occur. And they were able to show that the in silico model predicted the failure pretty much spot on. So I do think there is a benefit in silico, but at the moment, it can't replace human clinical data, which is probably the best way.

But I do see value, particularly if you're thinking about stents and what's going to be the worst cases. You know, we would perhaps start to expect manufacturers to produce this sort of in silico modeling to say, okay, we've looked at it, and even though we're getting very small, we still think it's within our acceptable failure rates, and our performance, and we can model a flow, and there were some very sophisticated modeling systems that we're going to see in the next few years.

Even from digital twins, where you effectively can create a population, and do a clinical trial on them virtually. And you may be able to enhance or pull out certain characteristics. So they may have a very small, you know, duct or a very small aortic valve, whatever it is you want to characteristic, you can then use that and model. And we know manufacturers are already doing this as part of their R&D because they won't waste time on designs that are never going to work. So it's already happening. What we have to find is a way of looking at it, validating the model, and accepting it as supportive evidence.

Matthias Fink - 53:00
Thank you. Enabling synthesis in silico is quite new and really rapidly evolving. A little bit on artificial intelligence: the MDR doesn't talk about artificial intelligence; now we have the AI Act. Do you think that maybe we're gonna get, maybe in the near future, let's say two or three years, an MDCG document that probably introduces, let's say, a little bit like you see with this orphan device? That we see probably the possibility that certain partially or maybe some in-syllable data can be used to partially replace or additionally support clinical data? Any discussions on that? Are you aware of any?

Nebojsa Serafimovic - 53:37
Well, no discussions on that in the CIE Working Group, in the Clinical Investigation Evolution Working Group. We have the matter addressed, I believe, either in the already published Q&A or in the upcoming Q&A.

This is the level of discussion; it's in the Q&A. Other than that, we in our member states do assess AI for the moment as software as it is, but yes, it's quite difficult because it's not exactly the same. In silico testing, yes, of course, it's supporting, and it's also mentioned in the guidance for orphan devices to be very helpful. 

The examples I have are stress-strain tests on tissues or even mechanic medical device parts that can also be done in silico. Or another example would be the long-term resilience or long-term life of devices that cannot be tested for a lot of years, and therefore, computer models can just extrapolate existing data into looking at what will happen in 30-40 years.

Other than that, I believe the orphan device status is something that has been recognized by the MDCG and was born out of the need. Artificial intelligence is a more complicated topic, I would say.

Matthias Fink - 55:06
No, certainly. And that's probably then part of another webinar. I think we have time for one last question from the audience. And I know we probably partially addressed this, but there's a specific question here.

What would be the study population size for a clinical study for a 4-off device approval? Any quick thoughts on that? I think there's no fixed number because I think it's a very complex topic.

But since we agree, it could be a lower quantity. And, of course, if you do a premarket investigation, the MDR requires statistical analysis. So, maybe any hints on how the statistic could be addressed in such a study for an orphan device?

Nebojsa Serafimovic - 55:48
If I may start, what is also outlined in the guidance is to postpone gathering data in the post-market clinical follow-up studies.

So, every device has to do its post-market clinical follow-up, but you are encouraged to plan studies to show them in your clinical evaluation plan for after your CE marking and after putting your device on the market. And therefore, I don't have an answer on a number for pre-market studies, but you can plan to gather information on specific areas that you are not able to show with your clinical evaluation pre-certification in your post-market plan, and therefore, you can a little bit, using the guidance, lower the numbers that are needed for your clinical evaluation for your assessment with the notified body, I would assume.

Gavin Quigley - 56:49
Yeah, I think what I would say as well is if Nebojsa said this already, if you're going before you've got a device in the market, go to the panel, have that sort of discussion with them, align the expectations, and then at least a notified body can be a list of keywords relevant to topics that may occur during the meeting. That was outlined really succinctly. You know, think about what your gaps are, how can you close them, and again you can go to the panel and get that advice and say, what do we need to try and close off this device, you know, and what is the sensible scientific justification given the limitations of the population we're treating, you know, it's not going to be perfect, and I think we all accept that.

But what we have to be able to look back on and say, if we look back at it from an auditing point of view, did we make a sensible decision? Is it in the best interest of the patient? And are we, you know, are we able to make those justifications?

And when you write your clinical evaluation reports, that’s the story that it should tell. Where do we start from? Why is this device important? What's the evidence we have? How are we going to cover the gaps, and what are we going to do in the post-market phase?

That should be a standalone story for your device that anybody with, you know, some scientific knowledge should be able to read and understand. That would be our usual advice.

Matthias Fink - 58:29
Thank you, thank you both. I think this really sums up very nicely what you just said. Whatever you do, whatever you present to the notified body, make sure it's sufficiently justified. While you do that, the notified body ideally can follow you. Thank you, everyone, especially thank you,  Nebojsa and Gavin, for the very interesting discussion.

I'm sorry, we could not - we tried to address most questions. We could not address all of these questions for the sake of time, but thank you, everyone, and a really important topic.