00:00:00
Voiceover: Hundred Days to Success. This is the podcast from G_NIUS, the Guichet national de l'innovation et des usages en e-santé. With Lionel Reichardt, meet e-health innovators and key experts to help you succeed in your projects.

00:00:21
Lionel Reichardt : Hello everyone! You're listening to One Hundred Days to Success, a podcast aimed at innovators and entrepreneurs in digital healthcare, but also at anyone curious about this field. This podcast is produced by G_NIUS, the Guichet national de l'innovation et des usages en santé of the French Ministry of Health and Prevention. In this episode devoted to CE marking. I'm delighted to welcome Valérie Hélin, Director of Scientific and Medical Affairs at Prédilife, to talk to her about the status of an innovative healthcare solution and how to successfully apply for CE marking. Prédilife was founded in 2004, and has developed an innovative personalized breast cancer screening solution. Valérie, thank you for agreeing to share your experience with us. Could you first tell us about your background, your training and how you came to join Prédilife?

00:01:10
Valérie Hélin : Hello, thank you for your invitation. So, I'm an engineer with a doctorate in biology. I started my career in a biotech company in genetics and joined Prédilife in 2007. Initially on a part-time basis to look after academic and institutional relations. Since 2016, I've been in charge of the company's scientific and medical affairs. So a few words about Prédilife. We develop predictive and preventive medicine solutions by collaborating with major academic centers, since we came out of Gustave-Roussy and, since 2019, we've been collaborating with the Institut Curie on the genetic part of one of our solutions, which we'll be presenting later.

00:01:49
Lionel Reichardt : Predilife does personalized screening. Can you tell us more about your background and your solution?

00:01:55
Valérie Hélin : So a little background on Prédilife. As I said, we were created in 2004 by Stéphane Ragusa, a polytechnician and statistician who worked on risk models. So Prédilife was first and foremost a meeting with an epidemiological researcher, Françoise Clavel-Chapelon, who set up the E3M cohort, 100,000 women tracked since the 90s on their lifestyle and state of health. From that meeting onwards, there have been ten years of research into risk factors. Quite quickly, we became interested in breast cancer, since it was a cohort of women, given the incidence and impact of the disease. So, ten years of research into risk factors on different models and we came up with a proprietary model, the nearest neighbors model. Next, we gained access to an American cohort of one million screening women through our collaboration with the Breast Cancer Surveillance Consortium. We finished developing the score and validating it on this cohort. We showed that it could be exported from one country to another.

00:02:57
Valérie Hélin : This is what enabled us, and we'll talk about this later, to be in charge of risk assessment in a major European project, the MyPeBS project. Then we met Suzette Delaloge, in charge of breast pathology at Gustave-Roussy, who helped us turn our products into solutions for use in clinical practice. So we then had a clinical study promoted by Gustave-Roussy, funded by ARC, which enabled us to show what I was saying about having a solution that could be used in doctors' surgeries. When we talk about medical use, we quickly get to the commercial side, so we had to look at the hidden regulatory side. We developed a global solution to assess the risk of breast cancer for personalized screening.

00:03:56
Valérie Hélin : So our solution is first and foremost a test that will help assess this risk. So, the MammoRisk test and, in parallel, we've developed another tool, DensiMammo, because one of the risk factors for breast cancer is breast density, which is an imaging marker that we evaluate using mammography. So it's a major risk factor. And today, there's standard golf. This is the radiologist's visual assessment, which has a problem of reproducibility between radiologists and between radiologists themselves. So, in order to offer a truly global solution, we developed a tool other than a medical imaging tool. So, we have two products for a complete personalized screening solution.

00:04:39
Lionel Reichardt : Fittingly, on that part there, Valérie you said it well. You have two tools: DensiMammo on the one hand, which allows you to evaluate breast density, and MammoRisk on the other, which contains the algorithm that calculates a woman's risk of developing breast cancer. How did you realize that these tools needed to be given a status? Were they regulatory tickets? Was it a desire to clarify the situation for doctors prescribing these solutions? How did you go about finding out what status your tools should have?

00:05:07
Valérie Hélin : So, very quickly, because of our network (we'll probably talk about it later), we called on the services of a firm, a regulatory consulting firm, to find out more about these issues. And so, on the basis of this advice, we drew up the available guides. We realized that one of the DensiMammo devices, the one that analyzes a mammography image. It was quite clear that this was a class 1 device. For the other, which is a risk prediction tool. Here, things were more tangential and so it was a strategic choice for the company to convert both products into medical devices to, as you say, have a certain credibility with doctors.

00:06:06
Lionel Reichardt : For entrepreneurs who would like to develop innovative solutions, can you tell us what are the major steps involved in bringing your medical devices into compliance? As you said, there was only one that had to be a class 1 device. I think you decided to do both. What were the stages of this compliance?

00:06:22
Valérie Hélin : Once the medical device category has been identified, in this case Class 1, it's self-certification, so relatively simple to obtain since you self-certify that you have a compliant file. So, the main task for this type of device is to draw up a technical document that will be available to the regulatory authorities. So we did this with the same consultancy firm. A technical document means taking over the software part of the project, including all the specifications, risk analyses, tests, etc. So, the bulk of the work lies with the same consultancy firm. So, the bulk of the work is here for our developers, documenting everything that has been done so far. And then, the other big part is the clinical evaluation. How does my product meet the medical claim I'm going to make of it, in this case for us, i.e. to assess the risk of breast cancer. How do I measure that and then give a correct breast density for the medical imaging tool?

00:07:37
Lionel Reichardt: What difficulties did you encounter in this stage? What were the facilitators? In terms of training, are there any organizations to contact or things to do? What advice can you give to an entrepreneur who, today, would like to switch to this class 1 status?

00:07:52
Valérie Hélin : There are some very good databases in France, notably the documentary database of the NSM, of the CNIL for another party, all the personal data. We're also a member of Snitem, the French medical device trade association, which also has a documentary database and offers interesting training courses, so there you have the whole package, the whole landscape of access to information. Frankly, we couldn't have done without it, the way we are, but it's true for a lot of companies, I think there are only a few of us, all dedicated to operations, whether it's the developers or the clinical side working with doctors, then sales. So, the time we have to devote to regulatory issues is relatively limited, although we also have a legal expert who has taken on some of this work. So, the support of a law firm was really important for us. But with a law firm, you also have someone who can suggest a strategy. It's also a good idea to go and look elsewhere, based on these documents, and to be able to compare opinions in order to develop a strategy that will have a major impact on the life of the company for some time to come.

00:09:18
Lionel Reichardt : So here, we're on compliance with European directives on medical devices. Directive 93-42 in 2015, then in 2017 came out a new European regulation on medical devices 2017-745 which was supposed to apply at the end of May 2020. With the health crisis, this has been postponed a little. But how, as a company, are you apprehending these changes, which are very structuring?

00:09:42
Valérie Hélin : That's right, the impact is indeed major, as our two solutions move up to class 2A. The main difference between Class 1 and Class 2A is that you have to be accredited for your quality management system. So, there's a management system that's already been set up, but it's relatively light. Now, you have to set up this quality management system, get audited and accredited. And then, on the other hand, the technical file I mentioned earlier will also be audited by a notified body. There are a certain number of devices like ours, which go from class 1 to class 2A, and so there was a sort of bottleneck in making an appointment with these notified bodies. So, as you said, we're in 2017, and the application was due in May 2020. You need to make an appointment at least a year in advance, because there's a risk that you won't pass on the first try. You need to know that when it passes, in May 2020 if you're not in class 2A, you're no longer allowed to sell. So, the impact on society is still major.

00:10:54
Valérie Hélin : So, making an appointment a year before turned out to be very, very complicated. Very few notified bodies. There was a lot of stress in the company. We had to get things up and running with the consulting firm and our teams, so we had to set up a quality management system and get our technical documentation in order. And then prepare for this audit with organizations that were either not yet accredited for this new regulation, or one or two were completely overwhelmed and we couldn't get access to them. So, 2018, 2019 we were a bit in the dark, knowing that we had recruited, to get things moving, a work-study student. We didn't have the means to dedicate someone to a permanent contract for this part of the project. So we opted to continue working with the consulting firm, and then a work-study student dedicated to setting up the quality management system, then preparing for the audit.

00:11:58
Valérie Hélin : So, that we started at the end of 2018 2019. We still have the stress of the audit, setting the deadline, etc. And then, I think it was at the end of 2019, we started to hear rumors that the date for software like us, for all class 1 medical devices that were going to be upgraded to class 2A, we were going to have a bit of an extension. So we started to take a breather, knowing that it's a bit difficult because we're hoping, we want to take a breather, but telling ourselves that if it doesn't go through, we still need to be ready for May 2020. And then, at the start of 2020, there's an announcement that class 1 devices will benefit from an additional four years to move up to class 2A if there are no major changes to the device. So, obviously, we had a blast. The idea, the work has largely been done. The idea is not to stop everything and wait until May 2023 to redo the audit. But it does allow us to keep our quality management systems alive, to continue with the consulting firm and a work-study student to prepare for this audit and to comply on all fronts.

00:13:21
Lionel Reichardt : As you mentioned, upgrading to Class 2 is very structuring. You also have to be ISO 13485 certified. Do you have any idea of the cost involved for a structure like yours to adapt to this compliance?

00:13:35
Valérie Hélin : Yes, yes, we have a very good idea of the cost, which is significant. As I said, there's already the support provided by the consulting firm plus the alternant. Then the company can choose to recruit someone else, but I think the costs are roughly comparable, if not higher. And then there's the cost of the audit. So, as I said, we have to have an audit for the quality part, i.e. ISO 1385. An audit for the technical file. And so, we had quotes from notified bodies in the region of 40,000 euros, bearing in mind that the ISO 1385 quality part is done annually. Once you've been accredited, you have an audit every year, so 20,000 euros, maybe 15,000-20,000. It was a bit cheaper the following years, we'll say 10 to 15,000 euros each year.

00:14:28
Valérie Hélin : So those are the direct costs, which are very clear. Then you have to add the time spent by the teams, i.e. the development teams who have to feed the technical file. As I said earlier, as soon as there's a small change, everything has to be documented. And then, as soon as there's a small change, we analyze the risk it has on the software. Then we run tests. Then there's the clinical part, which takes up a lot of my teams' time, i.e. clinical evaluation. Once we've got the file, once we've done the clinical evaluation, there's also everything to do with the post-market study. So you need to have indicators and studies in place to ensure that the tool is assessed, and the medical device evaluated, once it's been put on the market. All this, obviously, takes a lot of team time.

00:15:25
Lionel Reichardt : Are you wondering about the status of your innovative solution and CE marking? Find out from Thierry Sirdey, Director of Medical Devices, Cosmetics and In Vitro Diagnostic Medical Devices at ANSM, the French National Agency for the Safety of Medicines and Health Products. Hello Thierry. Could you start by giving us a definition of what we mean by a medical device? What does it encompass today?

00:15:50
Thierry Sirdey: So medical devices are a very, very vast field. It's a healthcare product in the same way as others, such as medication. Its very definition is that it is a healthcare product that is not a drug. So, like a drug, it can treat. It can compensate for handicaps. It will be able to make diagnoses, and we'll come across medical devices in just about every shape and form for the general public and for healthcare professionals. And then we'll cover just about every industrial speciality, i.e. we'll have medical devices made of plastic, others of metal. I'll also come across software medical devices and electronics. Medical devices are usually presented as a family, ranging from tongue depressors and plasters you can buy at the supermarket to nuclear magnetic resonance imaging (MRI) and hip prostheses. New regulations were published in 2017 and will be mandatory in 2021. And it's precisely this new regulation has given a stronger place than the previous one, which was, in its first draft made in the early 90s and which gives a stronger place to new information and communication techniques, either software solutions or connected solutions.

00:17:19
Thierry Sirdey: And inside, we're going to find definition elements, classification elements because we're also going to take into account the risk linked to the use of these applications. So we'll have different classification rules, then requirements that apply to these devices. And today, there are specific requirements for connected or software solutions. So, the manufacturer's first task is to find out what regulations are available. It's a language, it's a text, it may be a regulatory and legal text, but it's accessible to scientific project leaders, to people with scientific backgrounds such as engineers or healthcare professionals. Today, the advantage that an entrepreneur in 2020 has over someone who had their first software applications in the early 2010s, etc., is that there is a clear advantage for them. There's a lot more literature out there, and there are a lot of regulatory guides to help project developers. So, we have guides in France, we have guides in Europe, we have guides at the international level too, which enable us to position ourselves well, to understand well beyond the regulations what the spirit of these regulations is.

00:18:37
Thierry Sirdey: That's really something you have to do. An entrepreneur, a project leader who chooses to enter the healthcare field, I'd say has to make this investment effort to enter the healthcare field and understand that in fact this investment has to be made in the regulatory field, because we're dealing with a healthcare product, as I said earlier. We're not dealing with a drug, but a medical device is a healthcare product which has its benefits, but which also admits risks, and therefore needs to be made safe. Safety means regulation, and so this investment will have a fairly rapid return, because the company will have positioned its product. This will be an advantage towards the users he'll be targeting, whether they're professionals, the general public, or patients, and also for him, as an industrialist, as a project developer, it's going to secure him and enable him to have this compliance demonstration process that's as fair as possible in relation to his device and in relation to the use for which he intends it.

00:19:50
Lionel Reichardt : Predilife is also present in the United States. We had to comply with FDA regulations in this territory. You were in charge of this. How did you deal with this difference with the FDA in terms of registration, French or European compliance? Are there any big differences?

00:20:07
Valérie Hélin : So there were indeed big differences since we had a class difference. The risk assessment tool, the MammoRisk test in the United States, at least for the time being, is not a medical device. So we're outside the regulations imposed by the FDA, so there's nothing specific for this MammoRisk solution in the United States. On the other hand, the DensiMammo medical imaging tool was directly a class 2A device. So, we quickly came up against the FDA. We filed a 510k dossier, which means that when we have a comparable device on the market, the 510K consists in comparing ourselves to this device and showing equivalence. And if we are equivalent, we are entitled to be marketed on the territory. There was indeed an equivalent device, equivalent on the medical claim. In other words, a tool to assess breast density, but obviously not equivalent in technical terms. And there, you have to convince the FDA that the technique is correct and provide a really well-documented file. So, I think we've started to get a little idea of what's going to happen with the Class 2A audit for us in 2023, with what we've done with the FDA.

00:21:38
Valérie Hélin : So my feeling is that we're dealing with very specialized teams. It took us over a year, more than a year of exchanges with the FDA, which ended well, but also cost us money. We wanted to go ahead with a French consultancy firm that had some experience. Well, that was a bad idea. You have to work with local people, with local experts. You're in the United States, so you triple or quadruple the cost of consulting. But that's what it took to get the approval. So, experience. I think it will also help us for the future, for this 2A class in Europe.

00:22:28
Lionel Reichardt : Thank you for those elements, Valérie. Question that naturally follows: when is the refund?

00:22:33
Valérie Hélin : So for us, it's a bit special. Because when we talk about reimbursement, we talk about medico-economic files and impact analyses. And when it comes to risk prediction, we have to work on large prospective cohorts, and that can only be done with a lot of money and a lot of partners. That's the aim of the MyPePS study, funded by Europe and promoted by Unicancer. The Principal Investigator is Suzette Delaloge. There are currently 6 countries that will include 85,000 women, half of whom will undergo personalized risk-based screening, with Prédilife in charge of the risk assessment platform. The other half will undergo conventional screening as currently practiced, i.e. the same for all women aged 50 and over. The aim of the study is to compare these two types of screening, and to provide answers about the medico-economic impact, and the fact that it will enable all women, whatever their social class, to take part. So there are a lot of things to consider when applying for reimbursement that will only be possible with this type of study. So we're waiting for the results of these studies to see. There's no point in starting to approach the health authorities, who are going to ask us for the results of this type of study anyway. So, in the meantime, we're working on a non-reimbursed product that has all its utility at the individual level.

00:24:10
Lionel Reichardt : Valérie, it's clear that this journey, this step for a company is an extremely costly one in terms of time, energy and resources. This podcast was recorded as part of the launch of G_NIUS. If G_NIUS had existed during your development, what would you have hoped or expected from it?

00:24:25
Valérie Hélin : So, I think we couldn't have done without a consulting firm. On the other hand, it could save us time by facilitating access to the platform, which will bring together all the different resources we might need in terms of regulations and quality. So here's a tip, an additional opinion and a time saver I think.

00:25:00
Voiceover: Those who are making e-health today and tomorrow are on the G_NIUS podcast and all the solutions to succeed are on gnius.esante.gouv.fr