[00:01:12] Aleksandra: Welcome everyone. Big picture. So if you are working in drug development or are even remotely interested in digital pathology and tissue image analysis, and know that you need a lot of data for deep learning, you’ve probably heard this initiative, big picture. Everybody’s talking about collecting images for developing generalizable algorithms that can help accelerate drug development and improve healthcare in general patient care as.
[00:01:39] A big initiative. And if you’re not part of it, you probably heard of it peripherally, know what this initiative is about, that it’s going on. Um, that, and it’s called big pictures. So there’s gonna be many pictures of pathology slides. Um, this is why the project is called like that, but for those not directly involved, we meet here.
[00:02:02] With two of the project that are here with me. Welcome. I have here and Julie Black. Yaron is a full professor at rabu university in the Netherlands. And, uh, he is also the co organizer of the famous chameleon challenges, IMiD analysis, challenges for detecting breast cancer, metastasis in lymph nodes and Julie LER.
[00:02:25] She’s a veterinary pathologist as myself and she has over 20 years of experience, both in CROs and pharma, uh, in the non-clinical part of drug development. She is leading the non-clinical part of digital pathology and computational pathology efforts at Novartis. Welcome. Both of you. I’m super happy to have you here today.
[00:02:50] Julie: Thank you, Alex on. Thanks for inviting us.
[00:02:52] Aleksandra: Thank you. So I hope my intro to the big picture initiative was not too confusing, but even if it was we’re gonna. About what that is. We’re gonna answer the question. What is it? Why is it even organized? Who’s doing it. And how are you guys doing it? It’s a long-term initiative.
[00:03:12] Once we covered the basics, then what, why and who we’re gonna talk about what has already been done. So without further, do I, uh, leave the stage to. So Yaron, before we dive into the big picture, do you wanna say a couple more words about
[00:03:30] Jeroen: yourself? Sure. Yeah. Thanks Alexander. Yes. So my name is and as you said, I’m a professor of computational pathology in, in the Netherlands university medical center.
[00:03:41] I am also a gas professor in Dean chipping in Sweden. And I, I am the, the CSO of, uh, spinoff company, Amon, actually, I’m not a pathologist like you guys. I am, um, I’m a computer scientist. I studied computer science and I have been working in a pathology department for close to 30 years now. So a long time and almost all the time I have been developing,
[00:04:05] Aleksandra: you’ve put up with pathologists for 30 years.
[00:04:08] Yeah, it’s amazing. Isn’t it? that is a, an achievement
[00:04:12] Jeroen: in itself. I should put this on my CV. I think, yes. My, my research is really on, so on the development of computation, the computerized analysis methods of software for histopathology, which used to be in niche thing, but in the last say, 10 years now we have whole side imaging and we have modern AI technology.
[00:04:32] It it’s really become a mainstream thing in mythology. What we realized more and more actually is that yeah. As you. To be able to, to do this and to develop AI for, for pathology, we need lots of data and that’s where big picture comes in.
[00:04:46] Aleksandra: Okay. Thank you. Joran for the intro and yeah, 30 years, uh, in pathology department is a huge achievement.
[00:04:53] Julie, a couple of words about yourself from yourself.
[00:04:57] Julie: I’m Canadian line now live in Switzerland and work for Nova in Switzerland. I’ve worked in several places on the planet, so Canada, us, and also in Europe. And, uh, same year I, I was always an early adapter of new technologies, including digital pathology.
[00:05:15] So I’ve been using digital pathology since several years, probably since the, since it’s in fancy 15 years ago or so. So I remember starting in Novasis and already playing with digital images, but at that time we were not as ambitious as we are now. So those images were used for very specific use case.
[00:05:37] And, and, and now we are, uh, seeing this as a really a way. Completely revolutionize the way we will perform pathology in the future and improve our, uh, life being more efficient. For example, being better also this way, we really see that as a kind of help for the pathologist. That’s why I’ve been depending this project.
[00:06:01] So favorly, since it’s its beginning and that’s why I wanted to be involved as well. So let’s start
[00:06:08] Aleksandra: with what
[00:06:09] Julie: is big picture in a nutshell. So big picture, there is a big, uh, represe of all slide images and its metadata. And it’s a, um, a very big project, 70 million. It’s gonna include 3 million slides. So, uh, 2 million of which will come from the, the non-clinical node and 1 million will come from the clinical node.
[00:06:33] It’s gonna be very diverse, so different species, a little different disease by doing so then we are, uh, becoming the catalyst for, uh, the development of AR models, but having this repository that is open and that, that can be, that is accessible also. And
[00:06:53] Aleksandra: why did this start? The general? Why? I understand, because everybody wants to develop those algorithms and.
[00:07:03] No place to find all those slides, but why now and why in this form that you are doing it.
[00:07:10] Jeroen: So why? I think for me, I found that increasingly we, we have seen, so we, we organize for instance, the ed challenge in which we share 1400 scan slides with pretty much everyone. And we’ve seen that over 1500 research groups downloaded those 1500 slide, 1400 slides.
[00:07:29] So we noticed there was a huge hunger for data in the digital pathology domain and people working on AI and pathology. And from our own experience, we found that we were increasingly trying to, to get to collaborations. My time was really, I was spending a lot of time on, on, uh, data transfer agreements, material transfer agreements, and.
[00:07:51] We found that increasingly people needed data and bigger data sets to produce AI and to validate AI for pathology. So we see on one hand the, the possibilities of AI for pathology, but also we see that the biggest hurdle for everyone in the field, not just for academics, but also for pharma research and for companies, small companies, large companies, It’s really the lack of good quality data and, and associated metadata.
[00:08:18] So we realized that AI can play a big role. And it says here we will be able to get better knowledge of diseases by studying all kinds of disease, tissue sections. It can lead to better treatments. It can improve our diagnostic, accuracy, efficiency. And especially of course, in the docs, uh, both worlds, it can help to replace, reduce and refine animal research, but we realize to really be able to develop meaningful applications.
[00:08:44] We need lots of data. So that’s why we started to set a big picture
[00:08:48] Aleksandra: and who is doing it. So did you initiate it year own who’s who said we’re
[00:08:54] Jeroen: doing this? It’s basically, it’s a combination. The, the granting scheme that big picture is part of which is called IMI. It’s a European, um, union granting scheme in which principally pharmaceutical companies to gather it, European union, the fight projects.
[00:09:11] And let’s say they, they, uh, get out a call in which they say, okay, we would like to have people, uh, apply for a grant, which will do this and this and this. So basically from the pharmaceutical company sites, there was a clear need to do this. We as academics, we sought a call, we’d read a call and we realized this was exactly what we were actually thinking about for a long time already.
[00:09:36] So we wrote project proposal for this, which was reviewed. There were number of competing proposals. We won a bit. So to say after that, the pharmaceutical companies address say the consortium and together we, we wrote the final text and we developed the project as it is now. So if you say who is doing this, so basically at a larger scale, we would like it to be very inclusive.
[00:10:00] So the community that we ate for is even much broader than the consortium that we’re doing. The community could be what you could see here. So pathologists, but also SMEs, governmental organizations, regulatory agencies, researchers, pharma, startup companies. The consortium itself is 46 parties.
[00:10:20] Aleksandra: Before we go to the consortium Yuran mm-hmm can you go back one slide?
[00:10:25] Tell me about your logo. Pink thing. I assume these are some nodes of a convolution neural network, the blue dots, but what’s the pink thing. Is this a cell?
[00:10:37] Julie: It’s, it’s a tissue. Ah,
[00:10:39] Jeroen: okay. It’s taken a lot of time to get to the local where Alex and actually it’s not a neural network. It’s a, a connection of, of different parties in the community.
[00:10:49] So it’s connecting pathologists companies, uh, in kind of network. But of course you can see many things in it. And the, the
[00:10:58] Aleksandra: big thing we misinterpreted,
[00:10:59] Jeroen: we jokingly called Pix the potato. It does look like a
[00:11:03] Aleksandra: potato .
[00:11:04] Jeroen: The previous version was even more the potato it’s it originally was a cell. So someone, so the designer that made it made a sell and within it was a network.
[00:11:14] And it was shaped a bit differently now, but it’s yeah, it’s supposed to be a cell and it has the H and E column scheme a bit.
[00:11:21] Aleksandra: Okay. If, um, you are listening on the podcast, then I’m gonna put this into show notes, the picture of the logo. And if you’re on YouTube, then use know exactly what we’re talking about.
[00:11:32] Okay. Let’s move on then.
[00:11:35] Jeroen: So the consortium that is doing the project now is 46 parties of each 10 hour pharmaceutical company. And 36 are, it’s a mix of small, medium enterprises, academic centers, but also smaller hospitals. We have regulatory, um, agencies in it. So it’s really very diverse mix of people.
[00:11:55] Because we want to tackle, um, many different aspects of this. So we also have legal, um, involved to look at, let’s say the legal landscape, what can we do with data? What, what are we allowed to do? It’s a very large
[00:12:06] Aleksandra: consortium, but this list, like the, the people who are involved in this consortium is already closed, right?
[00:12:13] There’s no. Oh, we would like to join, uh, as we missed the initial registration, but we wanna join you. That’s not an option anymore. Right? Or am I.
[00:12:22] Julie: Usually it’s not an option, but we have heard the interest of several additional partners and we are exploring to see if this is possible or not, but I don’t wanna have anyone have their hopes to why, but we are looking into this.
[00:12:39] Okay.
[00:12:40] Jeroen: But so the, the consortium is close in the sense that the people that get funded by the European union, that’s pretty much closed. So it’s going to be tough to add problems. But we definitely want to be a very inclusive consortium and community in the future. So any parties that would like to share their data with big picture, use data in big picture, share AI, build AI on top of big picture.
[00:13:07] Everyone is welcome because this is not a 46 party effort in the end. It should be much bigger effort. So people are really welcome to join us. But will not be able to get funded by the EU cause that’s yeah, that’s closed for
[00:13:21] Julie: now. The 46 partner are, are bound for six years and that’s where everything starts.
[00:13:27] But after the six years we have what we call sustainability. So we want this repository not to be functional only for six years, but we want it to be functional and, and even growing with time. So if these partners that have not jumped in the boat early on, can jump in the boat later.
[00:13:48] Aleksandra: Okay, so let’s then talk about how are you doing it?
[00:13:52] You said that it’s gonna take six years to build this repository. Mm-hmm . And then after the six years you have a maintenance program for it, when actually people can benefit, how are you organizing this? It seems to me a massive project. How
[00:14:08] Julie: are you doing it? What we would like to do, uh, ultimately is to make the data and the AI tools available.
[00:14:15] So they will coexist. And this is true, an international and inclusive community of experts. So don’t forget that from the beginning, this is a, a community based model. So as you have seen us, we are it’s multidisciplinary. We are not coming from the same, uh, field, even though ye coexist, 30 years with, with pathologists.
[00:14:39] So it’s really based on community and, and sharing these people are coming from different fields. They are multidisciplinary, and we really want to, to pave the way for the computational pathology. So the why it takes six years. To gather 3 million slides. So 4.5 terabytes it’s it takes time. We need to organize with different type of specialists and, and then we need to build a safe, um, repository in which we will be able to put those, uh, 3 million slides.
[00:15:14] And their metadata make sure that the metadata is a very high quality. It is searchable. So when you want to do your AI model, you make sure that you are doing it on the right slides. And so to do so we are developing also tools that will help. The a AI develop. Some of these tools for example, are Mizer.
[00:15:36] So you, we can take slides from any format and put them in a DICOM format. Is there, we have annotation tools, viewers, and then we are also discussing with the authorities to make sure that up to a point, the AI models that will be built and shared on the database and the repository will be accepted by, uh, authorities.
[00:16:00] So that’s how it is built. So there are six work packages that we are calling. So these are the six group within. The IMI big picture. So the first one work package one is the management group. So this is the one that and I are leading. And then there is the one for the repository. There is the node, the clinical node.
[00:16:22] So these are the, the pathologists that are gathering all these cases. So the, these 3 million slides very important deciding also what metadata is important, which one we will include and not include to be able to do those AI model. And then work packaged for is the, uh, these are the scientists that are doing the different tools and the AI models.
[00:16:45] There will be four use case that we would like to work on. There will be tumor, infiltrating lympho side. There is another case that is, I think it’s a kidney rejection detection model then for non-clinical we, we do an outlier detection model. So normal, less is abnormal. And also, I think the last one is content based image retrieval mm-hmm
[00:17:11] So these are the poor case that the, the work package for will be doing. So to make sure that the repository is a excellent quality so we can test it. And also this will be made available after afterwards, and can serve as a kind of, uh, accelerator for, uh, others to share their models. Uh, work package five is the one that is, uh, more dealing with the regulatory aspect and the legal aspect and the ethical aspect.
[00:17:39] And the work package six is the one that it will be developing our sustainability model.
[00:17:47] Aleksandra: So how far are you in the project now?
[00:17:51] Julie: We are at month, 15 years. So it’s very early on.
[00:17:56] Aleksandra: What have you done so far? I assume there is a lot of organizational work, uh, going on. So I don’t know if you’ve gone beyond that already, but what have you done in this first year?
[00:18:09] Julie: Theary was put in place and tested. So there was some pilot that that was going on. So the first year is where you build the blocks. And after that you put the blocks together to make the house right.
[00:18:23] Aleksandra: Already have slides, or you have just a place where to put them.
[00:18:28] Julie: We have a place where to put them. Okay.
[00:18:30] And, uh, we have a few test cases that we have put in the pilot with mock data, more or less mm-hmm but not a huge amount of them.
[00:18:40] Aleksandra: Okay. So the test cases would be the algorithms. For the
[00:18:43] Julie: models? No, the test case cases are, the test are just test the slides with their metadata. Okay. And we’ve put in, we have started also to discuss the honest broker aspects.
[00:18:56] So who’s gonna be data controller versus contributor. So what’s gonna be the legal base because I think we will act some clinical data in there and also intellectual properties. So it’s important that these legal aspect are discussed ahead of time. So we make sure that it is protected and it is not available to, to, to everyone.
[00:19:21] Jeroen: Maybe it’s good to add in Europe, every country has their own, uh, rules. Of course. So there is a European law, um, GDPR that regulates a lot of issues around personal data. Which counts for all of us, but still every country has their own taste of it. So one mm-hmm I think actually quite big effort is done by, uh, a legal firm that has looked at what are the differences in different European countries who made a big report, which for me as a computer scientist, tough to read, but.
[00:19:56] It contains a lot of data and a lot of information on what can you do with a slide scan slide in Belgium versus France, which could be completely different. Mm-hmm so those are things that are well, not really that sexy, but still it’s important that we solve this because we, yeah. In the app it’s needed to be able to share the slides in the end and to know exactly what we can do and what we can’t do.
[00:20:18] Aleksandra: Yeah. So did you discover something that, for example is gonna. Let some countries take advantage of that and that other countries will not be able to take advantage because of some regulations. Do you already have insights?
[00:20:32] Jeroen: No, I don’t ex expect that to, we are all governed by the same, uh, rules in this, in this area, but there are local tastes to it.
[00:20:41] I think, I don’t think that anyone will benefit more than anyone else. I, I don’t expect it. Mm-hmm but it’s really important to notice. And. Actually also to inform people that are going to upload slides, because many instances, people in pathology labs are also not experts and they often don’t know exactly what, what counts for them.
[00:21:01] Right. It’s, it’s a very complicated landscape. So any help that we can give people is of course, very
[00:21:07] Julie: important. So maybe one thing we did not mention is who was the instigator of this and I, yes. I wanna
[00:21:14] Aleksandra: know that who started this?
[00:21:16] Julie: Yeah. Did. PII is the one who made the proposal. So he’s, it’s it came from, from his idea.
[00:21:24] And I remember several years ago he had the same proposal and it was not accepted. So it looks like we were not yet ready, but when he asked for it, I think he, what year was it? Year one, 2018. When this everything started or 19, 19, I couldn’t. 2019, then people were ready. It’s funny to see it. Sometimes people are just ahead of their time and yeah, totally.
[00:21:53] I think that the European union was not yet ready for, for digital pathology, but now they are. So we are weak . So I have a
[00:22:01] Aleksandra: question because, uh, many of the members of the consortium are international pharma companies. How does that work? Because it’s a European consortium European project, but are you gonna be collecting slides from overseas as well?
[00:22:17] How does, how is that solve?
[00:22:19] Julie: Most of the slides are coming from partners that are in Europe, but several partners, for example, Novas also have counterparts in the us. Some of the slides might come from the us and also we have some of the, the. Help, you know, for example, computational pathology group that are helping scientists and things like that might come also from, from the us, but the it’s better for IMI projects to have most of the contributions that is in the Europe.
[00:22:51] And
[00:22:51] Aleksandra: once it’s done after the six years, Can then people from rest of the world apply and use this data as well. Or it’s also more, you are restricted because I imagine the maintenance of such a huge endeavor and this infrastructure is gonna cost money. How is it gonna be financed? Now it’s a grant, I assume, uh, that there has been some money that was allocated to this.
[00:23:16] It’s gonna be forever, or at least the maintenance is supposed to be long. How is it gonna be financed? And is there gonna be some like maintenance fees? How do you envision that it’s gonna happen and for Europeans and for people from
[00:23:33] Jeroen: outside Europe? I would think that, of course we will not restricted to Europe because we will face the same problems all over the world.
[00:23:40] I think everyone would benefit if it’s much bigger than Europe. And I think also, as you say, I think it’s important question. Of course. How do you sustain such a thing? Do we have a work package as Juli dedicated to that? So there’s one work package that stalled already on day. To think about ways to, to make money or not to make money, but to say, to require money, to keep this alive, they’re actively looking into this.
[00:24:03] We, we can think about even having spinoff companies of the big picture project, and there are many different ways of, of course, of getting the funds to, to keep this alive. And the bigger it is. And the more let’s say countries participate, or the more people participate, the, the bigger chance that we will be able to sustain.
[00:24:21] And I think we do not really have the solution at this point that we, we can’t say what it is, or probably there will be
[00:24:27] Aleksandra: you have another five years to come up with it. Yes, luckily we do, but I’m just, I’m super curious.
[00:24:34] Julie: Plan is due in three years. Okay. Yet year three already. We need to have a plan in place and start to put it in place.
[00:24:42] So in year six, It’s up and running.
[00:24:45] Aleksandra: Okay. And do you also have either in your work packages or in the design of the project some way to adapt to new technologies, which we probably don’t yet know whether it’s gonna be, I don’t know, maybe some ways of more efficient ways of storing data, more efficient ways of, uh, maneuvering data.
[00:25:06] Is there somebody overseeing that if it even is possible to oversee something that we don’t really know what it’s gonna. Sounds complicated to
[00:25:14] Julie: it’s a great question. Yeah, ,
[00:25:16] Jeroen: it’s good to realize that the, the repository is based on existing technology. So the EGA is genomics archive has been in existence for quite some time, and we are say adapting that technology for our infrastructure, the infrastructure.
[00:25:32] So basically what we are developing is a software layer and the underlying technology probably can change. So for instance, we would have different ways of storing our data. That’s still the software that would take, keep track of that and take care of that can probably remain safe or has to be adapted.
[00:25:51] But of course the online hardware for instance, could change. Um, it’s all open source. Everyone could benefit from it. And also everyone could contribute to it at some point. But of course, yeah, it, it’s not easy to take things into account that we don’t know yet. It it’s very, we try to be very adaptable, very open, but I think, yeah, which often is more that we can.
[00:26:11] Aleksandra: So you say it’s open sourced. The tools that you’re building are gonna be open sourced. How do you manage to have the open source and community and open to everyone component with, uh, legal compliance and regulations, different in different countries and another thing. Regulating access to this. Is it something that somebody can take wherever they are and not let you know, uh, that they’re using it or is there gonna be a process to access it and to benefit from this?
[00:26:49] Julie: So, There will be a process. So this is defined by what we call the honest broker mechanism. Mm-hmm they will be the one doing the regulation of who is putting data in and who is taking data out. And, uh, our, who is putting. Algorithm in are always taking algorithm out. Cuz as you, as we have said earlier, it’s they coexist together on the repository.
[00:27:16] In a nutshell, there will be what we call a data access committee. So the, they will be the one deciding who has access or not, depending on the, on the case that you will provide. For example, Alexandra, you would like. Do a model to detect salt WWI hypertrophy in rat liver. So you will write a, a little, you will submit it and then they will revise it.
[00:27:44] Make sure that you are. Legitimate you are not an foster, and then they will provide you with access to the data that you need to be able to app your objective. That is
[00:27:57] Aleksandra: fulfilled mm-hmm so basically like an application process to whoever is gonna be taking care of this. Okay.
[00:28:06] Julie: And we need to make sure that this is not too cumbersome.
[00:28:09] Also, as we want this to be open. Being secure, but not to converse some. So we don’t want it to take six months before you are granted access. Cuz we don’t. We want to be a catalyst, not a break to, uh, to
[00:28:23] Jeroen: development. Yeah. I asked the exact same question that you just asked. Um, even before we started the project to one of the people in the, in the structure group.
[00:28:32] And I said, how is it possible to have opensource software and still know that it’s secure? It feels as if there is a contradiction. And he said, well, basically, because open source, everyone in the world can try to hack it and can have a look at it. And all the vulnerabilities will just come out because so many people have access to it.
[00:28:51] Yeah. And he, he gave the example of Linux, which is completely open source and is used by many companies for their server software. So actually having it open source means that so many people can have look at it and try to find weaknesses in it. It’s the best guarantee to get very strong, secure software.
[00:29:10] It feels like a contradiction, but it’s actually a very good way of doing it. I
[00:29:13] Aleksandra: think. Yeah, totally like self improvement of a product because the users can contribute to, to improving it. Yeah. This is amazing. So guys, and our question, where are you gonna have all those slides? And now the discussion is for every organization that is going digital.
[00:29:33] What are we doing with those slides? Hard drives is not an option. Can we have a big server? Can we have clouds? Like where are they gonna. This is gonna be a huge amount of data. If you can say, I don’t know. That’s a good question. So we have two, but it’s an open project I’m asking those questions.
[00:29:53] Julie: Yeah, sure.
[00:29:54] So there will be two server. There will be mirrored. So to make sure that it’s a kind of security and a recovery security, if something happens to one of the two and there will be one in Finland and one in Sweden. So
[00:30:07] Jeroen: the, the basic setup is that we will have, uh, two copies of all the data as Julie. But the repository will be built.
[00:30:15] The technology will be such that it can also be a federated system. So in the future, we could also have notes that keep their own data, but are still connected to big picture and are accessible through big picture. And if I’m correct, I think in Finland, they have really a huge bunker in which they will store all the data in Sweden.
[00:30:35] It’s a local cloud solution, but so there are, there is slightly different technologies behind it. Luckily, we have people that take care of that and we don’t have to why the hard risks ourselves.
[00:30:48] Aleksandra: Oh yeah. You would have to buy enough hard dress. No, that’s obviously not an option, but I’m just fascinated because it is like people think, oh, digital not physical anymore, but at some point it becomes physical because you are storing this digital information on hard.
[00:31:09] and this hardware has to get bigger and I don’t know, larger, more secure, double, because you wanna back up because it’s digital, it can disappear. You don’t want it to disappear. No totally crazy initiative. And that leads me to another question. So there are many packages, six years, many people involved, but you all have, uh, your other jobs.
[00:31:34] How are you managing this project on a daily basis? Because do you have people that are, full- employed by this consortium and by EMI or you guys work for your jobs? So how does
[00:31:46] Julie: it happen for myself? Probably past 30% of my time on the IMI big. And in navs, because we are involved in a lot of different IMI project.
[00:32:00] There are people that are, their job is just IMI just to do that. Okay. Yes. And we have some FPF partners that have, for example, some FTEs that are dedicated to IMI as well. And some are just for IMI big picture. We have
[00:32:16] Jeroen: a company, uh, Ture that is a, let’s say professional project management company. So they, they have a few people.
[00:32:23] Probably not completely full time, but redated people that spent a lot of their time on managing the project. And in addition, we have a lot of people with different partners that spent either their entire time or part of their time on this. So we have, for instance, a communication manager, someone that works one and a half days a week just for communication.
[00:32:44] So yeah, we have a lot of people it’s a, as Judy said, a 32 million Euro funded project. So we have a lot of people that work for the, for the project.
[00:32:54] Aleksandra: And does the funding come, uh, partially from, from the, the members of the consortium and partially from the European union. Or how does that work? So Julie, who’s paying for your
[00:33:07] Julie: 30%.
[00:33:08] So usually the IMI project, the way they are funded, alphabet is coming from the pharmaceuticals. And the other app is come, is coming from the European commission. So it’s a kind of, uh, the matching, the European commission will match the amount of Inkin contribution that the pharmaceutical will be providing.
[00:33:30] So us, our in kind for my big picture is in terms of slides and also people working on the project and then up to a 35 million and then 35 million was provided by the European commission to pay for the private, the private partners that are involved in the project. So the 36 other partners involved in the project.
[00:33:55] Aleksandra: And how does that work for academia? Yeah, so
[00:33:57] Jeroen: we get funded. So we, different parties have different roles in the project, which has been defined up front. Every part is for instance, responsible for certain deliverables, for certain tasks and we get funded. So we get different amounts for, for hiring people for I myself, I am funded one day a week for this project and actually that we have a budget
[00:34:18] Aleksandra: for, do you only work one day a week or do you work more.
[00:34:21] Jeroen: That I work one day a week for big picture officially. Um, of course I don’t track the time that that, but officially I’m paid for one day a week for big picture. And actually we have a budget for slides, portals providing slides can get reimbursed for the cost they make. So they may have to scan. They may have to hire people to upload slides, to find metadata.
[00:34:41] So we have some budgets, uh, for
[00:34:44] Aleksandra: data. Okay. So at the moment we said we are 15 months in. Let’s say it’s end of the planning, end of the organization, a little bit of proof of concepts on different, um, aspects. What do you expect to have done by next year? Because I’m gonna invite you again next year and we’re gonna be talking.
[00:35:05] Okay. What did you do, guys? What did go well, what didn’t go so well, what’s the plan to have accomplished. Months 24.
[00:35:13] Julie: So we have a big milestone coming up, which is the first clinical data that will be put in the database. This is one of the, the big milestone that it is coming up, the tangible milestone, cuz we have performed a lot of things in the first year, but it’s not the things that people see.
[00:35:32] Mm-hmm it’s a little bit like when you are doing renovation and they are a lot of things you do behind the walls, but nobody sees. So this is the first real data in the real database, not in the pilot. So it’s, we will be learning a lot from that. And at one 30 is when the non-clinical are uploading the data.
[00:35:52] The clinical group are our Guinea pigs and we will be the second in line, which is great. Cuz then they will have paved away. So it’s gonna be way easier for us. So this is coming for month 30. Okay.
[00:36:09] Aleksandra: This is exciting. So I’m gonna be inviting you next time when this milestone is met, or maybe before that.
[00:36:17] Thank you so much for taking this time and explaining this. I think as the project evolves, it’s gonna be. Maybe clear and more tangible, like you said, Julie, for people to understand, okay, what’s happening. There is software. There is slides. There is access. There is like so many things. So I’m super curious how it’s gonna evolve and thank you so much for joining me today.
[00:36:43] Julie: Thank you. I think you will need to re invite us probably.
[00:36:45] Aleksandra: I will. No worries. No worries. You have the invite for next year. No problem.
