The Power of Spatially Resolved Staining

Good afternoon, everybody. Welcome to the Life Science Symposium Series. I am Fiona Smith from Leica Biosystems, Marketing Manager for UK and Ireland. I'm here today with my colleague Gareth.

Hi, I'm Gareth and I am the Life Science Account Manager for the UK and Ireland for Leica Biosystems and have the pleasure to be joined by Eleanor and Brad today from GSK who will be doing a presentation on digital pathology and multiplexing. I'd like to now hand over to Brad and Eleanor.

Hi, my name is Elena Miranda and I'm the director for the non-clinical histology team at GSK in Stevenage.

Hi, I'm Bradley Spencer Dean. I'm a subject matter expert in next generation histology in non-clinical histology department in Stevenage.

Okay, so great to meet you both and thank you for coming today. I'd just be interested to know, firstly, the focus of your work currently.

Yeah. In my role, I lead the team. I need to focus on making sure that the team has everything to be healthy and productive and deliver the experiment and make sure that we meet all the timelines for the different projects.

My part of my remit is to introduce new technologies and to address new questions or improve workflows and then to get those introduced into the repertoire of what the lab does.

And what are the current drivers and the changes that have brought about the focus of your teams at GSK?

There are both internal and external drivers. We, the goal of GSK is to deliver medicine to patients. We need to make sure that we have, we respond timely to different projects and we cover different medicines and different targets. Sometimes there are external drivers as well. We have all seen with COVID. That was a big one. Many pharmaceutical companies respond to that. The external drivers depend on how severe the disease is and how we are sure that we can help the patients at the end.

The vast majority of what we deal with in the lab is tissue-based research and tissue-based samples. We're always trying to get the most information from those samples. Sometimes we don't have much choice in the way they're fixed or whether they're frozen or FFPE. Part of the work I'll be talking about is how to maximize information you can get from a limited amount of starting material. That will tie into reducing timelines and improving the whole efficiency of what we're producing. They'll then go on to, say, pathologists to report.

Thanks, Brad. It's interesting to hear about some of the challenges there with your formal and fixed paraffin embedded tissue. I'd be interested to hear about how you decide on innovation projects and how you choose and what to focus on.

Much of the time we're being asked to detect proteins by immunohistochemistry primarily. Often we won't have an antibody that's been fully optimized or we'll need to combine that with something else. In a situation like that, say for a secreted protein or a long, long coding RNA or something where there's just no good IHC optimized antibody, we need to bring in alternative technologies and maybe look at messenger RNA as complementary staining modalities so that we can really maximize the data we can generate from a small amount of material as quickly as possible.

That's interesting. And how does the digital fit in with your work? Do you work together?

Yes, we are creating with the use of all these scanners massive amount of digital files. And so, these digital files are data. Data is the new golden mine for everything that we do. We can see that in all aspects of our lives and that is replicated in the lab as well. It is a very complex project because we need the input from IT data, metadata expert bioinformaticians. It's very complicated to try to work on one single project altogether. There are different aspects in the lab that now are becoming digital, more automated. This is part of the project that then goes in the digital histopathology project.

That's great. That's interesting. I'm looking forward now to hearing the presentations from you. 

One recent set of experiments that I've been working on has been really to incorporate a new co-detection assay whereby on formalin-fixed paraffin-embedded tissues we can detect both an mRNA target and then also sequentially an immunohistochemistry protein target as well. So it's not multiplexing in the sense of looking at dozens and dozens of targets, but it's allowing us to really combine immunohistochemistry and in-situ hybridization together and try and maximize the information that you can generate from a limited set of tissues. This is also an assay that's available on the Leica BOND RX or manually. And that's why ideally, we'll always try and incorporate assays that can be run in an automated, high throughput, reproducible way, preferably overnight as well. That is a much more efficient use of our time and helps with reproducibility.

That's another big benefit of this assay. So, the primary objectives of this work were to set up and establish the assay and to look for different mRNA targets, I'll come on to those towards the end, and then protein cell markers. So in this instance, I'm going to demonstrate detection of mRNA chromogenically and then use antibodies to come in and detect various different cell lineages so that you can assign or the pathologist can assign the RNA signal to immune cells or tumor versus stroma, that kind of thing. The ideal way would be to run this fully or semi-automated overnight and then that saves technician time. The workflow is based on RNAscope.

So, this is an assay from ACD, now Bio-Techne. And really to take you through the workflow, the tissue section is that purple square on the slide, or you could have cells that are cultured adherent cells or cell pellets. But in any case, it needs to be on a positively charged glass slide. This is formalin fixed material. And then the first thing we must do is to break those formalin cross-links and then permeabilize the tissue in some way to enable the next set of reagents to really bind and do their stuff.

Hybridization, so this is where we're looking at complementary probes that are specifically going to bind to your RNA target of choice. And these are designed in a special way. These are double Z probes that I'll come on to a little later. And then the way the assay works is that once those probes are bound specifically, an amplification tree is built up that gets decorated with, in this case, alkaline phosphatase. And then the signal is then detected with fast red. So, you have a nice chromogenic signal. And then that's visualized typically in this punctate manner. So, you see this kind of spotty signal. And then another beauty of this, it's not just a qualitative result, but it's fully quantitative with different image analysis software.

We use Halo a lot in the lab. But the idea is that you get that opportunity to get quantifiable data as well. So just very briefly about what makes the assay special. So, RNAscope's been around now for over 10 years and really the beauty is that the probes are designed with this double Z structure so that the bottom part of the Z is 18 to 25 bases long and that's fully complementary to your mRNA target. And then there's a link here and at the top there's half of a binding site for preamplifiers. And the idea is that you must have two of these Zs bind next to each other, giving you about 50 bases of specific binding, creates this landing site, and that allows that amplification tree to build up. And then typically, you'd be looking at about 1,000 nucleotides worth of specific binding, so 20 of these double Zs, and that's where you get this real amplification of signal. So that's the principle of the RNA side of things.

And the other thing I wanted to mention is that there are different flavors of this. So typically, if you're looking for a target of choice that is nicely expressed, it's well fixed, the mRNA target is not highly homologous to other family members, so a nice sequence probe can be made. That's where you would use RNAscope. Here you're looking at approximately 20 pairs. Another flavor, if you like, or another variant is base scope. And this is where you're looking at much more difficult targets or things like point mutations or very highly homologous sequences where you've got a much more limited region of specificity to use. And then finally, mRNAscope. And this is where you'd be looking at microRNAs, antisense oligonucleotides, short hairpin RNAs. And the reason I mention these is really because we've had the opportunity and the now success in using all three of these variants of in-situ hybridization in combination with this co-detection assay.

I'm going to show you some examples soon of different assays, these assays in combination with protein markers. It's just really to show you the range of targets you can combine. All of these are automated, so that makes it great for us. Traditionally, those of you that are familiar with these assays, until fairly recently, if you were going to incorporate a sequential workflow, it was always the same that the RNAscope, the RNA in situ hybridization step always had to come first and had to be done to completion. And that would then be followed by immunohistochemistry on the premise that the RNA is a much more delicate target and is more likely to get degraded. And therefore, the protein detection would come after that. The problem with that is that as part of the assay, there's always this kind of dual target retrieval. There's a heated target retrieval in a basic solution. And then that's followed typically with a protease digestion step. And that will have a negative impact on any immunohistochemistry that you want to do. It's going to damage some epitopes, and some antibodies you might want to use are just not going to be up to the job. And you're not going to, you know, so sometimes this would work, but a lot of the time the antibody staining wasn't good enough.

The new assay that's come in, this co-detection assay that it's now called the Integrated Co-detection Workflow, ICW, changes things up a little bit. And what it really does is it moves the primary antibody incubation step higher up the workflow. So, you'll still start with the same, your bacon dewax section, so these are paraffin sections, but then the heated target retrieval step will occur. So that's broken as formalin cross-links. But now you'll put the primary antibody on. After that, you have a chance to bind, you'll cross-link that with neutral buffered formalin, and then you'd incorporate this requisite protease digestion step to really enable the in-situ hybridization workflow to occur and probes to bind, et cetera. And what we found is that that's very protective for a lot of those bound primary antibodies, such that when you've completed the in-situ hybridization step, and you begin the detection aspect of the immunohistochemistry with a different chromogen, it works very well. And certainly, all the antibodies we've tried so far in the lab have worked very nicely. And so that's how this workflow has really been brought in. And there's a couple of prerequisites. So, it's strongly recommended that you use the ACD co-detection antibody diluent. that seems to have been validated specifically for this assay, so we stick to that. And then typically, we would still always pre-validate and optimize any antibody we wanted to use for immunohistochemistry on that tissue fixed in that way so that we've already got a good starting point.

I'm going to show you a couple of examples now of the kind of staining you can achieve. One thing I should say is certainly for the chromogenic co-detection assay. The RNA is always going to be detected with alkaline phosphatase and is always going to be decorated with this fast red chromogen. And then the choices have been a bit more limited in what you could detect the antibody with. So that's really where I guess where Leica have come in because recently, they've brought out their green HRP chromogen, which was something we really wanted to try, to give us this nice contrast with the red, because even though you could use dab brown to detect the antibody, red and brown are not always the best contrast. So red and green we thought would be great. And then as you can see, hopefully from here, what we've been able to do is take, in this case, formalin-fixed paraffin embedded colon tissue. The red staining, you can see this red punctate staining is the mRNA. In this case, it's a housekeeping gene, PPIB. So, we would use this typically to pre-qualify tissue that the mRNA quality is suitable for any downstream targets we want to use. So, this would often be done as a first screen of tissue to make sure it's good enough so that you don't get any false negatives. And then I've combined this with an antibody to CD3 to pick up T-cells, in green. And then you've got hematoxylin counterstain. So, in this situation, you'd be able to detect which cells are expressing both the mRNA target and then would be CD3 sort of T-cells. That would be the principle.

Another example, and this is something that we've used a little bit more often, hereby, so again, the example is with PPIB as the red mRNA target. But in this case, we wanted to check, we wanted to really detect tumor versus stroma. So, we've used a pan cytokeratin marker, CAM 5.2, but certainly you can use others to really decorate and highlight the tumor epithelium in green and leave the stroma alone. So, in this case, you could look at your target of choice and see whether it's tumor associated or not, for example. And then certainly we found that the contrast, you can see that the green color, this HRP green chromogen is good. It's very robust and contrasts very, very well. And I should also say that for the image analysis side of things, those colors contrast very well and can be very easily detected and segmented. That was an assay that had already started to come online and was developed maybe a few months ago. But something that's a lot more recent is this micro RNAscope assay. So, you're still detecting an RNA. You're still detecting it with alkaline phosphatase and fast red chromogen. But in this case, your targets are not nice juicy mRNA transcripts that are beautifully fixed and available.

Here you're looking at a much more difficult target. So, these could be endogenous microRNAs. These could be antisense oligonucleotides or siRNA targets. Things that traditionally were very difficult to detect and certainly very difficult to detect in concert with anything else. And that's why we were really delighted to give this a go and see if we could see some of these targets in conjunction with a protein cell lineage marker. So, the assay is very similar. I've managed to get it working on both formalin fixed paraffin embedded. So, this is the blue side, but I've also shown it also works very well in this sort of pink side for fresh frozen as well. So, if that's all you're able to get, fresh frozen tissue, the assay works very well as well. Slightly different premise, so there's no heated target retrieval in that case. And all the target retrieval, if you like, is done by protease digestion. But the assay, as you say, as you can see there, takes about 10 hours. So typically, we would set this up, you know, 4 or 5 o'clock in the afternoon and then come in the next day. And then, you know, most of the assay has been completed. The one part that is not fully automated yet, but certainly something that we know that the new software that's available that we're hoping to have installed soon will allow us to do all of this in a fully automated way is the detection step with the green chromogen.

Here you've got that same workflow for formalin fixed paraffin embedded tissues. You've, again, you've got your antibody binding and then cross-linked after that initial target retrieval, your hybridization of your specific probe, and then you've ended up detecting that with the fast red, with the red refined detection system. And then instead of just finishing it with a hematoxylin, we create a sequential set of a secondary staining protocol where you have the peroxide post-primary polymers that are on the refined kit. So, this is the HLP refined kit that comes from Leica. And then instead of using the DAB that comes with that kit, we would effectively take the slides off at that point. So that's what we would take off in the morning, rinse them in water, and then make up that green chromogen, incubate it for about 10 to 15 minutes, and then blue in tap water, well, counter stain it in 50% Gill’s hematoxylin, blue it in tap water. And then because we're using fast red that's alcohol soluble, we need to use a vector mount or eco mount as a mountain. So just to show you what some of those look like, so previously on the earliest samples I've shown you, we used RNAscope, and it's very punctate. Typically with micro RNAscope, it can be a very strong signal. So again, this is RNU6. So again, this is a housekeeping positive control. And this is really to establish that this tissue, whatever it is, however it was fixed, however it was processed, is suitable for any downstream analysis. So, we would use this housekeeping long coding RNA to establish really that the signal is good. And in this case, every nucleus is staining. That's what we would expect, what we would hope for. And that tells me that the tissue is good to go, that any signal I see afterwards with this sample is believable. You're not going to get false negatives.

This is to demonstrate co-detection. And this was something that we wanted to establish was whether we could co-detect our target of interest in formula fixed mouse kidney and establish whether the signal was in proximal tubular epithelium primarily or only as opposed to in the glomerulus or the distal. And this was using an antibody called LRP2 or Megalin. And again, this antibody, we'd already been established in the lab and validated immunohistochemistry, but I was able to combine in the sequential manner a micro RNAscope signal in red and then sequentially with green. So, this example just shows you the positive control, but certainly we're hoping to combine this now and we're in the process of demonstrating our target of interest with this cell marker.

And just to demonstrate how sensitive it is. So even though primarily in that assay I wanted to see proximal tubular epithelium, certainly the antibodies and the assay are sensitive enough to pick up specific cell lineages and targets that are much lower expressed. You can just see, this is taken from this paper, that LRP2 is expressed at very low levels in certain immune cells. And this is in the mouse spleen. You could see that it's being picked up very nicely. That's just to demonstrate a very new assay and then combining that with this co-detection to maximize what we can get from that tissue that gives you a nice definitive answer that can be easily read by a pathologist or research scientist and then fully quantified and used in image analysis as well. We feel very confident that the assay works well. It's been established within the lab. It's fully transferable between labs, anywhere we have a Leica BOND RX or an RXm. I've demonstrated targets that are housekeeping controls and really used for RNA integrity as examples, but certainly we've had a lot of success with our specific targets as well, both RNAscope, base scope, and now with our micro RNAscope as well.

This combination of the Leica HRP green, chromogen, and fast red works extremely well with hematoxylin. They don't tend to mask each other. And it works well with image analysis software for quantification. And we're very aware that the newest version of the software would allow this whole process to be done fully in a fully automated manner. So that's something we're very much looking forward to incorporating into the lab. We're working on seeing how well this works on tissue and samples that are prepared in other ways. So certainly, fresh frozen and fixed frozen as well. And, some of you may be aware that these assays, that the co-detection assays are also available in a fully fluorescent option. And that's something we've certainly had already now initial success in the lab as well, whereby you'd be detecting your mRNA target with one fluorophore and then coming in and detecting your antibodies of choice and your protein targets of choice with other fluorophores. So, the entire thing is then fluorescent rather than chromogenic. So again, that's also done on the Leica in a fully automated manner. So that's something else that we're working on. Just to confirm really that all the human and animal tissues that have been used with full and all the appropriate ethics and human tissue act. Thank you very much for your attention.

In this presentation I will talk a little bit how we are deploying digital histopathology in the non-clinical histology lab. In GSK, digital histopathology has the potential to improve our ability to discover disease mechanisms and to identify and categorize patients in different diseases and to improve their, predict their disease outcome and to create more targeted therapies. Before the diagnosis was based on the pathologist looking down to the microscope, checking the slides, and deciding if it was a tumor, if it was autoinflammatory disease, all this kind of thing. Now, the diagnosis is really becoming a complex mechanism. There is multiple pieces of information that the clinician puts together to give the patient the right diagnosis and the right therapies.

Digital histopathology and histopathology in general started a long time ago. H&E seems, seems very strange, but H&E was first introduced 150 years ago and the first was introduced 100 years ago. So that shows a little bit how we are based all our knowledge on something that was created really a long time ago and remains stable for all this period. But something that really changed is how we are approaching and how we are visualizing and the data that we are getting from all these slides. So, the terms artificial intelligence was first used in 1956. And from 1956 to, let's say, 1990s, people started approaching microscopy and taking images in a slightly different way. So, they started putting camera on microscope and it was possible to get these images. And there were other terms related to artificial intelligence or image analysis that were used in those years. So, for example, convolutional network, neural network or deep learning were also invented in this period as well. But it's only from when we had in 1990 the first digital scanner that we could really implement and improve what we could do in artificial intelligence and image analysis. For the last 20, 30 years, then it was just exponential to grow.

We have different kind of scanners. We have scanners that are quicker, high throughput scanners. What is important then for the patient is that the regulatory agency, they approve something that we do with artificial intelligence and this happened two, three years ago with the FDA that approved the first artificial intelligence tools for medical diagnostic. The digitalization of rolled slide images streamlines the pathology workflow. It doesn't matter where the slide is created, it doesn't matter where the pathology is, you just need a screen and a scanner, and you can do your work. There are all these data that are collected in libraries. So, all these are sorts of information that then we can use to then characterize the patient and create diagnostic tools. We can apply artificial intelligence to all these images and determine information that can be used then for the patients.

With all this exponential growth, it's normal that all the investments are increased exponentially in the last period. Healthcare artificial intelligence projects are growing greater than any other sector in industry, more than mobile networking or anything else. In the last 18 months, we had 100 million invested in startups that are doing something related to pathology, artificial intelligence. And in 2019, the UK government invested 60 million in the UK Innovate Initiative that was to develop digital histopathology in the NHS. In 2018, there were the steam was like that 2 billion were invested in artificial intelligence project. This is going through triplicate in 2023 and triplicate again probably in 2025, and it might be even higher than that. With all this investment, it is normal that there are any venues from everything that is digital is pathology related are just growing exponentially in the last two, three years.

Let's see a little bit how the digital histopathology workflow works in the pharma industry. We have modern digital scanners, so the quality of the slides is better. The number of slides that we can scan is higher and we can do it faster and automatically. There are more tools for artificial intelligence or image analysis that are powerful and more user friendly. There is a change in the culture as well, because the more we can see the benefits, the more we might use these tools. The regulatory barrier is going down. At the beginning it was difficult to get something approved, but once that we have the first tool that was approved, now it will be a little bit easier. There is a challenge in business environment because with all these new tools, we need to use them and there is the pressure to use them.

The advantages of using digital histopathology are that reduce the disruption of home and work-related issues. It doesn't matter where you create the slides and where pathology is because you just need a video and you can do your own work. It is easier to connect with different pathologies, having collaboration and having the subject matter expert involved in the diagnosis as well. And then the two major things, there is a very high focus on quantitative data. It's not only the quantitative, qualitative, but the semi-quantitative, but it's quantitative down to the single cells. There is improved productivity and efficiency because there is automation and more investment.

As in everything, there are also limitations as well. The images that we create are very, very big. So, the amount of data that we produce and the amount of storage space that we need is massive. That's 99% of the time I see the cost of these scanners that we use. It is expensive. The training set of slides required to train the algorithm is becoming bigger and more complex. And as in all the aspects of the medicine development, when you have something that is rare, it becomes more difficult to have enough slides on a particular rare disease to train your algorithm to recognize that disease. Quick adaptation is required and it is a different work for pathologists. So, people that have learned their job on glass slides, it becomes difficult to adapt to digital files. But this shift is progressing. So, we are moving forward. With all the situation with COVID, all the last 18 months have been training virtually. So, the new generation of pathology will be just used to that.

In GSK, we have been engaged with digital histopathology project since 2004. It was normal at some point we were going to arrive in a situation where we were going to have a big project on this to standardize it. And in order to standardize a workflow, you need to 1st identify which kind of data or metadata you want to collect, how you are going to collect them, and then you're going to collect all the qualitative data that you produce from these images, and you need to make sure that everything that you are collecting as a sort of library, you can reuse it. The workflow starts from really the origin, so it's like there are study metadata, there are protocol metadata, there are animal metadata, and then there is what we do in the Histology Lab. So, the sample metadata that we are going to collect, and then it goes to the scanner, so the digital scanner metadata and the image analysis metadata. The way that I imagine this is like a train. You start in a certain point with just a locomotive and then you attach into the different station, different coach. These coaches are just bringing new data that are then will be associated with your digital image, with your digital file. All this train and all these image files are put together in a library and somebody in the future will just go back and search for a specific characteristic and then this person will be able to see all the digital files and all the data associated with that file. The idea is to have this sort of libraries for both internal and external images. So that is at a different level of complexity because we need to work also with our CRO partners to get all this solved.

For our digital histopathology workflow, we use this software called Prima. It is produced by a company, an American company called Fortelinia. It allows us to have this workflow in the lab. So we are at the little bit, we are starting with the software. There is not a full workflow that is all automatic. So, the initial capture of all the data coming from the study and the animals is done through a CSV file that is imported in the system. Everything that we do in the lab, so starting from the jar creation, the cassette, the slides and the stain slides is governed by a barcode. So, we just need to read this barcode, and all this barcode has all the information that we have on this tissue, on these organs and on the stain as well. The barcode is also read by the scanner and then there is software that puts all this metadata together with the digital file and then we collect this digital file in a long-term storage. These are just part of the equipment that we have in the lab and that we use in this workflow. We have cassette and slide printers, but don’t worry if you don't have them because you can have a label printer as well and you can just attach your paper label to your glass slides. We have a combination of tablets and computers or many PCs. There are people that prefer to use a tablet because there is a little bit of mobility and another advantage of using tablet is that if you have limited spaces, it's a little bit easier. I will talk about this switch box in a second. Then we have a slide printer again. We have used different barcode readers and we have found that what we were thinking, so for example, that the fact that sometimes cassettes are covered by wax is not really a problem.

The other point that is important is that you need to have a scanner that suits your purpose. What is important to do is that your LIMS system, so for example, in our case, Prima, is a scanner agnostic, because we have the P250 from 3D stack, but there are other companies that have Aperio or the NanoZoomer. In that case, we want to make sure that our system can have all the slides from all the images from the different scanners. So, it doesn't matter which kind of scanner you use. The Prima software works in a way that is similar to, for example, the Leica BOND. For example, before you writing what you're going to do in terms of your slides, on the Leica BOND, you register your reagents, you set up your protocols, and here it's similar. So, you have a control panel where you standardize your tissue, your organs, what you're going to do in a specific protocols, and then you just apply it to your slides.

There are other different modules, one is the workstation. So, there is the micro workstation, there is the microscopy workstation, and the other modules are one for the pathologist, the lab manager and the label creator. So, you can create your own template for the labels. This is a representation very like the software that we use in the Prima control panel. We add the tissue type, the protocol, so how the blocks will be organized, which kind of block we collect in a specific tissue. Registering data regarding the equipment is also important because you want to know which kind of model you have used, which kind of situation your equipment was in, especially if you're working in a regulatory environment. The imaginary naming folder is the landing zone where all your images are going. And the CSV is here just a representation of what we use in terms of is like an Excel file to introduce all our data in the system.

Our experience with this software is based on the last two years. You need to have IT involvement from the beginning because it is fundamental. You might think that you know your equipment, but there are connection, network, firewall, all the possible thing and you really need to have IT involved. It is better to have all the equipment already in place. One recommendation that my team had is don’t train virtually, but that for us was inevitable because we started all this during COVID, so it was not possible to do it in a different way. You need to use a sort of like decide where you want to use your LIMS system, if it's just toxicology or discovery. We have both and something that really helps us.

We started with the phase system and regulatory tox studies first. This is the reason why we use this printer switch boxes because as not everything that we were doing in the lab was going through Prima, we needed to switch from a normal not Prima system to Prima system and these boxes are doing exactly the same. It's up to you to decide the tablets and the other PCs that if you have space, if you want to have the flexibility, the super user. So, we divided, well, we selected few people in the team that became super user, and they have a role to test the first all the new version of the LIMS system and then to train other people in the team. It is very important that they have very high computer skills because to make sure that they can address all the problems that are important.

Another thing that is critical is your barcode quality because barcode readers are basing all the information on the barcode quality. A point that is important to stress is you need a file renaming convention because if not it becomes difficult to search for things. So the way that we approach this is we try to identify which kind of data were important to read in the file name to make sure that the pathologist or the histology that were doing the analysis knew exactly what they are going to analyze. So, we decided to use this convention to insert the study number first, the animal number, the tissue number, the style name, and eventually the level that we were going to section. You can select the one that makes more sense for you, but it needs to make sure that is something that everybody can understand.

Another key point, and we had a lot of discussion about this, is the problem with vocabularies. If you have only one company or one lab based in one single place, that is easy. But GSK is a global company, and we have labs both in the UK and the US. So, we had problems with British English versus American English, different names used in different places. So, you can call it colon, you can call it large intestine, you never know. We were the problem because we were also trying to work with the CROs. So, the CROs has a different way of recording tissue and organs and species and all this kind of thing. There is always the possibility that there are type errors, especially in username, because we were recording that. So, all this is something that you need to take into consideration when you do, when you approach this kind of list system.

Another thing that you need to consider is the image quality. I think that everybody has seen overexposed images, out-of-focus images, images that somehow you cannot quantify because there is a problem, or the contrasting is too strong and all this kind of thing. So, what we are trying to do is trying to automate image analysis check to get this information before the images go to the final analysis. And another thing that we need to do, especially for H&E and stain slides, is use color calibration. If I think that everybody has already tried to move to transfer protocol from one side to the other one, and they are never the same. The color can always change and to have an analysis that can really quantify and be specific and you apply the same algorithm to all the images created in different places, you need to have all the colors together.

There are challenges in establishing a digital histopathology workflow and the first one is the IT infrastructure. Sometimes it is outdated. The problem is that there are firewalls, security checks that sometimes are a little bit difficult because especially you come from a scientific background, not from an IT background. The metadata is important. I hope that I have shown you how they can be different even if two different people mean the same thing. They need to be standardized, and they need to be consistent. You need to have a very good relationship with your CRO partners because then all the slides that they are producing, they need to go through the same workflow. There is the problem with the accreditation for GLP, GCP. We don't have that problem, but when you use clinical sample, that is something that you need to consider and the user experience because it needs to be the same as glass slides or even better if not. Why the user needs to use that.

There are future challenges as well. The main one is that the regulatory agency, they need to accept the data generated from these tools. And we need to make sure that the trail of a collection of metadata and analysis is consistent. Another one is that you need to have always the right equipment and the data storage space, as I mentioned before. For example, this is the CIFAR-10, a database that collects 60,000 different images from different things. It can be dog, it can be roads, bridge, whatever. They are the size of 61 million pixels. This is not even half of one of the digital pixels that are included in one of our whole slides sections. This is also another database with images. It has been calculated that the number of pixels contained in ImageNet database, they are equivalent to 474 old slides that you can scan normally in histopathology lab. I think that we can scan 474 slides in two or three days easily. When we create thousands of these slides, the data storage becomes a very massive issue.

In summary, I hope that I show you that the digital images workflow is really at the renaissance of the pharma industry. There are a lot of industries that are using it now. A digital workflow makes histopathology easier and faster and then can improve the data quality and the type of data that we can produce. Image analysis tools or artificial intelligence are potentially very useful, and they are driving the scientific robustness and the objective of the decision making that we do for our drug discovery. The quantitative data needs to be verified and they are increasing their potential in the digital in the drug industry and there are a lot of challenges that you need to face if you want to establish a fit for post-digital histopathology workflow. This is not the work on one single group, but the work of many groups together. 

I would like to thank the, first of all, the Prima super user in the non-clinical histology team, the technical team together with their team, the data team that help us with the vocabularies and the definition of the renaming and pathology and safety as well. I didn't show any animal sample, but all the ones that we are using are ethical review and they are in accordance with all the regulations. Thank you for your attention and I'm happy to take any questions now.