Spatial Biomarker Panels for Clinical Trials

Hello and thank you for attending today's webinar, Spatial Biomarker Panels for Clinical Trials Presented by RareCyte, IO360 and Fierce Pharma.

My name is Janelle Karimoto, and I'll be your moderator. Before we begin, here are a few quick housekeeping items. To learn more about our speakers, please visit the Speaker Bio Window. If you accidentally close the window, you can reopen it by clicking its name in the top navigation bar. Additional resources can be found on the Handouts window.

This webinar is being recorded and will be available on demand within 24 hours. And finally, we will host a live Q&A session at the end of the webinar. Feel free to submit your questions at any time using the Submit questions window.

It is now my pleasure to introduce our speakers. Joining us today is Dr. Jennifer Bordeaux, Associate Director, Digital Pathology Solutions at Navigate Bio Pharma Services and Dr. Tad George, Senior Vice President of Biology, R&D at RareCyte.

All right, ready to start. Tad, over to you.

Tad -

Thanks, Janelle, and thanks for joining the webinar. We're looking forward to discussing really application of spatial biology kind of in the downstream space, particularly around Orion, which is a technology we've developed at rare sight that is deployed at, at in, at, you know, in CROs, including at Jen's Navigate Biopharma.

So really Orion has been designed, you know, for that, you know, translating spatial biomarkers into clinical value.

And so if we if we think about modern therapies, you know, like checkpoint inhibitors or things that are trying to modulate immune response or even cell based therapies or organ transplantation, a lot of them are designed to alter tissue microenvironments by recruiting kind of the right cells in the right state improve patient outcomes.

But measuring these alterations with confidence requires a couple things that are typically in conflict with one other, you know, biomarker panels that are large enough to resolve those tissue microenvironments, but also with sample throughput that allows you to analyze, you know, clinical trial level large sample cohorts in order to get statistical power right.

And really prior to Orion, investigators have really been forced to make a compromise choose, you know, one or the other of those. And so in a lot of the ways that people apply, you know, you know, protein based testing can fall into two, you know, broad camps that are at different extremes in terms of plex and throughput. A lot of ways to get more information is to immunofluorescence where you do multiple markers in around, but because you typically need lots of markers that in most traditional technologies will cycle that immunofluorescence. That gives you a tremendous amount of useful information, but it doesn't give you enough throughput for statistical significance.

There's lots of low flex technologies, IHC for example, where you get lots of statistics, you can get through large number of samples, but you don't have enough information to resolve those microenvironments. That was really prior to Orion. So you know, Orion was really built for that clinical translational zone where it gives you that essential information to resolve those microenvironments with the throughput that's required for those large trials. And that space is kind of, you know, in the 10 plus space. I'm sure Jen can talk a lot about that because you have a lot of clients that are, they're asking you for these types of assays, but need at minimum 30 samples and sometimes hundreds, right?

So that's really what Orion was built for the really the technological breakthrough was the, the ability of the system to in one single staining and imaging round address 20 channels of information. So it's 20 channels staining and imaging that gives you the speed necessary for the throughput. It also has really ultra-low run cost because it's very simple and it gives you high data quality for many reasons, but a lot of your tissues preserved it from a fundamental level. It's a whole specimen imaging device, so you don't need to know where to scan prior to it will scan the entire slide, your complete spatial context.

There's flexible panel design for, you know, pretty much any kind of panel for any kind of indication the reagent infrastructure, really you can combine all of the reagents in any, any combination that you need for to address your microenvironment you're measuring.

And really, with the recent release of the Orion HT device shown here, we've now brought high capacity 30 slide automation to spatial biology for those large cohort studies in, in multi user environments, you know, general talk about, I think your tumor landscape panel, you use it for multiple tissue types, really any, any tissue, any indication.

It's not limited to cancer, you know, autoimmunity, neurology, anything you want. The system works really well for any kind of tissue, any kind of indication. And the workflow actually is designed for high volume settings. You know, I'm sure Jen, as a CRO you're constantly receiving samples from, you know, all over the place for, you know, applying the different technologies to them. So you really want kind of a standard IHC level protocol with high throughput single round staining and imaging performed in parallel because typically you're going to be staining samples while you're imaging and analyzing things from last week, right?

You don't want to be in a, you know, doing things in series. And so really you prepare your slides on standard glass, your samples on standard glass slides. Typically it's FFPE but it can be fresh frozen. The staining is really simple. You just make a master mix of antibodies directly conjugated to ArgoFluors. Basically just like you would for flow cytometry style staining. You can easily stain dozens of slides per day. I think you guys use an auto stainer. You can do manual staining. It's really easy to do large batches of stains. You can scan them right away, but you can bank them for at least 12 months without loss of signal.

You can image them 20 on the system, 20 channels and a single scan and 24/7 on attended operation with 30 slide hopper where you can constantly add and remove slides throughout the day. And also at the same time, you're getting quantitative data output in parallel to the imaging so that you're, you know, it's giving you time to, to, to data as fast as you can possibly do to, to maintain the timelines that you need to for your, you know, for your pharma clients. Building panels are super easy. Essentially you use antibodies directly conjugated ArgoFluors. These are ultra bright photo stable fluors that are spectrally spaced to work really well and multiplex together.

The other important thing about the reagents is they're all validated to work on the Orion system with a single antigen retrieval condition. So our entire catalog, they're all, all the reagents are compatible with each other in any combination that you want. And I'll talk probably at the end, you know, once Jen goes through one of the analytical validation that she did with one of her panels, I kind of show you how we, we do that. We have off the shelf panels. We, we, we've offered internally, we've developed over 300 panels. It's super easy to do. If there's any reagents or biomarkers that we don't have antibodies for, it's super easy to make your own.

And I'll show you a little bit this panel designer which allows you to configure that your custom panels for any application. So again, really Orion is built, has been designed is an ecosystem designed for spatial biology of scale. It, you know that single round workflow really is what was transformational. I think with the addition of a high capacity automation that really is going to enable sort of even more high volume studies. You do not limited really with reagents. There's an extensive menu of antibodies and kits for independent development of very accurate reliable panels. You can also make your own reagents.

We also work really well with our clients. So, you know, you can develop your own panels, we can develop panels, we have panel transfer programs. You know, some CROs are working with pharma. So it signs we're working directly with pharma. So very flexible, easy to work with, with Rare sight as a partner as well, which I'm sure Jen can talk about. So, and with that, why don't I hand it over to you, Jen, and you can talk a little bit about kind of how you're, you know what you guys do with our system.

Jennifer Bordeaux

Thank you so much, Tad for the great introduction. I'm really happy to be here on behalf of Navigate Biopharma to talk about what we've been doing in the Orion space and our collaboration with RareCyte.

Before I get into that, I just wanted to take a moment to introduce Navigate Biopharma. Just for those that maybe have not heard of us before, we are a specialty CRO with our mission to really enable translation of innovative treatments into reality through our cutting edge biomarker solutions and high quality data delivery. We are a CAP CLIA accredited lab as well as having an ISO 15189 accreditation as well. And we really strive to enable this mission by focusing both on science and the patients that we're being trusted with their samples to actually execute these complex biomarker solutions.

And just briefly, we really are committed to high quality applications and we've contributed to a number of drug approvals over the last 25 years that we've been supporting clinical trial testing and we have a wide variety of GXP systems that are really designed to keep us in this regulatory space. We are really happy to be able to support the industry at large over a number of technology platforms that are really tailored to meet our customers needs. And this includes complex flow cytometry, genomics, really focusing in digital PCR and NGS Ligen binding and immunoassays. And then the space I'm really excited to talk to you about, about today is our immunohistochemistry and spatial biology, where we're really a leader in the fluorescence IHC multiplexing and excited to really bring this forward into the clinical trial space.

So focusing in, we're really going to talk today about how we've characterized the tumor microenvironment intended for clinical trial applications by looking at our 17 plex immunofluorescence and immunohistochemistry as well as combining this with cellular neighborhood analysis. So to start off, I want to just introduce our general workflow for our high plex fluorescence biomarkers at Navigate. We really focus on having a reproducible workflow across the board and it's embedded with our pathologists. They are instrumental and really QC check and upfront review of the samples that we are receiving.

And we do this through an H&E stain and evaluation to make sure that the sample quality is appropriate for the downstream application as well as identifying what the region of interest on that particular slide would be. According to the study of interest. Many times the space that we're in and what we're talking about today is focused in our oncology, but it could be really any disease indication where we have the tissue of interest. As Tad mentioned, we have automated our Orion staining using an the Intellipath system. This allows us to stain up to 50 slides at one time in a fully automated manner.

We then image these slides on the Orion or the Orion HT and bring it downstream into specifically tailored image analysis applications using Halo and Halo AI approaches. Any of the results from these assays are then QC’d by our subject matter experts in consultation with our pathologists. And as Tad alluded to earlier in his talk very nicely, there is a strong rationale for having high plex immunohistochemistry in clinical trials. It really historically has been somewhat of an unmet need. Even back in 2019 showed a comprehensive meta analysis looking at the right assays that could really best predict response to checkpoint and PD-1/PD-L1 therapy.

And what they saw through this analysis was that the multiplex IHC and IF signatures were really able to best predict and have the strongest area under the curve for showing which patients could respond to therapy even higher than conventional PD-L1 IHC or gene expression profiling and tumor mutation burden approaches really illustrating that the spatial relationships and protein co-expression are really important and understanding those cell subsets or what is driving the predictive power of these signatures.

Additionally, it's well established that many tumors may require more than 12 separate stains just for diagnosis. So if we're trying to support one of these disease indications, having the ability to stain multiple markers to even categorize and identify the tumor types that we're working in is very important. And as had also mentioned, many of the emerging investigational therapies that are coming out really require simultaneous understanding of that drug and diagnostic target expression and the surrounding immune landscape. And if you think about the tissue samples that were really receiving in clinical trials, many times these are small corneal biopsies.

If you're trying to profile this with a conventional immunohistochemistry, you may be looking at 20 different stains as a single marker. You're very quickly going to exhaust that precious tumor sample. So with that in mind, as we were looking for what approach we wanted to bring in to really meet this need for our customers, we quickly realized in collaboration with RareCyte that this spectral IHCFIHC approach really had the workflow we were looking for to put into the clinical trial space. That single round staining followed by a single round of imaging on the Orion is really key to establishing a reproducible workflow.

And it also was able to overcome potential chemical limitations of detecting multiple markers within same cell at the same time with other alternative cyclic detection methods such as TSA based chemistry that can be a challenge. In addition to that by the ability to put 17 markers on the slide, we're then able to consume a minimum of threefold less specimen than we would if we were doing a number of smaller plaque stains. While also being able to provide 1000 fold more data points. Because all of these markers are now on the same slide on the same scan. And you can look at interplay between the different cell populations as well as co-expression of all of these targets without the need to potentially be looking at image registration or co-registration or differences in the depth of sample between one section to another.

If you're doing a serial section, single stains for example, and we did validate this assay for exploratory use in a wide number of tissue indications, which is the data that I'm going to walk you through next, is how we really do that. So we strive to analytically validate any of the assays that we then put into a clinical trial space based on principles that are really outlined by CAP. And so in this case, we're really looking first at demonstrating sensitivity and specificity of all 17 markers that are in this panel. And we expect to see across the number of times that we evaluate each control that it remains a known positive tissue remains expressing the biomarker of interest.

Likewise, if it's expected to be negative tissue sample, we show that negativity across the board. What I'm showing here is the mean of five replicates for each of these samples. All as well as the standard deviation are shown in the error bars. And to pass our criteria here, we would expect our positive samples to be expressing at a greater than 5% positivity rate across all of the replicates of the sample. And therefore the converse for our negative is that it should be less than 5% in this instance. What I'm showing for positive control is primarily a tonsil and flame tonsil for the majority of markers in this with the expression of LAG3 and Sox10.

LAG3 because it is a fairly rare biomarker. We are actually using an overexpressing cell line control for this and for Sox10 because it is really intending to identify Melanoma tumor cells that would not be present in tonsil, we used a Melanoma tissue sample across the board. We use a normal heart as our negative control because as many of these are immune related markers, it is quite difficult to find a negative control that would be the disease of interest for example. And what we're showing here we really see across the board all of our markers are either low positive at the six or 7% up to potentially 70% positivity in tonsil depending and across the board the negative was less than 1%.

From there we move on to a precision reproducibility where we're looking at this across the disease indications of interest for the panel.

And what we're showing here is an example from each of breast cancer, colorectal cancer, melanoma, non-small cell and prostate where we evaluated all seventeen of these markers across intra assay precision, intra assay precision, intra operator precision and then quantify those expression patterns here.

And again we're showing the mean and the standard deviation across those runs. And what you can see from this graph is that really we saw the assay is very reproducible across all of these parameters that we tested for each of the various disease indications. Going into a little more detail on our assay validation criteria as it relates to precision, we're really looking at a reportable by reportable value to see that precision is maintained across a determined range. So we're looking for less than a 35% CV for any biomarker where we have a positivity greater than 5%. For values that are less than 5%, we are looking more in an absolute difference between those values due to low level positivity.

Just mathematically will show potentially a larger CV that may not be biologically relevant. And then across the board as we look at this, we need at least 85% of all of the parameters that we're putting into this validation to meet that criteria.

For this particular assay, we had 96% of all of those parameters meeting the acceptance criteria that we had set, which is again showing robust precision and reproducibility when looking either intra assay, inter instrument or inter operator across three different days, two instruments and two operators.

And then looking a little more deeply at the actual imagery that's generated from the Orion system. I'm showing here an example of non-small cell lung cancer where we're looking at each individual biomarker to see the dynamic range and we're able to really see high quality expression and localization of all the markers in the panel, both at a very low prevalence rate. So the FOXP3, the Granzyme B as well as high expressors such as the pan cytokeratin, the staining and image quality is very high. And because these are direct labeled antibodies, you can actually start to also see overall sub cellular localization of some of these targets and especially something like Granzyme B where we see it's really more polarized in the cells just based on the biology of marker.

You could actually see that at the resolution for these images. Our scans here are taken at 20X, but you do have the ability to also scan at 40X if needed. And if we look at the co-expression, some examples of this that I'm showing here, we're able with this panel to really faithfully define the tumor microenvironment. So the image on the left, we're looking at M2 macrophages as identified by CD163 that are also expressing PD-L1, and they're very close and adjacent to the tumor cells in this particular lung cancer, which are identified by Pan-CK in white. And then the right image is the same field of view.

And now we're looking also at the interplay of cytotoxic T cells that are CD8 positive, Granzyme B positive, as well as regulatory T cells, CD4 positive, FOXP3 positive, all in the single field of view. So one of the challenges with the ability to generate all of these data points is then also taking it down the line and how do we interpret it. So one approach that we've taken with this is to then also leverage cellular neighborhood analysis by using spatial cellular graph partitioning as a proof of concept on other ways to, you know, further refine and interpret the data that we're generating out of these panels.

So here we're just showing our workflow where we've taken our high plex images from the Orion, we've segmented into cells and phenotyped each cell of interest in Halo and then funneled this through the SCGP algorithm and 1st verified this on tonsil controls. And so doing this, we were able to pull out five main microenvironments that we would be expecting to see in the tonsil and those are illustrated in the middle panel color-coded. And then for context on the right, we're showing the Orion images of some of those main components where we would expect germinal centers to be heavy with B cells and that's what's being shown here.

And the T cell expression really in the follicle zones and then the crypt areas where we see pan cytokeratin expression. Taking this approach into the tumor samples, we then took a similar approach and determined that there were 12 unique microenvironments that could be identified with this across the five disease samples that I mentioned earlier. And this allowed us to really these microenvironments are illustrated here.

They ranged from just looking at T cell areas, the NK cells, myeloid components to more the tumor regions themselves as identified by Pan-CK or some of the more complicated co-localization areas where there were T cell infiltrating into the tumor region zones or areas where they were enriched for both T cells and myeloid, for example.

And here I'm just showing 2 examples of how these mapped onto the tumor resections that we were testing. And you can see that depending on the tumor and tumor indication, but likely also just the individual patient samples, you would see a range of these microenvironments across the sample of interest. We can see both where those regions are with the graphs on the left as well as plotting a percentage of the various cell types of interest in more of a heat, heat map fashion that are shown in the insects on the right. And so really our, our key takeaways is that we were able to validate the 17 plex to our immune landscape assay on a range of tissue types really meant for implementing in clinical trial applications.

Our assay is showing an optimal method to really analyze small clinical biopsy samples without really compromising the tissue integrity, which is a huge win for generating the data in these precious clinical samples. This particular panel, we validated 190 different phenotypes that could be pertinent to mechanism of action for many immune modulating therapies. And with the spatial cellular graph partitioning, we were able to identify 12 tumor microenvironments with differential expression across disease indications.

And given this limited availability of clinical trial samples and those challenges with collecting the biopsies to begin with, we really feel strongly that these types of high plex panels are going to provide a significant advantage by minimizing the tumor, the number of biopsy slides that are actually needed.

And we're really excited to be working with RareCyte to enable these high plex assays on the Orion HT as we do feel that this will provide the data at scale that we really need to support our clinical programs. So we're super excited to keep expanding on these offerings and the Orion platform in general. And as we move forward, we'll keep addressing more and more directed research questions. And I know Tad, happy to turn it back over to you to walk through how one of these panels could be developed.

- Tad -

Thank you, Jen.

Great to see the success that you've had with that panel that we actually transferred this panel to you, if I remember right, kind of right after your bought the system, I think a couple weeks after your training and, and you know, kind of your, we, we touch on a few things in terms of what it takes to, you know, do this high-plex analysis, which is critical to these therapies in clinical trial settings.

Really you touched on two of the bottlenecks. There's really three. One of them is actually developing a reliable panel is usually difficult, right? Of course, once you develop it, the word reliable is important because if it's not a very good panel, you're not going to be able to have the analytical validation success that you had with the system. And you're not going to offer that as a test, you know, to your clients, right? Because they're not going to want to spend their precious clinical trials on those panels if they're not giving you good results in a timeline you can meet. Because your clients, of course, are trying to, you know, determine if the therapy is helping or working right. And they need confidence in that the data is good.

So it all starts with developing a panel. So I briefly touched on sort of the, the reagent infrastructure. But, and all, you know, of course, as a vendor, we always say it's easy. In our case, it really is easy. So I'll give you an example of kind of how we designed the panel that we transfer to navigate. So to our tumor landscape panel here. If I come into this tool, the first thing I want to do is select biomarkers. Our catalog is very extensive. So I'm going to show you a couple things here. We can look just directly at the library of biomarkers that are commercially available from us that are directly conjugated to to ArgoFluors.

There's a lot there I can easily like. I know that you guys wanted melanoma, so Sox10 was important. So I can just search for that and add it to this list. So this is the one plex panel at this point. I can also leverage I mentioned we have pre-existing panels, off the shelf panels that we offer many of the markers that navigate wanted in their landscape panel were part of our tissue profiling core. So I can add basically all of those tissue profiling core markers over here. But you know, you don't need to use all these markers because we sell everything as individual reagents. So if you're not interested in CD31, I don't think you have CD68 in there or SMA.

You can just take those out. I think your slide indicated that you have two broad categories, right? Your phenotyping markers and your functional markers. We also group markers in terms of themes. So there was some deep phenotyping going on that was beyond the standards. You, I think, had CD4 cells, CD11B if I remember right, I think NK cells, you had plasma cells, right. So you can add these markers here. And then the other thing is not just phenotyping, but what functional state are they in is really important to your downstream analysis in terms of those, you know, those, those neighborhoods and what you expect those micro environments to be doing.

So functional markers are important. I think you had FoxP3 to look at regulatory cells, Granzyme B for cytotoxic potential checkpoint PD-L1 and Lag3, right? So it's as simple as just sort of picking biomarkers. So if we essentially go next any type of co-localization that's going to happen between your biomarkers, you can essentially say a CD3, CD4, CD8 co-localized, you know, maybe PD-1 on T cells as well. So you can sort of assign which biomarkers are spatially co-localized in the same pixel. The algorithm essentially decides which reagents to use. We'll take that into consideration.

So if I come here, this gets a little bit into how the system works. We have columns here which would indicate the channels on the system and these rows indicate clones that are associated with the biomarkers that are in the panel and the sort of blue squares or products that we sell. I can also clear the filter just to give you a snapshot into how large the catalog is today. It's extremely extensive, right? But anyway, we're filtering this based on the biomarkers that were in this panel. I can do things like exclude clones or exclude channels, or I can basically just say, hey, use the panel design tool to configure a panel for me.

What's going to happen here is it's going to assign these different biomarkers to different channels that are associated with products that we sell. So you can see here it's assigned CD163 to the first channel, CD11b to the second channel. This looks at the spectral overlap or crosstalk that is happening between these different fluors. The key ones that it points out is the space, the spectral overlap and spatially co-localizing biomarkers. And you can see this particular configuration, there's no crosstalk between them. So that's a pretty favorable situation to be in. So we can select this as the final panel.

This actually builds essentially once you're done with this, you can generate a quote request you can share with collaborators, etc. So typically once you receive those reagents, you would do initial sort of titration on a few of your positive and negative controlled tissues because you may want to adjust the concentrations that the reagents come in. And at that point you're, you can pretty much start your analytical validation. So I just wanted to point that out kind of what happens upstream of the analytical validation, this is in this type of thing customers can do on their own. They can actually have us help with.

Then we have panel transfer programs. You can make your own reagents quite easily in this panel design tool as a way to import your own custom built reagents. So anyway, I wanted to kind of give you a sense for, you know, how flexible a panel design really is. And yeah, so thank you very much.

Janelle –
All right, excellent presentations from both of you. Thank you, Dr Tad George. Thank you, Dr Jennifer Bordeaux. Audience, feel free to continue adding questions as you think of them. And we will be back in about 10 seconds for live Q&A.

All right, excellent. Let's jump into the slide Q&A. First question, you mentioned HT has 30 slide capacity. How many samples are practical to load for an unattended automation run?

OK, Yeah, that's a good question. Yeah. So as I, as I mentioned the HT you can load 30 slides on to the system maximum. It really depends on how large your tissue samples are. And it was designed really to make sure you could get through at least a long weekend with 24/7 unintended operation. So, you know, it takes about an hour per square centimeter to image in 20 channels. So if you have really small, like corn needle biopsies, you can easily load up several, you know, 15 to 20 of them get, you get through that in less than a day. If you have really large samples like colorectal cancer surgical resections, which might be 5 or 6 square centimeters, you'd probably just load, you know, enough to get through, you know, 24 hours or through the weekend.

The system also allows you to, you know, add and remove slides without interrupting operation. So it's real flexible. It depends on, you know, kind of the situation you're in. Yeah, that's that's kind of how the HT automation works. We've also verify one thing, the reagents, the fluors that we use, ArgoFluors are very photo-stable and we've demonstrated that they are stable at room temperature for at least three days, which is great because it can get you through a long weekend.

Janelle -
All right, excellent. Let's jump to the next question and audience. I see a ton coming in right now. Feel free to keep adding them. And I'm trying to combine a couple duplicates. This one is specifically directed at Tad.

I'm curious about single round of staining, which is very valuable. Do the dyes not overlap?

Tad -
Yeah, good question. So definitely that's one of the major advances that we did is, is certainly if you're going to put you know, 18 different fluor conjugated antibodies and stain them and scan them one round, there will be definitely spectral overlap. We screen probably 300 dyes for finding ones that were bright, photostable and spectacularly well spaced from one another. That actually led to us custom building a 9 laser module that goes into system to minimize, you know, maximize sensitivity and minimize crosstalk.

But there is still residual crosstalk amongst sort of adjacent fluors or nearby fluors. And so we also developed an extraction algorithm to isolate, you know, signal in, you know, for each biomarker into individual channels. So yeah, definitely a little bit of spectral overlap, but fully designed to minimize that and then it corrects it. Good question.

Janelle –
Getting the couple about cost, roughly how much does it cost per slide?

Tad -
You want me to take that one, Jen? So good for it. Yeah. So in terms of reagents, you know on average it costs somewhere around $25 to $30 for biomarker per slide if you're buying the reagents from us. So like, you know, let's say a 15 plex panel typically will run you, you know, less than $500 for slide. There's no other costs associated with it. There's no like staining gasket or all the reagents used for antigen retrieval or just normal lab reagents. So yeah, I would say you can easily get a high flex panel for less than $500 for slide.

Janelle –
All right. I, I did find one directed just for you, Jen. What are the challenges with cyclic IF methods?

Jennifer -
It's a good question, right, Because there are other ways to get at high plex biomarker assays like Tad had mentioned earlier in the talk, one of those beings like cyclic methods where you're coming back and doing repeated rounds of staining and then imaging. While that can work really great for potentially a research setting, when we're putting these into our clinical trials, the challenges come into play if something goes wrong during one of those rounds. So if the image itself is out of focus, for example, now you've potentially lost some of those biomarkers that are supposed to be part of that panel.

If you try to go back and repeat them at the end, now you're not running the same validated assay. So the unknowns that can come into play with troubleshooting the cyclic IF make it a challenging workflow for actually putting more into a later stage clinical trial space combined with just the length of time it could to also stain an image in that sequential round.

Janelle –
All right, new question just came in. Are the panels compatible with Leica automation?

Tad –
Yeah, yeah. So I think you saw, I think Jen, you guys use the biocare and telepath. We have multiple Leica Bond RX is in our lab. We run services.

So yes, it's fully compatible with most autostainers. The staining is very simple actually, we often do manual staining. But yes, you can, you can essentially incorporate you know this into sort of that high volume. This is kind of joining another question that I see related to sort of operationalizing this in high volume labs. It has a parallelized workflow where the staining is being done in parallel with the automated imaging and you can certainly incorporate automated staining if you if you want, if you're a high volume lab for sure. Good question.

Janelle –
Thank you. What is important when designing an assay for a clinical trial?

Jen –
I'll take this one. Really it's a number of the factors that we talked through the webinar. We really want that assay to be, you know, robust and reproducible to support the study. But also when you're thinking about the assay in question, really making sure that the targets of interest are incorporated early, thinking about maybe some of the deeper biological questions that you may wish to answer later on, and putting these together in a way that will robustly work for your study and really maintaining consistency over time. So all of those really come into play as you're thinking about the length a clinical trial could be running.

Janelle –
All right. Next, how do you handle batch effects when analyzing large cohorts over months?

Tad –
I can take that one, Jen. And you can certainly follow up if you have any comments because you guys probably do this a lot as well. So a couple thing, one is part of the reagent technology. We looked at various different types of technologies and really settled on simple amine conjugation chemistry and screened a bunch of sort of buffers for storing the reagents and settle on one that gave five year shelf life.

So one thing that's nice, if you have a multi year trial, you can actually sort of develop or get a sort of a single lot multi-year, you know, panel that you can do. Some of our clients have asked us to quarantine, you know, panels for them for over multiple years. So that's kind of one way to handle at least lots, lot variations. You don't have to do any lock bridge in that situation. In terms of specific staining batch effects, which, you know, typically what we do for our services is probably you do the same again is we'll accumulate samples for a trial and then stain them in large batches with a control sample, which expresses all the biomarkers in the panel.

And we sort of, you know, can bridge the performance that control sample from batch to batch to make sure that the staining intensity is, is good. So it's important to have like a staining run control during your batch staining runs.

Janelle –
All right, a bunch of new ones came in. Give me one quick moment to Jennifer. What's the timeline for 20 slides, 15 plex including run, Halo, QC and data delivery?

Jen –
That's a great question. Typically we're looking at something between 15 and 20 business days for a batch run of that size to really make sure that we take the time to fully analyze and QC the data.

Janelle -
Do you have specific panels for the MASH?

Tad –
That's probably for me. OK.

MASH, I think that's that's Metabolic dysfunction Associated Stato Hepatitis I think. I think so. So no, we don't, but what? So I think in this case this particular disease, you know the sample types, I think the biggest thing is getting the biomarkers necessary for the indication, but also the tissue itself being fatty liver. I think that's another thing that's important about doing single round staining is that particular type of tissue, the more fatty a tissue is I think the more risk you have of it falling off during a staining process. So I think it would be ideally suited for a single round approach.

You know we are biomarkers that I think we're up to about 500 conjugate products spanning about 200 biomarkers in a catalog. We do have some liver specific biomarkers, most were very reactive. We basically put things in our catalog that our customers are doing and we do have customers that have run liver panels with us. But I suspect you know, if we were doing a MASH panel, we would probably develop a few custom biomarkers specific for that particular application, which is pretty easy to do.

Janelle –
All right, thank you. Tad, can you share the panel builder tool?

Tad –
This sounds like from one of our customers I'm guessing. Yes, that it will become freely available on the portal very soon. So yeah, we're looking forward to it. I think it's going to help people design panels quite nicely. But yes, the answer is yes.

Janelle –
All right. Thank you so much. How many markers can be co-localized in same cell, in same cellular compartment?

Tad –
Probably another one for me, Jen, right.

So it's kind of related to an earlier question about spectral overlap. You do have the advantage with imaging and like in flow where you have spatial, you know resolution as well.

So I think when panel design and part of the panel designer actually handles this quite nicely, is you do, you can give input to the panel designer to indicate which biomarkers are spatially co-localized and the panel designer will typically space those away from one another to limit crosstalk in a spatially co-localizing pixel.

It's not required if the system will still subtract that pretty accurate, very accurately, if there's if there's overlap there. The main thing is, is that you limit your dynamic range on the high end of the camera if you have if you have co-localizing signal going to the same picture pixel. So the camera itself is 16 bits. So it's wide dynamic range. But the biggest thing is you want to avoid limiting the dynamic range on the high end. So I didn't directly answer the question how many markers may co-localize in the same compartment? I mean, you can design panels with easily 10, 12, 15 of those, you know, a lot of T cell panels. That's kind of what you're what you're doing, right? So, yeah,

Janelle –
All right. Do you, do you have to define focal points for each of the tissues or does the workflow for HT remove the time spent on defining the ROI and focal points?

Tad –
All right, good question. Wait, yeah, so the HT being a 30 slide automated imaging system, I guess the question is how do you, how does the actual auto focusing work? Right. So Orion essentially focuses on the center of the tissue, not on the glass and there's two ways to with the HT device to get good focus, focal plane. The preferred version of the more common version is the system will run through if you load say 15 slides in the hopper. The first pipeline step is a pre-scan at lo-mag 4X on the nuclear image. And those slides will be held and then remotely you can come in and draw essentially your focus points and then it will come back and automatically find the center of the tissue at various focus points and develop the contour for the imaging. That's the most common way to do it.

You can also manually load slides like you do on the Orion, normal Orion and set your focus points manually as well. We didn't really try to, you know, go with, you know, fully automated focusing because I think having your tissues in focus is critical and it's a pretty easy step to do kind of remotely with the HT. That's a good question.

Janelle –
More good questions are in here. Do we need auto fluorescence slides for every sample?

Jen –
I can take that one, Tad, if you want.

We, do not use a separate auto fluorescent slide for every sample, but certain disease indications tend to have different auto fluorescent profiles. So we may end up customizing based on the tissue type at large, versus each individual sample. But then if there are issues with the specific sample, there is the ability to go in and tune that further as needed on a sample basis. At least that's what we've found over here. So feel free to add Tad.

Tad –
Yeah, that's a good, that's an excellent answer. I mean, there's essentially A dedicated auto fluorescence channel in the system which I've sent because which handles most common autofluorescence that all tissues have. But yeah, as Jen mentioned, sometimes there's secondary auto fluorescence, particularly like a non small cell lung cancer example has some red shifted elastin fibers that you can dedicate secondary channels to and tune. But yeah, typically, you know, I think if you're designing a panel for a novel tissue from the ground up, it's a good idea to run an initial auto, you know, unstained control to see if you've detect that you need a secondary auto fluorescence channel and build your panel around it. But no, once you have those, those subtraction coefficients, like as Jen said, typically it works quite well and you can tune it afterwards if need be.

Janelle –
All right, we may have asked this one already. I apologize, but it's a question about can the Orion HT system image 3D tissues and generate 3D reconstruction views.

Tad –
Yeah, good question. So, so it, it doesn't do 3D, it's typically the way people do sort of 3D biology with the with the Orion is with multiple serial sections. So it's not a deep imaging technology. It's not like light sheet or anything like that. But we have had customers that basically, you know, take a block and do multiple serial sections and reconstruct it there. We, we typically are looking at 3 to 5 micron sections, but I think it is, but definitely not 3D imaging, good question.

Janelle –
All right, If you stay in tissue with 10 to 15 antibodies, can you strip them after scanning and then apply another 10 to 15?

Tad –
Let me take that one. Jen, you want me to go?

Yeah, yes, you can certainly cycle on the system we developed. We've basically screened a bunch of signal removal protocols and found one actually quite simple, really good. We've done internally in our lab, you know, 3 rounds of 17 to get a 51 plex. We also have demonstrated that you can take a slide that's been stained with say you know, 10 or you know, 10 or 15 plex and scanned and then stored for over a year in the freezer. You can take that out, remove the signal quite reliably and do another panel kind of a panel expansion maneuver. A lot of our customers actually use it that way.

But yes, the system is compatible with cycling, just at large steps per cycle, right? It's it's, you know, yeah, good question.

Janelle –
All right. What would be the fastest turn around time for analysis with three biomarkers

Jen –
Sure I can take this one. Right, like as in that kind of setting, if we're really talking about right perspective testing for clinical studies, that turn around time would come down drastically if we're only talking about, you know, a few samples at a time, staining is really fits within one day and then the imaging of the sample itself would be the next day and then you would move forward with analysis and QC from there, which would take another day or two.

So absolutely a low plex like that could fit into prospective testing for clinical studies if desired. But the main power and driver for this platform is really getting at that higher complexity biomarker testing, which if you're going to the 17 plex is unlikely to be something you might want in a, in a prospective testing approach. So it's really balancing right the needs of the study with the biomarker questions that you're asking.

Tad –
I think to add on to that, I mean, if you're getting towards like diagnostics right, where somebody, you know, is trying to get an answer from the sample, typically, and I think that's why the question was asked, typically as you move down that pathway, your plex does start to narrow down right to that three to five markers. So, so, yeah, it's a good question.

Janelle –
Another question, would this be a technology the hospitals and clinics could use routinely?

Tad –
Yeah, yeah, certainly depending of course on what the applications are. I think kind of that's related to the last question. I think I wouldn't put anything that's like high plex IF at the diagnostic stage now, but if you have, you know, high volume clinical labs in the hospital with access to samples where clinical research is being done, that's kind of how I would think about, you know, placing that. It's very commonly used in clinical trials where essentially, you know, samples are sent, you know, from a perspective clinical trial to a place like Navigate right, that does the analysis.

But certainly I think you know, clinical core labs doing even phase four stuff, we're looking at samples after the fact on, on retrospective studies. It's useful for. Yeah, I, I think it's, it's, that's certainly the technology I think has the potential to move in that diagnostic space, which is potentially what the flavor of this question was asking. And given that it's very simple reagents that are easy to QC, you can imagine making diagnostic because it's, they're very similar like flow cytometry grade reagents, right? The more complicated the reagent system, the less likely it's going to ever be, you know, a diagnostic. So I think it has potential, but you know, today it's the clinical research, I would say.

Janelle –
Amazing. Well, thank you so much, Doctor Ted George and Doctor Jennifer Bordeaux, and thank you to our audience for your active engagement. If we were not able to answer your specific question, we will do our best to follow up after the webinar. As a reminder, a recording of the session will be available on demand within 24 hours.

Thank you all for joining us today and we look forward to seeing you next time.