Hello. I’m Bettina Winkler from Visiopharm and want to welcome you to the second webinar of our masterclass series about the challenges of IHC standardization.

This week, we have Rasmus Røge with us, pathologist at the Aalborg University.

Rasmus is also part of the NordicQC organization and will talk about how EQAs are advancing diagnostic immunohistochemistry.

With that, I will hand over to Rasmus to start the presentation.

Thank you for the, kind introduction. My name is Rasmus Røge, and I’m a pathologist working at, Aalborg University Hospital. I have a PhD in immunohistochemistry with a small dab of, image analysis, but my main time is working as a hematopathologist in the upper pathology department.

And as a sideshow, I also work with the NordicQC organization, and that would be the main focus of my talk today. I’m so happy that Visiopharm invited me to give this talk. I always enjoy sharing the wonderful message about, external quality assessment in ISC.

So I wanted to start off with giving you a small example of why quality assessment and external quality control could be a big issue and is very important in IHC labs. This is two serial section stains for estrogen receptor in two different laboratories.

And as you can see, there’s more or less no difference between the IC quality of these two tumors. It looks as if both tumors have a very high, degree of estrogen receptor positivity and more or less all of the tumor cells are positive.

However, if we look at another tumor stained by the same two laboratories on the left, we see a lot of the tumor cells that still have some degree of positivity. However, on the other side, we see that more or less all the tumor cells are completely negative.

And thus, that actually means that this tumor will be categorized as estrogen receptor negative.

That would have been very hard to distinguish if you only had the tumors in one of the labs. However, if you compare labs as you do in a external quality assessment program, you will be able to detect these differences.

And another important thing that you can identify in this quality assessment program is which controls that actually should be used for the different stains. On the top, we now see uterine cervix and tonsil, And we can see that lab a stains all the uterine cervical cells beautifully, and also the germinal center follicular dendritic cells are positive for estrogen receptor. However, the lab b do not have the sufficient sensitivity of the protocol to identify these cells as positive.

So this kind of means that external quality assessment could be a very good idea to participate in for different pathology laboratories.

External quality assessment gives a way of providing objective testing of the laboratory performance.

And you can either do that by comparing the performance of a lab to a group of different laboratories or to a specific reference laboratory.

And in NordicQC, we have selected the last method where we compare the stains of the different participating laboratories to the reference laboratory in Aalborg, where we have made a lot of tests to validate the staining.

There’s a lot of benefits to participate in these quality assessment program. It will give you an evidence of the actual quality, and it can also be used in the more general sense to detect systematic problems with the associated kits if assessment programs identify some systematic problems with kids or specific antibodies.

However, there are also some drawbacks and that is that participation one or two times a year in a market that is only, examined a few times is only a snapshot of quality. And if that’s all you do for controlling your quality, it will give you a false sense of security. So you cannot alone rely on external quality assessment protocols.

But you also have to do some quality assessment in your own laboratory.

And now I will give you an introduction to the NordicQC program that I am part of. NordicQC is actually called Nordic Immuno Histo Chemical Quality Control and is an international academic proficiency testing program that is independent and not for profit organization.

NordicQC is located in Denmark, in the city of Aalborg on the top of Denmark, and we are physically located in the department of pathology at Aalborg University Hospital.

At the moment, we have seven people employed, and all of these people are more or less part time employees. This means that they also work a part of that time in the clinical laboratory. So we still have this leg in the clinical organization.

We have numerous collaborating partners, both pathologists and technicians throughout, especially Europe, but also in USA. And we, maintain, good relations with a lot of different companies that are involved in in IIT. This is an overview of the NordicQC program. Each year we have these different modules as we call it. And a module is a collection of of different markers. We have the general module, which contains markers for general surgical pathology laboratories. And we have some specialized markers for breast cancer, specifically HER2 estrogen receptor and progesterone.

Then we have the FISH module with HER2.

And the newest module is the companion module, which at the moment, more or less focuses on PDL one. But I suspect that, more companion markers will be included in this module in the in the coming years.

At the moment, we service almost seven hundred international laboratory service from almost sixty different countries. And you can see that, actually, it is the best cancer module that is, most popular, but we also have a lot of our choice participating in the other modules.

If we look on the general purpose of NordicQC, it is, of course, to assess the equality of immunohistochemical stains in the participating laboratories.

This is based on using standard process tissues that we circulate to the different laboratories.

Then we also want to identify optimal and insufficient results among these stains. And we want to correlate this to specific antibodies or protocols or even specific platforms to give an, idea on which protocols actually work.

Then we want to publish these results, on our homepage and specifically also try to publish important findings in scientific journals.

One of the most important thing is that we also want to give individually tailored recommendations for each laboratory on how to improve their quality. Finally, we want to collaborate with the industry in order to to improve the field of of IHC.

And if I move on to the specific method for NordeQC, this is very concrete on how we actually set up a proficiency run.

When we prepare for such a run, we construct a tissue microarray containing critical tissues for a specific epitope.

This means that for each epitope, we will decide on interesting tissues and select lot of different normal tissues and tumors from the archive of the department pathology and select the best tissues to be used in the block that should be circulated to the participants.

When the run is starting up, we ask all the participants of the NordicQC program to submit their staining protocol on the home page. This means that it’s obligatory for participants that want to to be part of or want to submit a staining protocol.

And this protocol is actually very detailed. We want to know specific temperatures and how long time the tissues have been exposed for, for antibodies and epitope retrieval. Once we have all these specific information, we circulate slides to the participants.

And then we ask the participants to state these slides and return them to an audio cue scene.

When all the slides have been returned, an expert group of pathologists, both internal and external, and expert technicians meet for assessment of these slides.

After the actual assessment, there will be a rather long time of data analysis of these, protocols and results in order to identify optimal and insufficient staining protocols that we can publish on NordicQC. And finally, we will return a assessment mark and a tailored protocol recommendation to each of the participants.

I’ll give you some examples. This is an overview from the NordicQC homepage where we can see the different parameters that we actually ask the participants to submit. We want to know the staining platform, the the primary antibody clone and lock numbers and dilution factor of the, antibody and which diluent offer how the epitope ritual was accomplished, if the laboratory used any kind of proteolytics and a lot of information about the the visualization system.

And this is an example of the number of flights that we need to cut for each run. We set up this block typically containing five to six different tissues, and these tissues have all been fixed according to the ASCO CAB guidelines with twenty four to forty eight hours of fixation in informaline.

The tissues are selected in such a manner that it contains both normal and clinically relevant tissues, both, normal tissue and tumor tissue. And we also want different levels of antigen expression. So some of the tissues will be very strong expresses, of the selected epitopes, while others will be moderately, low level expresses or even some negative tissues. And this is in order to identify both the sensitivity and the specificity of the the protocol. So we sent out two on stage slides for each market to the participants, ask them to stay in the protocol and return it to NordicQC.

And this is a picture from one of the COSI assessor meetings where we look through a microscope on a monitor to each of the slides. On the other monitor, we have a standardized staining, obtained in the reference slides and based on the expression patterns in the different, tissues, we provide each slide with the score, either optimal, which means it’s perfect or very close to the reference slides, good. If it’s fully acceptable in order to identify clinically relevant expression levels, However, there could be some optimization in order to obtain a better signal to noise ratio. And then the two insufficient cause, which we call borderline or poor, where we have a generally too weak, or or even false negative staining of one of the included tissues or a false positive staining reaction in in one or more of the tumors.

When we have compiled all the scores of all the laboratories and correlated this to the different parameters of the protocols, we are able to provide a general report of the assessment. And this is just the introduction to one of these reports. I I will show you an example later on the NordicQC homepage. And this shows the purpose of this PAX eight testing and which tissues we decided to include.

And if we move on, I will show you more or less the most important, but also the most complicated table of these reports where we can see the performance of each of the antibodies. And if we just see the concentrated antibodies, we can find the rapid monoclonal antibody SP three hundred and forty eight, which had a lot of labs that had optimal staining, some good, a few borderline, and then a few poor. This gave an overall performance of ninety one percent.

So that’s a very successful antibody. And if we find some of the Polyclonal that have a tendency to be insufficient, then we can see that not a lot of these actually performed very well.

So this gives us the information to give some detailed overview on the different problem. And these different categories, describe this.

So overall, if we have analyzed all these parameters, we are able to give some kind of information on parameters that would give you an optimal performance of IHC. And the most important is to select a successful antibody. And there are different problems with antibodies. It can be poor antibody that will never perform.

Some are less robust. That means that the range of, concentration where it performs optimally is very narrow. Then some, companies have poorly calibrated RTUs where the dilution is either too strong or too weak.

And then some antibodies can perform different on different stainer platforms. I’ll show you a few examples of these antibody problems.

This is two melanomas, a one and a two, the stains for Sox ten. And Lab one has used an optimal protocol with the antibody EP two hundred and sixty eight. And as expected, we see that more or less all the tumor cells are strongly positive. However, another lab stayed the exact same tumors with a polyclonal antibody and called false negative results in both tumors. Actually, the same antibody was also applied by a third lab that got much better staining. If we look at the tissue controls, we see that in skin, the Optimum laboratory got, strong staining of melanocytes and in appendix staining of the Schwann cells, the nerve sheath, cells.

But in the insufficient lab, it gave a much weaker staining of these melanocytes and and Schwann cells.

And the third lab that actually were able to stain the, Melanomas gave a a false positive staining of the, of the skin cells. And you can also see the epithelial cells at the appendix were positive.

So the polyclonal antibodies did not work. This clone did work, and that is also seen in the NordicQC assessment where we saw that the EP two hundred and sixty eight clone performed very good, but the poll polyclonal antibody all failed.

And, another example of MAM1 where the MAM1 p one antibody gave optimal staining with strong staining of plasma cells in appendix, while the clones BC five and MRQ forty three gave either aberrant cytoplasmic staining in the epithelial cells or an overall false positive in both the epithelial cells or muscle cells.

And that was also seen in the, NordicQC results.

Then we also see, and I will not show examples of this, but, some antibodies are very good, but the laboratories select a very low dilution, meaning that they will not be able to get a good sensitivity of the protocol. Perhaps they have insufficient or incorrect epitome material. Perhaps they use a low pH buffer when protocol should use a high pH buffer. Or perhaps a laboratory uses, proteinase epitope retrieval instead of, heat induced epitope retrieval.

Some labs select less sensitive visualization systems in a market that requires a very high sensitivity of the protocol. And of course, we can also look at the quality of the morphology of these stains. If lice have been boiled too much, it will give an impaired morphology that will be hard to read. So based on all these information, we will be able to give each lab a tailored recommendation on how to improve. This could be this is just different examples, but it could be that we want the laboratory to replace their primary antibody with another clone that could be be more successful.

Perhaps the lab should calibrate the antibody concentration to a higher or lower level. They should optimize their epitope retrieval or perhaps select another here time, temperature, or pH.

And I actually think that this is one of the most important things in NordicQC where we give specific recommendations to the laboratories on how to to perform.

This could be a typical, individual feedback to one of the labs where we have the overview of the polar, assessment here, the different epitopes, and how the laboratory performs. And for each of these insufficient results, we will have a specific comment on what went wrong, false negative reactions and how to improve the different protocols.

This is from a study we published in two thousand and seventeen concerning SUX-ten, where we saw the performance of different protocols when we compared monoclonal and polyclonal antibodies.

And as you can see in one forty five, we have a, an okay performance of the monoclonal, but a much worse performance of the polyclonal.

And that was more or less the same in forty eight. So over time, when more laboratories got the advice to to stop using the polyclonal, we have seen a decrease in the use of the polyclonal antibody and an increase in the use of monoclonal antibody. And this has resulted in an increased pass rate among the laboratories. So it actually helps to be part of, external quality assessment program.

This is an overview of the, performance of all the laboratories based on the drawing of all the labs, and when we split this up in the different modules, we still see that a lot of different slides become insufficient, either borderline or poor. It’s easy to see that the breast cancer module has the highest percentage of optimal results. And this is specifically related to the HER2 and estrogen receptor, a test that we have performed very many times. I think it’s more than thirty five times it has been repeated. So the laboratories have had a lot of possibilities to prove that. And it’s very nice to see that it’s actually possible to reach these higher levels. But in the general module, we still see that more or less, twenty five to thirty percent of these slides actually become insufficient.

And latest, we did a survey among the different participants where we asked them to rate us and we got a very high score from the laboratories on their overall performance of NordicQC. We wanted also some input this time on which markers we should select for the next one. So I think this will be an annual thing to ask the laboratories on both the performance of NordicQC, but also how we are supposed to continue working as an organization.

And now I want to switch to show you some of the online resources that are available on the NordicQC, homepage. So I’ll just switch to a browser.

So this is the front page of the, NordicQC homepage, and I want to specifically direct you to these, three areas of the home page. The first one is called assessment, and it gives you a list of the different epitopes that we have we have tested through the time. The last run number in here is listed here. So if we select once, it could be TTF and click on TTF. We can see a short description of both the marker and how it’s supposed to use. A very important area here is the selection of controls, and there’s also a specific link to the latest reports concerning this episode.

If we go to the protocols area, this could be the starting point when you are starting out with a new marker. I was just select PAX eight.

And as you can see here, we have seventeen available protocols for PAX eight eight. So you are able to select PAX eight, find your staining platform, or perhaps find your the clone that you want to to work with. And then you can specifically just click the PDF and have a and a start out protocol. Of course, it’s important that you validated this in your own laboratory, but it it will, at least give you a starting point.

And then finally, the controls where we have a list of suggested controls for specific immunohistochemical staining.

If we select, as an example, CDX two, the list suggest that you can use appendix, co pancreas, and tonsil as controls. And if we select this, we have an overview of the expression pattern that, expected in these different tissues. We have both a high expression tissue, a low expressive tissue, and then negative tissue. And this box will give you an description of the, staging pattern, but you’re also able to to see this, as a scan and click to enlarge it.

So you can get a closer look on how the, the tissues are are supposed to to be seen. So this is a very strong resource when you are testing IC protocols. I need to mention that it is not not all GANs that are available for non members, only the four stains of the primary panel. So we have to be a member to to get access to this.

If we have found these controls on the NordicQC homepage, you can simply select these different tissues and use them as part of your internal control tissue on on your slides.

So overall, moving on, to the conclusions, I think that it’s fair to say that external quality assurance provides some objective evidence of the performance, of different labs. But it’s especially this comparison between the different labs and identification of methodological errors that are strong point of of EQA programs.

And based on this knowledge that we obtain by seeing all these different protocols and states, we are able to provide directions for improvement and controls.

However, these results also indicate that improvement of IIT is still strongly needed. And EQA schemes, the industry and and experts, they must work together in order to obtain, better IIT protocols.

And we pathologists and working together with technicians and the industry, we we must describe these requirements for optimal IC performance in order to be able to to identify optimized protocols.

But still, twenty to twenty five percent of all IHC stains that are returned to Nordic are still insufficient.

It can be, in some ways, explained by, new labs participating and perhaps new antibodies. But one still must ask how many IC things that are actually produced both by labs that participate in EQA schemes, but also, in non participating labs that actually are insufficient.

And and how will this affect the patient and the possibilities of optimal diagnostic, pathology?

And also, how many scientific publications are actually based on these, insufficient, I see stains.

So that was, all for me. I want to thank all my collaborators in in NordicQC and also Visiopharm for giving me this excellent opportunity to address you all.