Genomically-Derived Antibodies in Diagnostic Immunohistochemistry

1. The origin of genomically-derived protein markers and the methods of their discovery.
2. Their practical application to diagnostic tumor immunohistochemistry.
3. The advantages and potential limitations of their utilization in histopathology.

MEHRDAD NADJI, MD:

The topic of presentation, as you can see, is genomically derived antibodies in diagnostic immunohistochemistry. And the reason for that is that molecular techniques in the past 10 to 15 years have increased the utility of genomically derived diagnostic and predictive biomarkers, and today, particularly our discussion will be limited to the diagnostic markers as the field of predictive biomarkers in rather vast and requires a separate session.

Introduction

Alright, this is the presentation outline. The outline is how, what, and when. And by that I mean how these proteins were discovered, what are we detecting by using this technology, and when do we use them in our daily practice. And as an example of that I'm going to show you illustrative cases, some of the common, and common tumors where these markers are used, and I will finish the talk with the values and limitations of all these technologies at this time. This is my disclaimer, and that’s that I have no interest, financial or otherwise, in the content and material presented in this educational activity.

The immunohistochemistry that we use with genomically derived antibodies we have to know how proteins are discovered first. As opposed to classic methods of producing antibodies, which normal cell and tissue proteins are isolated, an antibody was produced against them, the genomically derived antibodies take advantage of molecular abnormality detected, and those molecular abnormalities may or may not result in expression of an abnormal protein, or an abnormal expression of a protein. That’s where we go from here.

These all discovered by expression profiling, either gene amplification, gene translocation and other genetic abnormalities, gene fusion and gene mutations, which produce proteins that we can take advantage of them and use them in immunohistochemistry. So what are we detecting is basically very simple, is protein overexpression, and protein loss, or under expression, or the product of the abnormal protein, such as fusion and mutated protein products. Where do we use them, a large number of these genomically derived antibodies are used as actually predictive markers. In fact, it was started in the area of predictive medicine, and they were presented as the so called companion test for a number of targeted drugs. They are occasionally also used for prognostication. What we’ll be discussing today is diagnostic use of these markers for either tumor diagnosis or tumor classification. And I have to mention to you that the diagnostic markers, that means tumor diagnosis versus a benign or normal lesion, is very uncommonly use for immunohistochemistry of these markers. These markers, with a rare exception, are basically only used for tumor classification and cell lineage identification.

Before I start to give you examples of these usages I would like to share with you how I'm going to approach with these case examples. First, on the left upper corner is the name of that particular marker, and its available the abbreviation for that name, where this name is coming from. And then I'll take you into the mentioning it's a type of protein product, how it was produced, whether or not it is expressed in normal tissue, and the main practical application, at least in my experience, and other potential uses, and sometimes abuses.

Let's start case examples with those markers that show protein over expression, and there are a whole bunch of them. By the way, the examples I give you today are just the tip of the iceberg. There are hundreds of these particular markers that are available, and use them on a daily practice of diagnostic pathology, what I can just share with you a few of them. And what a better place to start than with DOG1, which is known to a lot of us.

DOG1

It was discovered by gene expression profiling. And the staining is in cytoplasmic and membrane staining, which is unusual. Usually most of these, as we can see later, are nuclear antigens. But this particular one is actually cytoplasmic, and cytoplasmic membrane antigen.

It is expressed in a large of number gastrointestinal stromal tumors, more than 95%. The importance of this marker is that it is expressed in KIT negative gastrointestinal stromal tumor, the ones which are related to PDGF-alpha abnormality. Occasionally other tumors also could be positive, such as a chondroblastomas, and a few epithelial tumors which may be focal positive for this particular marker. An example of the staining, these are hematoxylin using photomicrographs of the gastrointestinal tumor, with some epithelioid features, low and high power. And on the right side you can see the staining for DOG1 in the cytoplasm and cytoplasmic membrane. Similar to KIT, CD117, the reaction for DOG1 is usually diffuse and the so-called all or none phenomenon. Focal staining of these tumors is very uncommon.

I'll give you an example of how it may prove to be very useful, here is a biopsy of a gastric ulcer. The patient was bleeding and the clinician felt that there was an underlying large mass protruding through the mucosa and causing this ulcer and bleeding. And this biopsy, it's very difficult to see anything different except the inflammatory cells and so forth, but paying attention to this particular area in the middle of the center of this particular biopsy, there are some nondescript cells which could potentially be an epithelioid gastrointestinal stromal tumor; and hence, with that clinical differential diagnosis we performed a KIT, or CD117. In that area you can see is clearly negative. And the epithelial nature of that area is much better seen on this KIT staining. However, the staining for DOG1 clearly demonstrates that the tumor is actually an epithelioid gastrointestinal stromal tumor, which proved on the resection of the tumor to be the case.

SATB2

I'm going to go to the next category of these markers, and the one that I found very useful many times is SATB2, this is special AT-rich sequence binding protein 2. This also was discovered by protein expression and profiling. It is expressed normally in epithelial lower-GI tract, colon or occasionally appendix. The great of majority of colorectal carcinoma is a stain positive, and the marker, of course, is a nuclear marker. Unrelated, but interestingly, it is also expressed a number of osteosarcomas. As you know, these carcinomas normally don’t have too many markers. So for our part it's going to be useful for us. A particular use of this SATB2 is in those rare cases of colorectal carcinoma which are negative for CK20, or sometimes even for CDX2. And that includes the rare medullary type of colorectal carcinoma, and occasional high-grade neuroendocrine colorectal carcinoma, which look like a small-cell carcinomas. These two may be CK20 negative, but in order to prove their colonic origin, SATB2 would be very helpful.

Here on the top you can see two low and high-power magnification of this particular distal colon tumor. It looks like a small-cell tumor. On the bottom left you can see cytokeratin staining, which has a typical staining pattern for a small-cell malignant tumor, a pointy type of a staining. Next to that in the middle, synaptophysin, it shows that this is indeed a neuroendocrine tumor; however, cytokeratin 20 next to that is completely negative. Now the question, is this a metastatic tumor, is actually colorectal in origin. So here is SATB2 low power and high power showing a disuse nuclear staining universally for that particular marker, extremely helpful in these kinds of occasions.

This next slide shows an interesting case of sarcomatous carcinoma of the lung, or the so-called carcinosarcoma of lung, also known as metaplastic carcinoma of the lung. The epithelial component of the tumor not shown here was a fully differentiated squamous cell carcinoma. The mesenchymal component was a sarcoma with focal evidence of osteoid formation, as you can see with the red arrow there. And that area made us to believe that maybe the rest of the sarcomatous elements are also an osteosarcoma. And using this marker you can see that almost all the tumor cells are positive for nuclear expression of SATB2, including that area which looked like an osteoid. So not having too many markers for osteosarcoma, this may be very helpful for that differential diagnosis, unrelated to colorectal cancer.

SF-1

The next protein is SF-1, it is a steroidogenic factor 1. This is a member of nuclear hormone receptor family, just like the other that we know, like ER/PR, HER2, et cetera. Important for differentiation of steroidogenic tissues, and is expressed as you expect in adrenal cortex, and adrenal corticol neoplasms. We have plenty of markers for adrenal cortex and cortical neoplasms which are fairly specific, so we really don’t need that for that particular differential diagnosis, unless in unusual circumstances. But this is also expressed by gonadal sex cord stromal tumors, as these tumors are also mainly steroidogenic. Where this is most helpful is actually in some of the granulosa Sertoli-Leydig cell tumors with the unusual morphology, which may or may not be the same for the common forms of the markers, such as inhibin and calretinin. The hematoxylin and using the slides on the top show a cystic granules, a cell tumor of the ovary, which was negative for inhibin. That was unusual. The calretinin was positive focally. In order to rally this diagnosis we use the SF-1, and the staining, as you can see in the nuclei of many of these tumor cells.

Brachyury

My favorite marker is this marker, brachyury. And I have no idea where the name is coming from. If somebody knows I would be happy to hear from you. It is a transcription factor of T-box family. It is important for the development of notochord, and is not expressed normally in any cell or tissue. And you know notochord is the one which is because of the spinal cord, and whatever, the spinal column and what is around it. And that’s why this particular marker is a great marker for most of the chordomas. Almost all of them are stained chordoma for brachyury. And it is also particularly helpful in unusual variants. And as you know, chondrosarcomas are sometimes in differential diagnosis of chordomas, but chordomas positive for epithelial markers, chondrosarcomas are not; therefore, in that differential are less helpful, but it could be helpful when you have an usual form of chordomas that one does not think are chordomas, or there is no clinical evidence to think about it. In addition to the chordomas, it could be positive in some epithelial tumors. The staining, however, in those few epithelial tumors is very focal; whereas, the nuclear expression in chordomas is generally is universal.

The hematoxylin used in slides on the top of a paraspinal tumor on T4, and frankly, by the hematoxylin using a stained slide, and the location, you were thinking of either hepatocellular carcinomas or even a renal cell carcinoma with that morphology, and neither of them actually reacted for the usual markers; however, as you can see, cytokeratin staining was positive, so it was an epithelial tumor. I don’t know what prompted us to use brachyury in this case, but I believe it was probably the x-ray findings suggested that. And here is the nuclear staining for brachyury in this usual form of chordoma, supporting that clinical, and finally the pathological impression.

Now, if you want to see some unusual ones, look at these two slides up there, which are actually a metastatic chordoma in lymph nodes, and without any clinical history it would be unlikely that anybody can think about them, because it doesn’t have the usual morphology of chordomas. And chordomas in lymph nodes are not that common an entity. Our understanding with knowledge of clinical history proved that yes, these tumor cells are basically positive for brachyury and they are metastatic chordomas. The interesting thing is that in the original description of brachyury we said is negative in dedifferentiated chordomas, and we have had at least two examples of these dedifferentiated chordomas which proved to be positive for brachyury, so again, it is an extremely useful marker for that differential diagnosis when it arises.

SALL4

Another useful marker protein is sal-like protein 4, also known as sall4. This is a transcription factor important for embryogenesis, and that’s why nuclear expression seen in my embryonic tissue. In adult tissue, however, it is not present except for the gonadal germ cells, and that’s the only place you find them. Tumor-wise, the nuclear expression is seen in most malignant germ cell tumors. Please remember that it is not expressed in benign germ cell tumors, much mature teratomas. Now, there are some high-grade carcinomas that may show positivity for sall4 occasional cells, and that is probably for the embryonic stem cells type of reactivity of sall4, as you know, is involving embryogenesis, so it is not usual to think that this marker could stain some embryonic stems cells in some high-grade tumors. It is not present in really low-grade tumors.

Now, I share with you this particular case, which as very difficult, as the core small biopsy of an abdominal mass in a young individual, I believe it was a male. And it shows some connective tissue, what appears to be inflammatory cells here and there. And on the top part of it one can see areas which look like clusters of cells, and these clusters of cells could be anything from lymphoreticular cells, to carcinoma cells, to sarcoma cells, et cetera. Oral markers for lymphoreticular and epithelial cells were negative. Since gastrointestinal stromal tumor was in the differential diagnosis we used CD117. And many of these cells stained, as you can see. As you know, CD117 or KIT is also a marker for germinomas in general, ceruminomas, and dysgerminomas, which are keratin negative. They are also D240 positive. But sall4 positivity of the nuclei, as you can see in the right lower corner, I identified as actually tumors of the germinoma origin.

NKX3

Another marker which has been very useful also in the differential diagnosis of tumor is NKX3, it's a big family. It's it is a prostatic tumor suppressor gene. It is a transcription factor located on chromosome 8. It plays an important role in prostatic development. It is down-regulated in most prostatic carcinomas; however, it its actually nuclear expression can be seen in more than 95% of prostatic carcinomas. It's a very good marker for prostatic carcinoma. Rare expression is reported in some androgen driven breast tumors, occasionally carcinomas of the breast have been shown to be positive for this one. None of them normally are in differential diagnosis of prostatic cancer, fortunately. We have the hematoxylin in the slides a poorly differentiated prostate cancer in a bladder, as you can see. It's very difficult to say is it epithelial or prostatic in origin. Prostatic specific antigen, or PSA is a fairly specific marker for prostatic carcinomas, but it may not be as sensitive. It may be focal staining of a particularly poorly differentiated tumors, and in a small biopsy; therefore, it could be potentially negative.

The NKX3 however is normally universally expressed in tumors of the prostate origin regardless of the degree of differentiation. And is also very helpful in metastatic areas, such as this case, which was a rectal tumor, and showed like a poorly differentiated carcinoma in the rectum, which in a male I always tell a resident to think about an extension of the primary prostate to rectum, and of course, this kind of poorly differentiated tumor could be of colonic origin. This was CK20 negative, also SATB2 negative. PSA was positive, as you can see, but NKX3 positivity was basically universally staining all the nuclei of the tumor cells. It's an extremely helpful marker in that differential diagnosis.

The next category of the case example includes those which we take advantage of protein loss, or under expression in tumor to come up a differential diagnosis. To start with, this is one of the most useful ones, which is INI1, integrase interactor 1.

INI1

This is a tumor suppressor gene product. The nuclear localization is seen in all normal cells. There is loss by gene mutation or chromosomal deletion in a few tumors, and those are the ones which we can use this particular marker very effectively. It is lost in all malignant rhabdoid tumors by mutation, whether the rhabdoid tumor is renal, extrarenal, it doesn’t matter. It's a very good marker. The other tumor that this proves extremely helpful is atypical teratoid rhabdoid CNS tumors. And these tumors are uniformly negative for INI1. This also can be lost in some epithelioid sarcomas by deletion, and occasional loss in some other tumors, which is usually is focal loss in these tumors, and not the universal loss, as you can see it in the rhabdoid tumor, or teratoid rhabdoid tumors.

This particular example is a malignant rhabdoid tumor which was not renal, it was a mediastinum, and you can see the loss of a nuclear reactivity for INI1. Your internal control, which are the endothelial cells of capillaries, and also some host lymphocytes are positive; whereas, the tumor nuclei are completely negative. A good example of when the lost antigen helps in the differential diagnosis.

BAP1

Another one of these is BAP1, which is BRCA1 associated protein 1. This particular marker is a suppressor gene product, and again, is a nuclear stain. Loss of expression is either by somatic or germline mutations. Somatic mutation is not uncommon in melanocytic tumors. And also where we use it usually is in malignant mesotheliomas. This is particularly helpful in fusion cytology with the understanding that the sensitivity is not 100%. I would say about half of these proven malignant mesotheliomas, whether in tissue or in cytology, may be negative. Occasional expression is also found in some tumors such as clear cell renal cell carcinomas. And I'll give you an example of this particular marker in fusion cytology. From the cell block you can see that those are normal cells are showing nuclear and cytoplasmic reactivity for calretinin. And then when you use your BAP1 you can see the nuclear loss which favors malignancy. But I can tell you that this is not as clear staining. You can have some areas that there are cytoplasmic staining for this particular marker which should not consider as evidence because cytoplasmic marker is not valued as a true loss of the nuclear antigen. So it is a nuclear loss of this particular antigen which is useful. In the absence of that, of course, we cannot be positive about it. It is somehow helpful, but it is not a very sensitive particular marker at this time, at least with the antibodies which now available.

I'll also share you a little interesting case, as we can see in the hematoxylin and using this slide, this was a solitary fibrous tumor (SFT) of pleura. This is a very large mass, and on the right side of it you can see a number of glandular structures. The question was whether this is an adenocarcinoma there, or is a malignant mesothelioma involved in there. And of course, solitary fibrous tumors, the marker for that is STAT6, which we're going to shortly discuss, but we can see that those glandular structures in the lower right are negative. Then, you see that these are mesothelioma or not, calretinin was positive in these, but BAP1 positivity universally note, these glands showed me that this is really a reactive process and not a neoplastic process. And as you can see, every cell is positive, including the stromal cells.

FLI-1

And our case example related to abnormal protein products, such as the fusion proteins, these are a large group of markers. And we're going to start with slide one, which is friend leukemia integration 1, it's a ETS family of transcription factor, nuclear expression in normal endothelium as your control. Expressed in more than 95% of EWING sarcoma PNET by fusion of EWS and FLI-1. It's also expressed in a great majority of endothelial tumors. And there's a weak expression in a few other tumors. Sometimes when you’re dealing with a small biopsy, such as this, this was a neck tumor in a patient who already has a diagnosis of vascular malignancy of unknown nature, that came to us with this particular tumor, which is not very well preserved. And you can tell you that when dealing with endothelial markers, CD34 is really needed a specific, not sensitive enough for this particular tumor. CD31 is to some extent specific, but also reacts with some hematopoietic cells, so; for example, if there is inflammatory exudate or lymphocytic exudate in there you can find that to be positive and it makes the interpretation of results fairly difficult. Factor 8 related antigen is more specific, but unfortunately, is not as sensitive. The reason I'm mentioning that, that as in this particular case with that history, the other markers were not helpful. But had sort of a pointed staining for factor 8 related antigen that those arrows will show you next to nucleus of many of these cells, and we have described this entity many years ago in one of our IHC Atlases that I share with you this photograph. As you can see, there is paranuclear staining, pointed pattern for the factor 8 in high-grade angiosarcoma. But the availability of the FLI-1, one can use this one and it's a clean as it could be, nuclear staining. You can see in almost all the tumor cells in this particular small biopsy, which was otherwise very difficult to interpret the results.

As I mentioned, in addition to endothelial tumors, a PNET can be positive, EWING and PNET can be positive for FLI-1 in more than 90% of the cases. And this is particularly useful in the areas which are not commonly PNETs are seen; such this one in the wall of the uterus. And positive for FLI-1 confirmed that diagnostic impression, which was also confirmed by CD99, et cetera. Molecularly, of course, these can easily confirmed by molecular techniques.

The next one that we mentioned shortly was STAT6, is a signal transducer and activator of transcription 6. It's a somatic fusion of NAB2 and a STAT6 gene. In more than 95% of solitary fibrous tumors this can be seen. It is also expressed in some differentiated and some dedifferentiated liposarcomas, and occasionally other soft tissue tumors. Interestingly, in solitary fibrous tumors the reaction is limited to the tumor nuclei, whereas in other tumors it may be in both cytoplasm and the nucleus.

Here is an example of a solitary fibrous tumor of the actual orbit, and are all the immunohistochemical stain for that particular tumor was CD34, which is universally positive. Now it's STAT6, you can see a nice nuclear staining for almost all the cells in this, confirming that diagnostic entity.

TFE3

One more of these particular useful markers is TFE3, transcription factor binding IGHM enhancer 3. This again, is a nuclear expression in tumors with TFE3 fusion. Interestingly, it is expressed in some of the rare renal carcinomas with Xpe11 translocation. And is also expressed in a small subset of PEComas. The main use of it is in alveoli soft part sarcomas in more than 95% of them. As you know, alveoli soft part sarcomas do not have so far a very reliable immunohistochemical marker. Usually the diagnosis is by exclusion. The availability of this marker has increased our efficacy of diagnosing this entity. The expression of this particular protein is diffuse in alveoloar soft part sarcomas, and also in those rare cases of Xpe11 renal cell carcinomas.

Let me share with you two cases. One of them was actually a primary alveoloar soft part sarcoma in bone, which is not a common finding. And the staining for TFE3 supported that particular diagnosis among the whole host of other markers that we use to identify this, first we tell you is an epithelial malignant tumor which was shown that it is really a sarcoma. TFE3 positivity proved that this is an alveoloar soft part sarcoma; although, it is a very difficult differential diagnosis.

The next case is lung excision of a tumor biopsy in a patient who already had the diagnosis of alveoloar soft part sarcoma. And the alveoloar soft part sarcomas, as you can see, are very epithelioid, and one has to rule out carcinoma of the lung, which is more common, of course. And as you can see in the left lower corner, the tumor is clearly negative for cytokeratin, but as the lung tissue is positive. With that clinical history, TFE3 showed that tumor cells are uniformly positive. The nuclear staining confirming that metastatic alveoloar soft part sarcoma in the lung.

TLE1

TLE1 is a transducin-like enhancer of split 1. This one also was discovered by gene expression profiling. In more than 90% of synovial sarcomas that had these SYT/SSX fusion is seen, the nuclear reaction is a strong and diffusing tumors. Less common and weak expression seen in the solitary fibrous tumor and hemangiopericytoma (HPC) cytomas, and occasional expression in epithelioid sarcomas. It was interesting, when this, like any other marker which is described, first was it was supposed to be a very, very specific marker for synovial sarcomas. And then a larger and larger studies showed that that is not true. A number of other tumors can be positive; therefore, in such it is the use of all these markers that we discussed in a panel. Of course, addition of these markers to the panel helped us tremendously because we didn’t have these before several years ago. This is an example of a synovial sarcoma with a hemangiopericytoma pattern. These tumors can be totally positive for cytokeratin and EMA. And normally they are also positive for BCL2 and CD99. So that’s where the question came that whether this particular one is a synovial sarcoma. And TLE1 positivity is seen in almost all the nuclei of the tumor cells. So this marker has helped me personally a lot in some of the unusual tumors and unusual locations, particularly in the soft tissue of the younger adults.

That was the good news. The bad news is this is an inguinal tumor in a lady, which by hematoxylin and inhibin and looks like smooth muscle type of tumor, and we confirmed that actually, by staining positive for h-caldesmon and desmin, so there was no question that this a leiomyosarcoma in that area. But as you can see, TLE is positive in a number of the nuclei of the tumor cells, not universally, but a number of them. And of course, that has to be taken into consideration, that this particular marker is not truly a specific for synovial sarcomas.

IDH1

Part of the mutated protein products, IDH1 is the isocitrate dehydrogenase 1. Mutation are all at codon: 132, the antibodies against the mutation is specific protein. IHC is reported actually by some that is more sensitive than sequencing the gene. Mutation is common in low-grade gliomas, and that’s where we have a problem with non-neoplastic gliosis. Those are always negative. And the diffuse gliomas, as you know, sometimes in a small biopsy are very difficult to differentiate from gliosis, and this particular marker helps in that differential diagnosis. In fact, some of the higher grade gliomas can also be positive for IDH1, and those are the ones which can do—have a better outcome, actually. And that’s one reason that IDH1 now is routinely performed on all the gliomas, regardless of differentiation. And here is an example of the low-grade diffuse astrocytoma, positive for IDH1, nuclear and cytoplasmic. And here is a higher grade tumor, which is negative, in the lower corner.

Conclusion

I’d like to close by telling you why we do immunohistochemistry as opposed to molecular technique where most of these particular abnormalities are actually genomic abnormality. So first of all, similar to molecular techniques that use for differential diagnosis of these tumors, the genomic abnormalities are also shared. It's not specific for a specific tumors, as we always thought it was; therefore, the protein for that also in us. So immunohistochemistry with these antibodies is almost as sensitive as a specific as the molecular technique, not always but sometimes is actually as sensitive, but definitely as specific.

The turnaround time is extremely important, because this allows us to do these tests much faster, and of course, is going to be much, much cheaper. But the very important issue is availability of tissue for doing these tests. And as you know, that the molecular tests require a considerable amount of tissue or tumor volume for testing; whereas, immunohistochemistry a few tumor cells can give us some clue about the line of differentiation. And finally, to us pathologists, immunohistochemistry is more familiar that molecular technique, and also molecular techniques are not always widely available in a number of institutions, and they have to be sent out, et cetera, et cetera. So overall, this immunohistochemistry can be very helpful, but is not going to be replacing molecular in a number of circumstances.

What are the limitations? Most genomic alterations you have to remember do not have protein product markers. On the other hand, proteins can be expressed without gene abnormality. And as I mentioned, questionable results have to be confirmed by genomic assays, so that’s a given. Also, as we discussed already, consider overlap exits in protein specificity. Variation in antibody sensitivity and specificity, each clone may have its own sensitivity and specificity so we have to be careful. I also want to caution you when you’re reading the paper of those papers which are studies of these markers are based on tissue microarrays. As you see in many of two tumors actually express the antigens focally; therefore, a small sample on a tissue marker may be negative, and falsely negative, and you can get larger samples of the tumor may be positive, so then their specificity of this particle markers become less as what is reported in the literature.

So, the final message for this particular part is the immunohistochemistry for this genomic derived antibody should be used always in a panel. It's no different than any other markers that we use for a diagnosis, so these are helpful, important, but again, they should be used in a panel, not as solitary agents.

Q&A

Does SATB2 plus medullary type CRC include both MSI unstable and stable tumors? And the second part of this question is, do you have any information regarding MSI status in the CK20 negative tumor you presented?

DR. NADJI: I can easily tell you no on both questions. I don’t not have that. That tumor is, by the way, not very common, and I'm sure some gastrointestinal pathologist can answer that better. So I don’t have any experience with MSI in these kind of tumors. And that’s about it.

Is the heterogeneity of the tumor a limitation?

DR. NADJI: It is always a limitation. And that’s why I brought up the issue of why false negative, particularly in a small biopsy such as tissue marker, may not be very useful in evaluating a new antibody for its sensitivity. The larger the sample, of course, the better. And however, these markers are the same as other markers that you have seen in the past, can be focally expressed. And focal expression of these markers, along with a number of other ancillary markers in a panel can be helpful. In isolated case, just for a single marker, obviously, the heterogeneity is going to cause a problem, and that’s a fact of immunohistochemistry.