Expires: February 3, 2023
Sounak Gupta, M.B.B.S., Ph.D., is an Assistant Professor of Pathology in the Department of Laboratory Medicine and Pathology at Mayo Clinic in Rochester, Minnesota.
Contact us: email@example.com.
Hi, I’m Matt Binnicker, the Director of Clinical Virology and Vice Chair of Practice in the Department of Laboratory Medicine and Pathology at Mayo Clinic. In this month’s "Hot Topic," Dr. Sounak Gupta discusses tissue selection for molecular testing, specifically tissue metrics and requirements, which differ based on test platforms. I hope you enjoy this month’s Hot Topic, and I want to personally thank you for allowing Mayo Clinic the opportunity to be a partner in your patient’s health care.
Thank you for the introduction.
I have no disclosures.
The topic of this presentation is selecting tissues and tissue considerations for molecular testing. We will discuss tissue metrics and requirements, which differ based on test platforms. And it is important to keep in mind that the details of tissue metrics change over time as they continue to evolve with the technology. We will review a number of examples and touch on a number of additional tissue considerations.
There are a number of details to keep in mind when selecting tissue for further molecular testing: the source of the tissue to be used for molecular testing as well as how it is collected and processed, including what fixative is used, the total amount of tumor in that tissue, and the percent tumor nuclei. Each of these will be discussed in more detail in a few slides.
Sources for obtaining tissue for molecular testing include various surgical procedures including small biopsy specimens, cytology preparations and even liquid biopsies for solid tumors. The tissue obtained from any of the 4 first procedures can be processed in a number of different ways. While fresh or frozen tissue often leads to the best quality nucleic acid, it is fairly rare in routine practice and difficult to obtain and store long term. Fixed tissue is by far the most common source of nuclei acid for molecular testing of solid tumors. A number of fixatives and fixation protocols are routinely used, and are also variables that need to be considered. Once tissue is fixed, there is more than one way to get nucleic acid from a tissue block (for example, taking cores from the block or scraping from unstained slides).
This slide highlights the variability in the amount of tissue that may be available. The examples of H&E slides and tissue blocks range from large tissues from resection specimens on the left, to different sized biopsies in the middle and right of your screen. Once the tissue is in this format, the molecular workflow really begins.
In the slide review process we are essentially looking at tissue for 2 requirements for each test: first, the amount of tumor tissue and second, the percent tumor nuclei. This is based on each test’s performance metrics, specifically the limit of detection. So, for example, our Lung Cancer-Targeted Gene Panel with Rearrangements, LNGPR, currently requires 30 nanograms of DNA. Generally, this can obtained from 0.4 x 0.4 cm of tissue scraped from 5-micron thick unstained slides or, approximately 5,000 cells from cytology specimens. As far as the tumor percent, the metric for this test is at least 20% tumor nuclei.
Looking at this table of test requirements, FISH is something of an outlier. Because we are not extracting DNA, we can actually tolerate a smaller tumor percent as the intact architecture helps identify the cells to be tested, so the cell count, rather than amount or cellularity, is the more relevant metric for FISH. In terms of NGS-based assays the different methods shown have widely varying requirements for tumor percent and tissue amount in order to obtain the necessary quality DNA for each test.
The tissue amount and cellularity requirements correlate with the amount of DNA needed for the test. So it is useful to keep in mind that normal human diploid cells have approximately 6 picograms of DNA. Often a test protocol states how much DNA is needed in nanogams and not how much tissue is needed in cm or mm or how many cells, so a little conversion is helpful. One nanogram equals 1000 picograms and at 6 picograms per cell, that is approximately 167 cells. So, if you have about 5,000 cells, that is approximately 30 nanograms; in comparison, with 500 cells, you really only have about 3 nanograms of DNA. Keeping in mind that no extraction method is 100% efficient. The cellularity and amount of tumor are not a perfect predictor of how much nucleic acid an extraction will yield or the quality of the extracted DNA (as there are other factors involved), but given some biologic limits, it is an useful and important metric.
The tissue used for testing is further refined by the process of macrodissection, including optimizing tumor percentage. With a large piece of tissue, such as the picture on the left, it is easy to overlay the unstained slide on top of the circled H&E slide, line-up the unstained tissue, and scrape enough tissue from the selected areas, thus avoiding the surrounding non tumor tissue. However, when you get a smaller tissue specimen, like the picture on the right, it can be problematic to obtain the amount of tumor tissue needed, even though we can and do scrape multiple slides from the same block.
As we have stated the amount of tissue required depends on the particular test and is related to the amount of nucleic acid needed for that test. This is not the only metric that impacts the quality and quantity of DNA; we will discuss others later. As mentioned before, the amount of DNA needed for different tests is evolving with technology, making this something of a moving target.
What if there is not enough tissue? Or DNA? The problems with a small amount of tumor and low cellularity in a sample, is that you may not get enough DNA for the test, which results in a higher chance of failing the test’s quality metrics. Also, a limited amount of template DNA can lead to sequencing artifacts and false positive or false negative results. Aside from problems with the PCR reaction itself, something to keep in mind on a biological level is when you only have a small sample of cells, you have limited sampling of the tumor itself and, as we know, there is heterogeneity within tumors, so you are less likely to see the spectrum of alterations that are truly present. Now, let’s move on to some examples.
This resection tissue has fantastic tissue cellularity and a high tumor percent, so this is likely a specimen that would yield enough nucleic acid and would be acceptable for molecular and cytogenetic methods.
When looking at small biopsies it is not just the size of the tissue, but the cellularity and tumor percentage that matter. In these examples, the biopsy on the left has high cellularity and high percent tumor nuclei and is likely acceptable for most molecular and cytogenetic methods, despite the fact that the tissue itself is rather small, only a couple of millimeters. The biopsy on the right, however, while being similar in size actually does not have many nuclei present within the tissue and even fewer tumor nuclei, and so, it would be inadequate for most cytogenetic and molecular methods.
Cytology preparations should be evaluated for the same metrics. This particular smear has excellent cellularity and numerous tumor cell clusters resulting in a high percentage of tumor nuclei and this would be acceptable for most methods.
In comparison this cell block has very poor cellularity, the bulk of what we are seeing on the slide is simply red blood cells and, of the nucleated cells that are present, the vast majority are white blood cells, with only very rare tumor cells present. So, not only do you have a low cellularity overall, but you also have a low tumor percent, so that this specimen would be inadequate for molecular and cytogenetic methods.
As the previous slides illustrated, tissue size and cellularity are not the only metrics we need to consider. Tumor percentage or percent tumor nuclei is a critical metric as well. As I have said before, what is acceptable varies by the test method. What may be sufficient for one assay may not be sufficient for another. The main reason we worry about this is that if the tumor percent or the percent tumor nuclei is too low, but you have sufficient nuclei to yield enough nuclei acid to pass the test quality metric, you risk a false-negative result.
Here is an example of a larger section of tissue; however, there is not much tumor. The percent tumor nuclei is low. I would say that this is probably less than 5% tumor nuclei. While this would be inadequate for most molecular tests, this would actually be sufficient for FISH testing.
Here is another large piece of tissue from a resection specimen, and while we do not see the large areas of lymphoid nuclei as in the previous slide, there is quite a bit of fibrotic stroma, and we simply do not have a large number of tumor nuclei present. Macrodissection can help optimize to about 10%, which would be enough for some molecular methods and FISH.
Now, in this resection specimen the bulk of the cells that are present on the slide are neoplastic, thus it has a very high tumor percent, approximately 90% or greater, making this sample acceptable for most any method.
With small biopsies, we frequently run into inadequate tumor percentage. The biopsy itself is actually a very good size, at least 1 cm, if not more in length; however, the majority of the tissue is actually benign liver. There is only a small area that contains tumor. So, if you look at the overall biopsy, the percentage of tumor nuclei is quite low. We can try to macrodissect out an area, which would then have a much higher tumor percent. But, it would be an extremely small amount of tissue so it would have to be present on many unstained sections to accumulate the amount tissue necessary for testing. So, this specimen may be inadequate for some molecular techniques, but with over 100 tumor cells present it would be acceptable for FISH.
This slide is an example of a much better small specimen. This tissue has a little bit of stroma, and there is actually a little bit of necrosis. However, the vast majority of the nucleated cells are tumor cells, and this would be acceptable for most tests.
The same metrics apply when you look at cytology preparations. In this particular case, the cells in this block are mostly tumor cells, approximately 90%, and there are a lot of them, so this would be acceptable for most methods.
In this cytology preparation there are a lot of cells present, for fairly good overall cellularity, but the tumor percent is low, in the <5% range. The cells in the tumor clusters are larger and eye catching, as they have a fair amount of surface area. However, they are relatively few in number compared to the number of benign lymphocytes in the background. And, it is that ratio–the number of tumor cells to the number of background cells–that matters, not the size of the cells. So this would actually be inadequate for most tests and, if there are more clusters on the slide, then this would be sufficient for FISH since there are less than 100 cells in this particular field.
Shifting gears, aside from the amount of tumor present in the samples, there are other tissue considerations to keep in mind, such as fixation. By far, the most common fixative is formalin, but a number of other formulations of formalin, as well as various acids, heavy metals, and decalcification solutions, are frequently used.
Formaldehyde as 10% neutral-buffered formalin is one of the most common tissue fixatives, and even though DNA from formalin-fixed, paraffin-embedded (FFPE) specimens works well for a number of molecular methods, formalin does result in DNA damage. The average fragment of DNA that can be obtained from FFPE is about 200 to 300 base pairs, at best. You can try to improve the quality, being very careful about excessive time between surgery and fixation, and also trying to optimize the duration of fixation that is (avoiding inadequate or excessive fixation).
Nevertheless, we generally describe formalin as being acceptable, same with zinc-buffered formalin, and ethanol for cytology specimens (as alcohols are a very good cytology fixative, and do not cause significant DNA damage). Decalcification solutions are quite variable and generally decalcification is not acceptable. B5 and Bouins are examples of fixatives that, while they have important histologic effects and uses in various subspecialties, they do create problems for molecular testing and are not acceptable.
Other tissue considerations include recognition of potential inhibitors. Even with optimal size tissue and good cellularity and tumor percent, this does not guarantee that every specimen will work. We know that there are inhibitors present in tissue that can inhibit the PCR reaction or compete with the substrate. Good examples of natural inhibitors that we run into are melanin and calcium.
When it comes to melanin, it is easy to identify on H&E, on unstained slides, and in the block itself. However, even among heavily melanotic specimens, it is difficult to predict which will work and which will be inhibited and fail. Because of this unpredictable variability, they are generally considered worth trying.
Crush artifact is another issue, not so much because of what it does to the DNA or the PCR reaction, but because it makes it difficult to determine tumor percent, making it difficult to determine if the specimen is adequate. For FISH assays, this can be problematic, as it is difficult to define the area of individual nuclei in order to count signals. Generally, crushed tissue is inadequate for testing.
For necrosis, the effect is somewhat variable and it depends how much of the tissue is necrotic. A small focus of necrosis is usually acceptable, while extensively necrotic tissue (as pictured above) is inadequate. In this case, this issue affects the quality and quantity of the nuclei acid that can be obtained.
Cautery is another artifact that makes it difficult to determine tumor percent and is problematic to score FISH signals. Additionally, DNA that is obtained from such a specimen generally does not amplify well and would be considered inadequate.
As we have already mentioned, decalcification causes damage to the DNA above and beyond that caused by formalin, rendering these samples inadequate for NGS-based testing. The issue of decalcification is further complicated by the number of different decalcification solutions and procedures that are currently used in practice.
In summary, the success of the molecular and cytogenetic testing depends in large part on having an adequate amount of tumor (thereby sufficient DNA), having enough tumor percent, and minimizing potential tissue issues.
Thank you for your attention.