I received this very good question by email yesterday. Please feel free to send in your questions, and I’ll post the answers here.
Q. When a squamous cell carcinoma is designated as “poorly differentiated”, what other parameters/tests are performed to determine the tissue of origin?
A. Differentiation, for those of you who have just joined us, is a quality of tumors that has to do with how much the tumor cells resemble their tissue of origin. Well-differentiated tumors are composed of cells that closely resemble their tissue of origin, whereas poorly-differentiated tumors are composed of cells that have little resemblance to their tissue of origin. Anaplastic tumors are the least differentiated of all: they show no resemblance to their tissue of origin.
This concept is important for a couple reasons. First, the degree of differentiation of a tumor often has a bearing on prognosis. Well-differentiated tumors generally carry a better prognosis than poorly differentiated tumors. Second, when a tumor is totally undifferentiated (anaplastic), you have to resort to special tests in order to figure out its origin (is it a squamous cell carcinoma, an adenocarcinoma, a lymphoma, a sarcoma, etc.).
Back to our question: when you have a poorly-differentiated squamous cell carcinoma, how do you know it’s a squamous cell carcinoma (as opposed to an adenocarcinoma, for example)? If the tumor is poorly-differentiated, that means there are still some morphologic features (albeit few) that reveal the squamous nature of the tumor. If you look carefully, you should be able to find some of these features, which would then point you towards the diagnosis of squamous cell carcinoma.
Two clear-cut features of squamous cell carcinoma are intercellular bridging and keratin pearls (there are other, “softer” features indicating a squamous cell origin, but we’ll focus on the more definitive features). Intercellular bridging is a term describing the special connection between the epithelial cells of squamous epithelium (it’s not present in glandular epithelium). By light microscopy, you can see little horizontal hair-like connections between the epithelial cells in both normal squamous epithelium and in malignant squamous epithelium. Look closely between the epithelial cells in the above image of a squamous cell carcinoma. See the little connections between the cells (they look like little zippers connecting the cells)? Those are intercellular bridges.
Keratin pearls are whorl-shaped accumulations of keratin made by malignant squamous cells. In normal squamous epithelium, keratin lies in a nice flat layer on the epithelial surface. In malignant squamous epithelium, the tumor cells can grow in any direction they want, and so the keratin they produce often gets trapped inside the tumor, forming pink, glassy, spherical masses. There’s a nice keratin pearl at about 2 o’clock in the image above.
So, if you have a tumor that you think might be a squamous cell carcinoma, but the cells aren’t showing clear squamous cell differentiation, look closely for epithelial bridging and keratin pearls. If you find either of these, it’s a good bet that you’re dealing with a squamous cell carcinoma. There are other little clues that point towards other types of tumors (like adenocarcinoma, or melanoma, or sarcoma), but that’s for another post.
Sometimes, you’ll get a tumor that shows no defining morphologic features whatsoever – no interepithelial bridging, no keratin pearls, no signs of differentiation along any other cell line. In these tumors (which would be described as “anaplastic”), you need to use a secret weapon to figure out what the cells are: immunohistochemistry. In this technique, you use a reagent consisting of antibodies against specific components of cells (there are lots of these specific components: squamous cells have cytokeratin in them, muscle cells have actin in them, etc.). These antibodies are bound to a substance that appears brown under the microscope. The concept is simple: you apply the reagent to the tissue in question, allow it to bind to the cells, then wash off the excess reagent and look at it under the microscope. If the tumor cells appear brown, that means they possess whatever antigen the antibodies in your reagent are directed against, and that information can help you figure out what type of cells your tumor contains.
So, if you have an anaplastic tumor, you might choose to apply immunohistochemical stains for cytokeratin, actin, and CD45 (an antigen present on lymphoid cells). If the cytokeratin stain comes back positive (brown), and the actin and CD45 stains come back negative, the tumor is most likely a squamous cell carcinoma. There are tons of immunohistochemical stains for all different types of cells. Generally, these stains are pretty specific for one cell line, but it’s not always totally straightforward. Some tumors stain positive for markers from cell lines other than their cell of origin, and some tumors show only weak staining with the stains that are supposed to be nice and positive. So you really need to use a panel of a bunch of different stains to make the best diagnosis.
The bottom line, then, if you have an undifferentiated appearing tumor: look for little morphologic clues (like keratin pearls) first. If you see few or no clues, then your next step is immunohistochemical staining, which will almost always reveal the origin of the tumor. If that doesn’t work, there are still other tests you can do – like cytogenetics or molecular diagnostics – and we’ll talk about those in future posts.