By Rodney T. Miller, M.D., Director of Immunohistochemistry   

 

Squamous differentiation is a common finding in tumor pathology. As any experienced surgical pathologist knows, many   different types of tumors can have “squamoid” morphologic features. However, appearances can be deceiving, and many   “squamoid” tumors are not true squamous proliferations. When I was a pathology resident, I was taught that there are only   two reliable morphologic features that reflect squamous differentiation: identification of intercellular bridges and tumor   cell keratinization. Alas, these criteria are also imperfect. I have seen a number of urothelial carcinomas that show   beautiful intercellular bridges (the first of which I misdiagnosed as a squamous carcinoma in my younger days), and when I   look too hard at high magnification, I seem to be able to find them in many different tumors (just like “melanin pigment” in   tumor cells). In some cases distinction of true keratinization from “pseudokeratinization” can be a challenging task, and I   have seen tumor necrosis misinterpreted as tumor keratinization on multiple occasions (secondary to necrotic cell debris   that I like to call “necrokeratin”). This month, we briefly review immunohistochemical features of squamous differentiation,   in the hopes that this information will aid readers in recognizing true squamous tumors and distinguishing them from mimics.

 

One characteristic feature of squamous carcinomas is strong and diffuse expression of high molecular weight cytokeratin (clone 34E12) (assuming of course, that your high molecular weight stain has good sensitivity). I have never seen a   squamous carcinoma that has lacked this finding. It is not specific for squamous carcinoma, as other tumors may also   demonstrate this as well. However, if a tumor does not express strong and diffuse high molecular weight cytokeratin, you are   safe in concluding that it is not a squamous tumor.

 

Expression of low molecular weight cytokeratin (keratins 8 and 18, using   clones 5D3, CAM5.2, Zym 5.2, or equivalent) is highly variable in squamous tumors. Many of the more poorly-differentiated   squamous carcinomas express low molecular weight cytokeratin extensively. In contrast, the highly differentiated squamous   carcinomas (that are morphologically so obvious that you would never think of having to resort to immunostains) may express   little or no low molecular weight cytokeratin. The same can be said for expression of cytokeratin 7 in squamous tumors. As   such, another clue to a squamous tumor may be a marked predominance of expression of high molecular weight cytokeratin over  low molecular weight cytokeratin.

Expression of cytokeratin 5 (or cytokeratin 5/6) is analogous to the expression of high molecular weight cytokeratin in   squamous tumors. With few exceptions, squamous tumors typically express strong and diffuse cytokeratin 5. An important point   to keep in mind is that tumors besides squamous carcinomas may also express strong and diffuse cytokeratin 5, including   mesothelioma, cutaneous basal cell carcinoma, cutaneous adnexal tumors, myoepithelial tumors, and some high grade breast   carcinomas (see the July 2005 edition of this newletter for more information on these so-called “basal-like” or “basaloid”   breast cancers). As such, strong expression of cytokeratin 5 by itself is not sufficient to interpret the tumor as squamous   in nature. Many different kinds of tumors express focal or patchy cytokeratin 5, and to my knowledge this finding does not   have any particular significance (although in cases with squamous differentiation, cytokeratin 5 highlights these areas very   nicely).

 

Nuclear expression of p63 is also a characteristic feature of the large majority (probably 80-90%) of squamous tumors.   Again, it is not specific, as other tumors may also show this finding, including cutaneous basal cell carcinomas, cutaneous   adnexal tumors, myoepithelial tumors, and urothelial carcinomas. In contrast to their frequent expression of cytokeratin 5,   mesotheliomas are negative for p63.   When a squamous tumor is in the differential diagnosis, I always order both cytokeratin 5 and p63. In the appropriate  diagnostic setting, strong coexpression of cytokeratin 5 and nuclear p63 is virtually diagnostic of squamous   differentiation. There are a few other things that may coexpress these two markers in a strong and diffuse fashion,   including basal cell carcinoma, cutaneous adnexal tumors, and myoepithelial tumors, so if those possibilities are in the   differential diagnosis, additional markers may need to be employed.

 

In summary, squamous tumors virtually always show strong and diffuse expression of cytokeratin HMW, and they may show a   predominance of high molecular weight over low molecular weight cytokeratin. Additionally, they typically show strong and   diffuse expression of cytokeratin 5 and nuclear p63. Use of these markers should be of value in the identification of   squamous tumors when they enter into the differential diagnosis of a problem case.

Date of last revision: June 2005.