I am the founder of bone/soft tissue sarcoma pathology and dermatopathology discussion groups on Facebook. I also actively use Twitter to discuss these topics. This is an actively-growing compilation of my answers and comments to frequently asked questions on these topics. DISCLAIMER: This is not medical advice or an official consultation. These are just general comments that represent my personal views and experiences on these topics. I may be wrong about some of these things. Some of my views may (and hopefully will) change over time. My experiences DEFINITELY will change with time. Today I might say "I've never seen a case of XYZ behave like that." By next month, I may have seen a case of XYZ behave exactly like that. These comments are for educational purposes only and should never be used for patient care in place of peer-reviewed references and your own clinical judgement. If you are diagnosing a patient, you are their doctor and you are responsible. The same rule applies to me when I am diagnosing my own patients, too! Best wishes, Jerad M Gardner, MD
Special thanks to my former medical student Julie M Youngs for compiling, organizing, and editing my comments into one place here on this page. This would still only be a great idea in my mind without her hard work! :-)
General Comments and Questions
- For comprehensive Dermatopathology textbooks, McKee is my favorite because it has many great pictures, but it is expensive. Weedon is also great. For introductory dermpath books try "Practical Dermatopathology" by Ronald Rapini, "Dermatopathology by First Impression" by Christine Ko, and "Requisites in Dermatology: Dermatopathology" by Dirk Elston.
Q: What is on your differential for rhabdoid cell-type neoplasms?
A: Here are some rhabdoid things I have seen in skin and soft tissue: melanoma, carcinoma, rhabdomyosarcoma, myoepithelioma or myoepithelial carcinoma (co-expression of S100 plus keratin), proximal type epithelioid sarcoma, malignant extra-renal rhabdoid tumor, poorly differentiated synovial sarcoma, and rhabdomyoma. There are others but those are just a few off the top of my head that I have personally seen.
A: The "3 Ms" of Herpes viral cytopathic effect are: multinucleation, nuclear molding, and chromatin margination (a dark rim of chromatin around each nucleus, but pale "ground glass" center of the nucleus...like a kid outlined the nucleus with a big purple crayon, as Dr Dina Mody always likes to say).
Q: What is necrobiosis?
A: It is a bit confusing. I regard necrobiosis as just a fancy name for necrotic/degenerated collagen. But really it is basically a type of necrosis. We use the term necrobiosis when we are specifically referring to certain diseases like rheumatoid nodule and necrobiosis lipoidica.
Q: What do you use to take histology pictures?
A: I have an Olympus BX51 scope with 2x, 4x, 10x, 20x, and 40x. Love it. I have a mounted Olympus camera but I much prefer the modified adapter that Martin microscope makes that allows a cheaper and better SLR mass production Canon camera to be used on your scope, which can be found here. Or as I often do these days, I just take pictures with my iPhone freehand. Here is a video tutorial on how to do this at home. I edit pics on my iPhone with the Enlight app for white balance and watermarking.
General Questions About Working with Patients and Clinicians
Q: How do you deal with difficult surgeons (or other doctors in general)?
A: The best long term approach is persistent polite conversations about the issues, focusing on the literature and medical knowledge and not attacking the personality of the other doctor. Over time, most have come to respect that I generally know what I'm talking about and that I care about their patients and about making a good diagnosis just as much as they do. That usually has bought me credibility and then with mutual respect we can form a professional relationship over time that lets us speak openly about difficult issues or points of contention without causing offense. Persistent respect, even when not deserved, gives the best chance for us as pathologists to make a positive impact on our clinical colleagues over the long term and that is what is best for our patients. It takes hard work, but it's worth it. Of course, there are some times when we have to get angry and loud or assertive in order to prevent a serious error from occurring. But I try to avoid those situations if I can. And also doctors are humans and suffer from the same problems anyone can have...personality disorders, poor communications skills, bad tempers, or bad attitudes. Sometimes people are just jerks and no matter what you do, you can't help them see the light or behave appropriately. Fortunately, the vast majority of doctors (of all specialties) that I've known are NOT like that.
Q: Do you see many misdiagnosed sarcoma cases?
A: What I've learned in dermpath and sarcoma path... EVERYTHING is clinically a "cyst" (until it turns out not to be). I have a box of cases that were clinically cysts but histologically all sorts of weird tumors or other things. One day I'll make a slide show of "it's not a cyst." I work to teach my students and residents to stop blowing off every skin colored bump as "just a cyst." In our survey study of over 200 DFSP patients, the majority of patients were misdiagnosed clinically as a cyst, often delaying treatment for several years.
General Questions on IHC
Q: When do you use MDM2 for diagnosis?
A: For MDM2 (or CDK4, CPM1, etc), here is when I like to order it:
- (1) a retroperitoneal mature fatty tumor with no atypia at all (if MDM2 is negative, it is likely retroperitoneal lipoma)
- (2) a soft tissue tumor that looks like lipoma but is very large (greater than 10 cm), or has recurred multiple times, or has possible atypia that is not clear cut enough to be diagnostic of WDL/ALT
- (3) a retroperitoneal mass that looks like myxoid liposarcoma (that is such an unusual location for MLS, and WDL in retroperitoneum can look exactly like MLS), and
- (4) looks like a pleomorphic lipoma but in the wrong location, wrong age, or is very large.
In summary, MDM2 is very useful for difficult cases where WDL is suspected but where the histologic features are not perfect. I prefer the FISH to the IHC. During fellowship, we ordered FISH for MDM2 more than any other soft tissue FISH. Adipocytic tumors are much more common than other soft tissue neoplasms.
Q: I see scattered single cells with pale staining for pancytokeratin. They don't look like carcinoma cells. What are they?
A: You may be seeing mast cells that are non-specifically picking up pancytokeratin (AE1/AE3) stain. I have encountered this phenomenon relatively often with a variety of different antibodies, most commonly pancytokeratin AE1/AE3, desmin, MART-1, and myogenin (the cytoplasmic mast cell staining with myogenin can mimic nuclear tumor cell staining from low power!). I encounter this not only in lymph nodes but also in skin biopsies and tumors that have background mast cells. An even scattered distribution of the cells, fried egg cytology, and fine granular, less intense staining are all helpful clues that these are just mast cells and not real staining. I've heard several suggested explanations as to why this phenomenon occurs, but I'm not sure which is correct. I use the same antibody clones every day, but only see this in a subset of cases. I also haven't noticed it on negative controls before (although I do not routinely use those in my practice). And for many antibodies I've never noticed this phenomenon.
Q: For pleomorphic spindle cell tumors in skin, some use multiple keratins, some pathologists use one keratin and either p63 or p40, and others do no keratins and just p63 or p40 alone. Why are there so many different approaches?
A: This is an area with lots of different styles of practice. Some like multiple keratins. Tim McCalmont had a nice JCP editorial a few years ago advising against that. Honestly, I think just p63 or p40 without keratin is enough to exclude Spindle SCC. But I've seen a few SCC that were keratin positive but negative for p63/p40. And I've seen a few adenocarcinoma metastases to skin that were keratin positive and obviously p63/p40 negative. So I still do both. But for me, p63 or p40 is essential as Spindle SCC often loses keratin expression. And other random malignancies can also express keratin but not p63 or p40.
Alveolar Soft Part Sarcoma
General comments on appearance and identification:
- Notched "bite cells" are typical of ASPS, as well as cellular areas lacking alveolar architecture having a very zellballen appearance and looking an awful lot to me like paraganglioma. People don't usually talk about ASPS and paraganglioma being in the differential with one another, but I think you should always at least think about the possibility of a non-alveolar area of ASPS whenever you think you have a paraganglioma or see zellballen pattern (obviously once you see alveolar spaces it clearly can't be paraganglioma). I actually just recently learned that one of the old obsolete names for ASPS was "malignant nonchromaffin paraganglioma."
- The abundant, dense pink cytoplasm of ASPS can call to mind pleomorphic rhabdomyosarcoma, epithelioid soft tissue tumors, and RCC or other carcinomas, and doing immunostains to exclude those is probably a good idea in general if you have any doubt about the diagnosis. I've personally never seen a pleomorphic rhabdomyosarcoma with an alveolar architecture like ASPS has. Pleomorphic rhabdo is usually strikingly pleomorphic and floridly mitotically active, way more so than ASPS would be. And I often hear people mention alveolar rhabdomyosarcoma whenever I show a case of ASPS during a lecture. Take a quick second to google image search alveolar rhabdomyosarcoma. Seriously...go do it right now. See the difference? Alveolar RMS is always a ROUND BLUE cell tumor. It does make alveolar spaces usually, but the cytology is completely different. Because it is a translocation sarcoma the nuclei are going to be uniform and monotonous, not pleomorphic (for this same reason ASPS usually has monotonous nuclei albeit with big central nucleoli also).
Q: How do you confirm a diagnosis of AFX?
A: If ugly spindle cells are found in the dermis of old, sun-damaged skin, I do S100, p63, and pancytokeratin. If all are negative, I call AFX, provided it does NOT invade the subcutis. If it invades subcutis I classify it as superficial undifferentiated pleomorphic sarcoma since once in the subcutis, a lesion has metastatic potential (Fletcher calls these pleomorphic dermal sarcomas, if I understand correctly, but I think that name is confusing). But If it is a shave that is transected at the base, as it almost is, I will call AFX and add a comment that UPS could look and stain identically, and examination of excision specimen to exclude subcutis extension is essential. I don't need any positive marker for AFX/UPS as none of them are specific anyway. I just make sure I've excluded other possibilities.
Q: For pleomorphic spindle cell tumors in skin, some use multiple keratins, some use one keratin and either p63 or p40, and some do no keratins and just p63 or p40 alone. Why so many different approaches?
A: This is an area with lots of different styles of practice. Some like multiple keratins. Tim McCalmont had a nice JCP editorial a few years ago advising against that. Honestly, I think just p63 or p40 without keratin is enough to exclude spindle SCC. But I've seen a few SCC that were keratin positive, but negative for p63/p40. And I've seen a few adenocarcinoma metastases to skin that were keratin positive and obviously p63/p40 negative, so I still do both. But for me, p63 or p40 is essential as spindle SCC often lose keratin expression. And other random malignancies can also express keratin, but not p63 or p40.
Atypical Lipomatous Tumor/Well-Differentiated Liposarcoma (ALT-WDL)
- Sharon Weiss taught me that if you are having difficulty finding atypical cells in a suspected case of ALT/WDL, look around the blood vessels. They often are located there.
Q: What is the difference between ALT and WDL?
A: They are identical tumors histologically and molecularly, but with very different outcomes. If they are located in the retroperitoneum, mediastinum, or spermatic cord/inguinal canal, I call them WDL. They will usually recur multiple times and eventually kill the patient (if they don't die of something else first... this is usually a disease of elderly and recurrences/dedifferentiation may happen over a period of 10-20 years). If they are located in the subcutis or in deep muscle of an extremity they usually are cured by excision and only rarely have dedifferentiation. We call them ALT in those locations because death from disease in those sites is rare. By itself, ALT/WDL does not ever metastasize unless it first dedifferentiates. But multiple bulky recurrences in the retroperitoneum can eventually compromise vital structures by mass effect and cause mortality.
Q: Do you ever use CD10 in diagnosing AFX?
A: CD10 stains everything. Seriously, if it is spindled, it will probably stain with CD10. Normal dermis often has abundant background staining. In the world of soft tissue tumors, I find CD10 next to vimentin in regards to it's usefulness (ie - almost completely worthless...see this satire video about vimentin: https://www.youtube.com/watch?v=UDnp14nnNC4 ). It may play an important role in other specialties like gyn or GU or hemepath, but not in soft tissue in my opinion. I know it stains AFX, but contrary to previous reports, I have often seen it stain SCC and spindle melanoma as well as cutaneous PEComa, angiosarcoma, and many other things.
Basal Cell Carcinoma
Q: Do you distinguish adenoid-type BCC in your reports?
P: The adenoid pattern of basal cell carcinoma has mucin-filled spaces, giving a pseudo-glandular and pseudo-cribriform pattern. This can be very striking and cause confusion with adenoid cystic carcinoma or adenocarcinoma. I usually don't mention "adenoid type" in my pathology report unless it is a consult case that was sent to me by a pathologist who was specifically concerned about the gland-like spaces. For dermatologists, there is no clinical significance and I don't want to add potential confusion by bringing up a not often mentioned subtype.
Q: How is basosquamous carcinoma different from BCC?
A: Basosquamous carcinoma (a term I use only rarely) is a carcinoma with areas that look just like typical SCC and other areas that look just like BCC. Sometimes collision tumors of BCC growing next to an SCC can mimic basosquamous carcinoma. Typical BCC often has some areas of keratin pearl formation or pink squamous cytology (especially common in ulcerated BCC). In these cases, I just call BCC and may comment that there is some squamous metaplasia (mostly for the benefit of the Mohs surgeon so he is not confused on frozen section when he sees squamous pearls but the diagnosis was BCC)
Calcifying Aponeurotic Fibroma
- round cells and calicified cartilage islands with fascicles of fibroblasts
- mostly found in acral sites
- benign, but often recur locally
- usually seen in children, but I've also seen them in young adults where they can have abundant cartilage overgrowth and resemble chondroma of soft parts
Q: What stains do you use for cellular neurothekeoma?
A: For me, if it looks like cellular neurothekeoma I feel comfortable making diagnosis once other things (like melanocytic) are excluded. So usually if s100 is negative I don't pursue additional markers like NSE, PGP 9.5, NKI/C3. All of those are sensitive but very non specific so I don't feel like they help me very much in confirming the diagnosis.
Chondroid Syringoma, Mixed Tumor
Q: What mutation is characteristic of these tumors?
A: Myoepithelial and mixed tumors of salivary origin often have PLAG1 mutations whereas those of soft tissue tend to have EWS rearrangements. The literature is mixed (no pun intended) regarding these lesions in the skin. Some suggest they are more akin to salivary ones molecularly, whereas others say they are like the soft tissue ones. I think more studies are needed to help sort this out more clearly.
Q: What are the distinguishing features of chrondromyxoid fibroma?
A: Mark Edgar taught me that the nuclei for chrondromyxoid fibroma look like pennant flags. That little pearl works for me. The zonation into a hypercellular periphery and hypocellular center also helps. They can also occur in soft tissue (saw one in sinus recently that was eroding adjacent bone, but not based in bone).
Q: Can you distinguish chrondrosarcoma from normal cartilage by morphology?
A: Grade 1 chondrosarcoma can look nearly identical to normal cartilage or benign cartilage tumors. That's why I still often struggle with diagnosing them even though I had special training in sarcoma pathology. Cartilage tumors are one of the more difficult types of bone tumor that I deal with in my job.
Q: Do you use chrondrocyte nuclear features in diagnosis?
A: Chondrocyte nuclei are often hyperchromatic and small, but the purple appearance of the surrounding cytoplasm and the pale outer lacunar space can give the impression that the entire cell is one big dark nucleus. I do often find chondrocyte nuclei hard to assess for atypia. So when making a diagnosis of chondrosarcoma, many other criteria have to come into play. Also, reactive cartilage, as in a fracture callus for example, can look wildly atypical. I've seen callus misdiagnosed as grade 3 chondrosarcoma before. Beware!
Q: Can dermatofibromas occur on the face?
A: DF can definitely occur on the face, but I do always think twice about the diagnosis, and if it is an elderly sun-damaged face and there is some atypia I will sometimes do p63, pancytokeratin, and S100 protein just to put my mind at ease. I usually do no inmunostains on dermatofibroma, but this is one time I may make an exception.
Q: How do you distinguish between induction and basal cell carcinoma over a dermatofibroma?
A: I have a hard time believing that most of those are true BCC. And we don't have any perfect IHC marker or molecular to prove it. If it looks like superficial BCC and is over a DF, I almost by default believe it is induction unless it is obviously a collision tumor clinically. I've never seen a case that convinced me of true BCC over DF or that clinically behaved like BCC. If it exists, I think it must be exquisitely rare.
Q: Can DFs transform into DFSP?
A: Dermatofibromas do not turn into DFSP. I firmly believe that. I have seen very, very, very rare cases where dermatofibroma turned into a different type of sarcoma, but it definitely was not DFSP. Dermatofibroma usually looks different than DFSP under the microscope, but in some cases it can mimic DFSP and vice versa, especially on a small biopsy. Immunostains and DNA molecular tests can be used to sort out the diagnosis in those cases (if the pathologist realizes that the tumor is trying to trick him!)
Dermatofibrosarcoma Protuberans (DFSP)
Q: When do you do molecular testing (FISH, RT-PCR) to confirm t(17;22) COL1A1-PDGFB gene fusion in DFSP?
A: I usually only do FISH for t(17;22) in cases where: 1. I suspect DFSP but am not certain or 2. I am pretty sure something is NOT DFSP but I still feel worried that I'm missing it (so I do the FISH to sleep well at night in that case). If I had a small biopsy from a skin mass that looked like "fibrosarcoma" (cellular and herringbone) but no obvious DFSP component and IHC were negative for S100 and keratin, I would probably call it "sarcoma, not further classified", raise differential diagnosis of fibrosarcomatous DFSP in the comment, and state that further classification will be based on complete excision specimen. I've never seen a case personally where there was no DFSP component at all on the complete excision specimen.
Q: How do you identify fibrosarcomatous transformation?
A: I make a diagnosis of fibrosarcomatous transformation based only on the histologic features, namely fascicular "herringbone" growth often coupled with increased cellularity, and sometimes with increased mitoses and larger (but still uniform) nuclei. These areas often lose expression of CD34, but I do not use the lack of CD34 as a criteria for diagnosing fibrosarcomatous transformation. In fact, I rarely use CD34 when diagnosing DFSP since it is usually a straightforward diagnosis on H&E.
Q: How do you interpret positive beta-catenin stains when diagnosing fibromatosis?
A: For desmoid fibromatosis, only NUCLEAR beta-catenin counts as positive. Cytoplasmic staining is non-specific. This is why I dislike using beta-catenin stain for spindle cell tumors. The high background non-specific cytoplasmic staining is easy to confuse with true nuclear staining and can lead to overdiagnosis of fibromatosis. I have seen this happen multiple times. I feel that beta-catenin adds further diagnostic confusion far more often than it actually helps.
Digital Papillary Adenocarcinoma
- A biphasic appearance with solid sheets of oval or round cells with tubules, ducts, and/or cysts lined by columnar cells is classic. Papillary figures may be prominent or may be subtle. These usually have little nuclear atypia in my experience.
- An adnexal tumor in the digit should be considered digital papillary adenocarcinoma until proven otherwise.
- A recent paper by Suchak et al suggested we drop the "aggressive" from the name as these are not significantly more aggressive than many other carcinomas. About 25% metastasize and in their series only 1 out of 23 patients died of disease.
Epithelioid Hemangioma (Angiolymphoid Hyperplasia with Eosinophilia)
- Epithelioid hemangioma (EH) can be quite epithelioid and cellular. It can even have sheet like areas. Try SMA to highlight pericytes and CD31 to highlight endothelial cells. In cellular vascular proliferations/tumors I find that combo really helps me see the pattern better. Often those sheetlike areas are not actually solid but rather are composed of many vascular channels packed closely together so that their lumina are compressed. It's a useful trick. Also the well formed channels with very plump epithelioid cells is another useful clue for EH. Cytoplasmic vacuoles are often seen and can lead to confusion with epithelioid hemangioendothelioma (EHE). The presence of plump endothelial cells arranged into well formed vascular channels with lumens usually argues against EHE (which has cords/nests with minimal or no lumen formation in most cases...exceptions exist, of course).
Q: Can fibrosarcoma be identified by herringbone fascicular pattern?
A: Lots of things have herringbone fascicular pattern and most of those entities are NOT fibrosarcoma (examples include MPNST, synovial sarcoma, dedifferentiated liposarcoma, cellular schwannoma, mesoblastic nephroma, melanoma, etc). Fibrosarcoma is a vanishingly rare entity and diagnosis of exclusion. About the only time I ever use that word is: (1) when it is part of the name of a specific entity (e.g. dermatofibrosarcoma protuberans, sclerosing epithelioid fibrosarcoma) OR (2) for fibrosarcoma arising in DFSP OR (3) infantile fibrosarcoma. And both 2 and 3 can be confirmed by molecular testing if needed. Aside from that I advise my trainees to never use the term "fibrosarcoma."
Q: Is CD117 alone sufficient for a diagnosis of GIST?
A: I NEVER use CD117 alone to diagnose a GIST... always use it in conjunction with S100 protein. The reason is that melanoma is often CD117 positive and can metastasize to the GI tract and mimic a malignant GIST. The co-expression of CD117 and S100 is a clue to think of metastatic melanoma. DOG-1, of course, would be negative in melanoma so that would solve the problem too. A very good pathologist who is much smarter than me told me that he missed this once (he called it GIST but it was actually melanoma). I've never forgotten that story!
Q: What IHC stains do you use in the identification of glomus tumor?
A: Collagen IV shows each individual tumor cell invested in a layer of basement membrane. This is characteristic of glomus. PAS stain also nicely highlights this basement membrane.
Q: What are the features of glomus tumors?
A: Normal glomus apparati/bodies (aka Canals of Sucquet-Hoyer) are arterio-venous anastomoses that supposedly play some role in thermal regulation. They are most frequently found in the dermis or subcutis on the fingers or toes, often near the nails. The center of the glomus body is a small vessel lined by bland endothelial cells. The uniform small round cells surrounding the vessel in concentric layers are the glomus cells. These are a form of pericyte, modified smooth muscle cells that exist around small vessels.
Glomus tumors are benign neoplasms made of these glomus cells. They recapitulate the features of the normal glomus apparatus and show numerous small round cells arranged in concentric rings around blood vessels. Glomus tumors are also most commonly seen on the digits near the nail. Glomus cells are positive for smooth muscle actin but usually negative for desmin, cytokeratin, and vascular markers like CD34 and CD31.
Q: What causes Hailey-Hailey?
A: HH disease is caused by mutation of the ATP2C1 gene resulting in an abnormal calcium channel ATPase pump found in the smooth endoplasmic reticulum. Darier disease has a similar abnormality caused by mutation of ATP2A1. The mnemonic that my derm residents have taught me to remember which gene is for which disease is "Oh how I want 2 C Hailey's comet!" Some use the clever term "channelopathy" to describe these calcium channel pump related diseases... I like that name.
- I just want to point out one confusing thing. There are a handful of different types of hemangioendothelioma (epithelioid HE, kaposiform HE, ES-like (pseudomyogenic) HE, Dabska type HE, retiform HE, composite HE). Although they are all blood vessel tumors that have "intermediate" behavior (not always acting like full-blown cancer but also not fully benign) and all have "hemangioendothelioma" as part of their name, they are otherwise unrelated. They have different molecular/genetic causes, look different microscopically, and present and behave differently in patients. So they are a group of different diseases that share the HE name rather than one disease with several different subtypes. I know this is confusing for most pathologists even, so if it seems hard to grasp, you are in good company! But it makes a big difference sometimes as some of these (like EHE) tend to behave much worse than the others.
- I try to never use the term "hemangioendothelioma" unless a tumor fits clearly into one of the well described subtypes: epithelioid, epithelioid sarcoma-like (pseudomyogenic), retiform, Dabska type, or kapsosiform. Otherwise, it becomes a wastebasket term for all vascular tumors that we find challenging to diagnose!
Q: What other methods can be used to identify Kaposi besides HHV-8?
A: When you don't have HHV-8, a useful trick that favors Kaposi is if the spindle cells (not just the slit-like vessels, but the trickling spindle cells in the dermis) are positive for CD31 or ERG. Other mimics of Kaposi may have thin vessels but either lack spindle cells or have associated spindle cells that are fibroblasts, but not endothelial.
Q: What are the stages of Kaposi?
A: There are 3 stages of Kaposi sarcoma based on their clinical appearance, and each has somewhat different microscopic features:
- 1) Patch stage: infiltrative interconnected vascular channels dissect dermal collagen; scattered plasma cells are present, a useful clue
- 2) Plaque stage : infiltrative vascular channels are still seen; also, there is usually a spindle cell proliferation infiltrating the dermis and invading/destroying eccrine glands
- 3) Tumor stage: nodules with intersecting fascicles of spindle cells; blood-filled slit/sieve-like spaces are present between spindle cells
All forms of Kaposi sarcoma are positive for HHV-8 virus which can be detected by the HHV-8 (LANA-1) immunostain.
Q: Does Kaposi sarcoma ever regress?
A: Kaposi sarcoma can sometimes resolve/regress spontaneously to varying degrees if the immune system situation improves. So if a patient comes in newly diagnosed with HIV/AIDS and is not yet treated with HAART anti-HIV drugs, and that patient has Kaposi sarcoma, his Kaposi sarcoma might partially or even mostly regress and go away just by treating the HIV and allowing the immune system to start working again so it can fight off the HHV-8 virus (at least for a while). Kaposi is very different from most cancers in that way.
Q: Kaposi sarcoma vs acroangiodermatitis (severe stasis) - what if I don't have HHV-8?
A few tricks to help when you don't have HHV-8:
- Look for plasma cells in the infiltrate. Almost always present in Kaposi sarcoma (KS)
- Look at the eccrine coils. If spindle cells and vessels are infiltrating eccrine coil, it is strongly supportive of KS. It's not always present, but when it is, it's very helpful to make diagnosis of KS.
- Do a regular vascular marker, like CD31 or CD34 or even D2-40. In KS, the SPINDLE CELLS will be positive for endothelial markers (usually lymphatic markers too), but in acroangiodermatitis the spindle cells are actually fibroblasts and myofibroblasts so they will be negative. CD34 is tricky as it can stain dermal spindle cells in normal skin, so it's hard to use in subtle early patch stage Kaposi but can work well for more cellular nodular lesions.
- Remember that angiosarcoma can mimic Kaposi sarcoma. Severe atypia or pleomorphism is a warning sign to think of angiosarcoma (although Kaposi can sometimes have severe atypia). Something that looks like Kaposi sarcoma but is not occurring in the usual clinical scenario is also worrisome for angiosarcoma (e.g. - looks like Kaposi but is on the sun damaged scalp of elderly patient who doesn't have HIV? That is probably angiosarcoma until you prove otherwise. Get HHV-8 if at all possible and/or get a consult from an expert!)
- More tips about angiosarcoma on my YouTube video about vascular lesions: https://youtu.be/E5QQ3ZOJhoM
Q: Can leiomyosarcoma be epithelioid in appearance and on IHC?
A: Leiomyosarcoma can be epithelioid. Focal actin staining can be positive, but can also be seen in many other things and is not specific. Try staining with desmin. If desmin is negative, the only way I will call something leiomyosarcoma is if there is very strong diffuse actin expression.
- Dedifferentiated liposarcomas, despite being high grade, often have a longer course than other high grade sarcomas, with local recurrence being more of an issue than distant metastases in many cases. Unfortunately, when they are in the retroperitoneum, most cases will eventually be fatal due to their impossible location for complete surgical resection.
- I greatly prefer MDM2 FISH over immunostains when working up lipoma vs. well differentiated liposarcoma/atypical lipomatous tumor. It does cost more and take longer, but it gives me a discrete quantifiable ratio of MDM2 to centromere 12, allowing for more dependable assessment of true amplification. In general, I do not like it when one single immunostain decides whether a tumor is benign vs. malignant. Immunostains are very useful but they are not perfect; interpretation is not always straightforward.
Malignant Peripheral Nerve Sheath Tumor
Q: What is the S100 staining pattern for MPNSTs?
A: MPNST is negative for S100 in about 50% of cases and is only weak and patchy positive in the other 50%. Rarely ever is it strongly S100 positive, with the exception of epithelioid MPNST. Also, I don't make a definitive diagnosis of MPNST outside of one of these three settings: (1) sarcoma arising in a nerve, (2) sarcoma arising in a neurofibroma or other benign nerve sheath tumor, or (3) sarcoma arising in a patient with Neurofibromatosis 1.
Merkel Cell Carcinoma
Q: What are the staining patterns of Merkel Cell?
A: Although dot-like CK20 is the buzzword, it doesn't have to be dot-like, it can be diffuse. Also, other keratins like AE1/3 can have a dot-like pattern in Merkel. Neurofilament also works nicely as it stains most Merkel cells and is negative in most other neuroendocrine carcinomas.
Q: How do you distinguish Merkel from BCC using IHC?
A: If I think a case is BCC but just want to be sure it's not Merkel (i.e. low index of suspicion), I'll do CK20 and just sign it out as BCC if negative. If it looks classic for Merkel, I usually also just do CK20 alone. If positive, I sign it out as Merkel. If CK20 is negative, but I'm still pretty concerned about Merkel morphologically, I might pursue further with additional stains like CK7, synaptophysin, chromogranin, and neurofilament. NSE is very non-specific and I never use it. If 7 and 20 are both negative, be sure to do a pancytokeratin to prove that it truly is carcinoma (remember, CK7 + CK20 ≠ PanCK!). You might also consider doing S100 to rule out small cell melanoma (I see melanoma all the time and almost never have seen one that looked much like Merkel, but it's theoretically possible). TTF-1 is used by some, but I would urge caution. It is very sensitive for small cell carcinoma of the lung but it is not totally specific, as it can be seen in pretty much any neuroendocrine carcinoma, even away from the lungs. So if TTF-1 is negative (and even better, if CK7 is also negative), you can be pretty sure that it is not small cell lung carcinoma. If TTF-1 is positive, I would raise the possibility of metastatic small cell carcinoma in my report and then they can scan the patient and look for a lung mass. I have only seen one single example of small cell lung carcinoma metastatic to the skin, and it was from a palliative amputation in a patient with known stage IV disease. I asked a room full of hundreds of dermatopathologists (ASDP annual meeting in 2015) if any of them had ever seen metastatic pulmonary small cell carcinoma in the skin as the presenting sign of an occult lung primary, and only a couple of people raised their hand. Thus, I believe this scenario, although not impossible, is EXCEEDINGLY rare.
Q: How do you distinguish molluscum from myrmecia?
A: Molluscum usually doesn't occur on actual sites nor will it be as large or of as long duration as a verruca. On H&E, molluscum has viral bodies that are relatively uniform in size whereas myrmecia wart has viral aggregates of varying size and shape as well as large chunky purple hypergranulosis and a papillomatous architecture.
I get this question so often that I decided to make a video about it: https://youtu.be/G31xlSgYSlw
Q: How do you confirm MF in uncertain cases?
A: The absolutely most important thing there is to ask the dermatologist: (1) is there a proven history of MF or if not, then (2) does background skin AWAY from nodules look like patch or plaque stage MF. If so, then biopsy those areas and try to find more obvious MF. Cellular nodules of atypical T cells in skin can definitely be tumor MF but could be other things too like LYP, ALCL, etc. The history and clinical correlation is essential. Also beware as the clinician or chart may say "biopsy proven history of MF" from an outside hospital, but then once you actually review that material is just another nodule like you have now that was called MF without clinical correlation. We see this often.
Q: Do you see MF in children?
A: MF is rare in kids, but can happen. And when it does, it is often the hypopigmented form that leaves pale or white patches. BUT many other diseases can cause pale or white patches, and several of these can mimic MF under the microscope. Seeing pediatric dermatology is a great idea. A second opinion on the biopsy might be helpful. Oftentimes getting a few different biopsies over time (over months up to a couple of years, even) can really help sort out the final diagnosis. I know the idea of more than one biopsy for a young child is not pleasant (I have 3 daughters, all under the age of 4 years). But sometimes it helps. Discuss with your pediatric dermatologist and ask about a second opinion. Some early MF biopsies are just very difficult no matter how much of an expert one is.
Q: What are the characteristics of adult solitary myofibroma?
A: Myofibroma was originally described as multiple tumors occurring in young kids (myofibromatosis). But we now know they are often solitary and can commonly occur in adults too. Histology shows two zones: (1) myoid nodules which often have a bluish "pseudochondroid" appearance and (2) cellular areas between nodules composed of bland spindle cells with ectatic staghorn vessels. IHC is not usually needed. These are totally benign and rarely recur even if margins are positive. Some myofibromas can be quite cellular and have lots of mitoses but still behave well.
- There are three clinical forms: (1) a single skin or subcutaneous nodule, most common, often on the head or neck but can be anywhere, (2) multiple skin, subcutaneous, or bone nodules, with good prognosis, and (3) generalized with the internal organs also involved. The last has a poor prognosis and is usually seen in infants and kids, but is commonly seen in adults too.
- There are two morphologic zones in these tumors: (1) myoid nodules often with a “pseudochondroid” appearance (even though they are myofibroblastic, I think the nodules look very bluish and chondroid, which confuses many pathologists), and (2) cellular zones of bland spindle cells with staghorn vessels, usually between the nodules. These have "atypical features"--high cellularity, minimal myoid nodules, increased mitosis, and may have necrosis but no usually no pleomorphism. They may be confused with sarcoma, yet these "atypical" features have uniformly good prognosis (don't overdiagnose).
- IHC is usually not needed to identify these, but they are actin positive and usually desmin negative.
- There is no routinely useful positive molecular finding.
- Prognosis is generally good, as even with incomplete excision recurrence is rare. The exception is with internal organ involvement in the generalized form.
Q: Where are these tumors found?
A: These are essentially myxoid variants of undifferentiated pleomorphic sarcoma. Typically they arise in the subcutis (or deep muscle) on the extremities of elderly adults
Q: Do you use IHC in the diagnosis of myxofibrosarcoma?
A: I usually do not do any IHC for myxofibrosarcoma unless there is something unusual about it. IHC almost never helps in those cases, and the only other thing that really looks much like it is myxoinflammatory fibroblastic sarcoma, or perhaps myxoid areas of pleomorphic liposarcoma, but IHC doesn't help much with either of those.
Q: How do you describe myxoid liposarcoma morphologically?
A: I like "chicken feet" or "delicate branching/plexiform" vessels as descriptors, personally. I also point out to residents that the vessels of myxoid liposarcoma are very very thin, with a wall that is about one endothelial cell thick and a lumen that allows about one red cell through at a time. Many other things have branching vessel pattern, but not many have such delicate vessels.
Nerve Sheath Myxoma ("Conventional" Neurothekeoma)
Q: What is the relationship between nerve sheath myxoma (conventional neurothekeoma) and cellular neurothekeoma?
A: They are not related in my opinion. Nerve sheath myxoma (previously known as "conventional" neurothekeoma) is a benign nerve sheath tumor likely related to schwannoma. I currently only use the "neurothekeoma" term to refer to cellular neurothekeoma, which interestingly is not related to this entity nor is it even a nerve sheath tumor (unknown histologic line of differentiation...maybe myofibroblastic or fibrohistiocytic). The terminology here is very vexing and makes literature (even recent literature) challenging to interpret sometimes. Regarding IHC for this differential, I usually just use s100. It is positive always in nerve sheath myxoma but always negative in cellular neurothekeoma. Cellular neurothekeoma is also positive for a wide range of markers but they are all non-specific so I don't generally use them (MiTF, NKI-C3, NSE, S100 A6, etc).
Q: Are Wagner-Meissner bodies specific for diffuse neurofibroma?
A: Wagner-Meissner bodies are classically seen in diffuse neurofibroma, although I've seen them in other things, particularly neurotized nevus. I've also made a diagnosis of diffuse neurofibroma without WM bodies just based on infiltrative growth in the subcutis.
Q: What are the criteria for plexiform neufibroma?
A: Plexiform neurofibroma should only be diagnosed when the gross appearance of the mass has the classic "bag of worms" appearance. Even if there is low power multinodular appearance, I am very wary of diagnosing plexiform neurofibroma without the classic gross findings or a known history of NF-1. The reason for such strict criteria is that by diagnosing plexiform neurofibroma, you are essentially giving the patient NF-1, a heritable genetic disease with a significant risk of transmission to their offspring and a significant lifetime risk of developing sarcoma (MPNST). It's a benign tumor, but the implications for the patient are enormous. So I'm always very cautious here.
Q: Does IHC have a use in a cellular neurothekeoma diagnosis?
A: The long list of potential immunostains for cellular neurothekeoma are pretty much all non-specific, so I don't use them. Many of those will label histiocytic and/or melanocytic lesions too. I used to like MiTF and that's what Sharon Weiss used when I was in fellowship with her. But having now used MiTF a lot in Dermpath, I've seen it stain histiocyte nuclei all the time. New literature has shown this as well. So pretty much the only setting that I currently use MiTF for is to evaluate the intraepidermal growth pattern of a melanocytic lesion or to rule out subtle melanoma in situ in a background of solar lentigo and actinic keratosis. I do not trust MiTF for neurothekeoma diagnosis (nor do I trust it when evaluating for nodal or distant melanoma metastasis, by the way).
Q: How do you approach a diagosis of cellular neurothekeoma?
A: If it looks like one histologically (a dermal proliferation of histiocytoid or pseudo-Spitzoid cells arranged in cellular nodules with intervening fibrosis) and is negative for s100 (mainly to exclude Spitzoid melanocytic lesion, which it can closely mimic), then I usually feel okay calling it cellular neurothekeoma without any positive stains. Remember that cellular neurothekeoma commonly has mitoses, sometimes many, and may even have marked nuclear atypia or, rarely, atypical mitotic figures. Myxoid change is also commonly seen and is a useful clue.
Nuchal Type Fibroma & Gardner Fibroma
- Dense collagen and few spindle cells makes me think of nuchal type fibroma. Posterior neck in a young MALE patient is a good clinical setting for it. Gardner fibroma (associated with FAP colon cancer syndrome) looks very similar but usually presents in childhood.
- I think you can routinely and reliably make the diagnosis from 2x magnification in most cases without needing to look at high power for emperipolesis (I always do look anyway, but learning how to diagnose from 2x is a good skill). My former fellow loved to say "Pink and blue, baby!" because that is precisely what you want to see at 2x: sheets of big pink cells with multiple blue lymphoid aggregates. At higher power, you will see big histiocytes with large clear vesicular nuclei and prominent nucleoli, many plasma cells, and emperipolesis. S100 stains these tumors nicely and highlights the vacuoles containing inflammatory cells (these cells are intact and whole, unlike hematophagocytosis where the cells are being destroyed in the histiocytes and there is karyorrhexis), but I do not routinely do S100. This is an H&E diagnosis. RDD often resolves on its own usually within a few years.
Q: Is emperipolesis necessary for the diagnosis of Rosai-Dorfman?
A: I feel emperipolesis gets a lot more weight than it should. I don't require it for diagnosis of RDD. If I see alternating pink sheets of pale histiocytes and blue lymphoid aggregates from low power and then see large histiocytes with big vesicular nuclei, prominent nucleoli, and abundant pale cytoplasm at high power...those things are enough to call RDD for me. The histiocytes' nuclei are so distinct and unique that they confirm the diagnosis for me in most cases. If I'm unsure, I'll add A100 and hunt around for emperipolesis.
- Unlike neurofibromas, schwannomas rarely undergo malignant transformation. When they do become malignant, two of the patterns seen are (1) angiosarcoma arising in schwannoma and (2) small round blue cell sarcoma (Ewings/PNET like) arising in schwannoma.
Sclerosing Epithelioid Fibrosarcoma
Q: How do you confirm a diagnosis?
A: Sclerosing epithelioid fibrosarcoma is a serious diagnosis as about 40% of cases get metastases. Muc4 stain and FISH for EWS are helpful ways to confirm it.
Spindle Cell Hemangioma
- Spindle cell hemangioma is a benign vascular tumor that was formerly known as spindle cell hemangioendothelioma, but is now known to be totally benign. There are two zones: one zone displays large cavernous vascular spaces resembling cavernous hemangioma, and the other zone displays more cellular areas of both spindle and epithelioid endothelial cells bearing some resemblance to Kaposi sarcoma. Small clusters of endothelial vacuoles resembling miniature adipoctyes are often present in the cellular areas and are a very characteristic feature. These tumors usually occur on the distal extremity in the subcutis, and they may be multiple. Rarely, they may be associated with Maffucci syndrome (multiple enchondromas plus spindle cell hemangioma). To my knowledge, this case was solitary and not associated with Maffucci.
Spindle Cell Lipoma
Q: How do you distinguish the low-fat variant of spindle cell lipoma from fibromyxoid sarcoma and other translocation sarcomas?
A: These tumors are often very myxoid which raises lots of bad tumors in the differential diagnosis and prompts a consult. The vague palisading ("parallel arrays of nuclei") is a very useful feature. Low grade fibromyxoid sarcoma is always a great thing to think of in a case like this as it usually lacks mitotic activity and pleomorphism (translocation sarcomas are usually monotonous, not pleomorphic!). However, palisading, small size, focal adipocytes, ropey collagen, and a uniformly myxoid appearance present can distinguish spindle cell lipoma from LGFMS.
Spitz Nevus, Spitzoid Melanoma, & Everything in Between
Q: How do you deal with Spitzoid melanocytic lesions with atypia?
A. This entire area of dermpath is fraught with confusion and controversy. Here is my way of dealing with this difficult grey zone currently. I'm sure my approach will change over time, and I'm sure there are multiple appropriate ways to deal with these tough cases. I use the term "atypical Spitzoid neoplasm" for cases that are Spitzoid and atypical and where I am not totally certain the lesion is a benign Spitz nevus. I include a comment explaining that this is a difficult case and the differential diagnosis is atypical Spitz nevus versus a more aggressive lesion (if I'm really worried, I'll mention Spitzoid melanoma by name). I often do Ki-67/MART-1 double immunostain, molecular analysis, and/or FISH to help provide additional data points that push me in one direction or the other, but in the end, I usually still call it "atypical Spitzoid neoplasm" with a long comment. I often recommend re-excision and close follow up if I am. Some are in favor of doing sentinel lymph node biopsy on cases like this, but since even conventional Spitz nevi can involve regional nodes on occasion, I'm not sure what it really means prognostically to find an atypical Spitzoid neoplasm present in a lymph node. I'm still not convinced that means the lesion is melanoma just because the node is invovled. I interviewed Phil LeBoit for the ASDP a couple of years ago, and he gave me some great advice: If you are not sure it is benign, don't use the term 'nevus'. You can watch the video here: https://www.youtube.com/watch?v=uPvLykdBMrQ. I've altered my terminology on "grey zone" melanocytic lesions ever since giving that interview. I still try to favor benign or malignant when I can even in tough cases, but it seems ethically wise (and probably legally wise, at least in the USA!) to be honest when there is significant uncertainty about benign vs malignant melanocytic lesions. Disclaimer: I'm not a lawyer...I just try to use common sense and do the best job I can to take care of my patients.
Squamous Cell Carcinoma
Q: How do you determine between SCC and reactive changes in a lesion with some atypia and extensive involvement?
A: These cases can be hard to decide. I would do stains for invasive fungus, polarize the tissue for foreign bodies, check the dermis to make sure there is no underlying granular cell tumor - all could cause changes like this. Then I would call clinician to discuss. I had case like this before that looked very much like SCC but patient had extensive leg involvement. Clinical pics and discussion with dermatologist swayed us to favor stasis with reactive epithelial hyperplasia over SCC. They were considering amputation if we had called it SCC. Fortunately that was avoided and patient began to improve on conservative therapy. Really eye opening case!
Q: What features make you concerned about SCC?
A: My basic tip: if a lesion is a complex and endophytic squamous lesion on sun-damaged elderly skin, and if I see parakeratosis down inside cystic spaces in that setting and/or if I see abundant eosinophilic cytoplasm developing at the base, then I worry about SCC. And in that setting if I see a verrucous or cystic squamous lesion with some atypia but cannot see the base because it is a transacted shave biopsy, I tend to use terms like "atypical endophytic squamous proliferation" with comment that differential includes SCC and if the lesion persists, recurs, or doesn't respond to conservative management, or if clinician still concerned for malignancy, then to repeat the biopsy or excise.
Q: What is basaloid SCC?
A: Basaloid SCC is an CC that has a blue basaloid appearance. This is relatively common in the oropharynx or lung, but quite uncommon in the skin for some reason
Q: Can the FNCLCC grading system be applied to synovial sarcoma?
A: FNCLCC grading works well for many but not all sarcomas. Synovial sarcoma can indeed have very bland areas with low mitoses that would be technically FNCLCC Grade 1. But grade is not a very good predictor of behavior for synovial sarcoma. Better prognostic factors are size, depth/location, and presence of poorly differentiated (round blue cell or large epithelioid cell) components.
Syringocystadenoma Papilliferum (SCAP)
Q: Where are SCAPs found?
A: These and a variety of other benign adnexal tumors (and probably some true BCC, as well) have a tendency to arise within nevus sebaceus. They show dilated, branching cleft-like channels invaginating down from the epidermis. They are lined by a bland double layer of cuboidal to columnar epithelium. Papillary structures protrude into these glandular spaces. Numerous plasma cells are present in the cores of these papillae and in the adjacent dermis; this is very useful clue.
Tendon Sheath Fibroma
Q: What are the morphologic features of tendon sheath fibroma?
A: A classic example of fibroma of tendon sheath is a circumscribed nodule with a dense collagenous background, scattered bland spindle cells, and cleft-like vessels as well as cleft-like cracking artifact of the collagen. Older, mature lesions have these features, whereas "younger," more cellular lesions look very much like nodular fasciitis, a variant called "cellular fibroma of tendon sheath." If you see something like nodular fasciitis on an acral site, is is very probably cellular fibroma of tendon sheath instead.
Q: How do you distinguish trichilemmoma from true BCC and SCC?
A: Classic trichilemmoma features are seen with warty surface change, thickened epidermis with smooth bulging protrusion into dermis, clear glycogenated cells, and basal palisading. Desmoplastic areas show infiltrative nests and cords of clear cells. Infiltrative areas with extensive squamous metaplasia are also present. Taken out of context it is easy to see how these areas could be mistaken for infiltrative BCC or SCC. Recognizing the classic trichilmemmoma component is how to avoid this pitfall. Small shave biopsies from desmoplastic trichilemmomas on the face of elderly sun damaged adults are easily overcalled as BCC or SCC.
Q: Are trichoblastic carcinoma and basal cell carcinoma the same thing?
A: I used to think they were the same thing, but I've since seen two situations where I think the term trichoblastic carcinoma is appropriate: (1) a tumor that looks like trichoblastoma and has bland cytology, but also has infiltrative growth and/or aggressive clinical behavior, and (2) a high grade carcinoma that doesn't really look like typical BCC, arising out of what looks obviously like a benign trichoblastoma (sort of a carcinoma ex trichoblastoma if you will). To me, both of those situations clearly look trichoblastic in origin, clearly are malignant, and yet clearly are not BCC, thus I have nothing else to call them aside from trichoblastic carcinoma. I think my colleague and I have collectively used that term maybe 3 times since we both started practice 2 years ago, and we see a lot of odd consult cases aside from our everyday case load.
Trichoepithelioma & Trichoblastoma
Q: How do you distinguish trichoepithelioma from trichoblastoma?
A: The distinction is merely academic in my opinion, as both are benign follicular tumors. Also, in my experience it seems like follicular tumors exist on a spectrum and show overlapping features often. Rapini says it is easier to classify snowflakes than follicular tumors!
I tend to call it trichoepithelioma in these settings:
- when it is a relatively small superficial lesion
- when basaloid nests are small
- when there is abundant cellular stroma between nests with nice papillary mesenchymal bodies
I tend to call it trichoblastoma in these settings:
- when it is larger and deeper
- when there are large nests or sheets of basaloid cells
- when there is a trabecular/rippled pattern
- when the cellular stroma is less pronounced and papillary mesenchymal bodies are less obvious
Ackerman, if I understand his views correctly, believed that trichoepithelioma is merely a variant of trichoblastoma.
Undifferentiated Pleomorphic Sarcoma
Q: When do you diagnose undifferentiated pleomorphic sarcoma as opposed to high grade pleomorphic sarcoma?
A: On biopsy, I call “high grade pleomorphic sarcoma” as descriptive term when it is a pleomorphic spindled tumor with negative CK, S100, desmin, and (mostly) actin. But once the whole lesion has been completely resected and I can be sure there is no obvious line of differentiation histologically in the whole tumor (ie – no lipoblasts, well differentiated liposarcoma component, adjacent neural tumor, etc), then I call it undifferentiated pleomorphic sarcoma.
Q: What are the morphologic differences between VC and SCC?
A: It cannot be verrucous carcinoma if there is marked atypia. My understanding from my oral pathology colleagues is that by definition, verrucous carcinoma has minimal nuclear atypia. Verrucous surface profile, parakeratosis, broad bulbous rete with pushing invasion of the dermis, and increased mitoses along the basal layer are all features of verrucous carcinoma, but nuclear atypia must be minimal. If there is marked atypia, it is better classified as SCC.
- I've seen a few cases of early JXG that really resembled mastocytosis at low power. I had never before realized that those entities could mimic one another.