DIAGNOSTIC HISTOPATHOLOGY Gastrointestinal Tract, Soft Tissue and Skin

John R. Goldblum MD - Steven D. Billings MD Department of Anatomic Pathology

S. Apollonia Auditorium, Firenze, 14 - 16 May 2013

Ultima Cena - Museo del Cenacolo di S. Apollonia*- Andrea del Castagno (affresco, c.a. 1450)

SERVIZIO SANITARIO NAZIONALE – REGIONE LIGURIA ’ SANITARIA LOCALE N°5 “SPEZZINO” AZIENDA UNITÁ DIPARTIMENTO ONCOLOGICO S.C. DI ANATOMIA ED ISTOLOGIA PATOLOGICA E CITODIAGNOSTICA DIRETTORE: Dr FRANCO FEDELI DIAGNOSTIC HISTOPATHOLOGY Gastrointestinal Tract, Soft Tissue and Skin

John R. Goldblum, MD Chairman, Department of Anatomic Pathology Cleveland Clinic OH, USA Professor of Pathology, Cleveland Clinic Lerner College of Medicine

Steven D. Billings, MD Co-Section Head, Section of Dermatopathology Anatomic Pathology, Cleveland Clinic OH, USA Associate Professor in the Cleveland Clinic Lerner College of Medicine

S. Apollonia Auditorium – Firenze – Italy, May 14 – 16, 2013

PUBLISHED BY:

SERVIZIO SANITARIO NAZIONALE – REGIONE LIGURIA AZIENDA UNITÁ SANITARIA LOCALE N°5 “SPEZZINO” S.C. ANATOMIA ED ISTOLOGIA PATOLOGICA E CITODIAGNOSTICA DIRETTORE: DR FRANCO FEDELI [email protected]

COPY-EDITOR:

CLAUDIA MASSI - S.C. ANATOMIA ED ISTOLOGIA PATOLOGICA E CITODIAGNOSTICA GAVINO DELOGU - CENTRO STAMPA ASL5 AZIENDA UNITÁ SANITARIA LOCALE N°5 “SPEZZINO”

* Su concessione del Ministero per i Beni e le Attività Culturali - Divieto di ulteriori riproduzioni o duplicazioni con qualsiasi mezzo TABLE OF CONTENTS

Inflammatory Dermatopathology for the Surgical Pathologist Steven D. Billings, MD……………………………………………………………... I

Common Morphologic Patterns in Soft Tissue Tumors with an Emphasis on Useful Ancillary Diagnostic Techniques John R. Goldblum, MD…………………………………………………………….. II

Controversies in the Diagnosis of Barrett’s Esophagus and Barrett’s-Related Dysplasia. One pathologist’s Perspective John R. Goldblum, MD…………………………………………………………….. III

Fibrohistiocytic Tumors of Soft Tissue with a Focus on Fibrohistiocytic Tumors of Intermediate Malignancy John R. Goldblum, MD…………………………………………………………….. IV

Cutaneous Vascular Tumors Steven D. Billings, MD……………………………………………………………... V

IBD and IBD-Related Dysplasia John R. Goldblum, MD…………………………………………………………….. VI

Pleomorphic Cutaneous Spindle Cell Tumors Steven D. Billings, MD……………………………………………………………... VII

Cutaneous Soft Tissue Tumors of Uncertain Histogenesis Steven D. Billings, MD……………………………………………………………... VIII

Inflammatory Dermatopathology for the Surgical Pathologist

Steven D. Billings, MD

Co-Section Head, Section of Dermatopathology Anatomic Pathology, Cleveland Clinic OH, USA Associate Professor in the Cleveland Clinic Lerner College of Medicine [email protected]

- I -

Introduction Few specimens cause more difficulty to the general surgical pathologist than inflammatory skin lesions. The language seems arcane and impenetrable. There is significant histologic overlap between different entities. This often leads to the trap of an insufficiently precise descriptive diagnosis such as ‗chronic dermatitis‘ or ‗chronic inflammation‘. Overly general reports are little use to the clinician. This course will provide an overview for the approach to commonly encountered inflammatory skin diseases with specific tips on generating effective reports.

General Concepts Clinical Correlation It is vital to correlate the microscopic findings with the clinical information to generate appropriate reports on inflammatory lesions of the skin. The clinical information greatly influences the histologic differential diagnosis. Unfortunately, the quality of the clinical information varies widely. Some clinicians provide detailed descriptions and histories while others generically refer to everything as ‗lesion‘. In this latter setting it can sometimes be difficult to know if you dealing with an inflammatory dermatosis or a neoplasm let alone what type of inflammatory dermatosis. Furthermore when dealing with inflammatory dermatoses, the clinical information can exclude some diagnoses even if the histologic features are compatible with the diagnosis. As pathologists, it is important to utilize all tools at our disposal. After the microscope, the telephone has become my most important tool. Calling the clinician who submitted the biopsy on difficult cases can yield a wealth of information not present on the requisition form. It also helps establish a rapport with the clinician.

Ideal Inflammatory Dermatopathology Report Any pathology report should be considered a dialogue with the clinician. This is especially true when dealing with reports on inflammatory skin lesions. It is often not possible to provide a specific diagnosis, but it remains important to generate a report that is useful to the clinician and ultimately, and more importantly, the patient. An ideal report in this setting typically consists of three parts:

1. Microscopic description: A microscopic description is essential for an effective report on inflammatory lesions of the skin. A systematic approach is recommended. Most descriptions start from the outside in, from the epidermis to the dermis. The description should include any epidermal changes followed by description of the distribution and cellular components of the inflammatory infiltrate in the dermis. The description provides the foundation for the diagnosis as well as the explanatory note (see below). Dermatologists are well versed in histologic findings of inflammatory skin diseases. Unlike other consumers of pathology reports, they usually read the microscopic descriptions, and look to see specific features mentioned with different cutaneous diseases. The microscopic description can also help dermatologists formulate a new differential diagnosis when the histologic findings are at odds with the clinical impression.

2. Diagnosis: Obviously the diagnosis is the most important aspect of any pathology report. In the interpretation of inflammatory skin lesions, it is only possible to make a specific diagnosis when both the histologic and clinical information correlate. Often, however, the histologic features are

insufficiently specific or the clinical information is insufficient for making a specific diagnosis. In this situation, it is necessary to make a descriptive diagnosis. However, the descriptive diagnosis has to be couched in specific language useful to the clinician. Ideally the descriptive diagnosis correlates with an inflammatory reaction pattern, which will be the focus of the course (see below). The descriptive diagnosis also must be accompanied by an explanatory note.

3. Note: This part of the report is critical. This part of the report explains the thought process behind the diagnosis, especially when a descriptive diagnosis is rendered. It is a good idea to include your microscopic differential diagnosis and which diagnosis is favored and why. This is typically followed by the sentence: ―Correlation with clinical findings is recommended.‖ or ―Clinicopathologic correlation is recommended.‖ Depending on the clinical presentation, the dermatologist may believe the entity you thought was less likely on histologic findings to be a better fit for the clinical diagnosis.

Pattern Approach Like all surgical pathologists, dermatopathologists rely on the recognition of important low-power patterns that initiate the process of building an appropriate differential diagnosis. The pattern recognition approach was codified by Dr. Ackerman‘s seminal publication of his textbook on the subject in 1978. This course will cover the most common inflammatory reaction patterns encountered including the spongiotic, psoriasiform, interface, superficial perivascular, superficial and deep perivascular, interstitial, sclerosing, panniculitis, and bullous disease. The patterns can broadly be broken down into four groups: 1. epidermal patterns, 2. dermal patterns 3. panniculitis and 4. bullous disorders. As a general rule, epidermal changes trump other changes. In other words, if there are epidermal changes, the lesion most likely should be placed in the patterns associated with significant epidermal changes. After assessing the nature of the epidermal changes if any, the pattern and nature of the dermal inflammatory infiltrate should be evaluated. The pattern includes location within the dermis (i.e. superficial, or superficial and deep dermis), and the distribution (i.e. perivascular, lichenoid, interstitial, nodular). Then the nature of the inflammatory infiltrate is assessed whether it is composed of mononuclear cells (lymphocytes and histiocytes), mixed (mononuclear cells and granulocytes such as eosinophils), or predominantly granulocytes (neutrophils or eosinophils). It is important to remember that these patterns may show overlapping features and that certain entities may cross different patterns as will be discussed.

The primary patterns include 1. Epidermal Patterns a. Spongiotic pattern b. Psoriasiform pattern c. Interface pattern i. Lichenoid ii. Vacuolar with superficial or superficial and deep perivascular infiltrate

2. Dermal Patterns a. Superficial perivascular b. Superficial and deep perivascular c. Interstitial i. Nodular and diffuse ii. Palisading granulomatous d. Sclerosing 3. Panniculitis a. Septal b. Lobular 4. Bullous pattern a. Subepidermal b. Intraepidermal

Spongiotic dermatitis Discussion Spongiotic Pattern This is one of the most common patterns encountered. It is characterized by the accumulation of edema fluid within the epidermis. The resultant hydrostatic forces cause separation of the keratinocytes causing the epidermis to resemble the cut surface of a sponge. It is important to keep in mind that there are three phases of spongiotic dermatitis that exist along a continuum: 1. acute, 2. subacute, and 3. chronic spongiotic dermatitis.

1. Acute spongiotic dermatitis: In classic acute spongiotic dermatitis, the stratum corneum retains its normal basket-weave pattern and the epidermis does not show acanthosis (hyperplasia/thickening). The hydrostatic pressure can sometimes rupture the desmosomal attachments between the keratinocytes resulting in the formation of small blisters within the epidermis called spongiotic microvesicles. The keratinocytes often show pallor and there is edema in the papillary dermis. Within the dermis there is a variable inflammatory infiltrate composed of lymphocytes often admixed with eosinophils. Some neutrophils may be in the infiltrate as well, but this is not as common. The distribution is typically centered on blood vessels in the superficial dermis, but the distribution can be variable.

2. Subacute spongiotic dermatitis: As the inflammatory process evolves the epidermis undergoes additional reactive changes. There is increased proliferation resulting in acanthosis (thickening) and parakeratosis (retained nuclei in the stratum corneum). The granular layer may be diminished. As the process evolves there is less spongiosis in the epidermis and less edema in the dermis.

3. Chronic spongiotic dermatitis: In chronic spongiotic dermatitis there is hyperkeratosis with or without parakeratosis, acanthosis and little to no significant spongiosis. Frequently the granular layer is thickened. The dermal inflammatory infiltrate is often more mild in nature but otherwise similar to acute or subacute spongiotic dermatitis. The dermis may be fibrotic.

There is a range of inflammatory diseases that are characterized by the spongiotic pattern. Most forms of spongiotic dermatitis fall into the group broadly termed

‗eczematous dermatitis‘. Eczematous dermatitis refers to a group of inflammatory diseases that show papules, vesicles and variable crust scale, clinically. These most commonly encountered include atopic dermatitis, nummular dermatitis, contact dermatitis, dyshidrotic dermatitis, id reactions, and sometimes drug eruptions. The clinician may be able place them into one of the specific members of the eczema family based on clinical presentation, history, or ancillary tests in the setting of contact dermatitis.

For practical purposes the eczematous dermatitides are insufficiently distinct to reliably separate them into these more specific diseases histologically. Correlation with the clinical information is required to assign a more specific diagnosis. As these are so similar histologically, a brief discussion of the clinical features of these diseases is helpful for communication with submitting physicians.

Contact Dermatitis Contact dermatitis is a result of an exogenous stimulus and can be subdivided into allergic or irritant types. Allergic contact dermatitis is the result of a type IV hypersensitivity reaction that requires exposure to a specific antigen. The prototypical allergic contact dermatitis includes reactions to such substances as poison ivy or latex. Irritant contact dermatitis results from direct damage to the epidermis from the offending substance rather than an immune mediated response. Detergents are one of the most common causes of irritant dermatitis (so-called ‗dishpan hands‘). Histologically both show features of spongiotic dermatitis. Clinically, there are frequently clues to the diagnosis. For example, in poison ivy the eruption often has a linear arrangement corresponding to the edge of the offending leaf that brushed along the skin. Depending on the offending agent, there may be peculiar distributions such as with allergic reactions to latex gloves or nickel containing jewelry. A histologic clue to the diagnosis of allergic contact dermatitis is the presence of Langerhans cell microabscesses within the epidermis (not to be confused with the Pautrier‘s microabscesses of mycosis fungoides which are composed of neoplastic lymphocytes). Langerhans cell microabscesses are not always present in allergic contact dermatitis and they are not entirely specific. In irritant contact dermatitis, the inflammatory infiltrate tends to be less intense and there may be ballooning degeneration of keratinocytes in the epidermis.

Atopic Dermatitis This is a chronic relapsing pruritic dermatitis in patients with a familial history of atopy. Atopy is characterized by variable combinations of dermatitis, asthma, sinusitis, and allergic rhinitis. In children the eruption favors flexural areas such as the antecubital fossa. In adults, the presentation is more variable including very mild periorbital dermatitis to full body erythroderma.

Nummular Dermatitis Nummular dermatitis is characterized by round (coin-shaped) to oval patches variably composed of papules and vesicles usually on the extremities. As the eruption evolves, there may be central clearing, clinically resembling dermatophyte infection (tinea). In patients with atopic dermatitis, they may have flares of their disease presenting as nummular dermatitis. Nummular dermatitis frequently has a component of epidermal acanthosis.

Dyshidrotic Eczema (Pompholyx or palmoplantar dermatitis) Dyshidrotic eczema is characterized by a recurrent pruritic eruption of the palms, soles or digits. Clinically, the vesicles have a papular appearance. Over time, scaling and cracking can develop. In many patients this is a manifestation of atopy. A significant proportion of dyshidrotic eczema is the result of an allergic contact dermatitis. Spongiotic vesicles are a very common histologic feature. It is important to exclude dermatophyte infection, especially in eruptions from the feet (see below).

Id Reactions (auto-eczematization) Id reactions are the development of an eczematous dermatitis in regions away from the primary inflammatory focus. Dermatophyte infections of the feet (tinea pedis) and stasis dermatitis are two of the most common inciting conditions for id reactions. The patient can develop eczematous dermatitis on the upper extremities or trunk far way from the inciting process. In the case of dermatophyte-triggered eczematous dermatitis, no fungi are detectable in the dermatitis representing the id reaction. The id reaction component is difficult to treat without addressing the underlying trigger.

Eczematous drug reactions Drug reactions will be dealt with in more detail in a later section of the course. A minority of drug reactions <5-15% may be histologically indistinguishable from other forms of eczematous dermatitis. Association with new medications can help correlation with the diagnosis.

Practical Tips for Eczematous Dermatitis Descriptive diagnosis: Because of the histologic and clinical overlap, the most prudent way to sign-out these cases are with a descriptive diagnosis of ‗spongiotic dermatitis‘ followed by an explanatory note. An example of a stylized report:

Clinical diagnosis: Nummular dermatitis vs. contact dermatitis vs. psoriasis

Microscopic description: There is prominent parakeratosis overlying the epidermis that shows irregular acanthosis and mild spongiosis. Focal Langerhans cell microabscesses are present within the epidermis. Within the dermis there is a superficial perivascular lymphocytic infiltrate admixed with some eosinophils.

Diagnosis: Spongiotic dermatitis, see note

Note: The differential diagnosis could include an eczematous dermatitis such as nummular dermatitis or contact dermatitis. Given the Langerhans cell microabscesses, contact dermatitis is favored. The presence of eosinophils in the infiltrate makes psoriasis unlikely. Correlation with clinical findings is recommended.

Overlap with psoriasiform pattern: The epidermal acanthosis that accompanies evolving spongiotic dermatitis in the subacute and chronic phases results in overlap with the psoriasiform pattern (see below). Some prefer combining both terms when there is significant acanthosis (i.e. spongiotic psoriasiform dermatitis or psoriasiform spongiotic dermatitis) while others use the terminology of subacute or chronic spongiotic dermatitis. Either way is acceptable as long as the differential diagnosis/thought process is adequately discussed in the note section of the report.

Specific forms of Spongiotic Dermatitis There are some forms of spongiotic dermatitis that are sufficiently characteristic to allow a specific diagnosis. Of course this requires both adequate histologic findings in combination with the appropriate clinical setting.

Stasis Dermatitis Clinical Features Stasis dermatitis typically presents on the medial aspect of the lower extremities in association with evidence of venous insufficiency. Usually stasis dermatitis presents as pruritic, scaly plaques. Rarely it presents as a more circumscribed process clinically and can be confused with neoplasms. Acroangiodermatitis, a specific form of stasis dermatitis, presents as violaceous macules, nodules or plaques on the dorsal feet. It can be clinically and histologically confused with Kaposi sarcoma. Microscopic features The epidermis shows features of subacute or chronic spongiotic dermatitis as described above. The key differentiating features are found in the dermis. Within the papillary dermis there is a lobular proliferation of relatively thick walled vessels. There may be evidence of tissue edema or in long-standing cases fibrosis. There is extravasation of erythrocytes and associated perivascular siderophages to varying degrees. A perivascular lymphocytic infiltrate is present. The infiltrate is variable and can be mild or prominent. Differential diagnosis The differential diagnosis of most cases includes the forms of spongiotic dermatitis outlined above. In some cases there may be a combination of eczematous dermatitis and stasis dermatitis. In rare cases the vascular proliferation is so prominent as to mimic a vascular neoplasm such as Kaposi sarcoma. This form of stasis dermatitis is referred to as acroangiodermatitis. Careful attention to histologic features usually allows ready distinction. Acroangiodermatitis does not have the dense proliferation of spindled endothelial cells with slit-like vascular spaces, lacks the promontory sign of early Kaposi sarcoma and does not express latent nuclear antigen of HHV-8.

Practical tips Keep a high index of suspicion on biopsies from the lower legs, and remember the vascular changes are the most important feature.

Pityriasis Rosea Clinical Features Pityriasis rosea usually presents in young adults, though a wide age range may be affected. The eruption starts as a salmon colored herald patch that in the next 7-14 days is followed by a widespread, symmetric eruption of numerous small pink to red scaly plaques. The eruption usually starts on the trunk and then spread to the abdomen and proximal extremities. Microscopic Features The most characteristic feature of pityriasis rosea is the presence of discrete mounds of parakeratosis in the stratum corneum. The epidermis shows mild spongiosis and mild acanthosis. Within the dermis there is a superficial perivascular lymphocytic infiltrate. Eosinophils are rarely present. There may be extravasation of erythrocytes in the papillary dermis and exocytosis of erythrocytes into the overlying epidermis.

Differential Diagnosis The differential diagnosis includes subacute spongiotic dermatitis and it is often not possible to distinguish without clinical history. Guttate psoriasis (see below) also resembles pityriasis rosea. Guttate psoriasis often has mounds of parakeratosis surmounted by collections of neutrophils. Neutrophils are not a feature of pityriasis rosea. Practical Tips A specific diagnosis of pityriasis rosea is not possible without the appropriate clinical history. In the absence of a sufficient history, the case should be signed out descriptively as ‗spongiotic dermatitis, see note‘. The note could read: ―The biopsy demonstrates spongiotic dermatitis. Because of the presence of discrete mounds of parakeratosis, the possibility of pityriasis rosea should be considered. Clinicopathologic correlation is recommended.‖

Vesicular Dermatophytosis

Dermatophyte infections can sometimes present with prominent spongiosis. Usually there are neutrophils in the stratum corneum and eosinophils as part of the dermal infiltrate. Dermatophyte infection will be discussed in more detail in the next section.

Selected References 1. Weedon, David. Skin Pathology 3nd ed., New York: Churchill Livingstone, 2010. 2. McKee PH, Calonje E, Granter SR. Pathology of the Skin with Clinical Correaltions, 4th ed. Elsevier-Mosby, 2011. 3. Mark BJ. Slavin RG. Allergic contact dermatitis. Medical Clinics of North America. 90:169-85, 2006. 4. SD Billings and J Cotton. Inflammatory Dermatopathology: A Pathologist's Survival Guide 1st ed., New York: Springer 2011. 5. Bauer A. Rodiger C. Greif C. Kaatz M. Elsner P. Vulvar dermatoses--irritant and allergic contact dermatitis of the vulva. Dermatology. 210:143-9, 2005. 6. Beltrani VS. Beltrani VP. Contact dermatitis. Annals of Allergy, Asthma, & Immunology. 78:160-73, 1997. 7. Abramovits W. Atopic dermatitis. Journal of the American Academy of Dermatology. 53(1 Suppl 1):S86-93, 2005. 8. Houck G. Saeed S. Stevens GL. Morgan MB. Eczema and the spongiotic dermatoses: a histologic and pathogenic update. Seminars in Cutaneous Medicine & Surgery. 23:39-45, 2004. 9. Sivaram M. Chawla Y. Kumar B. Stasis dermatitis--a new cutaneous manifestation of Budd-Chiari syndrome. International Journal of Dermatology. 37:397-8, 1998. 10. Rao B. Unis M. Poulos E. Acroangiodermatitis: a study of ten cases. International Journal of Dermatology. 33:179-83, 1994. 11. Eslick GD. Atypical pityriasis rosea or psoriasis guttata? Early examination is the key to a correct diagnosis. International Journal of Dermatology. 41:788- 91, 2002. 12. Gonzalez LM. Allen R. Janniger CK. Schwartz RA. Pityriasis rosea: an important papulosquamous disorder. International Journal of Dermatology. 44:757-64, 2005.

13. Allen RA. Janniger CK. Schwartz RA. Pityriasis rosea. Cutis. 56:198-202, 1995. 14. Weaver J, Billings SD. Initial presentation of stasis dermatitis mimicking solitary lesions: A previously unrecognized clinical scenario. Journal of the American Academy of Dermatology. 61:1028-1032, 2009.

Psoriasiform Pattern The psoriasiform pattern is characterized by acanthosis (epidermal hyperplasia). This pattern frequently coexists with spongiosis. In certain settings, such as subacute or chronic spongiotic dermatitis, the decision to place an entity in the spongiotic or psoriasiform pattern is arbitrary, as described above. Some even combine the terms in this situation with a diagnosis of ‗spongiotic psoriasiform dermatitis‘. Entities in the psoriasiform pattern include: 1. Psoriasis 2. Lichen simplex chronicus/prurigo nodularis 3. Dermatophyte infection 4. Pityriasis rubra pilaris 5. Subacute to chronic spongiotic dermatitis (see above) 6. Stasis dermatitis (see above)

Psoriasis Psoriasis exists in three common clinical subtypes: psoriasis vulgaris (often referred to as just psoriasis), guttate psoriasis, and pustular psoriasis. Psoriasis vulgaris is the prototypical psoriasiform dermatitis.

Psoriasis vulgaris Clinical Features The common form of psoriasis usually presents in the 2nd or 3rd decade, but can present at any age. It presents as erythematous plaques with silvery scale. It commonly affects the extensor surfaces, scalp, gluteal cleft, and glans penis. Intertriginous areas can also be involved. This has been termed inverse psoriasis. Nail changes consisting of small pits and areas of yellow discoloration are frequently present. Psoriatic arthritis is seen in 1-5% of patients and its presence usually correlates with more severe skin disease. Microscopic Features Classic psoriasis vulgaris shows prominent, often confluent, parakeratosis overlying the epidermis. The epidermis shows uniform acanthosis with suprapapillary plate thinning and a diminished to absent granular layer. Within the stratum corneum and/or epidermis there are collections of neutrophils. Within the dermis there is a superficial perivascular lymphocytic infiltrate. Some neutrophils may be present, but eosinophils are typically absent. The papillary dermal blood vessels are dilated and often tortuous. Differential Diagnosis The differential diagnosis of psoriasis vulgaris includes eczematous dermatitides such nummular dermatitis and contact dermatitis, dermatophyte infection, seborrheic dermatitis, and pityriasis rubra pilaris. Nummular dermatitis and contact dermatitis often have eosinophils in the inflammatory infiltrate. These entities lack the suprapapillary plate thinning and the acanthosis is more irregular. They are also more likely to have a retained granular layer. Classically, they do not have collections of

neutrophils, but secondary impetiginization can result in neutrophils in the stratum corneum of these entities. Langerhans cell microabscesses are a feature often present in contact dermatitis but not seen in psoriasis. Dermatophyte infections of the skin have collections of neutrophils in the stratum corneum like psoriasis but the acanthosis is more irregular and there are usually eosinophils in the infiltrate. Dermatophytosis lacks the suprapapillary plate thinning. Special stains such as PAS or GMS will identify the fungal hyphae. Seborrheic dermatis has significant overlap with psoriasis but a more restricted clinical presentation on the scalp, central face and central chest. Pityriasis rubra pilaris lacks neutrophils and has alternating patterns of parakeratosis and hyperkeratosis.

Very rarely, eosinophils may be seen, but in general their presence argues against psoriasis, with one exception: drug-induced psoriasis from TNF-α inhibitors. The class of drugs paradoxically is used to treat psoriasis as well as other chronic inflammatory diseases such as inflammatory bowel disease. Some patients develop a rash that is histologically identical to psoriasis but also has conspicuous eosinophils in the infiltrate.

Practical Tips Classic psoriasis is easy to recognize. Unfortunately classic psoriasis is biopsied infrequently. The disease is often altered by partial treatment. In such cases the epidermis may lack some of the histologic features. Neutrophils may be lacking and the epidermis may show a retained granular layer and less prominent acanthosis. In excoriated psoriasis the granular layer may be retained. Similarly in psoriasis involving acral surfaces, the granular layer is almost always partially retained. A helpful clue to remember is that psoriasis typically lacks eosinophils.

In cases where the diagnosis of psoriasis is suspected but the histologic features are insufficient for an unequivocal diagnosis signing the case out descriptively as ‗psoriasiform dermatitis, see note‘ is a good strategy. The wording of the note can vary depending on the histologic features. An example: ‗The differential diagnosis in this case includes psoriasis vs. nummular dermatitis. Given the collections of neutrophils in the stratum corneum and the dilated papillary dermal blood vessels, the diagnosis of psoriasis is favored.‘ Another example: ―the differential diagnosis in this case could include psoriasis vs. nummular dermatitis. The presence of eosinophils in the inflammatory infiltrate argues against the diagnosis of psoriasis.‘

Psoriasis variants

Guttate Psoriasis Clinical Features Guttate psoriasis is characterized by a rapid onset of numerous small plaques. There is often a history of antecedent (streptococcal) pharyngitis. Microscopic Features Guttate psoriasis is characterized by discrete mounds of parakeratosis with associated collections of neutrophils overlying the epidermis. The epidermis typically does not show significant acanthosis. The papillary dermal blood vessels are often dilated similar to the vulgaris variant. Again, eosinophils are not a feature.

Differential Diagnosis The differential diagnosis can include the same entities as psoriasis vulgaris but pityriasis rosea as outlined above is the closest morphologic mimic.

Pustular Psoriasis Clinical Features Pustular psoriasis is also characterized by a widespread rapid onset of numerous pustules. It can be associated with pregnancy or discontinuation of systemic steroids in patients with psoriasis. Microscopic Features This variant is typified by large collections of neutrophils in the epidermis and/or stratum corneum. Because of the rapid onset there is often no significant acanthosis and the granular layer is only partially diminished or normal. Differential Diagnosis The differential diagnosis includes fungal infections such as dermatophytosis and candidiasis. PAS or GMS stains can help resolve this question. Acute generalized exanthematous pustulosis (AGEP), a peculiar form of drug eruption, can show striking resemblance to pustular psoriasis but the presence of eosinophils and the history of new medications (e.g. vancomycin) can help distinguish it from pustular psoriasis.

Selected References 1. Iizuka H. Takahashi H. Ishida-Yamamoto A. Pathophysiology of generalized pustular psoriasis. Archives of Dermatological Research. 295 Suppl 1:S55-9, 2003. 2. Ortonne JP. Recent developments in the understanding of the pathogenesis of psoriasis. British Journal of Dermatology. 140 Suppl 54:1-7, 1999. 3. Farber EM. Nall L. Childhood psoriasis. Cutis. 64:309-14, 1999. 4. Christophers E. Kiene P. Guttate and plaque psoriasis. Dermatologic Clinics. 13:751-6, 1995. 5. Trozak DJ. Histologic grading system for psoriasis vulgaris. International Journal of Dermatology. 33:380-1, 1994. 6. Farber EM. Nall L. Pustular psoriasis. Cutis. 51:29-32, 1993. 7. Farber EM. Nall L. Nail psoriasis. Cutis. 50:174-8, 1992. 8. Farber EM. Nall L. Strefling A. Psoriasis: a disease of the total skin. Journal of the American Academy of Dermatology. 12:150-6, 1985. 9. Puig L, Morales-Múnera CE, López-Ferrer A, Geli C. Ustekinumab Treatment of TNF Antagonist-Induced Paradoxical Psoriasis Flare in a Patient with Psoriatic Arthritis: Case Report and Review. Dermatology. 2012;225(1):14-7. 10. Al-Mutairi A, Elkashlan M, Al-Fayed HM, Swayed M. TNF-α inhibitor (adalimumab) induced psoriasis: a case report. Australas J Dermatol. 2012 May;53(2):157. 11. Hawryluk EB, Linskey KR, Duncan LM, Nazarian RM. Broad range of adverse 12. cutaneous eruptions in patients on TNF-alpha antagonists. J Cutan Pathol. 2012 May;39(5):481-92. 13. Steinwurz F, Denadai R, Saad-Hossne R, Queiroz ML, Teixeira FV, Romiti R. Infliximab-induced psoriasis during therapy for Crohn's disease. J Crohns Colitis. 2012 Jun;6(5):610-6.

Dermatophyte infection (dermatophytosis or tinea) Clinical Features The clinical presentation of dermatophyte infections can vary depending on the clinical location. Dermatophytosis often presents as an annular scaly plaques with central clearing. Microscopic Features The quintessential feature of dermatophyte infection is the presence of neutrophils in parakeratotic scale in the stratum corneum. The epidermis shows varying amounts of spongiosis and acanthosis. Within the dermis there is usually a perivascular mixed inflammatory infiltrate of lymphocytes and eosinophils. The fungal hyphae are often difficult to see on routine H&E stained sections. It is often necessary to perform special stains such as PAS or GMS stains. Differential Diagnosis The differential diagnosis includes spongiotic dermatitis and psoriasis as outlined above. Practical Tips It is important to keep a high index of suspicion for dermatophytosis. Whenever there are neutrophils in the stratum corneum, it is reasonable to consider special stains for fungi. Other situations to consider special stains for dermatophytes include spongiotic or psoriasiform lesions clinically described as annular or dermatitis that has had poor response to topical steroids. Selected References 1. Odom R. Pathophysiology of dermatophyte infections. Journal of the American Academy of Dermatology. 28:S2-S7, 1993. 2. Richardson MD. Diagnosis and pathogenesis of dermatophyte infections. British Journal of Clinical Practice. Supplement. 71:98-102, 1990.

Pityriasis Rubra Pilaris Clinical Features The most common form, or classical pityriasis rubra pilaris (PRP), presents in adults and is characterized by small follicular papules, confluent perifollicular erythema with islands of spared skin, and palmoplantar keratoderma. Patients may also have yellow discoloration of nails. Microscopic Features There is prominent hyperkeratosis and parakeratosis. The parakeratosis is characterized by the so-called ‗checkerboard‘ pattern in which the parakeratosis alternates with zones of hyperkeratosis, both vertically and horizontally. The granular layer is maintained and typically thickened. Follicular plugging is characteristic. Within the dermis there is frequently a mild superficial, perivascular lymphocytic infiltrate that may rarely include eosinophils. Differential Diagnosis PRP is often confused with psoriasis. However, PRP lacks the neutrophils in the epidermis or stratum corneum, and does not have suprapapillary plate thinning or a diminished granular layer. Chronic spongiotic dermatitis is also in the differential diagnosis. The checkerboard pattern of parakeratosis and the follicular plugging can help distinguish these entities. Practical Tips Biopsies of early lesions of PRP may be inconclusive. If there is a clinical suspicion of PRP, and the biopsy specimens do not show characteristic

morphology, a comment stating that a repeat biopsy from the most developed area of the eruption may be able to unequivocally establish a diagnosis. Biopsies from the follicular papules are often relatively non-specific. The presence of follicular plugging even in the absence of a checkerboard pattern is suggestive in the appropriate clinical context. Selected References 1. Cohen PR. Prystowsky JH. Pityriasis rubra pilaris: a review of diagnosis and treatment. Journal of the American Academy of Dermatology. 20:801-7, 1989. 2. Walsh NM. Prokopetz R. Tron VA. Sawyer DM. Watters AK. Murray S. Zip C. Histopathology in erythroderma: review of a series of cases by multiple observers. Journal of Cutaneous Pathology. 21:419-23, 1994.

Lichen Simplex Chronicus and Prurigo Nodularis Clinical Features Lichen simplex chronicus and prurigo nodularis are related entities that are the result of persistent scratching or rubbing. Lichen simplex chronicus presents as pruritic, scaly plaques and prurigo nodularis as pruritic nodules. The lesions may be ulcerated secondary to excoriation. As both are related to excoriation, it is important to remember that these lesions are only seen where the patient can reach. Common locations include nape of the neck, scalp (especially prurigo nodularis), shin, forearms, dorsal feet, and perianal/genital areas. Microscopic Features In lichen simplex chronicus, the epidermis shows prominent hyperkeratosis, focal parakeratosis, hypergranulosis, and psoriasiform hyperplasia. Within the dermis there is fibrosis of the papillary dermis that is characterized by vertically oriented thick collagen fibers (so-called ‗vertical streaking‘). Prurigo nodularis shows similar histologic features except that the epidermis has pseudoepitheliomatous hyperplasia rather than psoriasiform hyperplasia. There is often a mild superficial, perivascular lymphocytic infiltrate. Differential Diagnosis The differential diagnosis of lichen simplex chronicus is primarily chronic spongiotic dermatitis and psoriasis. Chronic spongiotic dermatitis shows less prominent psoriasiform hyperplasia and does not have the vertical streaking of the papillary dermal collagen. Eosinophils are also a typical component of the inflammatory infiltrate. As a caveat, lichen simplex chronicus is often superimposed on pre-existing spongiotic dermatitis. Patients with atopic dermatitis are at increased risk for the development of lichen simplex chronicus. Psoriasis has confluent parakeratosis and does not show a thickened granular layer or the dermal fibrosis. For Prurigo nodularis, the differential diagnosis often includes neoplasms such as squamous cell carcinoma. Prurigo nodularis shows reactive atypia, lacks atypical mitotic figures and often presents as multiple lesions. Practical Tips ―Hairy palm sign‖: The epidermal changes of both these entities resemble acral skin because of the prominent hyperkeratosis and hypergranulosis. However, lichen simplex chronicus and prurigo nodularis typically present on hair bearing skin. The presence of follicles in what otherwise looks like acral skin is a clue to the diagnosis of lichen simplex chronicus or prurigo nodularis. It is also important to realize that lichen simplex chronicus and prurigo nodularis have overlapping features. Sometimes it may not be possible to

distinguish them on a biopsy sample. In this situation the clinical presentation as a plaque or nodule should guide the diagnosis. In lesions with a prominent inflammatory infiltrate, biopsies with features of lichen simplex chronicus/prurigo nodularis may be superimposed upon other inflammatory conditions such as nummular or contact dermatitis. A dense infiltrate or eosinophils suggest the possibility of an underlying dermatitis with superimposed lichen simplex chronicus. Selected References 1. Gunasti S. Marakli SS. Tuncer I. Ozpoyraz N. Aksungur VL. Clinical and histopathological findings of 'psoriatic neurodermatitis' and of typical lichen simplex chronicus. Journal of the European Academy of Dermatology & Venereology. 21:811-7, 2007. 2. Chey WY. Kim KL. Yoo TY. Lee AY. Allergic contact dermatitis from hair dye and development of lichen simplex chronicus. Contact Dermatitis. 51:5-8, 2004. 3. Lee MR. Shumack S. Prurigo nodularis: a review. Australasian Journal of Dermatology. 46:211-18, 2005. 4. Accioly-Filho LW. Nogueira A. Ramos-e-Silva M. Prurigo nodularis of Hyde: an update. Journal of the European Academy of Dermatology & Venereology. 14:75-82, 2000.

Interface dermatitis Interface dermatitis is characterized by damage to the epidermis from the inflammatory infiltrate. Microscopically this is characterized by basal vacuolization with or without necrotic keratinocytes. Interface dermatitis can be broadly grouped into two subgroups based on the pattern of the inflammatory infiltrate: 1. lichenoid, or band-like, in which the infiltrate forms a dense layer parallel to the overlying epidermis and 2. perivascular in which the infiltrate is concentrated around blood vessels in either a superficial or superficial and deep distribution.

Interface dermatitis with lichenoid infiltrate

Diagnosis: Lichen planus

Lichen planus Lichen planus is the prototypical lichenoid interface dermatitis Clinical Features Lichen planus usually presents in adults as pruritic, polygonal violaceous papules. There is a predilection for extensor surfaces of the wrists and ankles but the eruption may be widespread. Lichen planus involves the oral mucosa, especially the buccal mucosa, in about 60% of patients. In the oral cavity, lichen planus usually presents as a lace-like eruption, but erosions and ulceration can also occur. Microscopic Features The stratum corneum shows compact hyperkeratosis but not parakeratosis. The granular layer is thickened, often with a wedge-shaped pattern. The epidermis may show mild acanthosis. Within the dermis there is a dense band-like pattern of mononuclear cells predominantly composed of lymphocytes. Eosinophils may be present, but are not prominent. Some admixed histiocytes may be present. There is a hypertrophic variant that show significant acanthosis. An atrophic variant has epidermal atrophy and often a less brisk inflammatory infiltrate than is usually seen.

Differential Diagnosis A common entity in the differential diagnosis is a benign lichenoid keratosis, also called lichen planus-like keratosis. In some cases the histologic features may be indistinguishable. However, the clinical presentation is quite different. Benign lichenoid keratosis is usually a solitary lesion that presents on the trunk. Clinically, it mimics basal cell carcinoma. Sometimes a recognizable component of solar lentigo is seen. Lichenoid drug eruption and a fixed drug can be confused with lichen planus, Lichenoid drug eruption may closely mimic lichen planus but in addition to features resembling lichen planus there are more eosinophils in the inflammatory infiltrate and usually some parakeratosis in the stratum corneum. Fixed drug eruptions lack the prominent epidermal changes and also have eosinophils in the infiltrate. Practical Tips One important tip is to make sure the biopsy is from a rash instead of a solitary lesion (benign lichenoid keratosis). Also remember that parakeratosis or numerous eosinophils are not features of lichen planus. If they are present, the possibility of a drug eruption should be considered. Selected References 1. Boyd AS. Neldner KH. Lichen planus. Journal of the American Academy of Dermatology. 25:593-619, 1991. 2. Soper DE. Patterson JW. Hurt WG. Fantl JA. Blaylock WK. Lichen planus of the vulva. Obstetrics & Gynecology. 72:74-6, 1988. 3. Conklin RJ. Blasberg B. Oral lichen planus. Dermatologic Clinics. 5(4):663- 73, 1987.

Lichenoid Drug Eruption Clinical Features The lesions of lichenoid drug eruptions clinically resemble lichen planus. However, they tend to be larger and are more frequently distributed on the trunk and lesions on the extremities are not limited to the flexural surfaces. Oral mucosa involvement is absent. Microscopic Features Most of the microscopic features are those of lichen planus with some key exceptions. There is often parakeratosis, a feature not seen in typical lichen planus. Eosinophils are usually numerous. Differential Diagnosis As outlined above the differential diagnosis is primarily lichen planus and a fixed drug eruption. Differentiating lichen planus from lichenoid drug requires identification of features not seen in lichen planus such as parakeratosis and frequent eosinophils. Fixed drug eruptions do not show the prominent epidermal changes of lichenoid drug eruptions. Practical Tips Parakeratosis and eosinophils are frequent features of lichenoid drug eruptions. Selected References 1. Halevy S. Shai A. Lichenoid drug eruptions. Journal of the American Academy of Dermatology. 29:249-55, 1993. 2. West AJ. Berger TG. LeBoit PE. A comparative histopathologic study of photodistributed and nonphotodistributed lichenoid drug eruptions. Journal of the American Academy of Dermatology. 23:689-93, 1990.

3. Van den Haute V. Antoine JL. Lachapelle JM. Histopathological discriminant criteria between lichenoid drug eruption and idiopathic lichen planus: retrospective study on selected samples. Dermatologica. 179:10, 1989.

Fixed drug eruption Clinical Features Fixed drug eruptions present as one or more violaceous plaques usually on the extremities or genitalia. On re-exposure to the drug the eruption recurs in the same locations. Common sensitizing agents include barbiturates, ibuprofen, and sulfa drugs. Microscopic Features The stratum corneum of the epidermis has a normal basket weave pattern. The epidermis may show ballooning degeneration of keratinocytes. Within the dermis there is a lichenoid infiltrate with interface damage to the overlying epidermis. The infiltrate contains eosinophils. Melanophages are present as the lesion evolves. Differential Diagnosis The differential diagnosis includes lichen planus and a lichenoid drug eruption (see above). A morbilliform drug eruption could also be considered. Morbilliform drug eruptions do not have a lichenoid pattern. In questionable cases, the characteristic clinical presentation is a key to the differential diagnosis. Practical Tips Unlike other lichenoid dermatoses, the epidermal change in fixed drug eruptions typically does not include the stratum corneum or granular layer. A phone call to the clinician is helpful in cases with inadequate history. Melanophages can be a clue to a recurrent fixed drug eruption. Selected References 1. Korkij W. Soltani K. Fixed drug eruption. A brief review. Archives of Dermatology. 120:520-4, 1984. 2. Masu S. Seiji M. Pigmentary incontinence in fixed drug eruptions. Histologic and electron microscopic findings. Journal of the American Academy of Dermatology. 8:525-32, 1983.

Interface pattern with perivascular infiltrate

Morbilliform drug eruption Clinical Features Morbilliform drug eruptions present as widespread erythematous, blanchable macules or papules. They can present shortly after initiation of the offending medication or it can take several months for the hypersensitivity reaction to develop. Microscopic Features The epidermis typically shows little change except for some mild basal vacuolization. Occasional necrotic keratinocytes may be present. In some cases there is no interface damage to the epidermis. Within the dermis there is a mild superficial perivascular mixed inflammatory infiltrate of lymphocytes and eosinophils. Differential Diagnosis In cases with interface damage, the differential diagnosis of morbilliform drug eruption includes acute graft vs. host disease, fixed drug eruption, lupus erythematosus, dermatomyositis, and viral exanthem. Acute graft vs. host disease typically occurs in the setting of bone marrow transplant and rarely in solid organ transplants. It usually occurs relatively soon after the transplant and most cases lack

eosinophils, though some eosinophils may be present. Connective tissue disease such as lupus erythematosus and dermatomyositis are characterized by interface change. However, they lack eosinophils and typically have increased dermal mucin. Viral exanthems usually lack eosinophils.

In drug eruptions without interface change the differential diagnosis includes dermal hypersensitivity reactions such as urticaria or so-called papular eczema. Practical Tips It is often not possible to make a definitive diagnosis of a drug eruption. A descriptive diagnosis of ‗interface dermatitis with mixed inflammatory infiltrate, see note‘ or ‗superficial perivascular mixed infiltrate, see note‘ are examples of appropriate descriptive diagnoses. The note can read as ―The histologic features are consistent with a drug eruption in the appropriate clinical context. Clinicopathologic correlation is recommended.‘ In cases without interface change, the note can read as ‗The histologic features are consistent with a dermal hypersensitivity reaction such as a drug eruption in the appropriate clinical context. Other forms of dermal hypersensitivity reactions such as urticaria or papular eczema could be considered. Clinicopathologic correlation is recommended.‘

Selected References 1. Weedon, David. Skin Pathology 3rd ed., New York: Churchill Livingstone, 2010. 2. McKee PH, et al. Pathology of the Skin with Clinical Correaltions, 4th ed. Elsevier-Mosby, 2011.

Erythema multiforme and toxic epidermal necrolysis These diseases can be viewed as ends of a spectrum Clinical Features In classic erythema multiforme (EM), the patient presents with episodic eruptions of macules, papules, or targetoid lesions (multiple forms) on the extensor surfaces, palms, soles, and/or oral mucosa. If there is extensive mucosal involvement the eruption can qualify for the designation of Stevens-Johnson syndrome. The eruption can be associated with herpes simplex virus infections (especially EM), mycoplasma infections, and drugs. Stevens-Johnson syndrome is typically associated with medications, with sulfa drugs being one of the most common triggers.

Toxic epidermal necrolysis (TEN) presents with widespread tender macular eruption with vesicles and bullae. Application of pressure to the skin can cause detachment of the epidermis (Nikolsky‘s sign). TEN is a medical emergency necessitating admission to a burn unit. The mortality ranges from 25-50%.

Microscopic Features EM and TEN have essentially the same histologic features. The epidermis is relatively normal with a basket weave stratum corneum. There is interface damage with necrosis of keratinocytes in association with a superficial perivascular lymphocytic infiltrate. Eosinophils are sometimes present, especially in cases related to medications. In TEN there is often full thickness necrosis.

Differential diagnosis The differential diagnosis can include morbilliform drug eruption, graft vs. host disease and connective tissue disease such as lupus erythematosus or dermatomyositis. The pronounced epidermal damage helps exclude a typical drug eruption. Graft vs. host disease has the appropriate clinical history. In connective tissue disease there are epidermal changes (e.g. parakeratosis, thickened basement membrane) that are not seen in the EM/TEN spectrum.

Practical Tips The hallmark is the degree of epidermal damage that is disproportionate to the density of the lymphocytic infiltrate. When the epidermis appears essentially normal except for prominent keratinocyte necrosis, EM or TEN should be considered. If there are large areas of full thickness necrosis, TEN is more likely. In cases where the history is inadequate and EM or TEN are the most likely possibilities, a descriptive diagnosis of ‗interface dermatitis, see note‘ followed by note stating: ‗The histologic features could be compatible with erythema multiforme or toxic epidermal necrolysis in the appropriate clinical context. Clinicopathologic correlation is recommended.‘

Selected References 1. Drago F. Parodi A. Rebora A. Persistent erythema multiforme: report of two new cases and review of literature. Journal of the American Academy of Dermatology. 33:366-9, 1995. 2. Ahmed I. Reichenberg J. Lucas A. Shehan JM. Erythema multiforme associated with phenytoin and cranial radiation therapy: a report of three patients and review of the literature. International Journal of Dermatology. 43:67-73, 2004. 3. Al-Johani KA. Fedele S. Porter SR. Erythema multiforme and related disorders. Oral Surgery Oral Medicine Oral Pathology Oral Radiology & Endodontics. 103:642-54, 2007. 4. Letko E. Papaliodis DN. Papaliodis GN. Daoud YJ. Ahmed AR. Foster CS. Stevens-Johnson syndrome and toxic epidermal necrolysis: a review of the literature. Annals of Allergy, Asthma, & Immunology. 94:419-36, 2005. 5. Wolkenstein P. Revuz J. Toxic epidermal necrolysis. Dermatologic Clinics. 18:485-95, ix, 2000. 6. Roujeau JC. Chosidow O. Saiag P. Guillaume JC. Toxic epidermal necrolysis (Lyell syndrome). Journal of the American Academy of Dermatology. 23:1039-58, 1990.

Lupus Erythematosus Clinical Features Cutaneous lupus erythematosus can be subdivided into chronic (discoid), subacute, and systemic (acute) forms. There is clinical overlap and patients with discoid or subacute lupus erythematosus can progress to systemic disease. Therefore patients may have cutaneous lesions characteristic of all three forms.

Chronic, or discoid, lupus erythematosus is characterized by sharply demarcated erythematous scaly plaques usually involving the head and neck often involving the face in a butterfly pattern. Lesions on the scalp can result in scarring alopecia. A

variant of discoid lupus erythematosus called tumid lupus presents as juicy papules and plaques on the upper trunk and head and neck. The tumid variant has less scale. Chromic forms of lupus erythematosus are usually not associated with underlying systemic disease. Progression to systemic disease is seen in roughly 5-10% of cases. Antinuclear antibody titers (ANA) are positive in approximately 70%.

The cutaneous lesions of subacute lupus erythematosus manifest as annular lesions or plaques in photodistributed areas on the head and neck, upper trunk, and upper extremities. The patients often have mild musculoskeletal symptoms. Central nervous system involvement is usually absent and renal involvement is variable. Traditionally renal involvement was not considered common but some reports have refuted this finding. Positive ANA titers are seen in about 50% of cases. Patients may develop lesions of discoid lupus erythematosus or progress to fully developed systemic lupus erythematosus.

Cutaneous lesions are present in about 80% of patients with systemic lupus erythematosus. The cutaneous lesions are less well defined as in the other forms of cutaneous lupus erythematosus. They present as erythematous patches with little scale. As in other forms of cutaneous lupus erythematosus, the cutaneous lesions are in photo-distributed areas, especially the malar face. Positive ANA titers are seen in approximately 90% of cases and >50% have anti-double stranded DNA antibodies.

Microscopic Features Similar to the clinical manifestations there is significant histologic overlap in the different clinical subtypes of cutaneous lupus erythematosus. From a practical standpoint the overlap may preclude subclassification based on histologic features alone. All are a characterized by interface change of basal vacuolization and a perivascular lymphocytic infiltrate with increased dermal mucin. Dermal mucin is variably identifiable on routine H&E stained sections; it depends on the slide preparation technique of individual laboratories. Colloidal iron stains can be helpful in highlighting dermal mucin when it is not evident on routine H&E stains.

In discoid lupus erythematosus, the epidermis shows hyperkeratosis, variable epidermal atrophy alternating with acanthosis and follicular plugging. The basement membrane is often thickened. The inflammatory infiltrate has a superficial and deep pattern and frequently involves adnexal structures. In older ‗burned out‘ lesions, there may be less active interface change. In such cases the evidence of epidermal change includes the thickened basement membrane, epidermal atrophy and melanophages in the upper dermis. In the tumid form, significant interface change is typically absent; the combination of a superficial and deep infiltrate with increased dermal mucin is an important clue. Subacute lupus erythematosus differs from the discoid form only slightly. There is usually a less intense inflammatory infiltrate and more prominent atrophy. In systemic lupus erythematosus, there is prominent basal vacuolization but necrotic keratinocytes are rare. The infiltrate is typically less intense and usually in a superficial perivascular distribution.

Differential Diagnosis In cases with relatively numerous necrotic keratinocytes, the differential diagnosis includes erythema multiforme. However, erythema multiforme is an acute process and does not show the other epidermal changes of lupus erythematosus such as

hyperkeratosis, atrophy, or basement membrane thickening. In cases with a dense inflammatory infiltrate, lichen planus could be considered, but the presence of dermal mucin and a deep inflammatory component are against lichen planus. In both instances clinical history is also helpful. Dermatomyositis can be remarkably similar to lupus erythematosus (see below). It also shows interface dermatitis with increased dermal mucin. The inflammatory infiltrate in dermatomyositis is generally mild and restricted to the superficial dermis. In some cases, it may not be possible to distinguish between these entities except by clinical history. Some cases of lupus erythematosus show reactive atypia in the keratinocytes of the epidermis. The reactive epithelial atypia can mimic the dysplasia of actinic keratosis or even squamous cell carcinoma. This is usually more of a risk in superficial shave biopsies. The clinical history and the presence of other findings of lupus erythematosus will allow for distinction. Practical Tips Eosinophils are not a typical feature of lupus erythematosus except in the rare cases of drug-induced lupus erythematosus. The presence of eosinophils raises the possibility of dermal hypersensitivity reactions such as an arthropod bite reaction or drug eruption. Tumid lupus erythematosus lacks interface change. The ‗actinic keratosis clue‘. Remember that some cases of lupus erythematous can superficially resemble actinic keratosis. If there is interface change and squamous atypia, consider the possibility of lupus erythematosus. Remember that biopsies from old lesions may not show active vacuolar interface change. Look for evidence of past interface damage such as atrophy, basement membrane thickening, and melanophages. Colloidal iron studies may help highlight the dermal mucin. Some cases of dermatomyositis and lupus erythematosus are histologically indistinguishable. In this situation sign out the case as ‗interface dermatitis, see note‘ with the note stating: ‗The histologic features are consistent with connective tissue disease such as lupus erythematosus or dermatomyositis.‘

Selected References 1. Patel P. Werth V. Cutaneous lupus erythematosus: a review. Dermatologic Clinics. 20:373-85, v, 2002. 2. Crowson AN. Magro C. The cutaneous pathology of lupus erythematosus: a review. Journal of Cutaneous Pathology. 28:1-23, 2001. 3. Sharon VR, Konia TH, Barr KL, Fung MA. Assessment of the 'no eosinophils'rule: are eosinophils truly absent in pityriasis lichenoides, connective tissue disease, and graft-vs.-host disease? J Cutan Pathol. 2012 Apr;39(4):413-8.

Dermatomyositis Clinical Features Dermatomyositis is characterized by the combination of muscle weakness and characteristic cutaneous findings of erythematous to violaceous slightly scaly lesions. The face, shoulders and extensor surfaces of the extremities are most commonly involved. Involvement of the face frequently takes the form of a periorbital heliotrope rash. Involvement of the shoulders is often diffuse causing the shawl sign. Periungual erythema and Gottron‘s papules are common findings on the hands. Muscle weakness, when present, involves proximal muscles. Cutaneous involvement

can precede muscle involvement by months to years, and some patients never develop muscle weakness (so-called ‗dermatomyositis sine myositis‘). Microscopic Features The histologic features are characterized by basal vacuolization, a superficial perivascular lymphocytic infiltrate and increased dermal mucin. The basement membrane may be thickened and melanophages may be seen in the upper dermis. The infiltrate is usually mild and restricted to the upper dermis. Occasional neutrophils may be present. Differential Diagnosis The primary differential diagnosis is lupus erythematosus. Unfortunately, it is not possible to unequivocally differentiate dermatomyositis from lupus erythematosus. (See also above section on lupus erythematosus). The same comments on differentiating lupus erythematosus from other forms of interface dermatitis apply to dermatomyositis. Practical Tips The infiltrate in dermatomyositis is usually mild and restricted to the superficial dermis. If there is a deep component, consider the diagnosis of lupus erythematosus. Some cases of dermatomyositis and lupus erythematosus are histologically indistinguishable. In this situation sign out the case as ‗interface dermatitis, see note‘ with the note stating: ‗The histologic features are consistent with connective tissue disease such as lupus erythematosus or dermatomyositis.‘ Colloidal iron stains may help highlight the dermal mucin. Selected References 1. Callen JP. Wortmann RL. Dermatomyositis. Clinics in Dermatology. 24:363- 73, 2006. 2. Gerami P. Schope JM. McDonald L. Walling HW. Sontheimer RD. A systematic review of adult-onset clinically amyopathic dermatomyositis (dermatomyositis sine myositis): a missing link within the spectrum of the idiopathic inflammatory myopathies. Journal of the American Academy of Dermatology. 54:597-613, 2006.

Graft vs. Host Disease Clinical Features Cutaneous graft vs. host disease (GVHD) usually occurs in the setting of bone marrow transplant but can sometimes occur in the setting of solid organ transplants. GVHD can be subdivided into acute GVHD and chronic GVHD. Acute GVHD typically occurs 2-4 weeks after transplantation but it can be quite variable and may be several weeks to months after transplantation. Another variable that is increasingly seen is the practice of donor lymphocyte reinfusion. The eruption of acute GVHD is characterized by an erythematous macular to papular eruption involving the face, posterior neck, ears, hands and feet. The eruption often starts with facial erythema that subsequently involves other parts of the body with a maculopapular eruption. Co-existing diarrhea is often present and may precede the cutaneous eruption. Laboratory tests frequently show elevated liver enzymes.

Chronic GVHD classically occurs greater than 6 months after transplantation. GVHD is subdivided into lichenoid and sclerodermoid forms. Classically, chronic GVHD first manifests with the lichenoid form with the sclerodermoid form following. Some patients present with both forms simultaneously. Lichenoid chronic GVHD presents

as polygonal violaceous papules reminiscent of lichen planus. Oral mucosal involvement is seen in approximately 90% of patients. Sclerodermoid GVHD presents as areas of dermal sclerosis similar to morphea/scleroderma (see below).

Microscopic Features Acute GVHD In acute GHD the epidermis is relatively normal, but some keratinocyte atypia may be seen as a result of prior chemotherapy. Within the dermis there is a sparse infiltrate of lymphocytes. Occasionally eosinophils may be part of the infiltrate. The interface change is manifested by basal vacuolization. As the eruption progresses necrotic keratinocytes are seen often with lymphocyte satellitosis. Severe cases can show cleft formation between the epidermis and dermis or even full thickness necrosis of the epidermis. Fortunately this is rarely seen. Acute GVHD is graded by the following scheme: Grade 0: Normal skin Grade 1: Basal vacuolization with a mild superficial perivascular lymphocytic infiltrate. Grade 2: Same features as Grade 1 with scattered necrotic keratinocytes and satellite cell necrosis. Grade 3:Same features as Grade 2 but with cleft formation between epidermis and dermis. Grade 4: Same as features as Grade 2 or 3 with complete separation of the epidermis from the dermis.

Unlike acute GVHD, there is no grading scheme for chronic GVHD. Lichenoid chronic GVHD shows epidermal changes of hyperkeratosis and hypergranulosis in addition to interface change of basal vacuolization with necrotic keratinocytes and satellite cell necrosis. The infiltrate in the dermis is usually mild but is often more dense than is seen with acute GVHD, and in some cases there is a dense band-like infiltrate similar to lichen planus. Rarely some biopsies may show transitional forms with histologic features of acute GVHD and lichenoid chronic GVHD in the same biopsy or different concurrent biopsies from the same patient. Sclerodermoid chronic GVHD resembles morphea or scleroderma (see below). There is epidermal atrophy with dermal sclerosis characterized by fibrosis with compaction of collagen fibers in the reticular dermis. There is a loss of adnexal structures. Differential Diagnosis In acute GVHD the histologic and clinical differential diagnosis is usually a drug eruption. In most cases of acute GVHD there are no eosinophils unlike typical drug eruptions. However, acute GVHD can sometimes have eosinophils as part of the infiltrate. Satellite cell necrosis is more common in GVHD, but unequivocal distinction may not always be possible. Erythema multiforme could be considered from a histologic standpoint but the clinical situation typically negates erythema multiforme from consideration.

In lichenoid chronic GVHD, the primary differential diagnosis is lichen planus. Lichen planus typically has a denser infiltrate. True satellite cell necrosis is more of a feature of GVHD. It is not possible to distinguish sclerodermoid chronic GVHD from morphea or scleroderma. Clinical information is essential.

Practical Tips It is rare to see acute GVHD before 14 days after transplantation. The histologic features may lag the clinical presentation. In very early biopsies of GVHD, the skin may show no histologic abnormalities. Subsequent biopsies may show classic GVHD. In cases with the clinical diagnosis of GVHD without histologic evidence a note stating that ―the histologic features of GVHD may sometimes lag the clinical presentation‖ may be helpful. Late onset acute GVHD (> 6 months after transplantation) may be seen in the setting of donor lymphocyte reinfusion, an increasingly common practice. Eosinophils may sometimes be seen in GVHD and does not exclude the diagnosis in the appropriate clinical setting. From a practical viewpoint, many of these patients may not have a sufficient immune system to mount a drug eruption. My bias is that the eruption in this clinical setting is GVHD until proven otherwise. Correlation with other clinical findings can be helpful in this setting (e.g. diarrhea). Multiple levels may be necessary. Additional clinical information (e.g. diarrhea or elevated liver enzymes) can help corroborate the diagnosis. Selected References 1. Schaffer JV. The changing face of graft-versus-host disease. Seminars in Cutaneous Medicine & Surgery. 25:190-200, 2006. 2. Canninga-van Dijk MR. Sanders CJ. Verdonck LF. Fijnheer R. van den Tweel JG. Differential diagnosis of skin lesions after allogeneic haematopoietic stem cell transplantation. Histopathology. 42:313-30, 2003. 3. Aractingi S. Chosidow O. Cutaneous graft-versus-host disease. Archives of Dermatology. 134:602-12, 1998. 4. Bridge AT. Nelson RP Jr. Schwartz JE. Mirowski GW. Billings SD. Histological evaluation of acute mucocutaneous graft-versus-host disease in nonmyeloablative hematologic stem cell transplants with an observation predicting an increased risk of progression to chronic graft-versus-host disease. American Journal of Dermatopathology. 29:1-6, 2007.

Pityriasis Lichenoides Clinical Features Pityriasis lichenoides is most common in young adult men and typically involves the extremities, trunk, and buttocks. Pityriasis lichenoides exists in two forms: pityriasis lichenoides et varioliformis acuta (PLEVA) and pityriasis lichenoides chronica (PLC). PLEVA presents as recurrent crops of skin-colored papules that become hemorrhagic or crusted, ulcerate, and leave behind varioliform (smallpox-like) scars. PLC is less hemorrhagic consisting of red-brown, scaly macules to papules. In PLC the lesions heal without scarring but there may be post-inflammatory changes.

Microscopic Features The microscopic features show significant overlap between the two entities. Both show epidermal changes of parakeratosis, basal vacuolization and necrotic keratinocytes. The epidermal changes in PLC tend to be more subtle and the interface change is focal in nature. The epidermal change is more pronounced in PLEVA. The interface change is more pronounced with more numerous necrotic keratinocytes. There is more prominent exocytosis of lymphocytes and sometimes erythrocytes. In

late lesions of PLEVA the biopsy may show ulceration of the epidermis. The dermal inflammatory infiltrate is predominantly composed of lymphocytes. In PLC the infiltrate may be restricted to the superficial dermis. In PLEVA the infiltrate is superficial and deep and often has a wedge-shaped configuration. Extravasation of erythrocytes is commonly present, especially in PLEVA, but fibrinoid necrosis of blood vessels is absent.

Differential Diagnosis For PLEVA, the differential diagnosis includes lymphomatoid papulosis (LYP)(discussed in detail in a later section). Both have similar clinical histories of recurrent crops of papules that can ulcerate and both can show a superficial and deep infiltrate with interface change. LYP usually has a population of large atypical CD30+ cells. Lupus erythematosus could be considered but the clinical presentation is different, the degree of epidermal damage is more pronounced in PLEVA, and PLEVA does not have increased dermal mucin. For PLC, the differential diagnosis includes pityriasis rosea. The latter has more discrete mounds of parakeratosis and lacks interface change. Obviously with the overlapping histologic features, PLEVA and PLC can be confused. Knowledge of the clinical presentation may be required to distinguish them since the histologic difference is one of degrees. Practical Tips Maintain a high index of suspicion. I The presence of interface change with hemorrhage is an important clue especially for PLEVA. Knowledge of the clinical history is especially helpful for the diagnosis of PLEVA or PLC. A conversation with the clinician can be very helpful if the initial history provided is inadequate. In cases where the biopsy findings could represent either PLEVA or PLC, sign out the case as ‗pityriasis lichenoides, see note‘ with the note stating that ‗the biopsy findings could be compatible with either pityriasis lichenoides et varioliformis acuta or pityriasis lichenoides chronica. Clinicopathologic correlation is recommended.‘ Selected References 1. Khachemoune A. Blyumin ML. Pityriasis lichenoides: pathophysiology, classification, and treatment. American Journal of Clinical Dermatology. 8:29- 36, 2007. 2. Ersoy-Evans S. Greco MF. Mancini AJ. Subai N. Paller AS. Pityriasis lichenoides in childhood: a retrospective review of 124 patients. Journal of the American Academy of Dermatology. 56:205-10, 2007. 3. Bowers S. Warshaw EM. Pityriasis lichenoides and its subtypes. Journal of the American Academy of Dermatology. 55:557-72, 2006. 4. Patel DG. Kihiczak G. Schwartz RA. Janniger CK. Lambert WC. Pityriasis lichenoides. Cutis. 65:17-20, 23, 2000.

Superficial Perivascular Pattern Entities in this pattern are characterized by the absence of significant epidermal change and the presence of an inflammatory infiltrate that is largely restricted to the superficial dermis around blood vessels. Not infrequently there may be some overlap with the superficial and deep perivascular patterns.

Post inflammatory pigment alteration

Clinical Features Post inflammatory pigment alteration is the result of a prior inflammatory process. Clinically the lesions may be hypopigmented or hyperpigmented.

Microscopic Features The epidermis is essentially normal with normal numbers of melanocytes. Within the dermis there are perivascular melanophages around the superficial vascular plexus. In some cases, the melanophages are not numerous. There may be an associated mild perivascular lymphocytic infiltrate, but inflammatory cells are nor prominent.

Differential diagnosis Vitiligo is one of the primary differential diagnoses. Vitiligo is defined by an absence or at least marked reduction in the number of melanocytes within the epidermis. The absence of melanocytes is best demonstrated with immunostains for melanocyte markers such as Melan-A or MiTF. Otherwise they can be remarkably similar with associated melanophage around the superficial vascular plexus. A regressed melanocytic lesion could be considered. In regressed melanocytic lesions, there is melanoderma similar to post inflammatory pigment alteration but usually there is effacement of the normal rete peg architecture and more prominent papillary dermal fibrosis. Pigmented purpuric dermatosis (see below) is associated with siderophages rather than melanophages.

Urticaria pigmentosa and other forms of mastocytosis

Clinical Features Cutaneous mastocytosis manifests in a variety of ways. The most common is urticaria pigmentosa, accounting for approximately 80% of cases. Urticaria pigmentosa usually presents within the first 4 years of life and resolves by puberty. The risk of systemic disease is low in this setting. Cases occurring in adulthood tend to persist and approximately 40% develop systemic disease. The typical clinical presentation is a generalized eruption of red-brown macules that urticate on stroking (Darier‘s sign).

Telangiectasia macularis eruptiva perstans (TMEP) primarily occurs in adults and presents as erythematous macules with telangiectasia on the trunk and proximal extremities. Systemic involvement is common.

Mastocytoma is the solitary tumor form of mastocytosis. It typically presents in children and most involute spontaneously.

Systemic mastocytosis, as mentioned above, develops in patients with urticaria pigmentosa or TMEP. Besides the skin, the bone marrow is most frequently involved. The bone marrow involvement may progress to mast cell leukemia.

Microscopic Features The histologic features are similar across all subtypes of mastocytosis with minor variation. In urticaria pigmentosa, there is usually a moderately dense superficial perivascular infiltrate of mast cells. In TMEP the infiltrate is typically sparser. In

mastocytoma the infiltrate is quite dense with sheets of mast cells. Some scattered eosinophils and lymphocytes may be present.

Differential Diagnosis Mastocytosis can sometimes be confused with drug eruptions and urticaria if eosinophils are present. Viral exanthems and non-specific perivascular lymphoid infiltrates are also in the differential diagnosis of urticaria pigmentosa and TMEP. Mastocytoma can be confused with intradermal nevi and rarely cutaneous lymphoma/leukemia. Essential to the diagnosis of mastocytosis is the recognition of the mononuclear cells in the infiltrate as mast cells. Special stains such as a Giemsa, toluidine blue or Leder stains highlight mast cells. Immunostains for tryptase and CD117 are also effective and likely more sensitive.

Practical Tips How many mast cells are too many? A rule of thumb is that between 10-15 mast cells per HPF is the upper limit of normal. In cases of TMEP, the infiltrate tends to be mild and it can be difficult to distinguish from normal skin. In some cases of TMEP, the quantitative mast cell count is around the upper limit of normal. In such cases as this the case can be singed out as ―superficial perivascular mononuclear cell infiltrate, see comment‖ with the comment stating: ―Quantitatively the mast cells are at the upper limit of normal and could be compatible with TMEP in the appropriate clinical context. A biopsy of adjacent normal skin for comparative purposes may be helpful‖. If you are unsure of the clinical diagnosis, a diagnosis of ―mastocytosis‖ is sufficient.

Selected References 1. Mihm MC, Clark WH, Reed RJ, Caruso MG. Mast cell infiltrates of the skin and the mastocytosis syndrome. Human Pathology. 4:231-9, 1973. 2. Soter NA. The skin in mastocytosis. J Invest Dermatol.96::32S-9S, 1991.

Urticaria Clinical Features: Clinically, urticaria classically presents as transient (<24 hrs.) erythematous plaques without scale (i.e. hives). In some cases the lesions may be persistent.

Microscopic Features: The epidermis is unremarkable. The papillary dermis usually shows evidence of edema. Within the dermis the nature of the infiltrate is similar to an arthropod bite reaction in that it is a perivascular and interstitial mixed infiltrate typically rich in eosinophils. The infiltrate may be primarily superficial or superficial and deep. In contrast to arthropod bite reactions and Wells‘ syndrome, the infiltrate is typically sparse. As a result of the sparse nature of the infiltrate, the biopsy may superficially resemble normal skin on low power examination. Neutrophils and lymphocytes are often a component of the infiltrate. Collections of neutrophils within vessel lumens may be a clue in some cases.

Differential Diagnosis The differential diagnosis of urticaria is as outlined above in the section on arthropod bite reactions.

Practical Tips: An unequivocal diagnosis of urticaria is not possible in the absence of good clinical information. The infiltrate is usually mild in nature. If a dense infiltrate is present consider entities such as an arthropod bite reaction. The infiltrate may be superficial or superficial and deep. A descriptive diagnosis of ―mild perivascular mixed infiltrate consistent with a dermal hypersensitivity reaction, see note‖ is helpful. In the note a statement that ―the differential diagnosis would include dermal hypersensitivity reactions such as urticaria, a drug eruption or papular dermatitis‖ should be included.

Selected References 1. Barzilai A, Shipiro D, Shapiro D, Goldberg I, Yacob-Hirsch Y, Diaz-Cascajo C, Meytes D, Schiby R, Amariglio N, Trau H. Insect bite reaction in patients with hematologic malignant neoplasms. Arch Dermatol. 135:1503-7, 1999. 2. Fung MA. The clinical and histologic spectrum of ―dermal hypersensitivity reactions‖, a nonspecific histologic diagnosis that is not very useful in clinical practice, and the concept of a ―dermal hypersensitivity reaction pattern‖. J Am Acad Dermatol. 47:898-907, 2002.

Diagnosis: Pigmented purpuric dermatosis (Schamberg’s disease)

Clinical Features The pigmented purpuric dermatoses (PPD) are a group of inflammatory skin diseases in which lymphocytic infiltrates around venules result in extravasated erythrocytes and hemosiderin deposition, manifesting clinically as purpuric macules, brown, or golden-brown patches. Depending on the color, size and distribution of lesions cases are classified into several types, including Schamberg‘s disease (the most common), purpura annulare telangiectoides of Majocchi, lichenoid purpura of Gougerot and Blum, lichen aureus and eczematoid-like purpura of Doucas and Kapetanakis. Schamberg‘s disease may occur in persons of any age, and clinically presents as irregular patches and plaques of orange-brown, cayenne pepper-like discoloration. The lesions are chronic and may persist for years. Lichenoid purpura of Gougerot and Blum mainly affects middle-aged men and is characterized by pigmented purpura with lichenoid change. Scaling and lichenification are also seen with eczematoid-like purpura. Lichen aureus and Majocchi disease are usually seen in children or young adults. In lichen aureus, the eruption is usually a solitary lesion or localized group of lesions that may affect any part of the body; the leg is most commonly involved. Majocchi disease is characterized by small annular plaques of purpura that contain prominent telangiectasia.

Microscopic Features The histologic differences between the clinical entities of the PPPD lie in the number, pattern, and distribution of lymphocytes and number of siderophages. In Schamberg‘s disease and in purpura annulare telangiectoides of Majocchi, there is usually a

perivascular and interstitial lymphocytic infiltrate admixed with extravasated erythrocytes and/or siderophages. The epidermis is usually unremarkable but may demonstrate slight spongiosis. In eczematoid-like purpura of Doucas and Kapetanakis epidermal spongiosis is more extensive than in other variants. The infiltrate is band- like and heavy in lichen aureus and lichenoid purpura of Gougerot and Blum.

Differential Diagnosis Because of the presence of hemorrhage, leukocytoclastic vasculitis is often considered in the differential diagnosis. However, in leukocytoclastic vasculitis the inflammatory component is composed of neutrophils with associated leukocytoclasis and the vessels show fibrin deposition and sometimes frank necrosis. Stasis dermatitis may show evidence of hemorrhage but the lobular proliferation of capillaries in the superficial dermis distinguishes it from pigmented purpuric dermatosis. Lichenoid variants of PPD may show overlapping histologic features with purpuric mycosis fungoides. Both entities may demonstrate solitary lymphocytes in the lower half of the epidermis. However, edema of the papillary dermis and extravasated erythrocytes are more commonly found in PPPD.

Practical Tips The absence of overt vascular damage helps distinguish pigmented purpuric dermatosis from leukocytoclastic vasculitis, Epidermotropism of large atypical lymphocytes and lack of extravasated erythrocytes favor mycosis fungoides over PPPD.

Selected References 1. Barnhill RB, Busam KJ. Vascular diseases. In: Elder D, Elenitsas R, Jaworsky C, Johnson Jr. B. Lever’s Histopathology of the Skin, 8th ed. Philadelphia: Lippincott-Raven. 1997. 2. LeBoit, PE. Simulants of Cutaneous Lymphoma. Scientific Symposiums, October 2005. 3. Magro CM, Schaefer JT, Crowson AN, Li J, Morrison C. Pigmented purpuric dermatosis: classification by phenotypic and molecular profiles. Am J Clin Pathol. 128:218-29, 2007.

Leukocytoclastic vasculitis

Clinical features Leukocytoclastic vasculitis (LCV) presents as palpable purpura usually on the lower extremities. The process can be triggered by allergic reactions to medications or by infections. There are a variety of different forms of LCV. The case presented is a case of Henoch Schönlein Purpura (HSP). HSP accounts for >90% of LCV cases in children, but it may also present in adult patients. In addition to palpable purpura, patients have some combination of arthritis, gastrointestinal involvement and nephritis/renal failure.

Microscopic features The epidermis is usually unremarkable. Within the upper dermis there is a superficial perivascular neutrophilic infiltrate associated with leukocytoclasis, hemorrhage and

fibrinoid necrosis of blood vessels. In a given biopsy, overt fibrinoid necrosis may be absent. By immunofluorescence, there are perivascular deposits of IgA in addition to complement C3 and fibrinogen. This helps confirm the diagnosis.

Differential diagnosis As mentioned above, pigmented purpuric dermatosis could be considered but that entity has a lymphocytic infiltrate and lacks overt vascular damage. Sweet's syndrome also has a neutrophil rich infiltrate with leukocytoclasis, but the infiltrate is denser and vasculitis is typically absent. The clinical manifestations are also different. See below.

Superficial and Deep Perivascular Pattern This pattern is characterized by a perivascular infiltrate involving the superficial and deep dermal vessels. Typically there is little to no epidermal change. There are some exceptions to this in early arthropod bite reactions that show epidermal spongiosis or late lesions of lymphomatoid papulosis that can show interface change.

Lymphomatoid papulosis (LYP)

Clinical Features LYP is characterized by a recurrent eruption of papulonecrotic lesions which typically follow a benign course despite atypical histologic features. Clinical criteria currently used to diagnosis LYP are: 1. Multiple papules or nodules. 2. Spontaneous regression or waxing and waning of lesions that heal with superficial scars. 3. Lesions do not grow to a diameter larger than 3 cm during 3 months of observation without treatment. 4. Absence of lymphadenopathy. In general, lesions are asymptomatic, have a predilection for trunk and extremities, and can occur at any age from infancy to the eighth decade. Approximately 10% of the patients may subsequently develop lymphoma (mycosis fungoides, anaplastic CD30+ lymphoma and Hodgkin‘s disease).

Microscopic Features Histologic features include a superficial and deep perivascular and interstitial mononuclear cell infiltrate with atypical lymphocytes. Three morphologic subtypes have been described. The vast majority of LYP cases are type A with large atypical CD30+ cells in a mixed background of neutrophils, eosinophils, histiocytes and small lymphocytes. Type B lesions are composed of small to medium sized cerebriform cells that may demonstrate epidermotropism (MF-like). There are few if any CD30+ large cells. The so-called type C lesion shows nodules of large CD30+ cells and is histologically indistinguishable from CD30+ anaplastic large cell lymphoma (ALCL). Note: histologic subtypes are clinically and prognostically irrelevant and CD30+ large cells may be absent in very early lesions, resolving lesions, and type B LYP.

Immunophenotypic Profile Large cells activated helper cells CD30+ CD4+ (less than 5% of cases CD8+) variable loss of CD2, CD3, or CD5 cytotoxic protein expression (TIA-1, granzyme) alk-1- CD15-

Differential Diagnosis Type A LYP Reactive CD30+ lymphoid proliferations (arthropod bite, scabies, drug eruptions)

Type B LYP Mycosis fungoides Pityriasis lichenoides (PLEVA--demonstrates single necrotic keratinocytes and lack large CD30+ cells)

Mycosis Fungoides—prototype of cutaneous T-cell lymphoma Heterogenous clinical presentation (patch, plaque and tumor stages) Classic microscopic features: 1. Upper dermal band-like lymphocytic infiltrate; +/- folliculotropism 2. Variable epidermal changes, often with parakeratosis 3. Halo perinuclear effect with minimal spongiosis 4. Epidermotropism 5. Dermal fibroplasia 6. Nuclear atypia (small to medium-sized lymphocytes) 7. Some eosinophils and plasma cells may be seen in advanced MF 8. Differentiation from LYP type B requires careful clinicopathologic correlation

Type C LYP Primary cutaneous CD30 + lymphoma (anaplastic large cell lymphoma/ALCL) MF with large cell transformation

Cutaneous CD30+ lymphoma (ALCL) Clinical Features: Can present at any age with one or multiple, often ulcerated nodules. If multiple lesions are present, they usually have a regional distribution. There are three categories: 1. Primary cutaneous variant in which by definition greater than 75% of tumor cells express CD30 antigen. Prognosis is excellent. 2. Secondary variant arising in background of pre-existent mycosis fungoides or lymphomatoid papulosis. Large cell transformation of MF carries a poor prognosis (5-year survival of 11 to 19%). 3. Cutaneous spread from a systemic CD30+ lymphoma. Immune staining with ALK and EMA indicate a high likelihood of a secondary CD30 lymphoma.

Microscopic Features 1. 75% or more of tumor cells express CD30. 2. In contrast to LYP type C, the tumor often extends into the subcutaneous fat or deeper tissues.

Comparison of LYP and CD30+ ALCL

FEATURE LYP CD30+ ALCL Number of lesions Numerous Solitary or a few Distribution of lesions Limbs and abdomen Usually limited to one region Size of lesions 1.0 cm or less Usually greater than 1.0 cm Regression Invariable Uncommon Configuration of infiltrate Wedge-shaped or Nodular and diffuse nodular/diffuse CD30+ cells Scattered Confluent sheet Background Typical, often admixture Less common, at inflammation periphery Necrosis Absent Present

Type D LYP Recently, a new LYP variant has been described. Histologically it resembles aggressive CD8+ cytotoxic T-cell lymphoma. Immunophenotypically, Type D LYP is betaF1+, CD3+, CD4-, CD8+, CD30+. Clinically, it has the same presentation of other forms of LYP with waxing and waning papules.

Practical Tips Clinical correlation is paramount for correct diagnosis. Often best signed out as ―CD30+ lymphoproliferative disorder, see comment‖. Be aware of mimics in the differential diagnosis!

Selected References 1. El Shabrawi-Caelen L, Kerl H, Cerroni L. Lymphomatoid papulosis: reappraisal of clinicopathologic presentation and classification into subtypes A, B, and C. Arch Dermatol. 2004:140:441-7. 2. Cepeda LT, Pieretti M, Chapman SF, Horenstein MG. CD30-positive atypical lymphoid cells in common non-neoplastic cutaneous infiltrates rich in neutrophils and eosinophils. Am J Surg Pathol. 2003:27:912-8. 3. Gallardo F, Barranco C, Toll A, Pujol RM. CD30 antigen expression in cutaneous inflammatory infiltrates of scabies: a dynamic immunophenotypic pattern that should be distinguished from lymphomatoid papulosis. J Cutan Pathol. 2002:29:368-73. 4. McKee, P.H. Cutaneous T-Cell Lymphoma. In: Practical Dermatopathology 2001: Brigham and Womens Hospital. 5. Willemze R, Meijer CJ. Primary cutaneous CD30-positive lymphoproliferative disorders. Hematol Oncol Clin North Am. 2003; 17:1319-32. vii-viii.

6. Cardoso J, Duhra P, Thway Y, Calonje E. Lymphomatoid Papulosis Type D: A NewlyDescribed Variant Easily Confused With Cutaneous Aggressive CD8-Positive Cytotoxic T-Cell Lymphoma. Am J Dermatopathol. 2012 Oct;34(7):762-5. 7. Saggini A, Gulia A, Argenyi Z, Fink-Puches R, Lissia A, Magaña M, Requena L, Simonitsch I, Cerroni L. A variant of lymphomatoid papulosis simulating primary cutaneous aggressive epidermotropic CD8+ cytotoxic T-cell lymphoma. Description of 9 cases. Am J Surg Pathol. 2010 Aug;34(8):1168- 75.

Arthropod Bite Reaction

Clinical Features Variable. Typically patients present with pruritic, excoriated papules and vesicles. Often the clinical diagnosis is ―papular urticaria‖. Lesions may be papulonecrotic in nature prompting a clinical differential diagnosis of pityriasis lichenoides (PLEVA) or lymphomatoid papulosis (LYP).

Microscopic Features The infiltrate is eosinophil-rich, is often wedge-shaped and extends into the subcutis. The punctum, or point at which the stinger enters the skin, is manifested pathologically by an intraepidermal spongiotic vesicle. However, the epidermal changes are quite variable. A spongiotic vesicle is typically only present in recent lesions. In older lesions the epidermis can show features similar to subacute or chronic spongiotic dermatitis.

Differential Diagnosis Distinguishing histologic features of PLEVA and LYP are discussed previously in the handout. Eosinophilic cellulitis (Well‘s syndrome) has been described as a condition characterized clinically by large erythematous plaques on the trunk or extremities and histologically by a dense diffuse infiltrate of eosinophils with flame figures (degranulated eosinophils). Since the original description of this disorder, flame figures have been described in a number of eosinophil-rich processes including drug eruptions and bullous pemphigoid. Most authorities now believe Wells is not an authentic disease as much as an exaggerated hypersensitivity response. Dermal hypersensitivity reaction is a nonspecific histologic term used to encompass a number of clinical disorders including arthropod bite reaction, urticaria (see below), or drug eruption. Histopathologic features include superficial and mid-perivascular lymphocytic infiltrates with eosinophils and minimal if any epidermal alteration.

Practical tips Dx: Superficial and deep perivascular mixed cell infiltrate with eosinophils, see comment: ―The histologic features are compatible with a dermal hypersensitivity reaction. The differential diagnosis would include an arthropod bite reaction. CPC recommended‖ Patients with chronic lymphocytic leukemia may have exuberant reactions to arthropod assaults. The infiltrate in arthropod reactions are typically denser than in urticaria or drug eruptions.

In arthropod assaults there is generally less interface change than with PLEVA

Interstitial Pattern This pattern can be subdivided into ‗nodular and diffuse‘ and ‗palisading granulomatous‘ patterns. There is significant overlap with the nodular and diffuse pattern and the superficial and deep perivascular pattern. The primary difference is that the inflammation is not just centered on vessels. Distinction in individual cases is admittedly arbitrary. The palisading granulomatous pattern is characterized by a lymphohistiocytic infiltrate that surrounds altered collagen fibers.

Sweet’s syndrome

Clinical Features Sweet‘s syndrome (SS), also known as acute febrile neutrophilic dermatosis, is characterized by the acute onset of fever and leukocytosis associated with arthralgias and erythematous plaques. Lesions are most often found on the extremities and face. This entity most often occurs in middle-aged women after a non-specific respiratory or gastrointestinal infection. An association with an underlying malignancy such as leukemia is seen in about 10% of cases. SS is often seen in patients with inflammatory bowel disease, connective tissue disease, and underlying infection. The pathogenesis is not known. The clinical course is usually short-lived, and lesions respond to systemic corticosteroids.

Microscopic Features Sweet‘s syndrome is a classic manifestation of neutrophilic dermatosis, a histologic pattern characterized by the presence of a heavy dermal infiltrate of neutrophils and variable leukocytoclasis. Despite the latter finding, vascular damage is not a characteristic feature of this syndrome. Although a recent report described the presence of some vascular damage in SS, this tends to be seen in older lesions and is presumed to represent secondary vascular damage due to the release of neutrophilic enzymes.

Interestingly, recent studies have demonstrated that some cutaneous lesions of SS are histopathologically characterized by an inflammatory infiltrate composed of histiocyte-like immature myeloid cells, not polymorphonuclear leukocytes as is the norm. Deemed ―histiocytoid sweet syndrome‖ by the authors, this variant of SS may be confused with histiocytic interstitial processes such as granuloma annulare.

Differential Diagnosis Other neutrophilic dermatoses to be considered in the differential diagnosis include bowel bypass syndrome, which presents as pustular lesions and arthritis in patients who have undergone bowel bypass surgery for obesity. Rheumatoid neutrophilic dermatosis represents a rare manifestation of rheumatoid arthritis characterized clinically by symmetric nodules on extensor surfaces of joints and histologically by neutrophilic infiltrates indistinguishable from Sweet‘s syndrome. Pyoderma gangrenosum, characterized by ulcers with a raised undermined border and diffuse dermal neutrophilic infiltrate, is a diagnosis of exclusion. Leukocytoclastic vasculitis refers to a histologic combination of intramural neutrophils, leukocytoclasis, and fibrinoid necrosis of vessel walls with extravasated erythrocytes. Granuloma faciale/

erythema elevatum diutinum are forms of chronic LCV characterized by grenz zone with no edema, eosinophils and lymphohistiocytic infiltrates. Both entities feature neutrophilic infiltrates and vessels containing fibrin in their walls.

Practical Tips No true vasculitis Ulceration uncommon in Sweet‘s AML-M3 after all-trans retinoic acid (ATRA) treatment can mimic Sweet‘s Dx: Neutrophilic dermatosis, see note: ―The histologic features are consistent with Sweet‘s syndrome; if there is a clinical suspicion of an infectious etiology tissue culture is recommended‖.

Sarcoidosis Clinical features Sarcoidosis is a disease of unknown etiology characterized by noncaseating granulomas. It can be a self-limited disease, chronic cutaneous disease or systemic disease. Cutaneous lesions occur - 20-35% of patients with systemic sarcoidosis. In this setting chronic skin disease is associated with adverse outcome. Patients with sarcoidosis frequently develop granulomas at sites of trauma. Erythema nodosum is also seen in a significant percentage of patients with sarcoidosis (see below).

Microscopic features Sarcoidosis is characterized by noncaseating granulomas with poorly developed lymphocytic cuffs. Focal necrosis is rarely seen in the granulomas.

Differential Diagnosis An infectious process must always be considered. In the absence of a known history of sarcoidosis, special stains for fungi and mycobacteria (e.g. GMS, Fite) are essential. Granuloma annulare and necrobiosis lipoidica may have sarcoidal granulomas, but zones of altered collagen help distinguish them from sarcoidosis. Metastatic Crohn's disease can be indistinguishable, but cutaneous Crohn's often has eosinophils and may have epidermal ulceration, features not seen in sarcoidosis.

Practical Tips Diagnosis of exclusion: consider special stains Look for altered collagen in palisading granulomatous disease

Cutaneous Lymphoid hyperplasia

Clinical Features Most commonly presents on the head and neck as erythematous to violaceous nodule. More common in women and Caucasians. Triggering factor usually unknown.

Microscopic features Within the dermis are nodular lymphoid aggregates, often with germinal center formation. The germinal centers resemble those in lymph nodes with a mixture of centroblasts, centrocytes and tingible body macrophages. Eosinophils and plasma cells are often present.

Differential Diagnosis The primary differential diagnosis is cutaneous marginal zone B-cell lymphoma and cutaneous follicle center cell B-cell lymphoma. In marginal zone lymphoma there are sheets of monocytoid B-cells and aberrant B-cell expression of CD43. Light chain restricted plasma cells are often present. In follicle center cell B-cell lymphoma, the germinal centers are not polarized and tingible body macrophages are absent to rare.

Practical Tips Approach diagnosis with caution. Architecture of germinal centers important Consider immunostains for kappa and lambda light chains: absence of light chain restriction favors benign reactive process Always recommend continued clinical follow-up

Selected References 1. Sweet RD. Acute febrile neutrophilic dermatosis. Br J Dermatol.535:349-56, 1964. 2. Callen JP. Neutrophilic dermatoses. Dermatol Clin 20:409-419, 2002. 3. Malone JC, Slone SP, Willis-Frank LA, Fearneyhough PK, Lear SC, Goldsmith LJ, Hood AF, Callen JP. Vascular inflammation (vasculitis) in sweet syndrome: a clinicopathologic study of 28 biopsy specimens from 21 patients. Arch Dermatol 138:345-9, 2002. 4. Requena L, Kutzner H, Palmedo G, Pascual M, Fernandez-Herrar J, Fraga J, Garcia-Diez A, Sanchez Yus E. Histiocytoid sweet syndrome. A dermal infiltration of immature neutrophilic granulocytes. Arch Dermatol 141:834-42, 2005. 5. Tozman, ECS. Sarcoidosis: clinical manifestations, epidemiology, therapy and pathophysiology. Curr Opin Rhematol.3:155-59, 1991. 6. Antonovich DD and JP Callen. Development of sarcoidosis in cosmetic tattoos. Arch Dermatol. 141:869-72, 2005. 7. Bergman R, Khamaysi K, Khamaysi Z, Ben Arie Y. A study of histologic and immunophenotypical staining patterns in cutaneous lymphoid hyperplasia. J Am Acad Dermatol. 2011 Jul;65(1):112-24. 8. Arai E, Shimizu M, Hirose T. A review of 55 cases of cutaneous lymphoid hyperplasia: reassessment of the histopathologic findings leading to reclassification of 4 lesions as cutaneous marginal zone lymphoma and 19 as pseudolymphomatous folliculitis. Hum Pathol. 2005 May;36(5):505-11. 9. Baldassano MF, Bailey EM, Ferry JA, Harris NL, Duncan LM. Cutaneous lymphoid hyperplasia and cutaneous marginal zone lymphoma: comparison of morphologic and immunophenotypic features. Am J Surg Pathol. 1999 Jan;23(1):88-96.

Palisading granulomatous dermatitis

Granuloma annulare Clinical Features Granuloma annulare (GA) is a common, usually asymptomatic dermatosis of unknown etiology. Six clinical subtypes are recognized: localized, generalized, perforating, subcutaneous, linear and papular. The localized variant is the most common type and usually occurs in young adults with a female preponderance. Skin

colored papules in an annular or actuate arrangement are most commonly located at aural sites, particularly the knuckles and fingers. Lesions tend to chronicity and recurrences are common. Generalized lesions, defined as occurring on the trunk and either upper or lower extremities, are seen in approximately 15% of patients. The subcutaneous variant is most often seen in childhood.

Morphologic Features GA most commonly involves the upper and mid reticular dermis. The sine qua non of GA is the palisading granuloma: a palisade of histiocytes surrounding a central zone of altered collagen fibers associated with increased dermal mucin.

Differential Diagnosis Granuloma annulare falls into the category of palisading and granulomatous dermatitis. The main entities to consider in the differential diagnosis are necrobiosis lipoidica, rheumatoid nodule, and sarcoidosis.

Practical Tips Low power examination Palisade not always well developed Interstitial pattern common; may be subtle Certain drug eruptions may have a GA-like pattern

Necrobiosis lipoidica

Clinical Features Necrobiosis lipoidica or necrobiosis lipoidica diabeticorum (NLD) most commonly presents as yellowish brawny indurated plaques on the lower extremities, particularly the pretibial areas. An association with diabetes mellitus (either Type 1 or type 2) has been reported.

Morphologic Features NLD is characterized by palisading necrobiotic granulomata involving the mid to lower dermis. Histiocytes surround eosinophilic degenerate collagen that often exhibits a hyalinized appearance. Most notably, NLD typically has a layered appearance with horizontal zones of necrobiosis separated by uninvolved collagen. Giant cells and well-formed sarcoidal granulomata may be seen. Plasma cells are frequently a component of the infiltrate, features not typically seen in granuloma annulare.

Differential Diagnosis In contrast to GA, NLD tends to involve the entire dermis, shows diffuse necrobiosis, is more often associated with increased numbers of giant cells and tends to show less well-developed palisades of histiocytes.

Practical Tips Low power examination is the key to recognizing the tiered pattern The entire dermis is usually involved Plasma cells favor necrobiosis lipoidica over GA

In some cases the differential between granuloma annulare and necrobiosis lipoidica is not clear. In these cases a descriptive diagnosis of ―granulomatous dermatitis, see comment‖ is helpful.

Rheumatoid nodule

Clinical Features Rheumatoid nodules are subcutaneous lesions that tend to occur over bony prominences such as the extensor aspect of the forearms, elbows, hands, feet and knees. However, they can occur at virtually any site. There seems to be a correlation between severity of arthritis and development of rheumatoid nodules. Morphologic Features Lesions are located in the deep dermis, subcutaneous fat or soft tissue. Centrally areas of acellular fibrin are surrounded by histiocytes and giant cells in a palisaded pattern. Variable numbers of lymphocytes, plasma cells and eosinophils may be present.

Differential Diagnosis Deep granuloma annulare may be histologically indistinguishable from a rheumatoid nodule. Abundant mucin favors deep granuloma annulare but definitive diagnosis requires careful clinicopathologic correlation.

Selected References

1. Muhlbauer JE. Granuloma annulare. J Am Acad Dermatol. 3:217-30, 1980. 2. Dabski K and Winkelman RK. Generalized granuloma annulare: clinical and laboratory findings in 100 patients. J Am Acad Dermatol. 20:39-47, 1989. 3. McDermott MB, Lind AC, Marley EF, Dehner LP. Deep granuloma annulare (pseudorheumatoid nodule) in children: clinicopathologic study of 35 cases. Ped Develop Pathol. 1:300-8,1998. 4. Muller SA and Winkelmann RK. Necrobiosis lipoidica diabeticorum: histopathologic study of 98 cases. Arch Dermatol. 94:1-10, 1966. 5. Lowney ED and Simons HM. ‗Rheumatoid‘ nodules of the skin. Archs Dermatol. 88:221-26, 1963.

Sclerosing Dermatitis Pattern The sclerosing dermatitis pattern is characterized by dermal sclerosis, usually with minimal apparent inflammation.

Morphea/Scleroderma.

Clinical Features Morphea is also known as localized scleroderma. It is characterized by localized indurated plaques, frequently on the trunk. Scleroderma is a multisystem connective tissue disease, which is divided into two major clinical groups. Group 1: limited disease involving hands, forearms and face, often have CREST syndrome Group 2: Diffuse cutaneous sclerosis and frequent visceral involvement

Patients have indurated skin, sclerodactyly, hyperpigmentation with perifollicular pigment retention and telangiectasias

Microscopic Features Morphea and scleroderma are both characterized by fibrosing inflammation in the dermis and are histologically indistinguishable. Early lesions of morphea are characterized by superficial and deep, perivascular and interstitial infiltrates of lymphocytes and plasma cells (eosinophils or neutrophils sometimes). Some lymphocytes may be seen in the basal layer (may resemble mycosis fungoides). Thickening of collagen bundles may be subtle. In fully developed lesions of morphea inflammatory infiltrates are less dense, thickened collagen bundles are seen in the reticular dermis and there may be homogenization of the papillary dermis. Late lesions of morphea are characterized by sparse inflammation and absent or degenerative appearing adnexa.

Differential Diagnosis Late stage lesions of necrobiosis lipoidica can simulate morphea. In the former, granulomatous changes can be subtle and fibrosis marked. Both conditions can have plasma cells. A clue to the diagnosis of necrobiosis lipoidica is that an elastic tissue stain will show near absence of elastic fibers, while those are mostly preserved in morphea. Nephrogenic systemic fibrosis, formerly named nephrogenic fibrosing dermatopathy, occurs in patients with renal disease, and most but not all (90%) are undergoing renal dialysis. Patients develop large symmetric areas of hardened skin on the extremities, often with brawny hyperpigmentation. Histologically biopsies show an increase in fibroblasts in the dermis and subcutis, associated with thickened collagen bundles. Unlike morphea, increased mucin may be seen. No infiltrates of plasma cells are seen. Radiation necrosis is a late complication of radiation exposure. Changes in the reticular dermis resemble those of morphea, but fibroblasts can be large or even have bizarre nuclei. Blood vessels may have fibrin deposits in the lumens. Changes that resemble morphea rather than radiation sclerosis can occur following radiation for breast carcinoma.

Practical Tips On low power ―square biopsy‖—but must differentiate from normal back! Decreased spaces between collagen fibers Entrapped adnexal structures Elastic stains useful in differentiating late stage necrobiosis lipoidica vs morphea

Selected References 1. Jaworsky C. Connective tissue diseases. In Elder D, Elenitsas R, Jaworsky C, Johnson Jr. B. Lever’s Histopathology of the Skin, 8th ed. Philadelphia: Lippincott- 2. Raven. 1997. 3. Cowper SE, Su LD, Bhawan J, Robin HS, LeBoit PE. Nephrogenic fibrosing dermopathy. Am J Dermatopathol. 23:383-93, 2001. 4. Schaffer JV, Carroll C, Dvoretsky I, Huether MJ, Girardi M. Postirradiation morphea of the breast—presentation of two cases and review of the literature. Dermatology. 200:67-71, 2000.

Lichen Sclerosus (et atrophicus)

Clinical Features: Lichen sclerosus presents as white plaques with overlying epidermal atrophy. There is a predilection for the anogenital region, but approximately 20% may present in other locations. Microscopic Features: Early lesions may resemble an interface dermatitis with basal vacuolization and a mild lichenoid infiltrate. As the lesions progress there is edema and homogenization of the papillary dermis. In later lesions the superficial dermis is sclerotic, resembling the changes of morphea. Differential Diagnosis: The differential diagnosis of early lesions is lichen planus. Established lesions with the characteristic dermal changes are generally straightforward. In some advanced lesions, the possibility of morphea could be considered. Practical Tips Consider the possibility of early lichen sclerosus in any interface dermatitis of genital skin. Remember that lichen sclerosus may occur outside the anogenital area. In cases with histologic overlap with morphea, a descriptive diagnosis of ―sclerosing dermatitis, see note‖ with a note explaining that the differential diagnosis would include lichen sclerosus or morphea is acceptable.

Panniculitis Panniculitis is broadly grouped into septal and lobular patterns based on where the inflammatory process is present (i.e. the subcutaneous septae or lobules). Practically speaking there is overlap in almost all cases. Therefore it is important to have a stepwise approach. Evaluation of panniulitides—stepwise approach 1. The first step is to look for evidence of vasculitis. If the overlying skin and subcutis have neutrophilic inflammation in the walls of medium-sized arteries, the systemic disease polyarteritis nodosa is the diagnosis rather than a primary panniculitis. 2. Next, one should try to determine whether inflammatory changes are primarily in a septal vs lobular distribution. Both types of localized primary inflammation of the subcutis can be identified by separating them into septal and lobular distributions combined with associated types and distributions of inflammatory cells. 3. Finally, some diseases may require additional studies for definitive diagnosis (e.g. gene rearrangement studies to detect clonal T-cell or clonal B- cell populations).

Erythema Nodosum

Clinical Features Erythema nodosum (EN) is the most common type of panniculitis. It occurs at any age, with the incidence peak between 20-30 years of age. Women are more frequently involved than men. EN is characterized by an acute onset of tender, erythematous nodules or plaques, with healing without atrophy or scarring. Lesions

are sometimes associated with fever, arthralgias, fatigue, respiratory and GI symptoms. Pathogenesis is unclear; probably hypersensitivity to underlying antigens as EN is associated with different infections, drugs, malignant and inflammatory disorders. Streptococcal infections are the most frequent etiologic factor for EN in children whereas drugs, sarcoidosis and inflammatory bowel disease represent the most commonly associated disorders in adults. Microscopic Features Vary with the stage of the lesion. In early lesions there is more inflammation and less fibrosis within the septum; the inflammatory cells are typically neutrophils. In later lesions, there is more septal fibrosis; inflammatory cells are lymphocytes, histiocytes, multinucleated giant cells, and eosinophils. Often there is peripheral lobular inflammation. Vasculitis is typically not present in EN. Miescher’s radial granulomas are considered the histologic hallmark of EN. These are typically seen in late lesions and are small well-defined aggregates of histiocytes around a central stellate cleft. They are mostly within septa and may be surrounded by neutrophils. In older lesions multi-nucleated giant cells are seen with clefts within their cytoplasm. Miescher‘s granulomas may be inconspicuous in some cases and serial sections may be necessary to visualize them. Finding Miescher's granulomas are not requisite for the diagnosis.

Differential Diagnosis Other forms of panniculitis including erythema induratum, lipodermatosclerosis and lupus panniculitis. Infectious panniculitis may be a consideration especially in earlier lesions, when neutrophils may be prominent. Sarcoidosis falls in the differential diagnosis in well-developed lesions of EN.

Practical Tips An adequate biopsy (preferably a deep wedge) is required for optimal evaluation of the subcutaneous adipose tissue Low power examination is important for dividing the inflammatory process into the septal or lobular patterns. The inflammation is at the periphery of the lobules.

Erythema Induratum (Nodular Vasculitis)

Clinical Features: Erythema induratum presents as recurrent painful nodules, most frequently on the calves. It is associated with an underlying tuberculosis infection in 30-80% of cases.

Microscopic Features: Erythema induratum is classically a lobular panniculitis but secondary septal inflammation is commonly seen. Within the lobules there is a granulomatous inflammation with some evidence of vasculitis. The vasculitic changes can show frank fibrinoid necrosis in early lesions to endothelial swelling and a mixed inflammatory infiltrate in the vessel walls.

Differential Diagnosis: Erythema nodosum, as discussed above, is the primary entity in the differential diagnosis. Polyarteritis nodosa is also in the differential diagnosis. In polyarteritis nodosa, the inflammation of the fat lobules is more restricted to the immediate area

around the damaged vessels rather than the more diffuse pattern of erythema nodosum.

Practical Tips Low power examination crucial. Look for evidence of vascular damage. The inflammatory process involves the entire lobule.

Lipodermatosclerosis

Clinical Features: Lipodermatosclerosis presents as indurated plaques involving the lower extremities. It is associated with venous/arterial insufficiency, and previous thrombophlebitis.

Microscopic Features: Microscopically, lipodermatosclerosis appears relatively non-inflammatory. There is septal and lobular fibrosis. Within the lobules there is formation of the characteristic microcyst. The microcysts are characterized by cystic spaces lined by an acellular eosinophilic material that has been described as having a feathery or crenulated appearance. Given the clinical setting, there is sometimes overlying stasis dermatitis.

Differential Diagnosis: The differential diagnosis can include end stage lesions of erythema nodosum or erythema induratum. Subcutaneous morphea is also in the differential diagnosis. The presence of the microcysts is the key to differentiating lipodermatosclerosis.

Practical Tips Lipodermatosclerosis is relatively non-inflammatory. Microcysts are the key diagnostic feature. Clinical history of venous insufficiency.

Selected References 1. Requena L, Yus ES. Panniculitis. Part I. Mostly septal panniculitis. J Am Acad Dermatol. 45:163-83, 2001. 2. Requena L, Yus ES. Panniculitis. Part II. Mostly lobular panniculitis. J. Am Acad Dermatol. 45:325-61, 2001. 3. Dahl PR, Su WPD. Panniculitis. In: Hood AF and Farmer ER. Pathology of the Skin, 2nd ed. McGraw Hill Publications. 2000.

Bullous Pattern Bullous diseases are subdivided into subepidermal and intraepidermal patterns. In subepidermal blistering disease, the split occurs at the dermoepidermal junction. In intraepidermal blistering disease, the blister forms as the result of acantholysis within the epidermis. This group is restricted to diseases in which blister formation is the primary manifestation rather than blisters as a secondary manifestation (i.e. Blisters secondary to contact dermatitis would not be considered in this category.) The more common entities in this group of diseases will be discussed.

Bullous pemphigoid

Clinical Features Bullous pemphigoid (BP) is the most common autoimmune blistering disorder. It generally affects elderly patients and is characterized by tense bullae arising on normal skin or on an erythematous base. There is a predilection for extremities and the lower abdomen. Oral mucosal involvement develops in 1/3 patients. Pruritus is a common feature at presentation. The urticarial variant of BP is characterized by erythematous macules and papules as well as urticarial plaques.

Etiology In BP autoantibodies are produced primarily to two BMZ antigens: BP180: intercellular hemidesmosomal protein (plakin family) P230: transmembrane protein, most common autoantigen (type XVII collagen)

Microscopic Features BP is characterized by a subepidermal blister with eosinophils and other inflammatory cells in and around the blister. Rarely the inflammatory process can be neutrophil predominant or cell-poor. Inflammation is usually confined to the papillary dermis. Urticarial forms of BP demonstrate eosinophilic spongiosis.

Direct immunofluorescence (method to detect in vivo tissue bound antibodies) 90% linear C3 at DEJ 60-90% linear IgG at DEJ

Indirect Immunofluorescence (method to detect circulating antibodies) About 70% of patients have circulating IgG autoantibodies directed against the epidermal basement membrane zone. Routine IIF studies use normal human skin or monkey esophagus IIF using salt split skin –BP patients have antibodies which bind to the epidermal side

Direct immunofluorescence on Salt Split Human Skin Since 20-30% of BP patients may not have circulating antibodies, the SSS IIF method has limited utility. This DIF method utilizes the SSS technique to map the in situ distribution of immunoreactants in the patients own skin

Differential Diagnosis The differential diagnosis of bullous pemphigoid includes other subepidermal immunobullous diseases including pemphigoid gestationis (bullous pemphigoid of pregnancy, aka herpes gestationis), epidermolysis bullosa acquisita, and dermatitis herpetiformis. Although pemphigus vulgaris is typically classified as an intraepidermal blistering disorder, eosinophilic spongiosis may occur in urticarial lesions, similar to bullous pemphigoid. These entities are all discussed below.

Practical Tips For accurate DIF studies biopsy perilesional skin! False negative DIF results may be seen in biopsies from the lower extremities and lesional biopsies. Always consider urticarial phase of bullous pemphigoid when a biopsy from an elderly patient demonstrates eosinophilic spongiosis. Eosinophils tagging basal layer common finding in pre-bullous phase.

Pemphigoid Gestationis (PG—aka herpes gestationis)

Clinical Features PG was originally named herpes gestationis on the basis of the morphological similarity to herpetic blisters; however, this term is a misnomer because PG is not related to or associated with any active or prior herpes virus infection. PG typically manifests during the second or third trimester of pregnancy and is characterized by an abrupt onset of pruritic urticarial papules and blisters on the abdomen and trunk (often periumbilical). The eruption may spread peripherally, often sparing the face, palms and soles.

Etiology PG is a pregnancy-associated autoimmune disease. Most patients develop antibodies against BP180 and BP230.

Microscopic Features The histologic features are histologically indistinguishable from bullous pemphigoid: subepidermal blister with an eosinophil-predominant infiltrate.

Immunopathology DIF: All patients have linear deposits of C3 at the DEJ IIF: Most patients (50-75%) will have circulating antibodies which recognize the 180KD bullous pemphigoid antigen Practical Tips Resembles bullous pemphigoid Pregnancy

Epidermolysis Bullosa Acquisita (EBA)

Clinical Features EBA is a rare, non-hereditary subepidermal bullous disorder. Patients classically present with acrally distributed blisters that heal with scarring and milia formation (noninflammatory). However, occasionally lesions may also be vesiculo-bullous in nature very similar to bullous pemphigoid (inflammatory).

Etiology EBA antibodies recognize a 290 KD molecule which has been identified as type VII collagen, a major component of anchoring fibrils.

Microscopic Features The histology of EBA is that of a subepidermal blister with fibrin and only a few inflammatory cells in the lumen. In non-inflammatory lesions there is a sparse

superficial lymphocytic infiltrate around the vessels of the superficial vascular plexus, while in inflammatory lesions there is a heavy upper dermal inflammatory infiltrate composed of lymphocytes, neutrophils and eosinophils. In older lesions there will be some dermal scarring and milia.

Immunopathology DIF: All patients have linear deposits of IgG at the basement membrane zone. Some patients also have linear deposits of C3 at the basement membrane zone. IIF: About 50% of patients have circulating IgG antibody directed against the basement membrane zone. IIF on SSS-the antibodies bind to the dermal side of split skin (vs epidermal side for BP).

Practical Tips Usually non-inflammatory. Blisters in trauma prone areas.

Dermatitis Herpetiformis (DH)

Clinical Features DH is a rare, chronic subepidermal blistering disorder characterized by intensely pruritic papules and vesicles. Elbows, knees and shoulders are the most common areas of involvement. The majority of patients have associated gluten sensitive enteropathy that is subclinical but demonstrable on bowel biopsy. 80% of patients have improvement of skin disease when put on a gluten-free diet.

Microscopic Features In early lesions of DH neutrophils are observed at the tips of dermal papillae (―papillary microabscesses‖). Fibrin may be present near the tips of dermal papillae giving a ―necrotic:‖ appearance. Subepidermal vesiculation with neutrophils is seen in older lesions. Blood vessels in the upper and mid dermis are surrounded by an infiltrate of lymphocytes, histiocytes, neutrophils and eosinophils.

Immunopathology DIF: all patients have IgA deposited in a granular fashion at the dermal papillary tips; some patients may also have C3 deposits. IIF: Indirect immunofluorescence is negative. Practical Tips Neutrophil predominant infiltrate. No evidence of vascular damage.

Pemphigus vulgaris (PV)

Clinical Features PV is a rare blistering disease which affects older individuals, males and females equally. The primary lesion is a flaccid (loose) blister. Blisters are fragile and break easily leaving large erosions. Frequent sites of involvement include scalp, face, axilla, and groin. Oral mucous membrane involvement is seen in the large majority of patients. PV is a severe disease that, if let untreated, can often lead to death.

Etiology PV antigens are identified as 130 KD (desmoglein III) and 85 KD (plakoglobin).

Microscopic Features Established lesions of PV are suprabasilar blisters with acantholysis. The basal cells remain attached to the dermis giving a ―tombstone‖ appearance. Rarely PV may present with histologic changes of eosinophilic spongiosis, similar to BP.

Immunopathology DIF: IgG deposited in the intercellular regions of the epidermis. IIF: Most patients have circulating IgG antibodies which bind to the keratinocyte cell surface. Practical Tips Acantholysis key feature

Differential diagnosis Grover‘s disease can closely resemble pemphigus vulgaris. Grover‘s disease clinically most commonly presents as small pruritic papules rather than flaccid blisters. It is more common in men and usually presents on the trunk. It is often associated with recent sun-exposure, frequently related to a vacation to a warm climate. Histologically, Grover‘s disease has pemphigus like acantholysis but also corps ronds and corps grains (round or rice grain like cells in granular layer with dense keratohyaline granules) similar to that seen in Darier‘s disease. DIF studies are negative in Grover‘s disease.

Selected References 1. Schmidt E, Zillikens D. Autoimmune and inherited subepidermal blistering diseases: advances in the clinic and laboratory. Adv. Dermatol. 15:113-57, 2000. 2. Jordon R. Atlas of Bullous Disease. Philadelphia: Churchill Livingstone. 2000. 3. Ambros-Rudolph CM, Mullegger RR, Vaughan-Jones SA, Kerl H, Black MM. The specific dermatoses of pregnancy, revisited and reclassified. J Am Acad Dermatol 54:395-404, 2006.

Inflammatory diseases clinically mimicking neoplasms

Rosacea Rosacea begins as an erythematous eruption on the central face, especially the cheeks and around the nose. Occasionally the appearance may cause a clinician to consider the possibility of a basal cell carcinoma. Microscopic Features The histologic findings of rosacea are variable, most often consisting of a perivascular and perifollicular lymphohistiocytic infiltrate. Occasional cases have prominent granulomas and may be confused with sarcoidosis. Differential Diagnosis The histologic features are often non-specific, but the pattern described above in the clinical setting of a suspected basal cell carcinoma is quite suspicious for the diagnosis of rosacea. In such cases it is prudent to obtain deeper levels to exclude that

possibility. In granulomatous rosacea, sarcoidosis or infection could be considered. Usually the granulomas of granulomatous rosacea have a more developed lymphocytic cuff. The clinical suspicion for basal cell carcinoma is a also clue that suggests rosacea over sarcoidosis. Practical tips: Keep a high index of suspicion for rosacea in biopsies from the central face if a basal cell carcinoma is suspected clinically and the biopsy shows nonspecific perivascular and perifollicular inflammation. In such cases, sign them our descriptively as "perivascular and perifollicular infiltrate, see note". In your note suggest that these findings, while nonspecific suggest the possibility of rosacea.

Chondrodermatitis Nodularis Helicis Clinical Features Chondrodermatitis nodularis helicis has a very specific clinical presentation presents as a crusted or ulcerated nodule on the ear in middle aged to older patients. In men it presents almost exclusively on the helix. In women the antihelix is the most common location. Clinically, it can be confused with squamous cell carcinoma or, less frequently, basal cell carcinoma. The etiology of this process is thought to be related to chronic trauma. Some consider it a localized pressure ulcer. Microscopic Features The epidermis is usually, but not always, ulcerated. Adjacent to the ulcer the epidermis shows pseudoepitheliomatous hyperplasia. Immediately beneath the ulcer there is the characteristic fibrinoid degeneration of the collagen. Beneath the fibrinoid material is a reactive vascular proliferation but with relatively little inflammation. Depending on how deep the process goes a reactive proliferation of perichondrial fibroblasts and degenerative changes of the collagen may be seen. Differential Diagnosis The most common entity in the differential diagnosis is squamous cell carcinoma. There may be some reactive atypia but hyperchromasia and atypical mitotic figures are not seen in chondrodermatitis nodularis helicis. The fibrinoid degeneration of the dermal collagen is the key histologic feature. The histologic findings are actually quite distinctive, and misdiagnosis is rare when one is aware of the features. Practical Tips Keep a high index of suspicion on any biopsy from the ear. Look for the fibrin in the dermis.

Stasis Dermatitis As previously mentioned, stasis dermatitis is sometimes confused with neoplasms, most commonly squamous cell carcinoma. Recognition of the characteristic vascular changes and absence of atypia allow for recognition.

Common Morphologic Patterns in Soft Tissue Tumors with an Emphasis on Useful Ancillary Diagnostic Techniques

John R. Goldblum, MD

Chairman, Department of Anatomic Pathology Cleveland Clinic Professor of Pathology Cleveland Clinic Lerner College of Medicine [email protected]

- II -

2

Recognizing patterns at low magnification is one of the keys to organizing differential diagnoses in soft tissue sarcomas. Each pattern elicits a differential diagnosis, and often a combination of light microscopic and immunohistochemical features will allow one to classify a neoplasm with a given morphologic pattern. Below is a description of some of the more common morphologic patterns seen in soft tissue sarcomas with some practical points on the approach to tumors with these morphologic patterns.

Well-Differentiated Lipomatous Tumors Well-differentiated lipomatous tumors are among the most common soft tissue consults I receive. Some of these cases are actually benign lipomatous tumors which are thought to be malignant (most commonly pleomorphic lipoma), but some are liposarcomas (well- differentiated liposarcoma, also known as atypical lipomatous tumor), which are thought to be lipomas. I always have a high index of suspicion of malignancy when the tumor is large and deeply situated, especially in the retroperitoneum, but on occasion, large deeply situated tumors are benign. Although careful morphologic assessment is the traditional (and most reliable) way to distinguish benign from malignant well-differentiated lipomatous tumors, newer ancillary testing is now available (MDM2 amplification by fluorescence in situ hybridization -- discussed below), which, in my opinion, is an extremely useful diagnostic adjunct for this problem.

The vast majority of well-differentiated lipomatous tumors that all pathologists encounter are run-of-the-mill lipomas, which really do not cause much of a diagnostic problem. Most of these lesions are small and superficial and, based upon those facts alone, are easy to recognize as benign. However, there are a number of benign lipomatous tumors that can cause diagnostic difficulty. For example, lipomas with fat necrosis or hibernomatous change can cause problems for those pathologists who are "lipoblast hunters." I virtually never look for lipoblasts in this setting for the following reasons: Lipoblasts are not even required to make a diagnosis of well-differentiated liposarcoma/atypical lipomatous tumor (WDL/ALT, discussed below). Lipoblasts actually may be found in benign lipomatous tumors. The identification of lipoblasts, like many things in pathology, is in the eye of beholder, as many cells/conditions can result in pseudolipoblasts (e.g. fat necrosis, fat atrophy, poor fixation, tumors that infiltrate fat, signet ring cells, etc).

3

This discussion will focus on some of the benign lipomatous tumors that can be easily mistaken for a WDL/ALT. Intramuscular lipoma is certainly one of these, since, by definition, it is deeply located and is very often quite large at the time of excision.1 Thus, the fact that intramuscular lipoma is large and deeply situated lends suspicion of malignancy. Histologically, intramuscular lipoma is characterized by close intercalation of mature adipocytes and skeletal muscle, not only at the periphery of the tumor, but also in the central zone, resulting in a characteristic "checkerboard" pattern. Moreover, the enlarged hyperchromatic nuclei found in WDL/ALT are not present, although one must be cautious not to confuse atrophic skeletal muscle fibers (which can look quite bizarre) with the aforementioned atypical cells of WDL/ALT. Finally, because they are deeply situated and, therefore, difficult to completely excise, intramuscular lipomas may locally recur. Simply because a large, deep-seated, well-differentiated lipomatous tumor recurs does not necessarily mean it is malignant. If complete excision of an intramuscular lipoma can be performed, it will have essentially no risk of local recurrence. In all honesty, I now have a fairly low threshold for utilizing MDM2 FISH testing to prove an intramuscular lipoma (expecting the result to the negative), but this will discussed below.

The other benign lipomatous tumor that seems to cause most difficulty is pleomorphic lipoma/spindle cell lipoma, which simply represent a histologic spectrum of the same entity. In 1981, Shmookler and Enzinger described 48 cases of a lipomatous lesion characterized by mature adipose tissue admixed with bizarre, pleomorphic, multinucleated giant cells.2 Many of these bizarre cells had a floret-like arrangement of nuclei, and were often associated with interlacing bundles of dense collagen. 83% of these lesions occurred in males, with a mean age of 57 years, and 78% of the cases occurred either on the shoulder, neck or back. None of the cases recurred, confirming the clinical benignancy of this lesion. Thus, pleomorphic lipoma is a unique variant of lipoma occurring in a particular clinicopathologic setting; that is, a well-circumscribed lesion in the subcutis of a middle-aged or elderly male in the shoulder, neck or back region.

Grossly, pleomorphic lipomas are typically well circumscribed and are sharply demarcated from the adjacent mature adipose tissue. Although the average size of this tumor is close to 4.0 cm, they may be significantly larger (up to 12 cm). Histologically, this lesion is characterized by multinucleated floret cells with a wreath-like arrangement of hyperchromatic nuclei. These cells are admixed with mature lipocytes and dense, birefringent collagen fibers. Occasionally, there is a prominent myxoid stroma composed primarily of hyaluronic acid. Lipoblast-like cells have been described in up to 50% of cases. Thus, this tumor does have overlapping features with WDL/ALT. However, given the characteristic

4

clinicopathologic setting and the superficial nature of the lesion, this tumor can be distinguished from WDL/ALT or other pleomorphic sarcomas.

In 1975, Enzinger and Harvey described a lesion (spindle cell lipoma) with a similar clinicopathologic setting as that seen in pleomorphic lipoma, but characterized histologically by a mixture of lipocytes and uniform bland spindled cells within a myxoid stroma and accompanied by dense collagen.3 Spindle cell lipoma is part of a spectrum with pleomorphic lipoma, given the similar clinicopathologic setting and overlapping histologic features. Up to 25% of cases of pleomorphic lipomas show areas indistinguishable from spindle cell lipoma. Furthermore, both spindle cell lipoma and pleomorphic lipoma typically stain diffusely for CD34.4 The spindle cells are often deposited in a prominent myxoid stroma composed primarily of hyaluronic acid, and in those cases in which the lipomatous component is inconspicuous, differentiation from other myxoid tumors, including myxoid sarcomas, may be difficult. Both can also have little or no lipomatous component (fat-free spindle cell5 or pleomorphic lipoma6), which certainly can add to the diagnostic confusion. Similar to pleomorphic lipomas, spindle cell lipomas are treated with local excision, and virtually never recur. Finally, cytogenetic studies have also linked spindle cell and pleomorphic lipoma, both of which generally show monosomy 16 or partial loss of the long arm of chromosome 16 in association with unbalanced alterations of the long arm of chromosome 13.7 As discussed below, these cytogenetic alterations differ from those found in WDL/ALT.

Liposarcoma is one of the most common types of adult soft tissue sarcoma. Although one could get bogged down in a complex classification of liposarcoma, I think it is easiest to conceptualize of liposarcoma as having three distinct subtypes, each of which is characterized by its own clinical, histologic and cytogenetic features. Well- differentiated/dedifferentiated liposarcoma can be thought of as a histologic spectrum ranging from a low-grade sarcoma (WDL/ALT) to what is typically a high-grade sarcoma (dedifferentiated liposarcoma), which represents a form of tumor progression that occurs over a period of time. As described below, these lesions have characteristic cytogenetic aberrations that are distinct from those found in other subtypes of liposarcoma. Similarly, myxoid/round cell liposarcoma can also be thought of as a spectrum of lipomatous tumors in which the round cell variant represents the poorly differentiated form of myxoid liposarcoma. Similar to the concept of well-differentiated/dedifferentiated liposarcoma, round cell liposarcoma is a form of tumor progression characterized by a more aggressive clinical course. Finally, one can think of pleomorphic liposarcoma as standing on its own, distinct from well-differentiated/dedifferentiated and myxoid/round cell liposarcoma. Pleomorphic liposarcoma is an aggressive high-grade sarcoma characterized by pleomorphic lipoblasts

5

that may arise in a high-grade spindle cell or round cell/epithelioid cell background. Each of these entities will be described in further detail below.

Well-differentiated/Dedifferentiated Liposarcoma The nomenclature of well-differentiated lipomatous tumors has been fraught with confusion, and a variety of terms have been employed for the same lesion. Before 1979, differentiated lipomatous tumors characterized by atypical stromal cells intermingled with mature fat and variable numbers of lipoblasts were all designated as well-differentiated liposarcomas (WDL), whether they were found in the subcutaneous tissue, the deep soft tissues of the extremity, or the retroperitoneum.8 However, in 1979, Evans et al proposed a change in nomenclature because of the variability of clinical behavior depending on site.9 These authors evaluated 30 well-differentiated lipomatous lesions, all of which were histologically similar but varied according to site. Nine cases were found within the subcutaneous tissue, and none of these cases recurred, dedifferentiated (to be discussed below), metastasized, or resulted in patient death. Of 13 lesions in the deep soft tissue (intramuscular) of the extremities, 9 cases (69%) recurred, although similar to the subcutaneous lesions, none dedifferentiated, metastasized, or resulted in patient death. Of the 8 retroperitoneal lesions, 5 recurred (62%), and although none of the cases dedifferentiated or metastasized, 3 patients (37%) died of their disease. Based upon this data, Evans et al proposed that the lesions in the subcutaneous tissue be called “atypical lipoma,” and the intramuscular extremity lesions be called “atypical intramuscular lipoma,” given their lack of associated morbidity or mortality. However, he proposed that the term “well-differentiated liposarcoma” be retained for histologically identical lesions of the retroperitoneum, given their propensity to recur and occasionally result in patient death.

In 1987, Azumi and colleagues performed a similar study on 69 well-differentiated lipomatous lesions.10 Similar to Evans' study, none of the 17 subcutaneous lesions recurred, dedifferentiated, metastasized, or resulted in patient death. Although 29% (7/31 cases) of the deep soft tissue lesions recurred, none resulted in significant morbidity or mortality. Of 21 retroperitoneal lesions, 14 recurred (67%) and 5 dedifferentiated (23%). Although none of the cases metastasized, 9 patients (43%) died of disease. Thus, these authors proposed that the subcutaneous and deep soft tissue lesions be called “atypical lipoma” and opted to retain the term “well-differentiated liposarcoma” for the retroperitoneal lesions.

In 1992, Weiss and Rao re-analyzed a large group of well-differentiated lipomatous tumors with a minimum follow-up of 2 years, and found that behavior was strongly influenced by

6

location, with retroperitoneal lesions having the worst prognosis, deep soft tissue lesions having the best prognosis, and inguinal lesions having a prognosis in between.11 Although their data on retroperitoneal lesions is similar to the other studies (recurrence rate: 91%; dedifferentiation: 17%; metastasis: 17%; death due to disease: 33%), 3 of 46 cases from the deep soft tissue of the extremities (6%) showed areas of dedifferentiation. These authors concluded that dedifferentiation is not a site-dependent, but rather a time-dependent phenomenon, and is observed in locations with a high likelihood of clinical persistence of disease. Thus, they recommend the use of the term “well-differentiated liposarcoma” for lesions in all locations, except those located in the subcutis, which are usually easily cured at initial excision and do not have the opportunity to dedifferentiate. They proposed the term “atypical lipoma” for these subcutaneous lesions. Others have proposed referring to all of these tumors as "atypical lipomatous tumor" (ALT), regardless of location.

Histologically, WDL/ALT is characterized by lipocytes that vary in size and shape often divided into irregular lobules by dense fibrous septa. Within these septa and/or between lipocytes, there are scattered enlarged, hyperchromatic atypical nuclei that are not seen in typical lipomas. Lipoblasts, characterized by cytoplasmic lipid vacuoles that indent or scallop the nucleus, can be identified but are few in number and are not absolutely necessary for the diagnosis.

The term “dedifferentiated liposarcoma” was coined by Evans in 1979, and refers to a well- differentiated liposarcoma that is associated with or has progressed to a high-grade sarcoma, typically resembling a high-grade MFH or fibrosarcoma.12 Dedifferentiation most commonly occurs in retroperitoneal lesions, occurs less commonly in lesions of the deep soft tissue, and exceptionally in subcutaneous liposarcomas,11,13,14 and this process appears to be a time-dependent phenomenon.11

Traditionally, dedifferentiated liposarcoma has only been diagnosed when the dedifferentiated area resembles a high-grade sarcoma. However, it is apparent that there is a spectrum of histologic grade and extent of dedifferentiation within this group of tumors, but the effects of these parameters on clinical outcome are largely unknown. To address this issue, Henricks et al studied 155 cases of dedifferentiated liposarcoma, 130 of which had clinical follow-up (median: 3.0 years).14 Although the majority of these tumors occurred in the retroperitoneum (68%), 21% of the cases were peripherally located, either on the trunk or within the deep soft tissues of the extremities. In most cases (86%), the dedifferentiated foci were present at the time of initial diagnosis (de novo dedifferentiation). In particular, the authors more fully describe the non-lipogenic areas which were more cellular than classic

7

WDL/ALT but had not yet reached the degree of cellularity, atypia and mitotic activity characteristic of full-fledged high-grade dedifferentiation, and coined the term "low-grade dedifferentiation" (LGDD). In many cases, the LGDD areas surrounded the high-grade areas, but in some cases, they represented the most histologically progressed area of the entire tumor, without any evidence of high-grade dedifferentiation. Overall, 41% of the patients developed recurrences, 17% developed metastases, and 28% died of disease. The authors found no significant difference in prognosis relative to the extent of dedifferentiation for the range examined (<25% of the tumor versus ≥25% of the tumor). Furthermore, there was no significant difference in survival between patients with low-grade versus high-grade dedifferentiation. Thus, the authors proposed that the traditional definition of dedifferentiated liposarcoma be expanded to include liposarcomas with areas of low-grade non-lipogenic dedifferentiation.

More recently, Harry Evans readdressed this issue in his study of 61 previously unpublished cases of WDL/ALT with and without dedifferentiation.15 Following careful histologic review, he categorized the 61 tumors as conventional ALT (n=15), cellular ALT (n=21), dedifferentiated liposarcoma (n=24) and ALT with a pleomorphic liposarcoma-like component (n=1). By definition, dedifferentiated liposarcoma had non-lipogenic areas of moderate to high cellularity with at least 5 MF per 10 HPF. Cellular ALT, by description, is very similar to LGDD, as defined by Henricks and colleagues. According to Evans, cellular ALT had non-lipogenic areas of increased cellularity not meeting the criteria for a dedifferentiation liposarcoma or pleomorphic liposarcoma-like component, usually of moderate cellularity (but occasionally highly cellular) and having fewer than 5 MF per 10 HPF. The most important prognostic factor was tumor location, as centrally located tumors (i.e. retroperitoneal) acted in a more clinically aggressive fashion than tumors at other sites. Tumors in this location were characterized by multiple local recurrences, often multiple, regardless of histologic category. Metastatic disease only occurred in patients with dedifferentiated liposarcoma (not in cellular ALT), although some of these tumors started as ALT and became dedifferentiated with subsequent metastases. Survival was significantly shorter for patients with dedifferentiation liposarcoma (median survival 77 months) than the other categories, and there was no significant difference in survival between cellular ALT (median survival 142 months) and conventional ALT (median survival 209 months).

As previously mentioned, WDL/ALT and dedifferentiated liposarcoma have cytogenetic alterations that differ from those found in other lipomatous tumors, showing giant marker or ring chromosomes derived from 12q13-15.7 Furthermore, these tumors are characterized by amplification of MDM2 and CDK4 genes on chromosome 12q13-15.16 Amplifications of

8

MDM2 and CDK4 genes may be demonstrated by a variety of techniques, including comparative genomic hybridization, quantitative PCR and fluorescence in situ hybridization (FISH). In addition, antibodies are available to both MDM2 and CDK4, potentially serving as adjunctive techniques in discriminating WDL/ALT from its mimics. For example, Binh et al found immunoexpression of MDM2 and CDK4 in 97% and 92% of WDL/ALT, respectively.17 Very few benign lipomatous tumors expressed either of these antigens (MDM2 - 5%; CDK4 - 2%). The sensitivity and specificity of MDM2 and CDK4 immunoreactivity in identifying WDL/ALT among other soft tissue tumors were 97% and 92%, and 83% and 95%, respectively. A very strong correlation was observed between MDM2 and CDK4 immunoreactivity and gene amplification status, as determined by FISH, and we have also found this to be true.18

9

References

1. Kransdorf MJ, Moser RP, Meis JM, et al. Fat-containing soft-tissue masses of the extremities. Radiographics 1991;11:81-106. 2. Shmookler BM, Enzinger FM. Pleomorphic lipoma: a benign tumor simulating liposarcoma. A clinicopathologic analysis of 48 cases. Cancer 1981;47:126-133. 3. Enzinger FM, Harvey DA. Spindle cell lipoma. Cancer 1975;36:1852-1859. 4. Suster S, Fisher C. Immunoreactivity for the human hematopoietic progenitor cell antigen (CD34) in lipomatous tumors. Am J Surg Pathol 1997;21(2):195-200. 5. Billings SD, Folpe AL. Diagnostically challenging spindle cell lipomas: a report of 34 "low- fat" and "fat-free" variants. Am J Dermatopathol 2007;29:437-442. 6. Sachdeva MP, Goldblum JR, Rubin BP, et al. Low-fat and fat-free pleomorphic lipomas: a diagnostic challenge. Am J Dermatopathol (in press) 7. Fletcher CDM, Akerman M, Dal Cin P, et al. Correlation between clinicopathological features and karyotype in lipomatous tumors. A report of 178 cases from the Chromosomes and Morphology (CHAMP) Collaborative Study Group. Am J Pathol 1996;148:623-630. 8. Enzinger FM, Winslow BJ. Liposarcoma: a study of 103 cases. Virchows Arch Pathol Anat 1962;335:367-388. 9. Evans HL, Soule EH, Winkelmann RK. Atypical lipoma, atypical intramuscular lipoma, and well-differentiated retroperitoneal liposarcoma: a reappraisal of 30 cases formerly classified as well-differentiated liposarcoma. Cancer 1979;43:574-584. 10. Azumi N, Curtis J, Kempson RL, Hendrickson MR. Atypical and malignant neoplasms showing lipomatous differentiation: a study of 111 cases. Am J Surg Pathol 1987;11(3):161-183. 11. Weiss SW, Rao VK. Well-differentiated liposarcoma (atypical lipoma) of deep soft tissue of the extremities, retroperitoneum, and miscellaneous sites: a follow-up study of 92 cases with analysis of the incidence of “dedifferentiation.” Am J Surg Pathol 1992;16(11):1051-1058. 12. Evans HL. Liposarcoma: A study of 55 cases with a reassessment of its classification. Am J Surg Pathol 1979;3(6):507-523. 13. Yoshikawa H, Ueda T, Mori S, et al. Dedifferentiated liposarcoma of the subcutis. Am J Surg Pathol 1996;20(12):1525-1530. 14. Henricks WH, Chou J, Goldblum JR, Weiss SW. Dedifferentiated liposarcoma: A clinicopathologic analysis of 155 cases. Am J Surg Pathol 1997;21(3):271-281. 15. Evans HL. Atypical lipomatous tumor, its variants and its combined forms: a study of 61 cases, with a minimum follow-up of 10 years. Am J Surg Pathol 2007;31:1-14. 16. Dei Tos AP, Doglioni C, Paccinin S, et al. Coordinated expression and amplification of the MDM2, CDK4 and HMGI-C genes atypical lipomatous tumours. J Pathol 2000;190:531-536. 17. Binh M, Sastre-Garau X, Guillou L, et al. MDM2 and CDK4 immunostainings are useful adjuncts in diagnosing well-differentiated and dedifferentiated liposarcoma subtypes: a comparative analysis of 559 soft tissue neoplasms with genetic data. Am J Surg Pathol 2005;29:1340-1347. 18. Weaver J, Downs-Kelly E, Goldblum JR, et al. Fluorescence in situ hybridization for MDM2 gene amplification as a diagnostic tool in lipomatous neoplasms. Mod Pathol 2008;21:943-949.

10

"MFH-Like" Pattern The concept of malignant fibrous histiocytoma (MFH) has undergone significant change over the past five decades. The term was first introduced in 1963 to refer to a group of soft tissue tumors characterized by a storiform or cartwheel-like growth pattern, which were believed to be derived from histiocytes on the basis of early tissue culture studies demonstrating ameboid movement and phagocytosis of explanted tumor cells.1,2 Ultrastructural studies both endorsed and refuted the histiocytic origin of these tumors, however. With the advent of immunohistochemistry and the accessibility of numerous monoclonal antibodies directed against various structural proteins of specific cell types, the phenotype of this tumor was shown to be more closely aligned with a fibroblast than a histiocyte.3–6 Furthermore, many, but not all, lesions labeled as “malignant fibrous histiocytoma” could, upon close scrutiny, be subclassified as lineage-specific sarcomas, an observation that led some to question the existence of MFH as a distinct entity.7 The extent to which such lesions can be subclassified as sarcomas of alternative type is, in large part, dependent on definitional criteria and the number of ancillary studies a pathologist is willing to bring to bear on the evaluation of a pleomorphic sarcoma. There is still no general agreement as to what percentage of pleomorphic sarcomas, when subjected to rigorous evaluation, remain unclassified. These discrepancies, nonetheless, underscore the fact that the criteria by which a pleomorphic tumor is provisionally labeled as a undifferentiated pleomorphic sarcoma (UPS/MFH) as well as the criteria by which some are reclassified differ from institution to institution. Whatever the true incidence of this lesion, there is agreement that the term MFH should be used synonymously with undifferentiated pleomorphic sarcoma (UPS) which, by a combination of sampling and immunohistochemistry, shows no definable line of differentiation and by electron microscopy manifests fibroblastic/myofibroblastic features.8 At this point, we use the term undifferentiated pleomorphic sarcoma in our diagnostic reports, but also state that this is synonymous with so-called MFH in parentheses so as to avert any misunderstanding with clinicians who continue to be familiar with that term. Whenever I encounter a soft tissue neoplasm with a "MFH-like" pattern, I consider several broad possibilities before concluding that the lesion is in fact an undifferentiated pleomorphic sarcoma (UPS). First, I consider whether the lesion in question is some type of pleomorphic sarcoma with a specific line of differentiation that can be identified through light microscopy and/or immunohistochemistry. The rationale behind attempting to more precisely classify a pleomorphic sarcoma is discussed below. Second, I consider the possibility that the lesion could be a component of a dedifferentiated sarcoma, particularly when dealing with a sarcoma in the retroperitoneum. Thorough sampling is often required in order to recognize

11

the low-grade sarcoma from which the dedifferentiated "MFH-like" areas arose. Importantly, I also want to exclude the possibility that the lesion is a non-mesenchymal neoplasm. The most common consideration is that of a sarcomatoid carcinoma, particularly when the lesion arises on a mucosal surface, skin or within a parenchymal organ. Other considerations might include (depending upon site and other clinical factors) sarcomatoid mesothelioma, melanoma and even anaplastic lymphoma. Finally, if all of these can be excluded, then one can arrive at the conclusion that the lesion is a UPS.

Pleomorphic sarcoma with a specific line of differentiation

A variety of pleomorphic sarcomas may have areas which resemble UPS. In some cases, determining the specific line of differentiation may rely on random sampling of a small area within a large tumor. Although a specific type of pleomorphic sarcoma may be suggested by histologic features, immunohistochemical stains are often required to confirm the diagnosis. Although it could be argued that subtyping pleomorphic sarcomas is nothing more than an academic exercise, there is some evidence to suggest that pleomorphic sarcomas with myogenic differentiation are more clinically aggressive than those without myogenic differentiation9–11 (see below). The only criterion for rendering a diagnosis of pleomorphic liposarcoma is the recognition of multivacuolated pleomorphic lipoblasts. The major difficulty in such cases is separating pleomorphic sarcomas that infiltrate fat and isolate individual cells from those with true lipoblasts. Pleomorphic leiomyosarcoma is composed of cells with distinct cytoplasmic eosinophilia. At least focally, most cases have areas with a fascicular arrangement and cells with blunt-ended nuclei with a perinuclear vacuole and deeply eosinophilic cytoplasm. Pleomorphic rhabdomyosarcoma is recognized by the presence of large cells with eosinophilic cytoplasm and cross striations which can be confirmed by the immunohistochemical demonstration of skeletal muscle differentiation (desmin, MyoD1, myogenin). A definitive diagnosis of pleomorphic malignant peripheral nerve sheath tumor can be difficult unless the pleomorphic sarcoma clearly arises from a benign nerve sheath tumor or arises from a peripheral nerve in a patient with type 1 neurofibromatosis. The only criterion for recognizing extraskeletal osteosarcoma is the production of osteoid or bone by cytologically malignant cells. As mentioned, several studies have suggested that pleomorphic sarcomas with myogenic differentiated are clinically more aggressive than those without myogenic differentiation.9–11 Fletcher and colleagues reviewed 100 cases diagnosed as "MFH" and concluded (based upon morphology and immunohistochemistry) that 70 of these tumors were non-myogenic pleomorphic sarcomas, whereas 30 showed evidence of myogenic

12

differentiation.10 Of these 30 cases, 20 were classified as leiomyosarcoma, 9 as myogenic sarcoma, not otherwise specified, and 1 as rhabdomyosarcoma. The pleomorphic sarcomas with myogenic differentiation were significantly more likely to be deep-seated when compared to those without myogenic differentiation (83% versus 63%; p=0.04), and they were also more likely to be grade 4 tumors (97% versus 67%; p=0.02). Similarly, in 2003, Deyrup et al compared 42 cases of pleomorphic sarcoma with myogenic differentiation to 50 pleomorphic sarcomas without myogenic differentiation.9 The only significant difference between these two groups was that the tumors with myogenic differentiation were significantly more likely to be grade 3 sarcomas (88% versus 74%; p=0.038). There was a significant difference in overall survival, as those sarcomas with myogenic differentiation had shorter overall survival. Overall survival was also directly correlated with the number of myogenic markers found to be positive, as those with 3 markers found in a given tumor were more aggressive than those with fewer markers. Massi et al evaluated 65 pleomorphic sarcomas of the extremities and compared 31 sarcomas with myogenic differentiation (22 leiomyosarcomas, 4 rhabdomyosarcomas and 5 myofibrosarcomas) with 34 non-myogenic pleomorphic sarcomas.11 The tumors were classified into diagnostic categories based upon morphology, immunohistochemistry and, in some cases, ultrastructural analysis. Tumor site and myogenic differentiation were both found to be independent predictors of disease relapse for localized tumors by multivariate analysis, but only myogenic differentiation was found to be an independent predictor of overall survival by multivariate analysis. In summary, there are at least three studies which strongly suggest the adverse prognostic effect of myogenic differentiation in a pleomorphic sarcoma. However, there are a number of practical issues which make the assessment of myogenic differentiation in a pleomorphic sarcoma rather difficult. First, these are relatively uncommon tumors, and there are inconsistencies with regard to diagnostic criteria. There is a relatively small number of cases subjected to multivariate analysis in order to control for tumor site, tumor size and tumor depth. Exactly how many myogenic markers should be positive to designate a tumor as having myogenic differentiation is unclear, and the extent of staining needed to designate a myogenic marker as positive is similarly unclear.

Pleomorphic sarcoma as a result of dedifferentiation The process of tumor progression or dedifferentiation involves the transformation of a low-grade sarcoma to a higher-grade sarcoma, which usually (but not always) resembles a UPS. The most common scenario is the progression of a low-grade well-differentiated liposarcoma to a pleomorphic sarcoma (dedifferentiated liposarcoma). Other low-grade neoplasms can also dedifferentiate, including chondrosarcomas, chordomas and parosteal osteosarcomas. Certainly in a limited biopsy specimen it can be impossible to prove that a

13

pleomorphic sarcoma is part of a dedifferentiated sarcoma if the low-grade component is not represented. However, in a retroperitoneal sarcoma where dedifferentiated liposarcoma is always a strong consideration, it can often be suggested that the high-grade sarcoma could be part of a dedifferentiated liposarcoma. There is some evidence to suggest that dedifferentiated sarcomas (dedifferentiated liposarcoma in particular) are more indolent tumors than de novo pleomorphic sarcomas arising in the same location. For example, in the study of 32 cases of dedifferentiated liposarcoma by McCormick et al, 19 of 32 dedifferentiated liposarcomas (59%) arose in the retroperitoneum or paratesticular region, most of which arose de novo (94%).12 These patients were followed for a mean of 5.6 years, and metastasis was detected in only 4 of 27 cases with follow-up (15%), a rate much lower than would be expected for de novo pleomorphic sarcomas in that location. Similarly, Henricks and colleagues studies 155 cases of dedifferentiated liposarcoma, most of which arose in the retroperitoneum, spermatic cord or scrotum (77%) and most of which showed de novo dedifferentiation (86%).13 Similar to the results reported by McCormick et al, these authors found a metastatic rate of only 18% for dedifferentiated liposarcomas of the retroperitoneum and an overall metastatic rate of 17% for dedifferentiated liposarcomas at all sites. Thus, it does appear that dedifferentiated liposarcoma has a lower metastatic rate than other pleomorphic sarcomas arising at the same site. If this is true, then it is logical to presume that it is of clinical importance to distinguish a dedifferentiated liposarcoma from a de novo pleomorphic sarcoma of some other type. The utility of MDM2 and CDK4 analysis in distinguishing dedifferentiated liposarcoma from other types of pleomorphic sarcoma is controversial, but the preponderance of evidence suggests that this analysis is useful. Inasmuch as both MDM2 and CDK4 are amplified and often immuno-positive in well-differentiated liposarcomas, it is reasonable to presume that these markers would be similarly expressed in dedifferentiated liposarcomas since the latter are derived from the former. In fact, there are a number of studies that confirm the frequent amplification and immunoexpression of these markers in dedifferentiated liposarcoma. For example, Binh et al found amplification of MDM2 and/or CDK4 in 53 of 55 (96%) dedifferentiated liposarcomas, using either RT-PCR or a-CGH.14 Using these same techniques, however, some of the simulators of dedifferentiated liposarcoma were also found to have amplification of these genes, including myxofibrosarcoma (8/13 cases; 62%), leiomyosarcoma (5/13 cases; 16%), MPNST (2/6 cases; 33%) and "MFH" (3/39 cases; 8%). By immunohistochemistry, almost all cases of dedifferentiated liposarcoma expressed MDM2 (52 cases), CDK4 (51 cases) or both markers (49 cases). However, some of the simulators of dedifferentiated liposarcoma also showed immunoexpression of these antigens. For example, 8 of 13 cases of

14

myxofibrosarcoma stained for MDM2, 3 stained for CDK4 and 2 stained for both markers. Thus, although assessment of gene amplification and immunoexpression in dedifferentiated liposarcoma is highly sensitive, it lacks complete specificity. I polled many of my soft tissue pathology colleagues in the United States and in other countries, and it is clear there is a lack of uniformity in the approach to utilizing these techniques in the distinction of dedifferentiated liposarcoma from other types of pleomorphic sarcoma.

Pleomorphic non-mesenchymal neoplasms resembling a pleomorphic sarcoma It can be exceedingly difficult to distinguish a UPS from a sarcomatoid carcinoma. A reasonable approach would be to assume a pleomorphic malignant neoplasm arising in the skin, mucosal surface or parenchymal organ is a sarcomatoid carcinoma, until proven otherwise. A battery of epithelial markers including broad-spectrum, low- and high- molecular weight cytokeratins is required, but equivocal results are not uncommon for several reasons. First, not all sarcomatoid carcinomas show the immunohistochemical expression of epithelial markers. Second, virtually any type of sarcoma, including UPS, can on occasion express cytokeratins. Strong and diffuse cytokeratin expression, especially with multiple antibodies, strongly supports a diagnosis of sarcomatoid carcinoma, as does the recognition of an intraepithelial/intramucosal dysplastic component. In the end, some cases are not resolvable and can only be diagnosed as a pleomorphic malignant neoplasm, sarcoma versus carcinoma. p63 can also play a role in this immunohistochemical work-up. While p63 is frequently expressed in a variety of sarcomatoid carcinomas, it is actually very rarely expressed in soft tissue tumors. In the study by Jo and Fletcher in 2011, p63 was expressed in only 9% of 650 soft tissue tumors tested.15 For example, p63 was found in only 1 of 20 cases of AFX, 0 of 20 cases of dedifferentiated liposarcoma and 0 of 20 cases of leiomyosarcoma. It should also be kept in mind that sarcomatoid mesothelioma, melanoma and anaplastic lymphoma can on occasion mimic UPS; a panel of markers including CAM5.2, S-100 protein, melanocytic markers such as HMB45 and Melan A, CD30 and ALK1 can help resolve these issues. Following the exclusion of the aforementioned scenarios, one is left with UPS as a diagnosis of exclusion, the details of which are described below.

Undifferentiated pleomorphic sarcoma: a diagnosis of exclusion Depending on definitional criteria, UPS still accounts for a significant proportion of sarcomas occurring in late adult life. It manifests a broad range of histologic appearances, although the most common form consists of a mixture of storiform and pleomorphic areas. UPS is characteristically a tumor of late adult life, with most cases occurring in persons between the ages of 50 and 70 years.16 Tumors in children are exceedingly rare, and this diagnosis should always be made with caution in patients under 20 years of age.

15

Approximately two-thirds occur in men, and whites are affected more often than blacks or Asians. The tumor occurs most frequently on the lower extremity, especially the thigh, followed by the upper extremity, usually as a painless, slowly enlarging mass. As mentioned previously, careful sampling and microscopic observation is necessary in retroperitoneal lesions to exclude dedifferentiated liposarcoma, a lesion which in our experience is far more common than de novo UPS in this location. Microscopically, the classic form of UPS has a highly variable morphologic pattern and shows frequent transitions from storiform to pleomorphic areas, although the emphasis in most tumors is on haphazardly arranged pleomorphic zones. Storiform areas consist of plump spindle cells arranged in short fascicles in a cartwheel, or storiform, pattern around slit-like vessels. Although such tumors resemble dermatofibrosarcoma protuberans, they differ by a less distinctive storiform pattern and by the presence of occasional plump histiocytic cells, numerous typical and atypical mitotic figures, and secondary elements including xanthoma cells and chronic inflammatory cells. Although this pattern is easily recognized, it is seldom seen throughout the entire tumor. Instead, most tumors have a combination of storiform and pleomorphic areas, with a preponderance on the latter. Least often, tumors have a fascicular growth pattern and resemble fibrosarcomas, except for scattered giant cells, and it may be an arbitrary distinction in some cases as to whether a given tumor should be designated as a UPS or fibrosarcoma. In contrast to the storiform areas, pleomorphic areas contain plumper fibroblastic cells and more rounded histiocyte-like cells arranged haphazardly with no particular orientation to vessels. Pleomorphism and mitotic activity are usually more prominent. The stroma and secondary elements vary considerably in the storiform and pleomorphic areas. Usually, the stroma consists of delicate collagen fibrils encircling individual cells but occasionally collagen deposition is extensive and widely separates cells. Rarely, the stroma contains metaplastic osteoid or chondroid material. If, however, bone or cartilage is extensive and/or appears immature, the tumor should be classified as an osteo- or chondrosarcoma. The vasculature, although elaborate, is seldom appreciated unless it becomes dilated and resembles that of a hemangiopericytoma. Some examples of this tumor have numerous giant cells, a lesion formerly referred to as the giant cell type of MFH or malignant giant cell tumor of soft parts.17,18 These tumors tend to be distinctly multinodular and composed of a mixture of spindled, rounded and osteoclast- type giant cells. Dense fibrous bands containing vessels often encircle the nodules, which frequently show secondary hemorrhage and necrosis. The cells display pleomorphism and prominent mitotic activity, and may contain ingested material such as lipid or hemosiderin. The nuclei of the osteoclast-type giant cells tend also to be of high nuclear grade. Focal osteoid or mature bone is present in up to 50% of these cases and is usually located at the

16

periphery of the tumor nodules.

Immunohistochemical findings

The role of immunohistochemistry in the diagnosis of UPS has traditionally been an ancillary one, primarily serving as a means to exclude other pleomorphic tumors. Thus, the diagnosis continues to presuppose thorough sampling and evaluation of hematoxylin-eosin-stained sections. Despite the limited diagnostic applications of immunohistochemistry aside from excluding other lesions, there is ample evidence that these tumors do not display features of monocytes or macrophages but, rather, fibroblasts/myofibroblasts.8,19,20 Many of these tumors show focal immunoreactivity for smooth muscle actin, but stains for desmin and h- caldesmon are typically negative.21 Some examples also show rare cytokeratin-positive cells, which can cause confusion with those tumors in which a sarcomatoid carcinoma is a real consideration.22–24 Thus, focal immunoreactivity for any number of intermediate filaments is, in my opinion, insufficient evidence of a specific line of differentiation and should not dissuade one from rendering a diagnosis of UPS. On the other hand, diffuse immunoreactivity is far more likely to reflect a specific line of differentiation.

Cytogenetic and molecular genetic findings

Over the past several decades, studies have reported a variety of cytogenetic abnormalities in "MFH" but of course the utility of this information is limited by the varying criteria for making this diagnosis.25,26 In general, pleomorphic sarcomas of all types are characterized by complex but nonspecific cytogenetic aberrations and, as such, this technique is not useful in distinguishing among these pleomorphic sarcomas.27 More recently, a number of studies utilizing comparative genomic hybridization have evaluated UPS and compared the findings to those of other pleomorphic sarcomas.28–31 Interestingly, several studies have found striking similarities between UPS and pleomorphic leiomyosarcoma suggesting a shared lineage.29,31

Clinical behavior

The vast majority of UPS are high-grade lesions having a local recurrence rate ranging from 19–31%, a metastatic rate of 31–35%, and a 5-year survival of 65–70%.32–37 Both local recurrence and distant metastases often develop within 12–24 months of diagnosis. Only a minority of patients develop metastases after 5 years, with the common metastatic sites being lung (90%), bone (8%), and liver (1%). Regional lymph node metastases are decidedly uncommon.

17

The factors that correlate consistently with metastasis, survival, or both are depth, tumor size, grade, necrosis, and local recurrence, although they are not necessarily independent variables. For example, size and depth appear to co-vary because large tumors tend to be deep tumors. In the study by Engellau and colleagues, necrosis and local recurrence were significant predictors of metastasis within the first two years of diagnosis and throughout a longitudinal follow-up period, whereas only tumor depth and local recurrence were significant predictors beyond 2 years.32

18

References:

1. Ozzello L, Stout AP, Murray MR. Cultural characteristics of malignant histiocytomas and fibrous xanthomas. Cancer. 1963;16:331–344.

2. O’Brien JE, Stout AP, Malignant fibrous xanthomas. Cancer. 1964;17:1445–1455.

3. Iwasaki H, Isayama T, Johzaki H, Kikuchi M. Malignant fibrous histiocytoma. Evidence of perivascular mesenchymal cell origin immunocytochemical studies with monoclonal anti- MFH antibodies. Am. J. Pathol. 1987;128(3):528–537.

4. Iwasaki H, Isayama T, Ohjimi Y, et al. Malignant fibrous histiocytoma. A tumor of facultative histiocytes showing mesenchymal differentiation in cultured cell lines. Cancer. 1992;69(2):437–447.

5. Roholl PJ, Kleyne J, Elbers H, et al. Characterization of tumour cells in malignant fibrous histiocytomas and other soft tissue tumours in comparison with malignant histiocytes. I. Immunohistochemical study on paraffin sections. J. Pathol. 1985;147(2):87–95.

6. Roholl PJ, Kleyne J, Van Unnik JA. Characterization of tumor cells in malignant fibrous histiocytomas and other soft-tissue tumors, in comparison with malignant histiocytes. II. Immunoperoxidase study on cryostat sections. Am. J. Pathol. 1985;121(2):269–274.

7. Fletcher CD. Pleomorphic malignant fibrous histiocytoma: fact or fiction? A critical reappraisal based on 159 tumors diagnosed as pleomorphic sarcoma. Am. J. Surg. Pathol. 1992;16(3):213–228.

8. Montgomery E, Fisher C. Myofibroblastic differentiation in malignant fibrous histiocytoma (pleomorphic myofibrosarcoma): a clinicopathological study. Histopathology. 2001;38(6):499–509.

9. Deyrup AT, Haydon RC, Huo D, et al. Myoid differentiation and prognosis in adult pleomorphic sarcomas of the extremity: an analysis of 92 cases. Cancer. 2003;98(4):805– 813.

10. Fletcher CD, Gustafson P, Rydholm A, Willén H, Akerman M. Clinicopathologic re- evaluation of 100 malignant fibrous histiocytomas: prognostic relevance of subclassification. J. Clin. Oncol. 2001;19(12):3045–3050.

11. Massi D, Beltrami G, Capanna R, Franchi A. Histopathological re-classification of extremity pleomorphic soft tissue sarcoma has clinical relevance. Eur J Surg Oncol. 2004;30(10):1131–1136.

12. McCormick D, Mentzel T, Beham A, Fletcher CD. Dedifferentiated liposarcoma. Clinicopathologic analysis of 32 cases suggesting a better prognostic subgroup among pleomorphic sarcomas. Am. J. Surg. Pathol. 1994;18(12):1213–1223.

13. Henricks WH, Chu YC, Goldblum JR, Weiss SW. Dedifferentiated liposarcoma: a clinicopathological analysis of 155 cases with a proposal for an expanded definition of dedifferentiation. Am. J. Surg. Pathol. 1997;21(3):271–281.

14. Binh MBN, Sastre-Garau X, Guillou L, et al. MDM2 and CDK4 immunostainings are useful adjuncts in diagnosing well-differentiated and dedifferentiated liposarcoma subtypes: a comparative analysis of 559 soft tissue neoplasms with genetic data. Am. J. Surg. Pathol. 2005;29(10):1340–1347.

19

15. Jo VY, Fletcher CDM. p63 immunohistochemical staining is limited in soft tissue tumors. Am. J. Clin. Pathol. 2011;136(5):762–766.

16. Weiss SW, Enzinger FM. Malignant fibrous histiocytoma: an analysis of 200 cases. Cancer. 1978;41(6):2250–2266.

17. Guccion JG, Enzinger FM. Malignant giant cell tumor of soft parts. An analysis of 32 cases. Cancer. 1972;29(6):1518–1529.

18. van Haelst UJ, de Haas van Dorsser AH. Giant cell tumor of soft parts. An ultrastructural study. Virchows Arch A Pathol Anat Histol. 1976;371(3):199–217.

19. Brecher ME, Franklin WA. Absence of mononuclear phagocyte antigens in malignant fibrous histiocytoma. Am. J. Clin. Pathol. 1986;86(3):344–348.

20. Wood GS, Beckstead JH, Turner RR, et al. Malignant fibrous histiocytoma tumor cells resemble fibroblasts. Am. J. Surg. Pathol. 1986;10(5):323–335.

21. Agaimy A, Gaumann A, Schroeder J, et al. Primary and metastatic high-grade pleomorphic sarcoma/malignant fibrous histiocytoma of the gastrointestinal tract: an approach to the differential diagnosis in a series of five cases with emphasis on myofibroblastic differentiation. Virchows Arch. 2007;451(5):949–957.

22. Lawson CW, Fisher C, Gatter KC. An immunohistochemical study of differentiation in malignant fibrous histiocytoma. Histopathology. 1987;11(4):375–383.

23. Litzky LA, Brooks JJ. Cytokeratin immunoreactivity in malignant fibrous histiocytoma and spindle cell tumors: comparison between frozen and paraffin-embedded tissues. Mod. Pathol. 1992;5(1):30–34.

24. Weiss SW, Bratthauer GL, Morris PA. Postirradiation malignant fibrous histiocytoma expressing cytokeratin. Implications for the immunodiagnosis of sarcomas. Am. J. Surg. Pathol. 1988;12(7):554–558.

25. Gazziola C, Cordani N, Wasserman B, et al. Malignant fibrous histiocytoma: a proposed cellular origin and identification of its characterizing gene transcripts. Int. J. Oncol. 2003;23(2):343–351.

26. Al-Agha OM, Igbokwe AA. Malignant fibrous histiocytoma: between the past and the present. Arch. Pathol. Lab. Med. 2008;132(6):1030–1035.

27. Mertens F, Fletcher CD, Dal Cin P, et al. Cytogenetic analysis of 46 pleomorphic soft tissue sarcomas and correlation with morphologic and clinical features: a report of the CHAMP Study Group. Chromosomes and MorPhology. Genes Chromosomes Cancer. 1998;22(1):16–25.

28. Kageyama K, Moriyama T, Hizuka N, et al. Hypoglycemia associated with big insulin-like growth factor II produced during development of malignant fibrous histiocytoma. Endocr. J. 2003;50(6):753–758.

29. Larramendy ML, Gentile M, Soloneski S, Knuutila S, Böhling T. Does comparative genomic hybridization reveal distinct differences in DNA copy number sequence patterns between leiomyosarcoma and malignant fibrous histiocytoma? Cancer Genet. Cytogenet. 2008;187(1):1–11.

20

30. Nishio J, Iwasaki H, Nabeshima K, et al. Establishment of a new human pleomorphic malignant fibrous histiocytoma cell line, FU-MFH-2: molecular cytogenetic characterization by multicolor fluorescence in situ hybridization and comparative genomic hybridization. J. Exp. Clin. Cancer Res. 2010;29:153.

31. Carneiro A, Francis P, Bendahl P-O, et al. Indistinguishable genomic profiles and shared prognostic markers in undifferentiated pleomorphic sarcoma and leiomyosarcoma: different sides of a single coin? Lab. Invest. 2009;89(6):668–675.

32. Engellau J, Anderson H, Rydholm A, et al. Time dependence of prognostic factors for patients with soft tissue sarcoma: a Scandinavian Sarcoma Group Study of 338 malignant fibrous histiocytomas. Cancer. 2004;100(10):2233–2239.

33. Le Doussal V, Coindre JM, Leroux A, et al. Prognostic factors for patients with localized primary malignant fibrous histiocytoma: a multicenter study of 216 patients with multivariate analysis. Cancer. 1996;77(9):1823–1830.

34. Salo JC, Lewis JJ, Woodruff JM, Leung DH, Brennan MF. Malignant fibrous histiocytoma of the extremity. Cancer. 1999;85(8):1765–1772.

35. Gibbs JF, Huang PP, Lee RJ, et al. Malignant fibrous histiocytoma: an institutional review. Cancer Invest. 2001;19(1):23–27.

36. Zagars GK, Mullen JR, Pollack A. Malignant fibrous histiocytoma: outcome and prognostic factors following conservation surgery and radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 1996;34(5):983–994.

37. Belal A, Kandil A, Allam A, et al. Malignant fibrous histiocytoma: a retrospective study of 109 cases. Am. J. Clin. Oncol. 2002;25(1):16–22.

21

Myxoid Soft Tissue Tumors and Tumor-Like Lesions The pathologist need not panic when one comes across one of these lesions because these entities can be separated from one another when one pays attention to certain key features when evaluating these lesions. For example, evaluation of the cellularity of the lesion, as well as the arrangement of the cells with respect to one another is absolutely critical in distinguishing these lesions from one another. While some lesions are characterized by extremely low cellularity (intramuscular myxoma), others are characteristically much more cellular (nodular fasciitis). Similarly, these cells may stand apart from one another, with little cell-cell contact (myxoid liposarcoma), or they may be arranged into nests or chains (myxoid chondrosarcoma). Nuclear pleomorphism, although somewhat subjective, is also useful in this evaluation, given that some lesions totally lack nuclear pleomorphism (intramuscular myxoma), whereas others are characterized by a high degree of cytologic atypia (myxofibrosarcoma). Another often underappreciated feature in this evaluation is the presence or absence of an underlying vasculature. While some lesions are characteristically of low vascularity (intramuscular myxoma), others are characterized by an intricate vascular network that allows one to recognize the lesion as malignant (myxoid liposarcoma and myxfibrosarcoma). Occasionally, the evaluation of the myxoid stroma using histochemical techniques may also be useful. Hyaluronic acid and chondroitin sulfate are the most common mucosubstances found in these lesions, and one or the other of these substances is often typical of a particular lesion. For example, intramuscular myxoma, myxoid liposarcoma and myxofibrosarcoma are characterized by hyaluronic acid-rich myxoid stroma, whereas myxoid chondrosarcoma and chordoma are characterized by a chondroitin sulfate-rich myxoid stroma. Although Alcian blue (pH2.5) stains both hyaluronic acid and chondroitin sulfate, pretreatment with hyaluronidase will result in the loss of Alcian blue positivity if the stroma is made up of hyaluronic acid. In contrast, chondroitin sulfate-rich stroma is hyaluronidase resistant. Benign myxoid soft tissue lesions that enter into this differential diagnosis include nodular fasciitis, intramuscular myxoma and angiomyxoma, among others. The malignant lesions may include myxoid liposarcoma, myxofibrosarcoma, extra-skeletal myxoid chondrosarcoma and low-grade fibromyxoid sarcoma (Evans' tumor).

Nodular fasciitis is a self-limited, reactive lesion often mistaken for a sarcoma due to its high cellularity, rapid growth and brisk mitotic activity.1 In Montgomery and Meis' series of 53 cases, only 43% of the cases were correctly diagnosed, and 21% of these cases were misdiagnosed as some type of sarcoma2. This lesion typically presents as a rapidly growing mass of short duration, and is most common in young patients (average age: 34 years), although essentially any age group may be affected. The upper extremity is the most

22

common site; other fairly common sites include the trunk, lower extremities and head and neck region (although virtually any superficial site in the body may be involved). Most lesions are less than 3.0 cm, but I have seen examples as large as 6.0 cm. Most cases are subcutaneous, but some are in skeletal muscle, causing further concern that the lesion is a sarcoma.

Histologically, nodular fasciitis has many appearances, varying dramatically in cellularity and amount of myxoid stroma both between lesions, and within the same lesion. Most cells are plump, immature “tissue culture” fibroblasts or myofibroblasts that show little nuclear pleomorphism. Mitoses are characteristically numerous, but atypical mitoses are not present. The cells are arranged into short, irregular bundles or fascicles, but never form long, sweeping fascicles, and are deposited in a hyaluronic acid-rich myxoid matrix that occasionally forms microcysts. Other helpful histologic features include cleft-like spaces, red blood cell extravasation, rare giant cells and, occasionally, metaplastic bone or cartilage. In older lesions, there is an increased amount of stromal fibrosis with a less prominent myxoid matrix. Immunohistochemically, the majority of cells stain for smooth-muscle actin, which may be a source of confusion if one is considering the possibility of leiomyosarcoma.2

The intramuscular myxoma typically occurs in middle-aged to elderly patients, and is extremely rare in childhood. These lesions present as a painless, palpable fluctuant mass within the deep soft tissues of the thigh, shoulder, buttocks or upper arm, although virtually any site may be involved.3 In addition, lesions with similar histology can occur in a cutaneous and juxta-articular location. Although usually solitary, multiple intramuscular myxomas have been found to be associated with fibrous dysplasia, and generally occur in the same anatomic region as the bony abnormalities.4 Rare patients also display melanotic pigmentation of the skin and endocrine abnormalities (Albright's syndrome). Myxomas occurring in a cutaneous location may be sporadic or associated with Carney's complex, characterized by an association with endocrine abnormalities, spotty pigmentation, cardiac myxomas, and psammomatous melanotic schwannomas, inherited in an autosomal dominant manner.5 The juxta-articular myxoma is another variant of myxoma, most commonly found in the area of the knee (90%).6 Males are affected significantly more commonly than females, typically between the third and seventh decades of life. In the series by Meis and Enzinger, 34% of the cases recurred, sometimes with multiple recurrences. Despite the tendency for these lesions to recur, these are best treated conservatively by local excision.

23

Grossly, the intramuscular myxoma appears to be well-circumscribed, although a true fibrous capsule is not present. Histologically, these lesions characteristically are of low cellularity, composed of bland spindled or stellate cells with cytoplasmic processes. The cells tend not to touch one another, but rather are separated by abundant myxoid stroma composed of hyaluronic acid. Although some cells with vacuolated cytoplasm may be present and may resemble lipoblasts, these are macrophages that have imbibed products of the myxoid stroma resulting in cytoplasmic vacuolization. In addition, although grossly well circumscribed, there is often some infiltration into the surrounding skeletal muscle, with entrapment of atrophic skeletal muscle fibers. Although scattered blood vessels may be present, there is relatively little vascularity, and the vascularity lacks the organization of many myxoid sarcomas. These lesions are essentially cured by local excision, and have little tendency to recur, even if incompletely excised.

Angiomyxoma (aggressive angiomyxoma) typically occurs as a large, ill-defined mass within the pelvis, perineum, or genital tract in women.7,8 Rare cases have also been reported in men.9 Histologically, the lesion is composed of spindled or stellate cells that generally do not touch one another, and are separated by an abundant myxoid stroma composed primarily of hyaluronic acid. The cells lack nuclear atypia, and mitotic figures are difficult to identify. Mast cells are frequently prominent. In addition, these lesions are characterized by a prominent vascularity with vessels of different caliber, including thin-walled vessels and thick hyalinized vessels. Although histologically bland, these lesions are characterized by a high rate of local recurrence; metastases have not been reported.

Myxoid liposarcoma is the most common subtype of liposarcoma.10 This is a tumor of adult life and typically occurs in the deep soft tissues of the extremity, especially the thigh and popliteal region. At low power, the most striking feature is the very characteristic delicate plexiform capillary pattern that is found throughout the neoplasm. The spindled cells between the capillaries are primitive mesenchymal cells, and vary little from one another, without significant nuclear pleomorphism. The cells are evenly distributed, and typically do not touch one another. Interspersed between the primitive mesenchymal cells are the diagnostic lipoblasts, which occur in varying numbers. By definition, a lipoblast is characterized by a sharply defined lipid droplet that usually pushes the nucleus to a peripheral location and indents or scallops the nucleus. Vacuolated cells indistinguishable from lipoblasts may be found in a variety of benign and malignant lesions. Therefore, an appropriate histologic background must be present in order to establish that cell as a true lipoblast. Unless there is a significant round cell component, myxoid liposarcomas are generally considered to be low-grade sarcomas.

24

Round cell liposarcoma is considered to be a poorly differentiated form of myxoid liposarcoma for several reasons. First, it is not uncommon to see mixtures of both myxoid and round cell liposarcoma within the same tumor. Furthermore, the characteristic translocation found in myxoid liposarcoma, t(12;16)(q13;p11), is also present in round cell liposarcoma,11,12 and can be detected by either fluorescence in-situ hybridization13,14 or polymerase chain reaction.15 At the molecular level, this translocation results in fusion of the DDIT3 (CHOP) gene on 12q13 with the FUS gene on 16p11.16-18 Rarely, the DDIT3 gene is fused to the N-terminal portion of the EWS gene.19,20

It has been suggested that tumors that have “hypercellular” or “round cell” areas within an otherwise typical myxoid liposarcoma pursue a more aggressive clinical course.21-24 However, it is difficult to know where on the spectrum of cellularity these cases actually lie. Furthermore, it is unknown whether there is a critical amount of these “round cell” areas that are predictive of a worse prognosis. Smith et al studied 29 cases of myxoid/round cell liposarcoma of the extremities and found that those patients whose tumors had greater than 5% round cell component were more likely to develop metastases or die from their disease.25 Round cell areas were defined as those areas with a marked increase in cellularity in which the cells were round and separated by little or no myxoid stroma. In these areas, the mitotic index was generally increased, and a plexiform vascular pattern was difficult to recognize secondary to the overgrowth of primitive round cells. In addition, transitional areas, defined as areas of increased cellularity compared to typical myxoid liposarcoma, but in which the cells remained spindled, did not have overlapping nuclear borders, and retained an easily discernible plexiform vascular pattern, were not found to worsen clinical outcome in the absence of a round cell component. Kilpatrick et al. found similar findings, although they found a cut-off point of 25% round cell component to be prognostically important by multivariate analysis.26

The myxoid variant of MFH was first characterized by Weiss and Enzinger in 1977, and was defined as an MFH that has at least 50% of the tumor cells deposited in a hyaluronic acid- rich myxoid stroma.27 However, it is apparent that there is a wide spectrum of lesions ranging from superficially located, hypocellular, low-grade myxoid lesions (low-grade myxofibrosarcoma) to those that are more deeply located, of higher stage and more biologically aggressive. Mentzel et al. recently evaluated the clinicopathologic features of 75 cases of so-called myxofibrosarcoma,28 now the preferred term for this entity. Almost 70% of their cases were superficially located, either in the dermis or subcutaneous tissues, usually on the upper or lower extremities, and characterized by a nodular growth pattern, a myxoid

25

matrix containing elongated, curvilinear capillaries and spindled or stellate-shaped tumor cells with hyperchromatic atypical nuclei. These superficially located low-grade lesions are the lesions that are most likely to be confused with benign myxoid lesions. Some of the cases were more deeply located and showed areas of increased cellularity and cytologic atypia, more typical of the classic “myxoid MFH” described by Weiss and Enzinger. The latter group of lesions was characterized by moderate cellularity in which the cells showed significant nuclear pleomorphism and hyperchromatism, with easily identified mitotic figures. Similar to myxoid liposarcoma, a characteristic vasculature was present throughout the neoplasm, although these blood vessels tended to be more coarse than those seen in myxoid liposarcoma. Frequently, the atypical cells condensed along the periphery of the blood vessels. Although the depth of the primary lesion did not influence the incidence of local recurrence, only those neoplasms of intermediate or high-grade metastasized. In addition, some cases of low-grade myxofibrosarcoma progressed to higher-grade lesions in recurrences.

In 1987, Evans described a tumor that typically presents as a large, well-circumscribed mass in the deep soft tissues, most commonly in the shoulder, thigh and inguinal region, which he termed “low-grade fibromyxoid sarcoma.”29 Histologically, this lesion is composed of bland spindled cells of low to moderate cellularity deposited in a fibrous and myxoid stroma. The cells often have a swirling growth pattern and occasionally condense in a perivascular location. Cytologically, there is little nuclear pleomorphism, and mitotic figures are difficult to identify. Similar to other myxoid sarcomas, low-grade fibromyxoid sarcoma often has a rich, regular vascular network that is useful in its distinction from a benign lesion. Also, despite the gross circumscription, there is microscopic infiltration of the surrounding tissues. Of the twelve patients described in Evans’ 1993 paper, nine had local recurrences, seven had evidence of distant metastasis, and four died of disease.30 Some authors have also reported histologic progression to a higher-grade lesion in recurrences.31 Low-grade fibromyxoid sarcoma (and the related hyalinizing spindle cell tumor with giant rosettes) is characterized by a t(7;16) with fusion of the CREB3L2 gene on chromosome 12 with the FUS gene on chromosome 16.32 Mertens and colleagues reported the presence of CREB3L2-FUS fusion in 22 of 23 (96%) cases of low-grade fibromyxoid sarcoma,33 and we utilize paraffin- embedded tissue for FISH, probing with a FUS breakapart probe to confirm this difficult diagnosis.

Similar to myxoid liposarcoma, extraskeletal myxoid chondrosarcoma also occurs primarily in the deep soft tissues of the extremities.34 Macroscopically, the neoplasm occurs as a multinodular, well-circumscribed mass, which frequently shows large areas of hemorrhage.

26

Histologically, this is a lesion of moderate cellularity in which the cells tend to touch one another and are arranged in small cords or strands. These cells show little nuclear pleomorphism, low mitotic activity, and have a moderate amount of eosinophilic cytoplasm. The vascularity is not prominent, in contrast to myxoid liposarcoma and myxofibrosarcoma. The myxoid matrix in this case is composed of chondroitin sulfate. Immunohistochemically, these cells may be S100 protein positive, although Dei Tos et al found that only 7 of 39 cases (18%) stained for this antigen,35 although it is usually unnecessary to perform immunostains. In addition, this tumor has been found to harbor a characteristic translocation, t(9;22)(q22;q12), which involves a rearrangement of the EWS gene on 22q12 with the NR4A3 gene (formerly known was CHN or TEC).36 The resultant NR3A3-EWS fusion transcript can be detected by RT-PCR or FISH.37 In the series by Meis-Kindblom et al, older patient age, larger tumor size and tumor location in the proximal extremity or limb girdle were adverse prognostic factors identified by multivariate analysis.38 Local recurrences and metastases were noted in 48% and 46% of patients, respectively. Some patients had prolonged survival even after the development of metastasis, although many of these patients eventually died as a result of tumor. Interestingly, this study did not identify a relationship between tumor cellularity and prognosis.

In conclusion, despite the fact that numerous benign and malignant soft tissue lesions may have a myxoid stroma, these lesions can be reliably separated from one another through the systematic evaluation of certain parameters, in conjunction with clinical features including age, site and rate of growth of the neoplasm, with little need for ancillary studies. However, molecular testing for translocations has become increasingly important.

27

References

1. Bernstein KE, Lattes R. Nodular (pseudosarcomatous) fasciitis, a nonrecurrent lesion: Clinicopathologic study of 134 cases. Cancer 1982; 49:1668-1675. 2. Montgomery EA, Meis JM. Nodular fasciitis. Its morphologic spectrum and immunohistochemical profile. Am J Surg Pathol 1991; 15:942-948. 3. Kindblom LG, Stener B, Angervall L. Intramuscular myxoma. Cancer 1974; 34;1734- 1744. 4. Ireland DCR, Soule EH, Ivins JC. Myxoma of somatic soft tissues. A report of 58 patients, three with multiple tumors and fibrous dysplasia of bone. Mayo Clin Proc 1973; 48:401-410. 5. Carney JA, Headington JT, Su WPD. Cutaneous myxomas. A major component of the complex of myxomas, spotty pigmentation, and endocrine overactivity. Arch Dermatol 1986; 122:790-798. 6. Meis JM, Enzinger FM. Juxta-articular myxoma. A clinical and pathological study of 65 cases. Hum Pathol 1992; 23:639-646. 7. Steeper TA, Rosai J. Aggressive angiomyxoma of the female pelvis and perineum. Report of nine cases of a distinctive type of gynecologic soft tissue neoplasm. Am J Surg Pathol 1983; 7:463-475. 8. Begin LR, Clement PB, Kirk ME, Jothy S, McCaughey WTE, Ferenczy A. Aggressive angiomyxoma of pelvic soft parts: A clinicopathologic study of nine cases. Hum Pathol 1985; 16:621-628. 9. Tsang WYW, Chan JKC, Lee KC, Fisher C, Fletcher CDM. Aggressive angiomyxoma. A report of four cases occurring in men. Am J Surg Pathol 1992; 16:1059-1065. 10. Weiss SW, Goldblum JR. Myxoid variant of liposarcoma. In: Soft Tissue Tumors, 4th Edition, CV Mosby, St. Louis, MO, 2001. 11. Gibas Z, Miettinen M, Limon J, et al. Cytogenetic and immunohistochemical profile of myxoid liposarcoma. Am J Clin Pathol 1995;103:20-26. 12. Tallini G, Akerman M, Del Cin P, et al. Combined morphologic and karyotypic study of 28 myxoid liposarcomas. Implications for a revised morphologic typing (a report from the CHAMP group). Am J Surg Pathol 1996;20:1047-1055. 13. Mezzelani A, Sozzi G, Pierotti MA, Pilotti S. Rapid differential diagnosis of myxoid liposarcoma by fluorescence in-situ hybridization on cytological preparations. J Clin Pathol 1996;49:308-309. 14. Aoki T, Hisaoka M, Kouho H, Hashimoto H, Nakata H. Interphase cytogenetic analysis of myxoid soft tissue by fluorescence in-situ hybridization and DNA flow cytometry using paraffin-embedded tissue. Cancer 1997;79:284-293. 15. Kuroda M, Ishida T, Horiuchi H, et al. Chimeric TLS/FUS-CHOP gene expression and the heterogeneity of its junction in human myxoid and round cell liposarcoma. Am J Pathol 1995;147:1221-1227. 16. Aman P, Ron D, Mandahl N, et al. Rearrangement of the transcription factor gene CHOP in myxoid liposarcomas with t(12;16)(q13;p11). Genes Chromosomes Cancer 1992;5:278-285. 17. Crozat A, Aman P, Mandahl N, Ron D. Fusion of CHOP to a novel RNA-binding protein in human myxoid liposarcoma. Nature 1993;363:640-644. 18. Rabbitts T, Forster A, Larson R, Nathan P. Fusion of the dominant negative transcription regulator CHOP with a novel gene FUS by translocation t(12;16) in malignant liposarcoma. Nature Genet 1993;4:175-280. 19. Panagopoulos I, Hoglund M, Mertens F, et al. Fusion of the EWS and CHOP genes in myxoid liposarcoma. Oncogene 1996;12:489-494. 20. Dal Cin P, Sciot R, Panagopoulos I, et al. Additional evidence of a variant translocation t(12;22) with EWS/CHOP fusion in myxoid liposarcoma: clinicopathological features. J Pathol 1997;182:437-441.

28

21. Azumi N, Curtis J, Kempson RL, Hendrickson MR. Atypical and malignant neoplasms showing lipomatous differentiation: A study of 111 cases. Am J Surg Pathol 1987;11(3):161-183. 22. Enzinger FM, Winslow DJ. Liposarcoma: A study of 103 cases. Virch Arch Pathol Anat 1962;335:367-388. 23. Evans HL. Liposarcoma: A study of 55 cases with a re-assessment of its classification. Am J Surg Pathol 1979;3(6):507-523. 24. Evans HL. Liposarcomas and atypical lipomatous tumors: A study of 66 cases followed for a minimum of 10 years. Surg Pathol 1988;1(1):41-54. 25. Smith TA, Easley KA, Goldblum JR. Myxoid/round cell liposarcoma of the extremities: A clinicopathologic study of 29 cases with particular attention to extent of round cell liposarcoma. Am J Surg Pathol 1996;20(2):171-180. 26. Kilpatrick SE, Doyon J, Choong PFM, Sim FH, Nascimento AG. The clinicopathologic spectrum of myxoid and round cell lipoma. A study of 95 cases. Cancer 1996;77:1450-1458. 27. Weiss SW, Enzinger FM. Myxoid variant of malignant fibrous histiocytoma. Cancer 1977; 39;1672-1689. 28. Mentzel T, Calonje E, Wadden C, et al. Myxofibrosarcoma: Clinicopathologic analysis of 75 cases with emphasis on the low-grade variant. Am J Surg Pathol 1996;20(4):391-405. 29. Evans HL. Low-grade fibromyxoid sarcoma: A report of two metastasizing neoplasms having a deceptively benign appearance. Am J Clin Pathol 1987; 88:615-619. 30. Evans HL. Low-grade fibromyxoid sarcoma: A report of twelve cases. Am J Surg Pathol 1993; 17(6):595-600. 31. Goodlad JR, Mentzel T, Fletcher CDM. Low-grade fibromyxoid sarcoma: Clinicopathological analysis of 11 new cases in support of a distinct entity. Histopathology 1995;26:229-237. 32. Panagopoulos I, Storlazzi CT, Fletcher CD, et al. The chimeric FUS/CREB3L2 gene is specific for low-grade fibromyxoid sarcoma. Genes Chromosomes Cancer 2004;40:218-228. 33. Mertens F, Fletcher CD, Antonescu CR, et al. Clinicopathologic and molecular genetic characterization of low-grade fibromyxoid sarcoma and cloning of a novel FUS/CREB3L1 fusion gene. Lab Invest 2005;85:408-415. 34. Enzinger FM, Shiraki M. Extraskeletal myxoid chondrosarcoma: An analysis of 34 cases. Hum Pathol 1972;3:421-435. 35. Dei Tos AP, Wadden C, Fletcher CDM. Extraskeletal myxoid chondrosarcoma: An immunohistochemical reappraisal of 39 cases. Appl Immunohistochem 1997;5(2):73- 77. 36. Hinrichs SH, Jaramillo MA, Gumerlock PH, et al. Myxoid chondrosarcoma with a translocation involving chromosomes 9 and 22. Cancer Genet Cytogenet 1985;14:219-226. 37. Brody, RI, Ueda T, Hamelin A, et al. Molecular analysis of the fusion of EWS to an orphan nuclear receptor gene in extraskeletal myxoid chondrosarcoma. Am J Pathol 1997;150:1049-1058. 38. Meis-Kindblom JM, Bergh P, Gunterberg B, Kindblom L-G. Extraskeletal myxoid chondrosacroma. A reappraisal of its morphologic spectrum and prognostic factors based on 117 cases. Am J Surg Pathol 1999;23:636-650.

29

Fibrosarcoma-Like Pattern Highly cellular spindled mesenchymal neoplasms arranged into a fascicular growth pattern are not uncommonly encountered in the deep soft tissues, mediastinum or retroperitoneum. When presented with such a lesion, the most common differential diagnosis usually includes cellular schwannoma, MPNST, leiomyosarcoma, monophasic synovial sarcoma and fibrosarcoma. Through a combination of light microscopy and immunohistochemistry, one can usually detect a specific line of cellular differentiation in such tumors. As a matter of fact, a diagnosis of fibrosarcoma, once a very common diagnosis to make in the realm of soft tissue pathology, has become a diagnostic rarity.

Cellular Schwannoma Cellular schwannoma is one of several variants of schwannoma that may cause diagnostic confusion, in this case because of its high cellularity, mitotic activity, and presence of bony destruction. These lesions typically occur in middle-aged patients (although the age range is broad), with a slight female predilection. They are most commonly found within the paravertebral region of the posterior mediastinum, retroperitoneum, and pelvis. A small number of patients have been found to have neurofibromatosis.

Grossly, these lesions are typically encapsulated, and may or may not be associated with an identifiable nerve, either grossly or microscopically. Degenerative changes, including cyst formation, hemorrhage and necrosis may be seen. Histologically, the neoplasm is composed of slender, elongated spindled cells with wavy contours that may be arranged in short intersecting fascicles, or in longer sweeping fascicles reminiscent of the herringbone pattern seen in fibrosarcoma. By definition, the lesion is composed almost entirely of Antoni A areas, and although abortive nuclear palisades may be seen, true Verocay bodies are not formed. Although at first glance this lesion may be difficult to differentiate from other spindle cell sarcomas, a variety of histologic features may serve as useful diagnostic clues. This lesion typically has cellularity that is disproportionate to the degree of mitotic activity and cytologic atypia that is present. It should be noted that the cellular schwannoma can have some mitotic activity, although it is typically less than that seen in MPNSTs (<4 MF/10 HPF). Similarly, focal cytologic atypia has been identified in a small percentage of cellular schwannomas, but is typically not to the degree both in quality and quantity as is seen in MPNSTs. Other useful features to recognize this lesion include (1) the presence of paracapsular and/or perivascular lymphoid aggregates; (2) the focal presence of Antoni B areas; (3) prominent perivascular hyalinization and, very importantly (4) diffuse and strong immunoreactivity for S100 protein.

30

It is critical to distinguish the cellular schwannoma from a MPNST, given the differences in both therapy and prognosis. Tumor cellularity is not useful in that both of these lesions are highly cellular. As mentioned previously, although nuclear pleomorphism and mitotic activity may be seen in cellular schwannoma, MPNSTs typically have more extensive nuclear pleomorphism and mitotic figures, including atypical mitoses. Divergent differentiation, often in the form of rhabdomyoblasts, is found in approximately 10% of MPNSTs1, but is not seen in cellular schwannomas. Finally, S100 protein immunoreactivity is a useful adjunct in this differential diagnosis, as cellular schwannomas show diffuse and strong S100 protein positivity, whereas only about 60% of MPNSTs show S100 protein positivity, typically in a focal distribution.2 Four large studies of cellular schwannoma with significant follow-up (total of 119 cases) have been published.3-6 Although up to 5% of tumors have locally recurred, none of the patients have developed metastatic disease or died due to their tumor. Importantly, erosion of adjacent bone has been noted in approximately 13% of the patients in these series, and may contribute to the erroneous diagnosis of a sarcoma.

Ultrastructurally, these cells have been found to have the characteristic features of schwann cells, with elongated bipolar cytoplasmic extensions, interdigitating cytoplasmic processes, and multilayering of basal lamina.7

Other Variants of Schwannoma Ancient Schwannoma Ancient schwannomas are recognized by their extensive degenerative changes, which include cyst formation, hemorrhage, hyalinization and calcification. Significant nuclear atypia may also be noted, but importantly, the degree of mitotic activity is not proportionate to the degree of nuclear atypia that is present, suggesting a degenerative phenomenon.8 These degenerative findings typically occur in lesions that have been present for a long duration, and thus are more commonly found in more deeply situated tumors.

Plexiform Schwannoma Rare schwannomas may also be arranged in a plexiform architecture, reminiscent of what one encounters in the plexiform neurofibroma. Frequently, the individual nodules closely resemble those seen in the cellular schwannoma. It is important to distinguish these lesions from plexiform neurofibroma, as they lack an association with neurofibromatosis type I, and also are not known to undergo malignant transformation.9,10

31

Glandular Schwannoma Rare peripheral nerve sheath tumors also contain clear-cut epithelial differentiation in the form of glands. Although most of these lesions have been described in patients with neurofibromatosis type I and presumably have arisen from neurofibromas, there have been several reports of glandular differentiation in schwannomas.11,12 However, the existence of this entity has been disputed, as some authors believe that these reports represent schwannomas containing entrapped skin adnexal structures.13 Woodruff and Christensen identified 11 cases of glandular peripheral nerve sheath tumors and found that 92% of the tumors were histologically malignant and 74% of the patients had neurofibromatosis type I. The authors did not identify any cases of glandular schwannoma.13

Multiple Schwannomas (schwannomatosis) Rarely, schwannomas may occur multifocally, either in the form of multiple cutaneous schwannomas coursing along a nerve14 or occurring in multiple different sites and often associated with intracranial tumors.15,16 While some authors believe that this entity (schwannomatosis) is distinct from neurofibromatosis type II (NF-2), others believe that this might represent an unusual variant of NF-2.17

Neuroblastoma-like Schwannoma Another rare variant of schwannoma is one in which the cells have a rounded morphology and are often centered around a collagen core forming rosette-like structures, giving a superficial resemblance to neuroblastoma.18 However, both ultrastructural and immunohistochemical examination reveal that these cells have the characteristic features of schwann cells, and this is supported by the benign clinical outcome in all cases.

Melanotic Schwannoma Another rare form of schwannoma has been variably referred to as the melanocytic schwannoma19 or the psammomatous melanotic schwannoma.20 Interestingly, over 50% of the patients with this unusual tumor have evidence of Carney's syndrome (myxomas in a variety of sites, spotty pigmentation and endocrine overactivity).20 The most striking histologic features are the heavy pigment deposits which stain positively for Fontana's stain and negatively for iron, as well as the presence of psammoma bodies, which may be extensive in some cases. Immunohistochemically, the cells strongly express S100 protein and HMB-45. Ultrastructurally, the cells resemble typical schwann cells, except for the

32

presence of premelanosomes and melanosomes. Although most of these tumors act in a benign fashion, rare cases may metastasize.20

Finally, the issue of malignant transformation in a schwannoma is an interesting one, as Woodruff et al could only find 9 acceptable cases reported in the literature.21 None of the patients had evidence of neurofibromatosis type I, and the majority of the patients died of their disease. Histologically, the benign component in these tumors was classical schwannoma, whereas the malignant component consisted of an epithelioid malignant peripheral nerve sheath tumor in 7 cases, and showed evidence of neuroepithelial differentiation in 2 cases. Thus, although exceedingly rare, schwannomas do have the capacity to undergo malignant transformation.

Synovial Sarcoma Synovial sarcoma is the third most common type of sarcoma (after so-called MFH and liposarcoma), and typically affects adolescents and young adults (most common between 15-40 years).22 By far the most common location is the extremities, particularly in proximity to large joints (especially the knees), but distal extremity synovial sarcomas are not uncommon. These tumors can also occur on the upper extremities, head and neck region and the trunk. Although these tumors are often intimately related to tendons, tendon sheaths and bursal structures, they are exceedingly rare within joint cavities.

Histologically, synovial sarcomas are composed of variable mixtures of epithelial cells, spindled cells and cells that have features intermediate between these two (transitional cells). Thus, the classic biphasic synovial sarcoma is fairly easily recognized, as it is composed of a distinct population of these three cell types. However, when the epithelial elements are predominant (epithelial type of synovial sarcoma), these lesions may be extremely difficult to recognize and separate from carcinomas, melanomas or mesotheliomas. At the other extreme, when the epithelial elements are difficult to identify or completely absent (monophasic fibrous type of synovial sarcoma), then this lesion becomes difficult to differentiate from other highly cellular spindle cell sarcomas.

The monophasic fibrous type of synovial sarcoma is probably the most common subtype and can usually be recognized through a combination of histologic and immunohistochemical features. At low magnification, one is often impressed with a “marbled” appearance that is due to alternating areas of low and high cellularity. The spindled cells are often arranged into irregular fascicles, but typically lack the regular herringbone pattern seen in fibrosarcoma. Although the cells are generally spindled, some may have more of an ovoid appearance.

33

Nuclear pleomorphism is typically minimal, and mitotic figures can be identified fairly easily. Other features that should suggest a diagnosis of monophasic fibrous type of synovial sarcoma include the presence of calcification or ossification, a conspicuous mast cell infiltrate, and a hemangiopericytomatous vasculature.

Immunohistochemistry is extremely useful in arriving at this diagnosis. Virtually all monophasic synovial sarcomas stain for cytokeratins, epithelial membrane antigen, or both.23-26 Guillou et al.26 found that all but one of 100 cases of synovial sarcoma stained for at least one of these epithelial markers. In their hands, a significantly greater percentage of cases stained for EMA, although in our laboratory, we have found AE1/AE3 or CAM 5.2 to be more consistently positive. In addition, Guillou et al. noted the relative frequent S100 protein immunoreactivity in all subtypes of synovial sarcoma. Thus, not all spindle cell sarcomas that stain for S100 protein are necessarily malignant peripheral nerve sheath tumors. Given the fact that rare cases of MPNST stain for cytokeratins (and in fact may be S100 protein negative), we have found cytokeratin subsets useful in this regard. Smith et al.27 found that virtually all monophasic synovial sarcomas stained for cytokeratins 7, 19, or both, whereas staining for either of these antigens is extremely rare in cases of MPNST. More recently, gene expression profiling of synovial sarcomas revealed consistent expression of the TLE1 gene.28 Subsequently, an antibody to TLE1 was developed which, according to Terry et al, shows a high degree of sensitivity and specificity for synovial sarcoma.29 In our experience, this antibody is exceedingly useful in confirming a diagnosis of synovial sarcoma, and it is very easy to interpret, since, in my experience, virtually all cells in every synovial sarcoma I have tested shows strong nuclear immunoreactivity. Finally, it must be kept in mind that a significant percentage of synovial sarcomas (including poorly differentiated synovial sarcomas) show membranous immunoreactivity for CD99, a feature which can cause confusion with the Ewing's family of tumors.30

A consistent, specific translocation, most commonly a balanced reciprocal translocation, t(X;18)(p11;q11), is found in virtually all synovial sarcomas, regardless of subtype.31 This translocation involves the fusion of the SYT gene (also known as SS18) on chromosome 18 with either the SSX1 or SSX2 gene on the X chromosome (both at Xp11) or very rarely with SSX4 (also Xp11).32-34 This fusion can be detected by RT-PCR or FISH. In our practice, we utilize an SYT breakapart probe utilizing paraffin-embedded tissue, an ancillary test used in virtually any case in which we suspect a diagnosis of synovial sarcoma.

Synovial sarcoma is generally viewed and treated as a high-grade sarcoma. However, in a large series of cases reported by Bergh et al, the authors identified features that could place

34

patients in either low or high-risk groups.35 Adverse prognostic factors with regard to metastasis and survival included older age, larger tumor size, the presence of poorly differentiated areas, high Ki-67 values, the presence of necrosis, vascular invasion and prior local recurrence.

Malignant Peripheral Nerve Sheath Tumor Another lesion that frequently enters into the differential diagnosis is malignant peripheral nerve sheath tumor (MPNST). Unless this lesion clearly arises from a nerve trunk, a neurofibroma, or occurs in a patient with von Recklinghausen's disease, it may be extremely difficult to recognize and differentiate from these other lesions. At low magnification, similar to the appearance of monophasic synovial sarcoma and cellular schwannoma, there are alternating hypo- and hypercellular regions resulting in a marbled appearance. The spindled cells are arranged into an irregular fascicular pattern, similar to that seen in fibrosarcoma, but other architectural patterns may be present, including areas of nuclear palisading, myxoid zones, and a perivascular targetoid growth pattern. The tumor cells are wavy or angulated, but show more nuclear pleomorphism than that seen in cellular schwannoma. Mitotic figures are usually numerous (>4 MF/10 HPF). Immunohistochemically, about 60% of MPNSTs stain for S100 protein, typically with only focal positivity.

35

References

1. Ducatman BS, Scheithauer BW. Malignant peripheral nerve sheath tumors with divergent differentiation. Cancer 1984;54:1049-1057. 2. Weiss SW, Langloss JM, Enzinger FM. Value of S100 protein in the diagnosis of soft tissue tumors with particular reference to benign and malignant schwann cell tumors. Lab Invest 1983;49:299-308. 3. Lodding P, Kindblom LG, Angervall L, Stenman G. Cellular schwannoma: A clinicopathologic study of 29 cases. Virch Arch Pathol Anat 1990; 416:237-248. 4. White W, Shiu MH, Rosenblum MK, Erlandson RA, Woodruff JM. Cellular schwannoma: A clinicopathologic study of 57 patients and 58 tumors. Cancer 1990; 66:1266-1275. 5. Fletcher CDM, Davies SE, McKee PH. Cellular schwannoma: A distinct pseudosarcomatous entity. Histopathology 1987; 11:21-35. 6. Woodruff JM, Godwin TA, Erlandson RA, Susin M, Martini N. Cellular schwannoma: A variety of schwannoma sometimes mistaken for a malignant tumor. Am J Surg Pathol 1981; 5:733-744. 7. Woodruff JM. Cellular schwannoma. Bone and Soft Tissue Specialty Conference, Case 1. USCAP Meeting, Toronto, Canada, 1995. 8. Dahl I. Ancient neurilemmoma (schwannoma). Acta Pathol Microbiol Scand 1977;85(6):812-818. 9. Fletcher CDM, Davies SE. Benign plexiform (multinodular) schwannoma: A rare tumor unassociated with neurofibromatosis. Histopathology 1986;19:971-980. 10. Hirose T, Scheithauer BW, Sano T. Giant plexiform schwannoma: A report of two cases with soft tissue and visceral involvement. Mod Pathol 1997;10:1075-1081. 11. Brooks JJ, Draffen RM. Benign glandular schwannoma. Arch Pathol Lab Med 1992;116:192-195. 12. Fletcher CDM, Madziwa D, Heyderman E, et al. Benign dermal schwannoma with glandular elements - true heterology or a local “organizer” effect? Clin Exp Dermatol 1986;11:475-485. 13. Woodruff JM, Christensen WN. Glandular peripheral nerve sheath tumors. Cancer 1993;72:3618-3628. 14. Buenger KM, Porter NC, Dozier SE, Wagner RF. Localized multiple neurilemmomas of the lower extremity. Cutis 1993;51:36-38. 15. Purcell SM, Dixon SL. Schwannomatosis: An unusual variant of neurofibromatosis or a distinct clinical entity? Arch Dermatol 1989;125:390-393. 16. Shishibo T, Niimura M, Ohtsuka F, Tsuru N. Multiple cutaneous neurilemmomas as a skin manifestation of neurilemmomatosis. J Am Acad Dermatol 1984;10:744-754. 17. Reith JR, Goldblum JR. Multiple cutaneous plexiform schwannomas: Report of a case and review of the literature with particular reference to the association with types 1 and 2 neurofibromatosis and schwannomatosis. Arch Pathol Lab Med 1996;120:399-401. 18. Goldblum JR, Beals TF, Weiss SW. Neuroblastoma-like neurilemmoma. Am J Surg Pathol 1994;18:266-273. 19. Fu YS, Kaye GI, Lattes R. Primary malignant melanocytic tumors of the sympathetic ganglia with an ultrastructural study of one. Cancer 1975;36:2029-2041. 20. Carney JA. Psammomatous melanotic schwannoma: A distinctive heritable tumor with special associations including cardiac myxoma and the Cushing syndrome. Am J Surg Pathol 1990;14:206-222. 21. Woodruff RM, Selig AM, Crowley K, Allen PW. Schwannoma with malignant transformation. A rare, distinctive peripheral nerve tumor. Am J Surg Pathol 1994;18(9):882-895. 22. Weiss SW, Goldblum JR. Synovial sarcoma. In: Enzinger and Weiss's Soft Tissue Tumors, 5th Ed. Elsevier, New York 2008.

36

23. Ordonez NG, Mahfouz SM, Mackay B. Synovial sarcoma. An immunohistochemical and ultrastructural study. Hum Pathol 1990;21:733-749. 24. Schmidt D, Thum P, Med C, Harms D, Treuner J. Synovial sarcoma in children and adolescents. A report from the Kiel Pediatric Tumor Registry. Cancer 1991;67:1667- 1672. 25. Fetsch JF, Meis JM. Synovial sarcoma of the abdominal wall. Cancer 1993;72:469- 477. 26. Guillou L, Wadden C, Kraus MD, Dei Tos AP, Fletcher CDM. S100 protein reactivity in synovial sarcomas - A potentially frequent diagnostic pitfall. Immunohistochemical analysis of 100 cases. Appl Immunohistochem 1996;4(3):167-175. 27. Smith TA, Machen SK, Fisher C, Goldblum JR. Utility of cytokeratin subsets in distinguishing monophasic synovial sarcoma from malignant peripheral nerve sheath tumor. Am J Clin Pathol 1999;112:641-648. 28. Nielsen TO, West RB, Linn SC, et al. Molecular characterization of soft tissue tumours: a gene expression study. Lancet 2002;359:1301-1307. 29. Terry J, Saito T, Subramanian S, et al. TLE1 as a diagnostic immunohistochemical marker for synovial sarcoma emerging from gene expression profiling studies. Am J Surg Pathol 2007;31:240-246. 30. Folpe AL, Schmidt RA, Chapman D, Gown AM. Poorly differentiated synovial sarcoma: immunohistochemical distinction from primitive neuroectodermal tumors and high-grade malignant peripheral nerve sheath tumors. Am J Surg Pathol 1998;22:673-682. 31. Dal Cin P, Rao U. Jani-Sait S, Karasousis C, Sandberg AA. Chromosomes in the diagnosis of soft tissue tumors. I. Synovial sarcoma. Mod Pathol 1992;5:357-362. 32. de Leeuw B, Balemans M, Olde Weghuis D, Geurts van Kessel A. Identification of two alternative fusion genes, SYT-SSX1 and SYT-SSX2, in t(x;18)(p11.2;q11.2)- positive synovial sarcomas. Hum Mol Genet 1995;4:1097-1099. 33. Fligman I, Lonardo F, Jhanwar SC, Gerald WL, Woodruff J, Ladayni M. Molecular diagnosis of synovial sarcoma and characterization of a variant SYT-SSX2 fusion transcript. Am J Pathol 1995;147:1592-1599. 34. Argani P, Zakowski MF, Klimstra DS, Rosai J, Ladanyi M. Detection of the SYT-SSX chimeric RNA of synovial sarcoma in paraffin-embedded tissue and its application in problematic cases. Mod Pathol 1998;11;65-71. 35. Bergh P, Meis-Kindblom JM, Gherlinzoni F, et al. Synovial sarcoma: identification of low and high-risk groups. Cancer 1999;85:2596-2602.

37

Round Cell Tumor Pattern The differential diagnosis of small round cell tumors is broad. On occasion, one can encounter a benign round cell tumor (e.g. glomus tumor, giant cell tumor of tendon sheath, cutaneous adnexal tumors of various types), which can be mistaken for a high-grade round cell sarcoma. However, most of the time when one encounters a round cell pattern, the differential diagnosis includes a broad group of malignant round cell tumors which includes the Ewing's family of tumors (EFT), alveolar rhabdomyosarcoma, neuroblastoma, lymphoblastic lymphoma, Merkel cell carcinoma, small cell carcinoma, poorly differentiated synovial sarcoma, mesenchymal chondrosarcoma, round cell liposarcoma, desmoplastic small round cell tumor, small cell osteosarcoma and others. Although the light microscopic features are useful in narrowing this differential diagnosis, in virtually every case, ancillary studies including immunohistochemistry and molecular genetic studies are required in order to more precisely classify the round cell tumor. Given that this has important therapeutic and prognostic implications, simply designating a given tumor as a round cell tumor, not otherwise specified, is not acceptable in most cases. However, there are rare cases that cannot be precisely classified.

Ewing’s/Peripheral Neuroectodermal Tumor (Ewing's Family of Tumors or EFT) A review of the literature over the past twenty years reveals a remarkable evolution in the concepts regarding classic osseous and extra-osseous Ewing's sarcoma. Through an accumulation of data, it has become apparent there is a spectrum of tumors that ranges from classic Ewing's sarcoma to classic peripheral neuroectodermal tumor (PNET). Since its initial description by James Ewing in 1921, Ewing's sarcoma was felt to arise only in bone, and it was not until 1975 when Angervall and Enzinger described the first cases of extra- osseous Ewing's sarcoma.1 PNET, since its initial description by Arthur Purdy Stout in 1918 as a round cell tumor of the ulnar nerve,2 has been documented in soft tissue unassociated with nerve,3 as well as within bone.4 As discussed below, these entities have histologic, immunohistochemical, ultrastructural, cytogenetic and molecular genetic features that are overlapping, supporting the histogenetic relationship among these neoplasms.

EFT: Clinical Features Most patients with EFT are adolescents or young adults, the majority of whom are less than 30 years of age.5 Although the mean age for PNET is similar to that of ES, there tends to be a broad age range for the former, with a significant number of patients over the age of 40 years. In contrast, patients with classic extraskeletal ES are rarely over 40 years of age. Both tumors are slightly more common in males than in females.

38

PNET most commonly arises in the extremities. In my experience, the most common anatomic sites are the upper thigh and buttock, followed by the upper arm and shoulder. Tumors intimately attached to a major nerve may give rise to signs and symptoms related to diminished neurologic function. The principal sites of classic extraskeletal ES are the paravertebral region and chest wall, generally in close association with the vertebrae or the ribs. These tumors may also arise in the soft tissues of the lower extremities and rare in the pelvic and hip regions, the retroperitoneum and the upper extremities. However, it is important to note that virtually every anatomic site has been documented to be involved by this family of tumors. In general, the tumor presents as a rapidly growing, deeply situated mass measuring between 5 and 10 cm. Superficially located cases do occur but are quite rare.

EFT: Pathologic Findings Classic ES is composed of solidly packed uniform small cells arranged in a lobular pattern separated by dense fibrous septa. At low magnification, there is nuclear uniformity, and nucleoli are inconspicuous, as are mitotic figures. The nuclear chromatin is fine and powdery, and there is a thin rim of pale cytoplasm often filled with glycogen. Hemorrhage and necrosis are prominent features. On the other end of the spectrum, classic PNET is characterized by more irregularity in nuclear size and shape, as well as a coarsening of the chromatin, more prominent nucleoli and increased mitotic figures. In addition, rosettes of varying types, including Homer Wright, Flexner-Wintersteiner and perivascular rosettes are typically seen. However, in between these ends of the spectrum is a variety of architectural and cytologic features in which it is unclear whether the lesion is best classified as a classic ES or PNET. The entity of "atypical Ewing's sarcoma" has been proposed to include some of these cases between the two ends of the histologic spectrum.6 However, the cut-off between ES and atypical ES, and between ES and PNET, is unclear and arbitrary. Fortunately, however, this distinction is not necessary since a number of more recent studies have shown that there is no significant difference in prognosis based upon where a given tumor lies on this morphologic spectrum.7

Ultrastructurally, typical ES is a primitive neoplasm composed of uniform round cells devoid of specific features and characterized by abundant deposits of cytoplasmic glycogen. On the other hand, PNET shows ultrastructural evidence of some degree of neural differentiation, including rare dense core granules, neuritic cell processes, neurofilaments and neural tubules. Similar to that seen histologically, there is a spectrum of ultrastructural features between the two extremes that can be seen.

39

EFT: Immunohistochemical Features For many years, a diagnosis of either ES or PNET was essentially an immunohistochemical diagnosis of exclusion. However, it is clear that the product of the MIC2 gene (CD99) is the most sensitive marker on this family of tumors.8 The MIC2 gene is a pseudoautosomal gene located on the short arms of the sex chromosomes, and its product is a membranous glycoprotein that can be detected immunohistochemically using a variety of differenti antibodies (including 12E7 and O13). Although initially believed to be highly specific for the EFT, it is apparent that virtually all other round cell tumors in the differential diagnosis, on rare occasion, show membranous immunoreactivity for the MIC2 gene product, including lymphomas, particular T-lymphoblastic lymphoma and precursor B-lymphoblastic lymphoma, Merkel cell carcinoma, small cell carcinoma, alveolar rhabdomyosarcoma, small cell osteosarcoma, desmoplastic small round cell tumor and mesenchymal chondrosarcoma.9-12 Notably, however, childhood neuroblastomas have not been reported to stain for this antigen. Thus, although immunostains for CD99 are highly sensitive for recognizing the EFT, this marker should be used as part of a panel of immunostains, given the lack of complete specificity.

There are a few other notable immunohistochemical findings that one should be aware of for EFT. For example, up to 20% of tumors have focal immunoreactivity for low-molecular- weight cytokeratins,13 although these tumors do not express cytokeratins 7 or 19, a useful finding for distinguishing EFT from poorly differentiated synovial sarcoma.14 Desmin may also rarely be found in EFT, but there is no ultrastructural evidence of skeletal muscle differentiation in these tumors.15

Cytogenetic and Molecular Genetic Findings Approximately 90-95% of EFT are characterized by rearrangements of the EWS gene on 22q12 and ETS-related oncogenes, most commonly FLI-1 on 11q24.16 Less commonly, the EWS gene is fused with other ETS-related genes, including ERG on 21q22,17 ETV-1 on 7p22,18 E1AF on 17q1219 or FEV on 2q33.20 The translocation breakpoints are restricted to introns 7-10 of the EWS gene and introns 3-9 on the ETS-related gene, with the most common fusion being between exon 7 of EWS and exons 5 or 6 of FLI-1.21 These translocations result in a novel chimeric gene that encodes for a chimeric transcript in protein, the function of which is largely unknown. Given the limitations of traditional cytogenetic techniques for detecting these translocations, the ability to detect fusion transcripts by molecular genetic techniques (including RT-PCR and FISH) using fixed, paraffin-embedded tissues has greatly facilitated the diagnosis of these tumors. In our

40

practice, we utilize an EWSR1 breakapart probe as a routine part of the work-up for a suspected EFT.

Alveolar Rhabdomyosarcoma Alveolar rhabdomyosarcoma (ARMS) is another important lesion to distinguish from EFET, given the different therapeutic modalities used to treat these tumors. Although there may be some overlap in the age distribution, ARMS often occurs in patients younger than seen in EFT. This tumor has a predilection for the deep soft tissues of the extremities, although it may arise in many other sites, including the head and neck, trunk, perineum, pelvis and retroperitoneum.

Histologically, ARMS is composed largely of ill-defined aggregates of poorly differentiated round or oval tumors cells that frequently show central loss of cellular cohesion and formation if irregular "alveolar" spaces. The individual cellular aggregates are separated and surrounded by a framework of dense, frequently hyalinized fibrous septa that surround dilated vascular channels. The cells at the periphery of the alveolar spaces are well preserved and adhere in a single layer to the fibrous septa, whereas those in the center of the alveolar spaces tend to be more loosely arranged or freely floating. These centrally located cells are often poorly preserved and show evidence of degeneration and necrosis. There are also "solid" forms of ARMS that lack an alveolar growth pattern entirely and are composed of densely packed groups of tumor cells resembling the round cell areas of EFT. These solidly cellular areas are more commonly encountered at the periphery of the tumor and probably represent the most active and most cellular stage of growth. However, even in these solid areas, there is a regular arrangement of fibrous septa that surround the primitive round cells. Rhabdomyoblasts may be found, but in some cases, they may be extremely difficult to identify.

Immunohistochemistry is extremely useful in making the diagnosis of ARMS. Although it is true that most of these tumors do express desmin and muscle-specific actin (HHF-35), there are some ARMS that do not express either of these antigens. In my experience, myogenin is the best marker in recognizing ARMS. Both MyoD1 and myogenin are members of the MyoD family of genes, which encode a series of DNA binding proteins that are involved in the initiation of myogenic differentiation.22 These genes are expressed at the earliest stages of commitment of a mesenchymal cell to striated muscle, and antibodies to these genes appear to be the most sensitive markers of skeletal muscle differentiation.23 Although myogenin is expressed by virtually all subtypes of rhabdomyosarcoma, it is clear that ARMS tends to express this antigen more diffusely and strongly than the other subtypes of

41

rhabdomyosarcoma. It is also important to recognize that some examples of ARMS may show membranous immunoreactivity for CD99.

Again, cytogenetic and molecular genetic features of ARMS may be extremely useful in confirming this diagnosis. The most common translocation is a t(2;13)(q35;q14), resulting in the fusion of the PAX3 gene on chromosome 2 with FOXO1a (formerly known as FKHR) gene on chromosome 13.24 Less commonly, ARMS may show a t(1;13), resulting in a PAX7- FOXO1a fusion. In our practice, we utilize a FOXO1a breakapart probe on fixed, paraffin- embedded tissues to detect this translocation. However, it must be kept in mind that only about 75% of ARMS have either a t(2;13) or t(1;13), and 25% of ARMS lack either of these translocations.

Once a diagnosis of rhabdomyosarcoma is established, it is important to properly subtype the tumor, given the significant prognostic implications. However, pathologists are not particularly good at subclassifying rhabdomyosarcomas, as there is a high degree of inter- and intraobserver variation in classifying these tumors.25 The Intergroup Rhabdomyosarcoma Study (IRS) has proposed the International Classification of Rhabdomyosarcoma (ICR), which seems to be the most reproducible classification scheme as well as the scheme which predicts prognosis best.26 Tumors having a superior prognosis include botryoid and spindle cell variants of embryonal rhabdomyosarcoma. The usual type of embryonal rhabdomyosarcoma has an intermediate prognosis, whereas ARMS has a poor prognosis.

Desmoplastic Small Round Cell Tumor (DSRCT) DSRCT is a relatively uncommon entity that typically involves the abdominal or pelvic peritoneum of young males and pursues an aggressive clinica course. Most patients with this tumor are 15 to35 years of age, although patients younger and older than this classic age range have also been reported. In a study of 109 patients with this tumor by Gerald et al,27 the patients ranged in age from 6 to 49 years, with a mean age of 22 years. Males far outnumber females at a ratio of approximately 4:1.

Most patients present with a large abdominal and/or pelvic mass with extensive peritoneal involvement, usually without an identifiable visceral site of origin. However, this tumor has been noted to arise in virtually every other anatomic location and does not necessarily arise in association with a mesothelial-lined surface. Although there was some speculation that this lesion could be related to a mesothelial neoplasm (mesothelial blastoma), given the fact that these tumors can arise in non-mesothelial locations and given the absence of

42

convincing immunohistochemical or ultrastructural evidence of mesothelial differentiation, the histogenesis of this unusual tumors remains unknown.

Histologically, the neoplasm is composed of sharply outlined islands of tumor cells that are separated by a desmoplastic stroma containing myofibroblasts and prominent vascularity. There is often a suggestion of peripheral palisading, occasionally with central necrosis. The individual cells are relatively uniform, small and round to oval with hyperchromatic nuclei, inconspicuous nucleoli and scanty cytoplasm. Mitotic figures are easily identified. Rarely, cells with peripherally located nuclei and increased eosinophilic cytoplasm with a perinuclear clear zone (rhabdoid morphology) are seen. Despite this classic histology, the morphologic profile of this tumor continues to expand, as Ordonez noted that up to one-third of these tumors have atypical histologic features.28

Immunohistochemically, these lesions have a characteristic polyphenotypic profile with coexpression of cytokeratins, vimentin, desmin and NSE. The pattern of desmin immunoreactivity is quite unique with a characteristic perinuclear globular pattern of staining. An antibody that recognizes the carboxyl terminus portion of the WT1 gene product has also been developed and is highly sensitive and reasonably specific in recognizing this tumor.29 It is also important to note that up to 20% of DSRCT do stain for CD99. Interestingly, a characteristic cytogenetic aberration has been associated with this neoplasm, t(11;22)(p13;q12), involving the Wilms' tumor gene on chromosome 11 and the EWS gene on chromosome 22. In virtually all suspected cases, we utilize an EWSR1 breakapart probe in an attempt to detect this translocation.

Neuroblastoma Neuroblastoma may also be difficult to differentiate from some of these other round cell tumors. Patients with neuroblastoma are typically younger than those with the other tumors, as 90% of patients are diagnosed before the age of 5 years. Neuroblastoma is exceedingly rare in adolescents and adults. These patients often have elevated catecholamine metabolite levels in their urine. Neuroblastomas arise from either the adrenal gland or extra- adrenal sympathetic ganglia, although metastases may be seen in virtually any location.

Histologically, neuroblastomas are composed virtually entirely of small round undifferentiated cells, typically with dark nuclei and clumped chromatin, inconspicuous nucleoli and scanty cytoplasm. The tumor cells are divided into small lobules by fibrovascular septa. Typically, some cells with peripherally located larger nuclei with vesicular chromatin and increased eosinophilic cytoplasm (representing immature ganglion cells) are seen. Neuroblastomas are

43

characterized by rosettes of various types, including Homer Wright rosettes, with the cells deposited in a fibrillary background. Calcifications are often seen, and mature ganglion cells may be present.

Immunohistochemically, neuroblastomas do not stain for CD99, actin, desmin or myogenin. Although the vast majority of neuroblastomas do stain for neural markers, particularly NSE, the expression of these neural markers is not specific. Miettinen et al reported a high degree of sensitivity of the monoclonal antibody NB84 in recognizing neuroblastoma,30 but we have had no experience utilizing this antibody in our clinical practice. Neuroblastomas are characterized by a consistent cytogenetic abnormality with deletion of the short arm of chromosome 1 (1p-).31 Amplification of the N-myc gene is frequently seen in neuroblastoma and is of important prognostic value.32

Mesenchymal Chondrosarcoma Mesenchymal chondrosarcoma is a rare neoplasm that typically occurs in young adults with a peak age in the second to third decade of life. Approximately 20% of these neoplasms arise in an extraosseous location and are most common in the cranial or spinal meninges, orbit and soft tissues of the thigh.33 Histologically, Histologically, the neoplasm is characterized by a biphasic appearance of small nests or nodules of well-differentiated cartilage intimately admixed with undifferentiated round or slightly spindled cells with hyperchromatic nuclei and scanty cytoplasm. Frequently, a hemangiopericytoma-like vascular pattern is present. Although the biphasic appearance is characteristic, this may not be appreciated on a small biopsy, making distinction from other round cell tumors difficult. Immunohistochemically, mesenchymal chondrosarcomas do not express cytokeratins or myogenic markers, but variably express neural markers, including S100 protein. Membranous CD99 immunoreactivity is found in the majority of these tumors.34 In addition, there have been rare reports of mesenchymal chondrosarcoma with the identical t(11;22) identified in the EFT, raising the possibility that these tumors are histogenetically related.35

Round Cell Liposarcoma Round cell liposarcoma is a poorly differentiated form of myxoid liposarcoma and typically behaves as a high-grade sarcoma. Histologically, the cells are relatively uniform, small and round with vesicular nuclei. The fine plexiform vascular pattern that is so characteristic of myxoid liposarcoma is inconspicuous in the round cell areas. This neoplasm may be very difficult to recognize, particularly on a needle biopsy, without a component of myxoid liposarcoma. In such cases, the application of FISH can be extremely useful, since (like

44

myxoid liposarcoma) round cell liposarcoma is characterized most commonly by a t(12;16) involving the DDIT3 gene on chromosome 12 and the FUS gene of chromosome 16, for which breakapart probes are commercially available. Like myxoid liposarcoma, some cases of round cell liposarcoma may also harbor a t(12;22), and in such cases, using an EWSR1 probe can be useful.

Poorly Differentiated Synovial Sarcoma Poorly differentiated synovial sarcoma is composed of small round cells with little cytoplasm, often separated by a hemangiopericytoma-like vascular pattern. Unless one identifies other areas of classic biphasic or monophasic synovial sarcoma of lower grade, this lesion may be extremely difficult to separate from some of the other round cell sarcomas. In addition, the poorly differentiated form of synovial sarcoma is less likely to express cytokeratins.36 Further adding to the difficulty, some cases of poorly differentiated synovial sarcoma express membranous CD99 immunoreactivity, making distinction from EFT difficult.37 We have found the use to cytokeratin subsets useful in this regard, as a substantial portion of poorly differentiated synovial sarcomas stain for CK7 and 19, while EFT rarely, if ever, stains for the antigens.14 Detection of the t(X;18) using either RT-PCR or FISH (for the SYT gene) may be exceedingly useful in recognizing this tumor.38

45

References

1. Angervall L, Enzinger FM. Extraskeletal neoplasm resembling Ewing's sarcoma. Cancer 1975;36:240-251. 2. Stout AP. A tumor of the ulnar nerve. Proc NY Pathol Soc 1918;1:2-12. 3. Seemayer TA, Thelmo WL, Bolande RP, Wigglesworth FW. Peripheral neuroectodermal tumors. Perspect Pediatr Pathol 1975;2:151-172. 4. Jaffe R, Santamaria M, Yunis EJ, et al. The neuroectodermal tumor of bone. Am J Surg Pathol 1984;8:885-898. 5. Dehner LP. Primitive neuroectodermal tumor and Ewing's sarcoma. Am J Surg Pathol 1993;17(1):1-13. 6. Navarro S, Cavazzana AO, Llombart-Bosch A, Triche TJ. Comparison of Ewing's sarcoma of bone and peripheral neuroepithelioma: An immunocytochemical and ultrastructural analysis of 2 primitive neuroectodermal neoplasms. Arch Pathol Lab Med 1994;118:608-615. 7. Terrier P, Henry-Amar M, Trich TJ, et al. Is neuroectodermal differentiation of Ewing’s sarcoma of bone associated with an unfavorable prognosis? Eur J Cancer 1995;31:307-314. 8. Stevenson AJ, Chatten J, Bertoni F, Miettinen M. CD99 (P30/32MIC2) neuroectodermal/Ewing's sarcoma antigen as an immunohistochemical marker: Review of more than 600 tumors and the literature experience. Appl Immunohistochem 1994;2(4):231-240. 9. Riopel M, Dickman PS, Link MP, Pearlman EJ. MIC-2 analysis in pediatric lymphomas and leukemias. Hum Pathol 1994;25:396-399. 10. Lumadue JA, Askin FB, Perlman EJ. MIC-2 analysis of small cell carcinoma. Am J Clin Pathol 1994;102:692-694. 11. Weidner N, Tjoe J. Immunohistochemical profile of monoclonal antibody O13: Antibody that recognizes glycoprotein p30/32MIC-2 and is useful in diagnosing Ewing's sarcoma and peripheral neuroepithelioma. Am J Surg Pathol 1994;18(5):486-494. 12. Perlman EJ, Dickman PS, Askin FB, et al. Ewing's sarcoma: Routine diagnostic utilization of MIC-2 analysis. Hum Pathol 1994;25:303-307. 13. Gu M, Antonescu CR, Guiter G, et al. Cytokeratin immunoreactivity in Ewing’s sarcoma. Prevalence in 50 cases confirmed by molecular diagnostic studies. Am J Surg Pathol 2000;24:410-416. 14. Machen SK, Fisher C, Gautam RS, Tubbs RR, Goldblum JR. Utility of cytokeratin subsets for distinguishing poorly differentiated synovial sarcoma from peripheral primitive neuroectodermal tumour. Histopathology 1998;33:501-507. 15. Parham DM, Dias P, Kelly DR, Rutledge JC, Houghton P. Desmin positivity in primitive neuroectodermal tumors of childhood. Am J Surg Pathol 1992;16(5):483- 492. 16. Delattre O, Zucman J, Melot T, et al. The Ewing family of tumors - a subgroup of small round cell tumors defined by specific chimeric transcripts. N Engl J Med 1994;331:294-299. 17. Sorensen PHB, Lessnick SL, Lopez-Terrada D, et al. A second Ewing’s sarcoma translocation, t(21;22), fuses the EWS gene to another ETS-family transcription factor, ERG. Nature Genet 1994;6:146-151. 18. Jeon I-S, Davis JN, Braun BS, et al. A variant Ewing’s sarcoma translocation (7;22) fuses the EWS gene to the ETS gene ETV-1. Oncogene 1995;10:1229-1234. 19. Kaneko Y, Yoshida K, Handa M, et al. Fusion of the ETS-family gene E1AF to EWS by t(17;22)(q12;q12) chromosome translocation in an undifferentiated sarcoma of infancy. Genes Chromosomes Cancer 1996;15:115-121. 20. Peter M, Couturier J, Pacquement H, et al. A new member of ETS family fused to EWS in Ewing tumors. Oncogene 1997;14:1159-1164.

46

21. de Alava E, Kawai A, Healey JH, et al. EWS-FLI-1 fusion transcript structure is an independent determinant of prognosis in Ewing’s sarcoma. J Clin Oncol 1998;16:1248-1255. 22. Weintraub H, Davis R, Tapscott S, et al. The MyoD gene family: Nodal point during specification of the muscle cell lineage. Science 1991;251:761-766. 23. Wang NP, Marx J, McNutt MA, et al. Expression of myogenic regulatory proteins (myogenin and MyoD1) in small blue round cell tumors of childhood. Am J Pathol 1995;147:1799-1810. 24. Shapiro DN, Sublett JE, Li B, Downing JR, Maeve CW. Fusion of PAX 3 to a member of the forkhead family of transcription factors in human alveolar rhabdomyosarcoma. Cancer Res 1993;53:5108-5112. 25. Asmar L, Gehan E, Newton WA, et al. Agreement among and within groups of pathologists in the classification of rhabdomyosarcoma and related childhood sarcomas: Report of an international study of four pathology classifications. Cancer 1994;74:2579-2588. 26. Newton WA, Gehan EA, Webber BL, et al. Classification of rhabdomyosarcomas and related sarcomas: Pathologic aspects and proposal for a new classification - an Intergroup Rhabdomyosarcoma Study. Cancer 1995;76:1073-1085. 27. Gerald WL, Ladanyi M, de Alava E, et al. Clinical, pathologic and molecular spectrum of tumors associated with t(11;22)(p13;q12): desmoplastic small round-cell tumor and its variants. J Clin Oncol 1998;16:3028-3036. 28. Ordonez NG. Desmoplastic small round cell tumor. I: a histopathologic study of 39 cases with emphasis on unusual histological patterns. Am J Surg Pathol 1998;22:1303-1313. 29. Barnoud R, Sabourin J-C, Pasquier D, et al. Immunohistochemical expression of WT1 by desmoplastic small round cell tumor: a comparative study with other small round cell tumors. Am J Surg Pathol 2000;24:830-836. 30. Miettinen M, Chatten J, Paetau A, Stevenson A. Monoclonal antibody NB84 in the differential diagnosis of neuroblastoma and other small round cell tumors. Am J Surg Pathol 1998;22:327-332. 31. Brodeur GM, Green AA, Hayes FA, Williams KJ, Williams DL, Tsiatis AA. Cytogenetic features of human neuroblastomas and cell lines. Cancer Res 1981;41:4678-4686. 32. Brodeur GM, Seeger RC, Schwab M, et al. Amplification of N-myc in untreated human neuroblastomas correlates with advanced disease stage. Science 1984;224:1121-1124. 33. Nakashima Y, Unni KK, Shives TS, Swee RG, Dahlin DC. Mesenchymal chondrosarcoma of bone and soft tissue: a review of 111 cases. Cancer 1986;57:2444-2453. 34. Granter SR, Renshaw AA, Fletcher CDM, Bhan AK, Rosenberg AE. CD99 reactivity in mesenchymal chondrosarcoma. Hum Pathol 1996;27:1273-1276. 35. Sainati L, Scapinello A, Montaldi A, et al. A mesenchymal chondrosarcoma of a child with the reciprocal translocation (11;22)(q24;q12). Cancer Genet Cytogenet 1993;71:144-147. 36. Folpe AL, Schmidt RA, Chapman D, Gown AM. Poorly differentiated synovial sarcoma: immunohistochemical distinction from primitive neuroectodermal tumors and high-grade malignant peripheral nerve sheath tumors. Am J Surg Pathol 1998;22:673-682. 37. Dei Tos AP, Wadden C, Calonje E, et al. Immunohistochemical demonstration of glycoprotein P30/32MIC2 (CD99) in synovial sarcoma. A potential cause of diagnostic confusion. Appl Immunohistochem 1995;3(3):168-173. 38. Shipley J, Crew J, Birdsall S, et al. Interphase fluorescence in-situ hybridization and reverse transcription polymerase chain reaction as a diagnostic aide for synovial sarcoma. Am J Pathol 1996;148:559-567.

47

Controversies in the Diagnosis of Barrett's Esophagus and Barrett's-Related Dysplasia. One Pathologist's Perspective

John R. Goldblum, MD

Chairman, Department of Anatomic Pathology Cleveland Clinic Professor of Pathology Cleveland Clinic Lerner College of Medicine [email protected]

- III -

One of the more common specimens encountered by both general surgical pathologists and gastrointestinal pathologists are esophageal biopsy specimens to "rule out Barrett's esophagus (BE)," as well as to assess for the presence or absence of dysplasia in patients with known BE. BE is associated with an increased risk of esophageal adenocarcinoma,1 and, thus, patients with this condition undergo periodic endoscopic surveillance with biopsies for early detection of dysplasia and malignancy. This discussion will be centered on some of the newer controversies related to BE and BE-related dysplasia. First and perhaps most important, the evolving definition of BE will be discussed. Given that this diagnosis implies an increased risk of esophageal adenocarcinoma and perhaps lifelong endoscopic surveillance, it is critically important to gain an understanding as to what constitutes BE. Second, this discussion will focus on the difficulties in recognizing BE-related dysplasia and, in particular, touch upon the controversial topic of so-called "basal crypt dysplasia." Finally, although there is often great attention paid to the difficulties of recognizing and agreeing upon dysplasia at the low end of the spectrum (negative for dysplasia versus indefinite for dysplasia versus low-grade dysplasia), very little attention has been paid to the upper end of the dysplastic spectrum, that is -- the separation of high-grade dysplasia from intramucosal adenocarcinoma in biopsy specimens.

The evolving definition of BE

The definition of BE has evolved over the years, and it is not necessary to delve too deeply into some of the past definitions of this condition. In 2008, the Practice Parameters Committee of the American College of Gastroenterology (ACG) published its recently updated guidelines for the diagnosis, surveillance and treatment of BE.2 The guidelines state the definition of BE to be a "change in the distal esophageal epithelium of any length that can be recognized as columnar type mucosa at endoscopy and is confirmed to have intestinal metaplasia by biopsy of the tubular esophagus."2 Thus, there are two basic components to this definition -- an endoscopic abnormality and a histologic alteration characterized by the identification of goblet cells. Although the job of the pathologist is fairly straightforward given this definition, it is amazing how much difficulty and confusion interpretation of these esophageal biopsy specimens can cause. BE can be further subdivided into long-segment (≥3 cm) or short-segment (<3 cm) BE, based on the length of involvement of the distal esophagus.3, 4 In addition, some patients show extremely short segments (<1 cm) of BE with very subtle irregularities of the squamocolumnar junction (so-called "ultrashort-segment BE"). The latter condition may be extremely difficult to distinguish from intestinal metaplasia of the gastric cardia. As stated in the excellent recent review by Riddell and Odze,5 " the basis for this definition relies mainly on the fact that the majority of adenocarcinomas of the esophagus develop in Barrett's esophagus mucosa with intestinal metaplasia, in retrospective cohort studies."6-10 This, therefore, implies that metaplastic, non-goblet-cell-containing columnar epithelium in the distal esophagus does not impart a similar risk of adenocarcinoma and, as such, a diagnosis of BE should not be rendered in the absence of intestinal metaplasia (IM). However, there is not universal agreement that IM should be required for a diagnosis of BE. For example, the British Society of Gastroenterology does not require IM to establish this diagnosis,11,12 and the Japanese do not require goblet cells to diagnose this condition, but rather simply require the identification of columnar-lined esophagus (CLE).13 There are several lines of thought which reasonably challenge the contention that IM should be required for a diagnosis of BE. First, there is some belief that most (if not all) patients with a CLE ultimately develop IM if followed for a long enough period of time.14 In truth, there are few data to strongly support this contention. However, there are data to

suggest that the majority of patients with at least 2 cm of CLE actually prove to have IM if enough biopsy samples are taken from the CLE segment.15 In the study by Harrison and colleagues, close to 75% of patients with CLE proved to have IM, but even with extensive biopsy sampling (up to 16 biopsies), about 25% of patients were not found to have evidence of IM. Kim et al found that up to 20% of patients with more than 3 cm of CLE had no evidence of IM.16 These studies clearly show that sampling error could be a confounding factor, in that some patients with significant lengths of CLE do not have documentation of IM. In reviewing the literature on this topic, I have reached several conclusions. First, the longer the segment of CLE, the more likely IM will be found. Second, the more biopsy samples are taken from the CLE, the more likely goblet cells will be sampled. Finally, since goblet cells tend to be found preferentially in the most proximal part of the BE segment,17 focused sampled from this location will give a higher yield of IM. For years, it has been assumed that only intestinalized mucosa predisposes to esophageal adenocarcinoma and hence the reason why the definition of BE (according to the ACG) relies upon the identification of goblet cells. Thus, by inference, this definition suggests that metaplastic, non-goblet cell columnar epithelium in the esophagus is not premalignant. There are recent data to suggest this may not in fact be the case. First, immunohistochemical studies have found that non-goblet cell columnar epithelium is essentially intestinalized, even though goblet cells are not identified.18-24 For example, metaplastic non-goblet cell columnar epithelium reveals positivity for markers such as CDX-2, DAS-1, villin and HepPar1, all of which are also expressed in intestinalized epithelium throughout the gastrointestinal tract, including esophageal IM. Second, several studies have detected molecular alterations in metaplastic non-goblet cell columnar epithelium which are similar to those found in intestinalized mucosa.25-27 Lui and colleagues, using image cytometry and high-fidelity DNA histograms, found similar DNA content abnormality in these two types of epithelia,26 including peak DNA index, aneuploidy rate and percentage of cells with DNA >5N. Both goblet cell and metaplastic non-goblet cell columnar epithelium in patients with BE showed alterations significantly different from those found in gastric controls. The last line of evidence which calls into question the entirely "benign" nature of metaplastic non-goblet cell columnar epithelium relates to the potential risk of dysplasia and carcinoma. Several studies have found a similar risk of progression in patients with and without goblet cells in CLE.14,28 For example, Kelty et al found a similar rate of development of esophageal adenocarcinoma in patients with CLE without goblet cells (3.6%) compared to those with goblet cells (4.5%).28 Similarly, Gatenby et al found a similar rate of progression to low-grade dysplasia, high-grade dysplasia or carcinoma in patients with CLE with (19.8%) and without (13.2%) goblet cells in their index endoscopy.14 Finally, in studies of esophagectomy specimens with esophageal adenocarcinoma, especially those with a very short segment of CLE, some of these tumors appear to arise in CLE without goblet cells, even with very detailed morphological study with extensive sectioning.29-31 In essence, these studies challenge the long-held belief that all BE-associated adenocarcinomas arise in the setting of intestinalized mucosa. In summary, as of the current time, the ACG requires the recognition of goblet cells to render a definitive diagnosis of BE. This, however, is not true in other parts of the world. Although many U. S. pathologists have embraced the absolute necessity of identifying goblet cells in biopsy specimens in order to make this diagnosis, an open mind is required, since some emerging data have shown CLE with and without goblet cells have immunohistochemical and molecular similarities. Should the definition of BE change to include those patients with CLE without goblet cells, this would obviously have enormous clinical and economic implications.

BE-related dysplasia

In our consultation practice, by far the most common question is the presence (and degree) or absence of dysplasia arising in a background of BE. Dysplasia can be defined as the presence of neoplastic epithelium that is confined within the basement membrane of the gland within which it arises.32 Unlike inflammatory bowel disease-associated dysplastic lesions, at least in my opinion, most cases of BE-related dysplasia do not closely resemble colonic adenomas. Rather, the typical form of BE-related dysplasia often arises in glands that retain their normal configuration and often lack nuclear stratification. Using the criteria as defined by Riddell et al. for dysplasia arising in inflammatory bowel disease, dysplasia in BE can be classified as either low-grade or high-grade based upon the degree of the abnormality present.32 Thus, the possibilities include 1) negative for dysplasia; 2) positive for dysplasia, either low-grade or high-grade; or 3) indefinite for dysplasia. As pointed out by Dr. Appelman,33 the criteria for defining dysplasia in BE was derived from those defined originally in inflammatory bowel disease. In fact, if one attempts to trace back the original criteria for BE-related dysplasia, it is a fruitless endeavor, and inevitably leads back to those defined for IBD in the seminal 1983 article published in Human Pathology.32 Nevertheless, an attempt will be made here to put into words what is extremely difficult to define. In low-grade dysplasia, crypt architecture tends to be preserved with only minimal distortion, and cytologically atypical nuclei are limited to the basal half of the crypts. The nuclei tend to show variable hyperchromasia, overlapping cell borders with nuclear crowding and irregular nuclear contours. Dystrophic goblet cells may be seen, although typically goblet cell numbers are markedly reduced. Separation of low-grade dysplasia from regenerative changes will be discussed below. Simply put, high-grade dysplasia shows more severe cytologic and architectural changes than are present in low-grade dysplasia. Architecturally, there tends to be more crypt distortion in high-grade dysplasia, sometimes with a villiform configuration of the mucosal surface and/or branched or cribriform crypts. Cytologically, the cells show more nuclear pleomorphism and hyperchromatism than is seen in low-grade dysplasia, and there often is nuclear stratification to the crypt luminal surface. A diagnosis of "indefinite for dysplasia," much to our clinical colleagues' dismay, is a legitimate diagnosis. The differentiation of regenerative changes from true dysplasia, particularly in a background of inflammation or ulceration, is at times difficult, if not impossible. Thus, if the pathologist is unsure as to whether the epithelial changes are regenerative or truly dysplastic, a diagnosis of indefinite for dysplasia should be rendered. In some cases, one may not be sure if one is dealing with regenerative or dysplastic changes, even in the absence of inflammation, and a diagnosis of indefinite for dysplasia is acceptable under these circumstances as well.

Distinguishing between regenerative changes and dysplasia

Because BE mucosa is metaplastic, there is a "baseline atypia" that is always present and in a sense must be overlooked in order to make a diagnosis of dysplasia. This baseline atypia is most pronounced in the glands at the base of the mucosa and does not involve the surface epithelium. In addition, biopsies from Barrett's mucosa are not infrequently inflamed, often with both acute and chronic inflammatory cells. As in the case of active chronic inflammatory bowel disease, neutrophil-mediated epithelial injury can induce regenerative cytologic changes that may be difficult to differentiate from dysplasia. Several rules that I follow in making a diagnosis of dysplasia include: 1) I am conservative about making a diagnosis of dysplasia in the face of neutrophil-mediated epithelial injury. Although neutrophils can be found within dysplastic epithelium, the changes have to be convincing in order for me to make a definitive diagnosis of dysplasia. Otherwise, I usually make a diagnosis of indefinite for dysplasia if I am unsure, or negative for dysplasia if the changes are clearly regenerative. 2) The low-magnification appearance of the mucosa is extremely important, as the cytologic changes are typically recognizable at low magnification. A focus usually "catches my eye" and causes me to go to a higher magnification to confirm a diagnosis of dysplasia. 3) I like to see cytologic changes on the surface epithelium, not just in the glandular compartment. In a well-oriented specimen, it is fairly straightforward to determine whether the cytologic changes involve the surface epithelium. On the other hand, in a tangentially sectioned biopsy specimen, this evaluation can be extremely difficult. The reader should be aware of the emerging concept of "basal crypt dysplasia" (BCD), as recently described by Lomo and colleagues.34 There is evidence to suggest that BE-related dysplasia begins in the crypt bases and progresses to involve the full length of the crypts and surface epithelium. Thus, BCD has been proposed to be a form of dysplasia that can be recognized prior to surface involvement (i.e. in the face of surface maturation). Morphologically, BCD has all of the features of traditional low-grade dysplasia, but is limited to the crypt bases and is devoid of active inflammation. In the study by Lomo et al, 47% of cases of BCD (in 15 patients) were associated with full-thickness dysplasia elsewhere.34 Molecular studies have also found similar alterations in BCD when compared to traditional dysplasia and different from those found in non-dysplastic epithelium.35 A myriad of features are useful in separating repair from dysplasia, although I believe some features are more important than others. Dysplastic epithelium tends to show variable nuclear hyperchromasia and pleomorphism. In other words, some cells look different from their neighbors, with some showing nuclear hyperchromasia and pleomorphism when compared to surrounding cells within the same crypt. In contrast, although both nuclear hyperchromasia and pleomorphism may be seen in repair, the changes tend to be less severe and more uniform, with cells resembling their neighbors within the same crypt or in adjacent crypts. Thus, the cytologic atypia associated with repair is more uniform than is seen in dysplasia. Dysplastic cells also tend to have a higher nuclear-to-cytoplasmic ratio as well as irregular nuclear contours. Although regenerative cells may have nuclear size similar to those seen in dysplasia, there tends to be a commensurate increase in the amount of cytoplasm such that the nuclear-to-cytoplasmic ratio is normal or only mildly increased. In addition, regenerative cells tend to have round and regular nuclear contours.

Sampling error and observer variation in the diagnosis of Barrett's-related dysplasia

In any given case, dysplasia may be diffusely distributed throughout a segment of BE, or may be extremely focal. Even using a rigorous sampling technique (4-quadrant biopsies for every

1-2 cm of BE using jumbo biopsy forceps), small foci of dysplasia can be left unsampled. The need for thorough sampling is further emphasized by the fact that many examples of high-grade dysplasia or early adenocarcinoma arising in BE are not associated with a grossly recognizable lesion. Another problem facing both the pathologist and the gastroenterologist (and for that matter the thoracic surgeon) is both the intra- and interobserver variation in the diagnosis of dysplasia. Given the subtle gradation of changes from baseline atypia to low-grade dysplasia to high-grade dysplasia, it is not surprising that this variation exists. Reid et al. found that this variation was most striking at the low end of the scale...in other words, distinguishing negative for dysplasia from low-grade or indefinite for dysplasia.36 A more recent study using kappa statistics by Montgomery and colleagues confirmed a high degree of intra- and interobserver variation in the separation of these diagnoses, even among pathologists with a special interest in gastrointestinal pathology.37

Can we separate bad from worse? The problem of high-grade dysplasia (HGD) versus intramucosal adenocarcinoma (IMC)

As treatment options broaden for patients with BE with HGD and "early adenocarcinoma," some advocates of non-surgical treatments (including intensive endoscopic surveillance, ablative therapy such as a PDT or endoscopic mucosal resection) have proposed that "surgical resection of the esophagus should be reserved for those patients in whom cancer has been documented by biopsy."38 These recommendations, therefore, rest on the assumption that pathologists can reliably discriminate between HGD and IMC in biopsy specimens. Given the fact that lymphatic channels are present within the esophageal mucosa, there is a small but definite risk of lymph node metastasis, even in patients with IMC.39 Recently, several studies have focused on observer variability in separating HGD from IMC in biopsy specimens. Downs-Kelly evaluated pre-selected slides from 168 patients who were treated by esophagectomy for a diagnosis of HGD, IMC or submucosal invasive adenocarcinoma (SMC).40 A pre-resection biopsy specimen from each case showing the most severe histologic abnormality with regard to HGD, IMC or SMC was selected for review by seven GI pathologists, using well-defined consensus histologic criteria. Each case was then placed into one of four diagnostic categories by each pathologist, including 1-HGD; 2-HGD with marked distortion of glandular architecture, cannot exclude IMC (so-called HGD/MAD); 3-IMC; and 4-SMC. HGD was defined according to the published criteria by Reid et al,36 Montgomery et al37 and Odze,41 incorporating both cytologic and architectural changes. Cytologic changes include stratified, hyperchromatic nuclei with nuclear enlargement and an increased N:C ratio, presence of prominent nucleoli, increased mitotic figures and loss of nuclear polarity. These cytologic changes involve both the crypt and surface mucosa. Architectural changes include increased crypt complexity with crowding, branching and cribriform formation along with variability of crypt size and shape. IMC is defined as a neoplasm that has invaded into the surrounding lamina propria or muscularis mucosae but not into the submucosa. However, well-defined histologic criteria for establishing this diagnosis have not been published. Four criteria were arbitrarily defined to establish a diagnosis of IMC in this study, including single-cell invasion of the lamina propria in more than one focus; sheets of cells obliterating the lamina propria; small angulated so- called abortive glands invading the lamina propria; or a never-ending anastomosing gland pattern similar to endometrial adenocarcinomas of the uterine corpus. Only one feature met the criteria for SMC -- unequivocal stromal desmoplasia.

Three criteria were arbitrarily used to establish a diagnosis of HGD/MAD, including glandular crowding with loss of intervening lamina propria; cribriform growth; or prominent (at least 3) dilated glands with intraluminal necrotic debris. The results of this study are intriguing. There was 100% agreement among all seven pathologists in only 12/163 (7.4%) cases. Overall, a majority (at least 4 of 7) of pathologists agreed in 85.9% of cases. However, following Kappa statistical analysis, the overall agreement for all four diagnostic categories was only fair (Kappa=0.30). Cases interpreted as HGD had the best agreement (Kappa=0.47, moderate agreement), while those with SMC had the lowest Kappa score (Kappa=0.14, poor agreement). HGD/MAD and IMC had Kappa scores between these two extremes (Kappa=0.21 and 0.30, respectively, or fair agreement). Based upon these results, the authors called into question management decisions based upon the distinction of these diagnostic categories in pre-treatment biopsy specimens. In a more recent study, Zhu and colleagues from the University of Michigan attempted to ascertain the prevalence of carcinoma in esophagi resected for HGD or HGD "suspicious for carcinoma" (HGD/S) diagnoses in pre-treatment biopsy specimens.42 Based upon the original diagnoses, carcinoma was found in 15 of 89 (17%) HGD cases and 28 of 38 (74%) HGD/S cases. However, these authors made an attempt to more precisely define the HGD and HGD/S categories. The latter category included those cases with solid or cribriform arrangements, ulcers occurring within the high-grade dysplastic mucosa, dilated dysplastic tubules containing necrotic debris, large numbers of neutrophils within the high-grade dysplastic epithelium and dysplastic tubules that were incorporated into overlying squamous epithelium. Following reclassification of the biopsy specimens into these categories, only 1 of 21 (5%) cases with HGD were found to harbor a carcinoma in the esophagectomy specimen. Of 25 cases reclassified as HGD/S, 18 of 25 (72%) had carcinoma, as did 17 of 23 (74%) reclassified as adenocarcinoma (used only for clearly invasive lesions, either clear-cut IMC or SMC). These authors concluded that it is actually quite rare to find an invasive lesion in an esophagectomy performed for HGD using rigorous criteria. HGD/S is a biopsy diagnosis that is more closely associated with an invasive lesion in the esophagectomy specimen; the more individual "suspicious" features found in the biopsy specimen, the more likely an invasive lesion would be found. Recently, Patil and colleagues from Cleveland Clinic attempted to compare the Cleveland Clinic (as defined by the Downs-Kelly study) and University of Michigan (as defined by the Zhu study) systems, as well as correlating pre-resection biopsy diagnosis with the final resection diagnosis with the hopes of identifying histologic features in biopsy specimens that might be predictive of invasive adenocarcinoma on esophagectomy.43 Using the Cleveland Clinic system, only those cases in which there was a consensus agreement (at least 4 of 7 pathologists agreed) were utilized for the purposes of this study. Thus, of the 168 pre-resection biopsy specimens, 112 biopsy samples had a consensus diagnosis, including 32 with HGD (29%), 32 with HGD/MAD (29%), 45 with IMC (40%) and 3 with SMC (2%). Of these 112 biopsy samples, 83 (74%) had adenocarcinoma found in the esophagectomy specimen, including 59 with IMC and 24 with SMC. Applying the University of Michigan system to the biopsy series, excellent diagnostic agreement was found between the Cleveland Clinic and University of Michigan systems (Kappa=0.86). Both systems showed significant correlation between preoperative biopsy diagnosis and esophagectomy diagnosis (p<0.001). The likelihood of finding adenocarcinoma in the resection specimen was significantly higher with HGD/MAD (using the Cleveland Clinic criteria) or HGD/S (using the University of Michigan criteria), compared to HGD alone. The presence of an endoscopic lesion, a "never- ending" glandular pattern, sheet-like growth, angulated glands, three or more dilated glands with intraluminal debris and one or more foci of single cell infiltration into the lamina propria increased the odds of finding adenocarcinoma in the resection specimen. The latter two

variables remained independent predictors of invasive adenocarcinoma in multivariable analysis. Thus, this study shows that there are specific histologic features which can be identified in pre-resection biopsy specimens that are highly predictive of finding an invasive adenocarcinoma in the esophagectomy specimen.

References

1. Wild CP, Hardie LJ. Reflux, Barrett's oesophagus and adenocarcinoma: burning questions. Mat Rev Cancer 2003;3:676-684. 2. Wang KK, Sampliner RE, Practice Parameters Committee of the American College of Gastroenterology. Updated guidelines 2008 for the diagnosis, surveillance and therapy of Barrett's esophagus. Am J Gastroenterol 2008;103:788-797. 3. Sampliner RE, Practice Parameters Committee of the American College of Gastroenterology. Updated guidelines for the diagnosis, surveillance and therapy of Barrett's esophagus. Am J Gastroenterol 2002;97:1888-1895. 4. Spechler SJ. The columnar-lined esophagus. History, terminology and clinical issues. Gastroenterol Clin North Am 1997;26:455-466. 5. Riddell RH, Odze RD. Definition of Barrett's esophagus: time for a re-think -- is intestinal metaplasia dead: Am J Gastroenterol 2009;104:2588-2594. 6. Haggitt RC, Tryzelaar J, Ellis FH, et al. Adenocarcinoma complicating columnar-lined (Barrett's) esophagus. Am J Clin Pathol 1978;70:1-5. 7. Smith RR, Hamilton SR, Boitnott JK, et al. The spectrum of carcinoma arising in Barrett's esophagus. A clinicopathologic study of 26 patients. Am J Surg Pathol 1984;8:563-573. 8. Skinner DB, Walther BC, Riddell RH, et al. Barrett's esophagus. Comparison of benign and malignant cases. Ann Surg 1983;198:554-565. 9. Rosenberg JC, Budev H, Edwards RC, et al. Analysis of adenocarcinoma in Barrett's esophagus utilizing a staging system. Cancer 1985;55:1353-1360. 10. Paraf F, Flejou JF, Pignon JP, et al. Surgical pathology of adenocarcinoma arising in Barrett's esophagus. Analysis of 67 cases. Am J Surg Pathol 1995;19:183-191. 11. Society BG. Guidelines for the diagnosis and management of Barrett's columnar-lined oesophagus. http://www.bsg.org.uk. In: Gastroenterology BSo, ed; 2005. 12. Playford RJ. New British Society of Gastroenterology (GSG) guidelines for the diagnosis and management of Barrett's esophagus. Gut 2006;55:442. 13. Ogiya K, Kawano T, Ito E, et al. Lower esophageal palisade vessels and the definition of Barrett's esophagus. Dis Esophagus 2008;21:645-649. 14. Gatenby PA, Ramus JR, Caygill CP, et al. Relevance of the detection of intestinal metaplasia in non-dysplastic columnar-lined oesophagus. Scand J Gastroenterol 2008;43:524- 530. 15. Harrison R, Perry I, Haddadin W, et al. Detection of intestinal metaplasia in Barrett's esophagus: an observation comparator study suggests the need for a minimum of eight biopsies. Am J Gastroenterol 2007;102:1154-1161. 16. Kim SL, Waring JP, Spechler SJ, et al. Diagnostic inconsistencies in Barrett's esophagus. Department of Veterans' Affairs Gastroesophageal Reflux Study Group. Gastroenterology 1994;107:945-949. 17. Chandrasoma PT, Der R, Dalton P, et al. Distribution and significance of epithelial types in columnar-lined esophagus. Am J Surg Pathol 2001;25:1188-1193. 18. Rogge-Wolf C, Seldenrijk CA, Das KM, et al. Prevalence of mabDAS-1 positivity in biopsy specimens from the esophagogastric junction. Am J Gastroenterol 2002;97:2979- 2985. 19. Glickman JN, Chen YY, Wang HH, et al. The phenotypic characteristics of a distinctive multilayered epithelium suggest that it is a precursor in the development of Barrett's esophagus. Am J Surg Pathol 2001;25:569-578. 20. DeMeester SR, Wickramasinghe KS, Lord RV, et al. Cytokeratin and DAS-1 immunostaining reveals similarities among cardiac mucosa, CIM and Barrett's esophagus. Am J Gastroenterol 2002;97:2514-2523.

21. Das KM, Prasad I, Garla S, et al. Detection of a shared colon epithelial epitope on Barrett epithelium by a novel monoclonal antibody. Ann Intern Med 1994;120:753-756. 22. Phillips RW, Frierson HF Jr, Moskaluk CA, et al. CDX2 as a marker of epithelial intestinal differentiation in the esophagus. Am J Surg Pathol 2003;27:1442;1447. 23. Chu PG, Jiang Z, Weiss LM, et al. Hepatocyte antigen as a marker of intestinal metaplasia. Am J Surg Pathol 2003;27:953-959. 24. Hahn H, Blount PL, Ayub K, et al. Intestinal differentiation in metaplastic, non-goblet cells columnar epithelium in the esophagus. Am J Surg Pathol 2009;33:1006;1015. 25. Chaves P, Crespo M, Ribeiro C, et al. Chromosomal analysis of Barrett's cells: demonstration of instability and detection of the metaplastic lineage involved. Mod Pathol 2007;20:788-796. 26. Liu W, Hahn H, Odze RD, et al. Metaplastic esophageal columnar epithelium without goblet cells shows DNA content abnormalities similar to goblet cell-containing epithelium. Am J Gastroenterol 2009;104:816-824. 27. Romagnoli S, Roncalli M, Graziani D, et al. Molecular alterations of Barrett's esophagus on microdissected endoscopic biopsies. Lab Invest 2001;81:241-247. 28. Kelty C, Gough M, Van Wyk Q. Barrett's oesophagus: intestinal metaplasia is not essential for cancer risk. Scand J Gastroenterol 2007;42:1271-1274. 29. Takubo K, Aida J. Naomoto Y, et al. Cardiac rather than intestinal-type background in endoscopic resection specimens of minute Barrett adenocarcinoma. Hum Pathol 2009;40:65- 74. 30. Clark G, Smyrk TC, Burdiles P, et al. Is Barrett's metaplasia the source of adenocarcinoma of the cardia? Arch Surg 1994;129:609-614. 31. Cameron AJ, Lomboy CT, Pera M, et al. Adenocarcinoma of the esophagogastric junction in Barrett's esophagus. Gastroenterology 1995;109:1541-1546. 32. Riddell RH, Goldman H, Ransohoff DF, et al. Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. Hum Pathol 1983;14:931- 968. 33. Appelman HD. Adenocarcinoma in Barrett mucosa treated by endoscopic mucosal resection. Arch Pathol Lab Med 2009;133:1793-1797. 34. Lomo L, Blount PL, Sanchez CA, et al. Crypt dysplasia with surface maturation: a clinical, pathologic and molecular study of a Barrett's esophagus cohort. Am J Surg Pathol 2006;30:423-435. 35. Zhang X, Huang Q, Goyal RK, et al. DNA ploidy abnormalities in basal and superficial regions of the crypts in Barrett's esophagus and associated neoplastic lesions. Am J Surg Pathol 2008;32:1327-1335. 36. Reid BJ, Haggitt RC, Rubin CE, et al. Observer variation in the diagnosis of dysplasia in Barrett's esophagus. Hum Pathol 1988;19:166-178. 37. Montgomery E, Bronner MP, Goldblum JR, et al. Reproducibility of the diagnosis of dysplasia in Barrett esophagus: a reaffirmation. Hum Pathol 2001;32:368-378. 38. Schnell TG, Sontag SJ, Chejfec G, et al. Long-term non-surgical management of Barrett's esophagus with high-grade dysplasia. Gastroenterology 2001;120:1607-1609. 39. Sabik JF, Rice, TW, Goldblum JR, et al. Superficial esophageal carcinoma. Ann Thorac Surg 1995;60:896-901. 40. Downs-Kelly E, Mendelin JE, Bennett AE, et al. Poor interobserver agreement in the distinction of high-grade dysplasia and adenocarcinoma in pre-treatment Barrett's esophagus biopsies. Am J Gastroenterol 2008;103:2333-2340. 41. Odze RD, Diagnosis and grading of dysplasia in Barrett's oesophagus. J Clin Pathol 2006;59:1029-1038.

42. Zhu W, Appelman HD, Greenson JK, et al. A histologically defined subset of high-grade dysplasia in Barrett mucosa is predictive of associated carcinoma. Am J Clin Pathol 2009;132:94-100. 43. Patil DT, Goldblum JR, Plesec T, et al. Prediction of adenocarcinoma on esophagectomy from pre-resection biopsies of Barrett's esophagus with at least high-grade dysplasia: a comparison of two systems. Mod Pathol 2010 (abstract).

Fibrohistiocytic Tumors of Soft Tissue with a Focus on Fibrohistiocytic Tumors of Intermediate Malignancy

John R. Goldblum, MD

Chairman, Department of Anatomic Pathology Cleveland Clinic Professor of Pathology Cleveland Clinic Lerner College of Medicine [email protected]

- IV -

Fibrohistiocytic tumors of soft tissue are classified into "benign," "intermediate malignancy" and "malignant" categories based on their potential for local recurrence and distant metastases.1 It has become clear over the past 30 years that the term "fibrohistiocytic tumor" is, as Calonje and Fletcher state, "best used as a descriptive term, with no histogenetic implications, to encompass a heterogeneous group of tumors that share histologic similarities."2

Benign Intermediate Malignancy Malignant Fibrous Histiocytoma Atypical Fibroxanthoma "Malignant Fibrous Histiocytoma"

Juvenile Dermatofibrosarcoma - storiform-pleomorphic Xanthogranuloma Protuberans - myxoid - giant cell Reticulohistiocytoma Pigmented DFSP (Bednar Tumor) - inflammatory

Xanthoma Giant Cell Fibroblastoma

Plexiform Fibrohistiocytic Tumor

Angiomatoid Fibrous Histiocytoma

Giant Cell Tumor of Soft Parts

The benign tumors rarely recur and do not metastasize. The tumors of intermediate malignancy frequently recur, but rarely metastasize, and the malignant tumors frequently recur and metastasize.

Benign Fibrous Histiocytoma Within the benign group of fibrohistiocytic tumors, the most common lesion is the benign fibrous histiocytoma (BFH), which has been known by a variety of different names including histiocytoma cutis, nodular subepidermal fibrosis, fibroxanthoma, dermatofibroma, and sclerosing hemangioma. Distinctive variants of BFH have been described, some of which have clinical implications (higher rate of local recurrence). The common fibrous histiocytoma, or dermatofibroma, typically presents as a solitary slow-growing nodule in young to middle-aged adults, with a female predilection. Histologically, these lesions occur predominantly in the reticular dermis, although they can extend into the superficial subcutis. The tumor is characterized by a proliferation of bland spindled cells admixed with histiocyte- like cells, with different lesions having different proportions of these two cell types. Secondary elements, including giant cells, foamy macrophages, siderophages and chronic inflammatory cells are typically found. At low magnification, the typical BFH has a heterogeneous appearance, unlike the monotonous low-magnification appearance of DFSP.

In addition, although there may be a well-formed storiform pattern in some areas of the lesion, this pattern is not as well-developed or monotonous as that seen in DFSP. Dense polarizable keloid-like collagen bundles are typically found at the periphery of BFH, in contrast to the thin, delicate nonpolarizable collagen fibers found at the periphery of DFSP.3 The overlying epidermis characteristically shows hyperplasia as well as basal cell hyperpigmentation, a feature which is typically absent in the epidermis overlying DFSP. Calonje et al described 74 cases of a distinctive variant of BFH they termed "cellular benign fibrous histiocytoma."4 Approximately 5% of BFH have the minimum histologic criteria (increased cellularity and mitotic activity) for a diagnosis of cellular BFH. Similar to the common BFH, cellular BFH typically presents as a solitary lesion on the extremities of young to middle-aged adults. However, in contrast to the common BFH, these lesions have a predilection for males (male/female: 2/1), have a larger average size, and occur in sites which are not typical for common BFH (face, ears, and distal extremities). More importantly, the local recurrence rate is significantly higher in cellular BFH (approximately 25%) than that seen in common BFH (< 2%).4 In addition, there are rare cases that metastasize (begging the question as to why the "benign" is in its name). Histologically, cellular BFH is characterized by high cellularity and increased mitotic rate, although atypical mitoses are not seen. The cellular regions may have a storiform and/or short fascicular growth pattern, and there is frequently involvement of the superficial subcutaneous tissue. At low magnification, these lesions may not appear to be as heterogeneous as the common BFH, since cellular polymorphism (giant cells, siderophages, foamy macrophages) is limited. In the series by Calonje et al, central necrosis was found in 12% of the cases. Thus, these lesions have enough worrisome features such that the potential for misdiagnosis of a sarcoma is reasonably high. The most common lesions with which this is confused are leiomyosarcoma and DFSP. Leiomyosarcoma is characterized by more plump eosinophilic spindle cells with cigar-shaped nuclei arranged in a more uniform fascicular pattern. Although a fascicular pattern may be seen focally in cellular BFH, it is not as well-developed or uniform throughout the tumor as is seen in leiomyosarcoma. Immunohistochemically, cutaneous leiomyosarcomas are typically diffusely positive for smooth muscle actin and desmin. Although smooth muscle actin may be found focally in cellular BFH, it is not as diffuse as is seen in leiomyosarcoma, and desmin and h-caldesmon are negative.4 Factor XIIIa positivity may also be helpful, as cellular BFH is characteristically strongly positive for this antigen. The distinction from DFSP will be described under that section. Another variant of BFH which may cause diagnostic confusion is the atypical BFH,5 which has otherwise been referred to as "dermatofibroma with monster cells"6 or "pseudosarcomatous" cutaneous histiocytoma.7,8 As its name implies, atypical BFH may be

difficult to distinguish from atypical fibroxanthoma (AFX) and superficial pleomorphic sarcoma. Histologically, atypical BFH shows many of the same features as the common BFH, but they are characterized by rare cells with prominent nuclear pleomorphism and large hyperchromatic nuclei. Although mitotic figures may be present, they are usually not numerous. However, in the study by Kaddu et al, mitotic counts ranged from 1 to up to 15 mitotic figures per HPF, including a number of cases with atypical mitotic figures.5 Some cases have other worrisome features including large size (>2 cm), involvement of the superficial subcutis, and geographic necrosis. In contrast, AFX usually arises in actinically- damaged skin of the head and neck of elderly patients.9 Histologically, there is frequent epidermal ulceration (as opposed to epidermal hyperplasia seen in atypical BFH), and there is usually marked pleomorphism throughout the lesion with numerous mitoses, many of which are atypical. It is possible that some of the lesions that have been diagnosed as AFX arising on the trunk of young adults may be examples of atypical BFH.9 In the study by Kaddu et al, follow-up data in 21 patients (mean of 51 months) revealed local recurrences in three patients; two patients developed distant metastases, and one patient died of disease.5 Aneurysmal BFH was originally described by Santa Cruz and Kyriakos as a distinctive variant of BFH characterized by central cystic hemorrhage, often leading to painful and rapid growth of the lesion.11 Clinically, these lesions present as solitary slowly-growing blue-brown nodules, typically in the extremities, and are often clinically misdiagnosed as vascular or melanocytic lesions. They tend to be larger than the common BFH, primarily due to the cystic change with extensive hemorrhage. Histologically, these lesions resemble the common BFH, except for the presence of large blood-filled spaces that are not lined by endothelial cells, but rather by flattened tumor cells. This lesion may be confused with the lesion that had been previously called "angiomatoid malignant fibrous histiocytoma," but is now referred to as angiomatoid fibrous histiocytoma (discussed below). Given this name change, there is apt to be confusion between the aneurysmal BFH (which is benign) and the angiomatoid fibrous histiocytoma (which is of intermediate malignancy). The histologic distinction of these two lesions will be discussed under the section on angiomatoid fibrous histiocytoma. Like the cellular and atypical variants of BFH, aneurysmal BFH seems to be associated with a higher rate of local recurrence than the common type.12 Epithelioid BFH is a rare and often unrecognized variant of cutaneous BFH, originally described by Wilson Jones in 1989.13 Clinically, this lesion typically presents as a solitary, polypoid cutaneous nodule in the extremities of middle-aged adults. The lesion occurs in the superficial dermis, is surrounded by an epidermal collarette, and by definition, over 50% of the cells have an epithelioid morphology. Fortunately, other areas usually show features of the common BFH, allowing for its distinction from other lesions with prominent epithelioid morphology, including Spitz nevus and reticulohistiocytoma.14 Although other variants of

BFH have been described, the ones mentioned above are the variants that tend to cause the most diagnostic confusion, and are thus important to recognize.

Dermatofibrosarcoma Protuberans Dermatofibrosarcoma protuberans (DFSP) is classified as a fibrohistiocytic tumor of intermediate malignancy most commonly arising in young to middle-aged adults with a slight male predilection.15 These lesions most often arise as slowly growing painless plaques, which over a period of time become nodular, and are most common on the trunk, proximal extremities and head and neck. These lesions recur in up to 50% of cases, and recurrences are often multiple, with shorter duration between recurrences. The risk of recurrence appears to be related to the size of the lesion as well as the adequacy of excision. In the study by Roses et al,16 41% of DFSP recurred with a surgical margin of 2 cm or less. Even with a surgical margin of 3 cm or greater, the recurrence rate was 20%. This most likely reflects the subtle pattern of infiltration of the underlying subcutaneous tissue, which is often more impressive than what is grossly appreciated. Thus, frozen section evaluation of the margins may be indicated to ensure complete excision. Gloster et al. found a recurrence rate of 6.6% in patients with DFSP treated with Mohs micrographic surgery, when compared to a 10% recurrence rate for patients treated with wide excision.17 In a review of the world literature, these authors found the average recurrence rate of DFSP following Mohs micrographic surgery to be 0.6%, when compared to an 18% recurrence rate following wide excision. Histologically, DFSP are characterized by a uniform population of plump spindled cells arranged in a monotonous storiform pattern. At low magnification, there is little heterogeneity in the lesion, as compared to that seen in BFH. Cytologically, the cells show little nuclear pleomorphism and typically there is low mitotic activity. Secondary elements, such as giant cells and xanthoma cells, which are commonly seen in BFH, are rarely found in DFSP. DFSP also infiltrate the underlying subcutaneous tissue in a manner that is not seen in BFH. In their study of 40 cases of DFSP, Kamino and Jacobson described infiltration of the underlying subcutaneous tissue in either a honeycomb or lacelike pattern (30% of the cases),or with a distinct multilayered pattern in which bundles of spindled cells were arranged parallel to the skin surface (60% of the cases).3 In contrast, of 185 cases of the fibrous variant of dermatofibroma with extension into the subcutaneous tissue, 28% of these lesions showed a smooth and well-demarcated deep margin with bulging into the subcutaneous tissue. However, 72% of the cases showed irregular extension into the subcutaneous tissue in a ray-like pattern, predominantly along pre-existing fibrous septa, giving the infiltrative margin a wedge-shaped appearance. Thus, thorough evaluation of the

deep margin is extremely useful in differentiating between BFH and DFSP. The distinction between BFH and DFSP may be particularly challenging when only a superficial portion of the lesion is present in a biopsy specimen, since the relationship with the subcutaneous tissue is not available for examination. In addition, when DFSP is in the plaque stage and does not significantly infiltrate the underlying subcutaneous tissue, these lesions may be even more difficult to differentiate from one another. The human hematopoietic progenitor cell antigen, CD34, is a 115kD monomeric transmembrane glycoprotein which is selectively expressed in hematopoietic progenitor cells in the bone marrow, as well as in some acute leukemias. In 1990, Nickoloff reported the presence of CD34-positive cells within the reticular and deep papillary dermis of the normal skin.18 Shortly thereafter, Cohen et al used a monoclonal antibody for CD34 to evaluate the expression of this antigen in a variety of benign and malignant vascular and spindled cell tumors.19 They found that 4 of 6 cases of DFSP stained positively with antibodies to CD34, but none of the 23 cases of fibrous-type dermatofibroma stained positively. Numerous other studies have shown DFSP to be almost uniformly CD34-positive. Factor XIIIa, or fibrin-stabilizing factor, is a tetrameric protein that plays an important role in the coagulation system. Cerio and colleagues found a population of dermal dendrocytes within the upper papillary dermis which stained with antibodies to factor XIIIa.20 These cells appear to be cytologically similar to the CD34-positive dermal dendrocytes found in the reticular and lower papillary dermis, but are immunophenotypically distinct. Abenoza et al found factor XIIIa immunopositivity in 40 of 40 cases of dermatofibroma. Conversely, 75% of the DFSP tested contained no or very few factor XIIIa positive tumor cells.21 In our experience, CD34 is an excellent marker of DFSP, and stains very few BFH. Similarly, the vast majority of DFSP do not show staining for factor XIIIa, except for the presence of scattered, presumably non-neoplastic dermal dendrocytes that are entrapped within the tumor. The vast majority of BFH tend to show diffuse and intense factor XIIIa immunoreactivity, with an accentuation of the peripherally located tumor cells.22 Some pathologists, however, view factor XIIIa as completely nonspecific and diagnostically worthless. Other immunohistochemical stains have been found to be useful in distinguishing BFH from other cutaneous spindle cell neoplasms. For example, Sachdev et al found CD163 expression to be consistently present in BFH but not in AFX.23 In contrast, however, a report from the University of Michigan found CD163 expression in 11 of 14 (79%) AFX, with CD163 immunoreactivity found in 24 of 29 BFH, including 8 of 8 cellular BFH and 6 of 9 epithelioid BFH.24 More recently, D2-40 was found to be a useful complementary immunostain to factor XIIIa and CD34 in cases in which distinction of BFH from DFSP was difficult.25 In this study,

all 56 cases of BFH demonstrated strong and diffuse immunoreactivity to D2-40, whereas none of the DFSP stained for this antigen.

Bednar Tumor Bednar tumor or pigmented DFSP accounts for fewer than 5% of all cases of DFSP, and present similarly and have a similar recurrence rate to DFSP.26 These tumors are characterized by scattered darkly pigmented, melanin-containing dendritic cells with long cytoplasmic processes. The number of melanin-containing cells varies greatly within these tumors. In some, the number of pigmented cells is so sparse that they can only be appreciated microscopically. The pigment has been confirmed to be melanin both by histochemical methods (Fontana stain) and ultrastructurally by the presence of mature membrane-bound melanosomes. These cells have numerous interlocking processes invested with basal laminae, suggestive of schwann cell differentiation; however, in the study by Dupree et al, S-100 protein was not found to be present within these cells.26 Due to the paucity of cases, it is difficult to assess the biologic behavior of this tumor. However, its strong similarity to non-pigmented DFSP suggests that the biologic behavior is comparable. The presence of melanin-bearing cells in a neoplasm with a monomorphic storiform pattern is virtually diagnostic of a Bednar tumor. Non-pigmented cases of DFSP may be more difficult to differentiate from other lesions including:

Benign Fibrous Histiocytoma. This lesion typically has a more heterogeneous appearance than DFSP, and often has a variable number of giant cells, xanthoma cells, inflammatory cells, and siderophages. Dense keloid-like collagen is often found at the periphery of the lesion. A grenz zone is often present, as well as reactive epidermal hyperplasia. Although BFH may infiltrate the subcutis at the base of the lesion, the pattern of infiltration characteristic of DFSP (as described above) is not present. Immunohistochemically, BFH is typically negative for CD34 and positive for factor XIIIa.

Atypical Fibroxanthoma. This lesion often has a storiform pattern similar to that of DFSP; however, the cells are typically more pleomorphic and hyperchromatic, and often show an increased number of mitotic figures, including atypical mitoses.

Neurofibroma (Diffuse Form). The diffuse form of neurofibroma may be difficult to differentiate from DFSP, particularly when it is in the plaque stage. The cells of neurofibroma have a distinct wavy appearance and are typically deposited in a loose myxoid background. Immunohistochemically, the majority of cells are S-100 positive.

Myxoid Sarcomas. The myxoid form of DFSP may be difficult to differentiate from other myxoid sarcomas, in particular myxoid liposarcoma and myxofibrosarcoma.27 However, myxoid DFSP has a more superficial location and lacks lipoblasts. It also lacks the nuclear pleomorphism characteristic of myxoid MFH.

Fibrosarcoma Arising In DFSP In a small percentage of cases of DFSP (10%), a portion of the tumor may resemble a fibrosarcoma (FS). Prior to 1997, fewer than 40 cases of DFSP-FS had been reported in the literature. In 1992, Connelly and Evans reported 6 cases of DFSP-FS, and found a similar rate of recurrence when compared to typical DFSP.28 However, 2 of the 6 patients in their series died of disease, one with metastasis. In their review of the literature, 6 of 28 patients (21%) developed metastases or fatal recurrence, a rate significantly higher than that seen in typical DFSP. Mentzel et al reported 41 cases of DFSP-FS,29 and they also found a significantly increased rate of metastasis and death in those patients with DFSP-FS, when compared to DFSP alone. However, only a few of the patients in this series were treated initially with wide local excision, as most of the patients were treated with either an incomplete excision or local excision. Thus, it is possible that more extensive initial surgery may decrease the risk of metastasis in those patients with DFSP-FS. In our experience of 18 patients with DFSP-FS treated with wide local excision and at least five years of clinical follow-up, none of the patients developed metastatic disease.30 More recently, Abbott and colleagues reported a metastatic rate of 10%, and fibrosarcomatous transformation was associated with gains of p53 mutations and increased proliferative activity.31 Histologically, a fibrosarcomatous component can be identified by its increased cellularity and arrangement into longer fascicles, often with a "herringbone" growth pattern, similar to that seen in deep soft tissue fibrosarcoma. In many cases, there is an abrupt transition between the less cellular monotonous storiform pattern of DFSP into the more cellular fascicular pattern of FS, often with the FS area in the deep portion of the tumor. In some cases, however, this transition from a storiform to a fascicular growth pattern is gradual, and may be difficult to recognize. The FS component has more cytologic atypia and mitotic activity when compared to the surrounding DFSP, and the FS component is often CD34 negative, or less intensely positive than the DFSP component.32

Giant Cell Fibroblastoma A lesion which has several features in common with and is certainly related to DFSP is giant cell fibroblastoma (GCF), first described in 1982 by Shmookler and Enzinger.33 Most

GCF occur in young patients (the majority of which are less than 5 years of age) with a male predominance, typically presenting as a painless, slowly enlarging subcutaneous mass. The anatomic distribution is similar to DFSP - that is, they occur most commonly on the trunk and less often on the extremities and head and neck.34,35 Review of the literature reveals that almost 50% of these lesions locally recur, although there have been no well-documented cases of metastasis (hence its designation under the heading of fibrohistiocytic tumor of intermediate malignancy). Histologically, these lesions typically involve the dermis and subcutis, and are characterized by hyperchromatic or stellate-shaped cells deposited in a collagenous to myxoid matrix. Scattered throughout the lesion are hyperchromatic multinucleated floret-like giant cells and characteristic "pseudovascular" or "angiectoid" spaces that are lined by the spindled and giant cells, as opposed to endothelial cells. In some areas, the lesion may closely resemble DFSP, with areas of increased cellularity and a more well-defined storiform pattern. Immunohistochemically, most cells express vimentin and CD34.35,36 There continues to be great controversy regarding the histogenesis of GCF, including theories of myofibrohistiocytic, fibrohistiocytic, fibroblastic and perineural origin. The possible relationship between GCF and DFSP was initially raised by Shmookler et al in 1989, as indicated by the title of their paper "Giant cell fibroblastoma: A juvenile form of dermatofibrosarcoma protuberans."34 Although GCF tends to occur in younger patients than DFSP, these two entities share many clinical characteristics, including a predilection for the trunk, and similar recurrence rates. In addition, there are many reports of tumors with histiologic features overlapping GCF and DFSP.35,37 Finally, both tumors characteristically express CD34, and both have been found to harbor an identical t(17;22) involving the COL1A1 and PDGFB genes,38 which can be detected by FISH or RT-PCR.39-41

Angiomatoid Fibrous Histiocytoma Previously termed "angiomatoid malignant fibrous histiocytoma," the World Health Organization Committee for the Classification of Soft Tissue Tumors has renamed this entity "angiomatoid fibrous histiocytoma" to reflect the rarity of metastasis and excellent prognosis.1 This tumor most commonly occurs in young patients, typically less than 20 years of age, and is rarely found in patients over the age of 40 years.42 The lesion most often occurs as a well-circumscribed nodule in the deep dermis or subcutis of the extremities. They may be painful, and occasionally patients manifest systemic symptoms such as fever, anemia, and weight loss, possibly due to the production of cytokines by the neoplastic cells. Histologically, angiomatoid FH is typically nodular or multinodular and is surrounded by a dense fibrous pseudocapsule. The most characteristic feature is the presence of

irregularly-shaped blood-filled spaces that are lined by tumor cells (not endothelium). The surrounding tumor cells may be spindled or have a round to ovoid shape, and resemble histiocytes, although their true nature remains unknown. These histiocyte-like cells have a faintly staining eosinophilic cytoplasm that often contains a dusky yellow-brown pigment, due to the phagocytosis of hemosiderin by tumor cells. Although most cases are characterized by a monomorphic population of these histiocyte-like cells, about 20% of cases show nuclear pleomorphism, usually as a focal feature.43 Mitoses are present, but are not numerous. The nodules of histiocyte-like cells are typically surrounded by a dense lymphoplasmacytic infiltrate that may have germinal centers, and may cause confusion because of the striking resemblance to a lymph node. The differential diagnosis for angiomatoid FH is relatively limited. Firstly, the presence of a peripherally located lymphoplasmacytic infiltrate with germinal centers may cause confusion with a metastasis to a lymph node. More commonly, this lesion can be mistaken for a benign fibrous histiocytoma, particularly the aneurysmal variant. As mentioned previously, aneurysmal BFH tends to occur in older patients (as opposed to younger patients with angiomatoid FH), is not associated with systemic symptoms, is typically located in the dermis (as opposed to deep dermis and subcutis), and most importantly, will show other areas of typical BFH. Angiomatoid fibrous histiocytoma does not have a characteristic immunophenotype, thus causing controversy regarding its histogenesis. In 1991, Smith et al evaluated the immunophenotype of 19 cases of angiomatoid FH and found almost 50% stained positively for CD68.44 As Smith et al suggest, it is likely that CD68 positivity represents the phenotypic expression of the acquisition of lysosomes and phagocytic activity, as indicated by the presence of hemosiderin in many of the tumor cells. Fletcher found desmin positivity in 5 of 6 cases, and suggested that this tumor may be of myoid origin.45 Fanburg-Smith and Miettinen evaluated the immunophenotype of a large number of cases of angiomatoid fibrous histiocytoma.46 Desmin positivity was noted in 51% of cases, most of which showed scattered similar desmin-positive cells in the surrounding lymphoid infiltrate adjacent to the tumor. Muscle-specific and smooth muscle actin were positive in 14% of cases. However, MyoD1 and myogenin were negative in all tumors studied. Forty-five percent of cases stained for CD99. Fifteen percent of cases were positive for CD68. All of the tumors were negative for CD21, CD35, S100 protein, CD34 and cytokeratins. Some cases may also stain for EMA.47 In Enzinger's original series (which he classified as angiomatoid MFH), 11 of 24 (46%) patients developed local recurrence, 5 patients (21%) developed metastatic disease, and 3 patients (13%) died of disease.48 However, in a much larger study of 108 patients by Costa et al, local recurrence developed in only 12% of patients, all of whom were cured by

re-excision.42 Only 5% of patients developed metastatic disease, and only one patient died of disease. These authors explain this difference in clinical behavior by the fact that their cases were diagnosed prospectively, and in most instances had been treated adequately with wide local excision. In the study by Fanburg-Smith and Miettinen,46 clinical follow-up on 86 patients indicated that only one patient was alive with a local nodal metastasis (1% frequency of metastasis) within one year, and two others had local recurrence, all over a mean follow-up period of six years. Recently, cytogenetic and molecular studies have demonstrated specific translocations in this tumor which, interestingly, are identical to those found in clear cell sarcoma. While some cases of angiomatoid FH do have a t(12;22) involving ATF1 gene on 12q13 and the FUS gene on 16p11, others have a t(12;22) involving an ATF1-EWSR1 fusion.49 However, several recent studies have found the most common gene fusion in angiomatoid FH to be a t(2;22) involving the CREB1 gene on 2q33 with the EWSR1 gene on 22q12.49,50 Interestingly, this same t(2;22) has been described in clear cell sarcomas involving the gastrointestinal tract. We now routinely FISH for translocation of the EWSR1 using paraffin-embedded tissues. In our recent study, 13 of 17 cases (76%) of AFH harbored rearrangements of this gene.51

References

1. World Health Organization Classification of Tumours: Tumours of Soft Tissue and Bone. Fletcher CDM, Unni KK, Mertens F (eds). IARC Press, Lyons 2002, 109-125. 2. Calonje E, Fletcher CDM. Cutaneous fibrohistiocytic tumors: An update. Adv Anat Pathol 1994; 1(1):2-15. 3. Kamino H, Jacobson M. Dermatofibroma extending into the subcutaneous tissue: Differential diagnosis from dermatofibrosarcoma protuberans. Am J Surg Pathol 1990;14:1156-64. 4. Calonje E, Mentzel T, Fletcher CDM. Cellular benign fibrous histiocytoma: Clinicopathologic analysis of 74 cases of a distinctive variant of cutaneous fibrous histiocytoma with frequent recurrence. Am J Surg Pathol 1994; 18:668-76. 5. Kaddu S, McMenamin ME, Fletcher CD. Atypical fibrous histiocytoma of the skin: clinicopathologic analysis of 59 cases with evidence of infrequent metastasis. Am J Surg Pathol 2002;26:35-46. 6. Tamada S, Ackerman AB. Dermatofibroma with monster cells. Am J Dermatopathol 1987; 9:380-87. 7. Fukamizu H, Oku T, Inoue K, Matsumoto K, Okayama H, Tagami H. Atypical ("pseudosarcomatous") cutaneous histiocytoma. J Cutan Pathol 1983; 10:327-33. 8. Beham A, Fletcher CDM. Atypical "pseudosarcomatous" variant of cutaneous benign fibrous histiocytoma: Report of eight cases. Histopathology 1990; 17:165-82. 9. Fretzin DF, Helwig ED. Atypical fibroxanthoma of the skin: A clinicopathologic study of 140 cases. Cancer 1973; 39:1541-52. 10. Dahl I. Atypical fibroxanthoma of the skin. A clinicopathological study of 57 cases. APMIS 1976; 3:183-97. 11. Santa Cruz DJ, Kyriakos M. Aneurysmal ("angiomatoid") fibrous histiocytoma of the skin. Cancer 1981; 47:2053-61. 12. Calonje E, Fletcher CD. Aneurysmal benign fibrous histiocytoma: clinicopathological analysis of 40 cases of a tumor frequently misdiagnosed as a vascular neoplasm. Histopathology 1995;26:323-331. 13. Wilson Jones E, Cerio R, Smith NP. Epithelioid cell histiocytoma: A new entity. Br J Dermatol 1989; 120:185-95. 14. Singh Gomez C, Calonje E, Fletcher CD. Epithelioid benign fibrous histiocytoma of the skin: clinicopathological analysis of 20 cases of a poorly known variant. Histopathology 1994;24:123-129. 15. Taylor HB, Helwig EB. Dermatofibrosarcoma protuberans. A study of 115 cases. Cancer 1962; 15:717-25. 16. Roses DF, Valensi Q, LaTrenta G, Harris MN. Surgical treatment of dermatofibrosarcoma protuberans. Surg Gynecol Obstet 1986; 162:449-52. 17. Gloster HM Jr, Harris KR, Roenigk RK. A comparison between Mohs micrographic surgery and wide surgical excision for the treatment of dermatofibrosarcoma protuberans. J Am Acad Dermatol 1996;35:82-87. 18. Nickoloff BJ. The human progenitor cell antigen (CD34) is localized on endothelial cells, dermal dendritic cells, and perifollicular cells in formalin-fixed normal skin, and on proliferating endothelial cells and stromal spindle-shaped cells in Kaposi's sarcoma. Arch Dermatol 1991; 127:523-29. 19. Cohen PR, Rapini RP, Farhood AI. Expression of a human hematopoietic progenitor cell antigen CD34 in vascular and spindled cell tumors. J Cutan Pathol 1993; 20:15-20. 20. Cerio R, Spaull J, Wilson Jones E. Identification of factor XIIIa in cutaneous tissue. Histopathology 1988; 13:362-64. 21. Abenoza P, Lillemoe T. CD34 and factor XIIIa in the differential diagnosis of dermatofibroma and dermatofibrosarcoma protuberans. Am J Dermatopathol 1993; 15(5):429-34. 22. Goldblum JR, Tuthill RJ. CD34 and factor XIIIa immunoreactivity in dermatofibrosarcoma protuberans and dermatofibroma. Am J Dermatopathol 1997;19:147-153.

23. Sachdev R, Robbins J, Kohler S, et al. CD163 expression is present in cutaneous histiocytomas but not in atypical fibroxanthomas. Am J Clin Pathol 2010;133:915-921. 24. Pouryazdanparast P, Yu L, Cutlan JE, et al. Diagnostic value of CD163 in cutaneous spindle cell lesions. J Cutan Pathol 2009;36:859-864. 25. Bandarchi B, Ma L, Marginean C, et al. D2-40, a novel immunohistochemical marker in differentiating dermatofibroma from dermatofibrosarcoma protuberans. Mod Pathol 2010;23:434-438. 26. Dupree WB, Langloss JM, Weiss SW. Pigmented dermatofibrosarcoma protuberans (Bednar tumor): A pathologic, ultrastructural, and immunohistochemical study. Am J Surg Pathol 1985; 9:630-39. 27. Reimann JD, Fletcher CD. Myxoid dermatofibrosarcoma protuberans: a rare variant analyzed in a series of 23 cases. Am J Surg Pathol 2007;31:1371-1377. 28. Connelly JH, Evans HL. Dermatofibrosarcoma protuberans: A clinicopathologic review with emphasis on fibrosarcomatous areas. Am J Surg Pathol 1992; 16(10):921-25. 29. Mentzel T, Beham A, Katemkamp D, Dei Tos AP, Fletcher CDM. Fibrosarcomatous variant of dermatofibrosarcoma protuberans: clinicopathologic and immunohistochemical study of a series of 41 cases with emphasis on prognostic significance. Am J Surg Pathol 1998;22:576-587. 30. Goldblum JR, Reith JD, Weiss SW. Sarcomas arising in dermatofibrosarcoma protuberans: a reappraisal of biologic behavior in 18 cases treated by wide local excision with extended clinical follow-up. Am J Surg Pathol 2000;24:1125-30. 31. Abbott JJ, Oliveira AM, Nascimento AG. The prognostic significance of fibrosarcomatous transformation in dermatofibrosarcoma protuberans. Am J Surg Pathol 2006;30:436-443. 32. Goldblum JR. CD-34 positivity in fibrosarcomas which arise in dermatofibrosarcoma protuberans. Arch Pathol Lab Med 1995;119:238-241. 33. Shmookler BM, Enzinger FM. Giant cell fibroblastoma: A peculiar childhood tumor. Lab Invest 1982; 46:76A. 34. Shmookler BM, Enzinger FM, Weiss SW. Giant cell fibroblastoma. A juvenile form of dermatofibrosarcoma protuberans. Cancer 1989; 64:2154-61. 35. Jha P, Moosavi C, Fanburg-Smith JC. Giant cell fibroblastoma: an update and addition of 86 new cases from the Armed Forces Institute of Pathology, in honor of Dr. Franz M. Enzinger. Ann Diagn Pathol 2007;11:81-88. 36. Goldblum JR. Giant cell fibroblastoma: A report of three cases with histologic and immunohistochemical evidence of a relationship to dermatofibrosarcoma protuberans. Arch Pathol Lab Med 1996;120:1052-55. 37. Zamecnik M, Michal M. Giant cell fibroblastoma with pigmented dermatofibrosarcoma protuberans component. Am J Surg Pathol 1994; 18(7):736-40. 38. O’Brien KP, Seroussi E, Dal Cin P, et al. Various regions within the alpha-helical domain of the COL1A1 gene are fused to the second axon of the PDGFB gene in dermatofibrosarcomas and giant-cell fibroblastomas. Genes Chromosomes Cancer 1998;23:187-93. 39. Takahira T, Oda Y, Tamiya S, et al. Detection of COL1A1-PDGFB fusion transcriptions and PDGFB/PDGFRBmRNA expression in dermatofibrosarcoma protuberans. Mod Pathol 2007;20:668-675. 40. Llombart B, Sanmartin O, Lopez-Guerrero JA, et al. Dermatofibrosarcoma protuberans: clinical, pathological, and genetic (COL1A1-PDGFB) study with therapeutic implications. Histopathology 2009;54:860-872. 41. Patel KU, Szabo SS, Hernandez VS, et al. Dermatofibrosarcoma protuberans COL1A1- PDGFB fusion is identified in virtually all dermatofibrosarcoma protuberans cases when investigated by newly developed multiplex reverse transcription polymerase chain reaction and fluorescence in situ hybridization assays. Hum Pathol 2008;39:184-193. 42. Costa MJ, Weiss SW. Angiomatoid malignant fibrous histiocytoma: A follow-up study of 108 cases with evaluation of possible histologic predictors of outcome. Am J Surg Pathol 1990; 14:1126-32.

43. Weinreb I, Rubin BP, Goldblum JR. Pleomorphic angiomatoid fibrous histiocytoma: a case confirmed by fluorescence in situ hybridization analysis for EWSR1 rearrangement. J Cutan Pathol 2008;35:855-860. 44. Smith ME, Costa MJ, Weiss SW. Evaluation of CD68 and other histiocytic antigens in angiomatoid malignant fibrous histiocytoma. Am J Surg Pathol 1991; 15:757-63. 45. Fletcher CDM. Angiomatoid "malignant fibrous histiocytoma": An immunohistochemical study indicative of myoid differentiation. Hum Pathol 1991; 22:563-68. 46. Fanburg-Smith JC, Miettinen M. Angiomatoid "malignant" fibrous histiocytoma: a clinicopathologic study of 158 cases and further exploration of the myoid phenotype. Hum Pathol 1999;30:1336-43. 47. Thway K. Angiomatoid fibrous histiocytoma: a review with recent genetic findings. Arch Pathol Lab Med 2008;132:273-277. 48. Enzinger FM. Angiomatoid malignant fibrous histiocytoma. A distinctive fibrohistiocytic tumor of children and young adults simulating a vascular neoplasm. Cancer 1979; 44:2147-57. 49. Antonescu CR, Dal Cin P, Nafa K, et al. EWSR1-CREB1 is the predominant gene fusion in angiomatoid fibrous histiocytoma. Genes Chromosomes Cancer 2007;46:1051-1060. 50. Rossi S, Szuhai K, Ijszenga M, et al. EWSR1-CREB1 and EWSR1-ATF1 fusion genes in angiomatoid fibrous histiocytoma. Clin Cancer Res 2007;13:7322-7328. 51. Tanas MR, Rubin BP, Montgomery EA, et al. Utility of FISH in the diagnosis of angiomatoid fibrous histiocytoma: a series of 18 cases. Mod Pathol 2010;23:93-97.

Cutaneous Vascular Tumors

Steven D. Billings, MD

Co-Section Head, Section of Dermatopathology Anatomic Pathology, Cleveland Clinic OH, USA Associate Professor in the Cleveland Clinic Lerner College of Medicine [email protected]

- V -

In this lecture, I will cover a selected group of cutaneous vascular tumors that poses diagnostic difficulty with an emphasis on recently described entities and their differential diagnosis.

Microvenular hemangioma Although not a new entity, Microvenular hemangioma is a significant source of diagnostic difficulty because of its infiltrative appearance and the fact that is frequently not discussed in surgical pathology textbooks. It typically occurs on the extremities of young adults. Microscopically, this variant causes significant diagnostic difficulty because of its infiltrative growth pattern as opposed to a circumscribed pattern seen in most benign vascular tumors. The tumor is composed of compressed slit-like or small ectatic vessels. Although the vessels may have an infiltrative appearance, they are well-formed with a distinct pericyte layer.

The differential diagnosis of microvenular hemangioma is predominantly Kaposi sarcoma. Kaposi sarcoma occurs in immunosuppressed patients. Although both may have an infiltrative growth pattern Kaposi sarcoma has much more irregular complex vessels lacking a pericyte layer. Microvenular hemangioma does not show the dissecting promontory sign seen in Kaposi sarcoma and lacks the solid spindled cells that are often seen in Kaposi sarcoma. Immunohistochemical stains for HHV8 latent nuclear antigen can be helpful in this differential diagnosis as microvenular hemangioma is negative for this marker. Immunostains for SMA can highlight the pericyte layer seen in microvenular hemangioma.

Selected References

1. Trindade F, Kutzner H, Requena L, Tellechea Ó, Colmenero I. Microvenular hemangioma-an immunohistochemical study of 9 cases. Am J Dermatopathol. 2012 Dec;34(8):810-2. 2. Hunt SJ, Santa Cruz DJ, Barr RJ. Microvenular hemangioma. J Cutan Pathol. 1991 Aug;18(4):235-40. PubMed PMID: 1939781. 3. Aloi F, Tomasini C, Pippione M. Microvenular hemangioma. Am J Dermatopathol. 1993 Dec;15(6):534-8.

Cutaneous epithelioid angiomatous nodule This is a relatively new entity first described in 2004. Clinically it usually presents in middle-aged adults may present in a broad age range with cases presenting in childhood and elderly patients. The majority of cases are small (<1.0 cm) solitary lesions but occasional multiple or eruptive forms have been described. They are usually erythematous to violaceous. The location is varied, most have been described on the trunk but a significant subset has also been described on the extremities and the head and neck area. Microscopically, cutaneous epithelioid angiomatous nodule is a circumscribed tumor. Recognizable vascular channel formation is usually focal in nature. They are largely solid and composed of epithelial endothelial cells with abundant eosinophilic cytoplasm. Intracytoplasmic vacuoles are usually apparent. The nuclei do not show significant pleomorphism but they typically are somewhat enlarged with a conspicuous nucleolus. These are mitotically active tumors with a mitotic rate up to 5

mitotic figures per 10 high power fields, but atypical forms are not seen. There is a variable inflammatory infiltrate usually composed of lymphocytes and histiocytes associated with cutaneous epithelioid angiomatous nodule. Occasional cases may show a significant eosinophil component. Cutaneous epithelioid angiomatous nodule is a benign tumor and is unrelated to underlying immunosuppression. The treatment is variable. Cases have responded to simple excision as well as topical steroids. There have been no reported cases of local recurrence or metastasis. The differential diagnosis of cutaneous epithelioid angiomatous nodule includes epithelial hemangioma (angiolymphoid hyperplasia with eosinophilia), and it has been suggested that cutaneous epithelioid angiomatous nodule may be related to epithelioid hemangioma. Epithelioid hemangioma usually presents in the head and neck area of young adults as solitary or locally multiple lesions. Microscopically, epithelioid hemangioma has more well-formed capillaries lined by epithelioid endothelial cells and is associated with lymphoid aggregates and numerous eosinophils. There is frequently a larger damaged vessel towards the central portion of epithelioid hemangioma suggesting that this may be a reactive process. Because of the epithelioid nature of the endothelial cells, epithelioid hemangioendothelioma can also be considered in the differential diagnosis. Epithelioid hemangioendothelioma is composed of cords to nests of epithelioid endothelial cells and does not usually have a solid sheet-like growth pattern of cutaneous epithelioid angiomatous nodule. Epithelioid sarcoma-like hemangioendothelioma could also be considered in the differential diagnosis. This is an unusual vascular tumor that closely mimics epithelioid sarcoma. It is less circumscribed that epithelioid angiomatous nodule and is negative for CD34.

Selected References

1.Sangueza OP. Walsh SN. Sheehan DJ. Orland AF. Llombart B. Requena L. Cutaneous epithelioid angiomatous nodule: a case series and proposed classification. American Journal of Dermatopathology. 30:16-20, 2008. 2.Brenn T. Fletcher CD. Cutaneous epithelioid angiomatous nodule: a distinct lesion in the morphologic spectrum of epithelioid vascular tumors. American Journal of Dermatopathology. 26:14-21, 2004.

Epithelioid hemangioendothelioma Epithelioid hemangioendothelioma typically occurs within adult patients. In the skin it has a nondescript appearance. Most present as skin-colored nodules and are not clinically thought to be a vascular tumor.

Approximately half are associated with pre-existing vessel and can be seen growing within and expanding the affected vessel. In these cases the tumor cells can be seen radiating out from the associated vessel. Intravascular growth is less common in cutaneous epithelioid hemangioendothelioma. The most typical growth pattern of epithelial hemangiomas is cords or nests of epithelioid endothelial cells. They typically not composed of well-formed vessels. However, the endothelial cells frequently show individual intracytoplasmic vacuoles (so-called "blister cells") that may contain red blood cells. Generally the nuclear features are relatively bland but

significant cytologic atypia may be seen in up to 25% of cases. The tumor cells are typically embedded within a myxohyaline/chondroid stromal matrix. They are immunoreactive for vascular marker CD31, CD34, and ERG. Rare cases may lose expression of CD34 or CD31. Importantly approximately 25% of cases will also show immunoreactivity for cytokeratin, but it is typically less intense than the immunoreactivity seen for the vascular markers.

There has been recent insight into the pathogenesis of this entity. It is now known that this tumor has a consistent recurring t(1;3) involving WWTR1 and CAMTA1. The first of these is a transcription factor highly expressed in endothelial cells, while the latter is a transcription regulatory protein usually expressed in the brain. How this fusion protein relates to the pathogenesis of this tumor is an area of active research, much of it being done by our colleague at The Cleveland Clinic, Brian Rubin. FISH to detect the t(1;3) can be a useful diagnostic tool in difficult cases.

Epithelial hemangioendothelioma recurs in about 10-15% of cases and metastasizes to regional lymph nodes or the lungs in 20-30% of cases. The overall mortality is about 10-20%.

The differential diagnosis of epithelioid hemangioendothelioma includes epithelioid sarcoma and epithelioid sarcoma-like hemangioendothelioma. However epithelial sarcoma typically has a more solid nodular proliferation of epithelioid cells often with central necrosis reminiscent of granulomas. They may also show more sheet-like proliferation of epithelioid cells. Epithelioid sarcoma shows strong keratin and epithelial membrane antigen immunoreactivity. Approximately 50% of epithelioid sarcomas may show immunoreactivity for CD34, but importantly epithelioid sarcoma is negative for CD31 and ERG.

Selected References

1. Weiss SW, Enzinger FM. Epithelioid hemangioendothelioma: a vascular tumor often mistaken for a carcinoma. Cancer. 1982 Sep 1;50(5):970-81. 2. Quante M. Patel NK. Hill S. Merchant W. Courtauld E. Newman P. McKee PH. Epithelioid hemangioendothelioma presenting in the skin: a clinicopathologic study of eight cases. American Journal of Dermatopathology. 20:541-6, 1998 Dec. 3. Requena L, Kutzner H. Hemangioendothelioma. Semin Diagn Pathol. 2013 Feb;30(1):29-44. 4. Tanas MR, Sboner A, Oliveira AM, Erickson-Johnson MR, Hespelt J, Hanwright PJ, Flanagan J, Luo Y, Fenwick K, Natrajan R, Mitsopoulos C, Zvelebil M, Hoch BL, Weiss SW, Debiec-Rychter M, Sciot R, West RB, Lazar AJ, Ashworth A, Reis-Filho JS, Lord CJ, Gerstein MB, Rubin MA, Rubin BP. Identification of a disease-defining gene fusion in epithelioid hemangioendothelioma. Sci Transl Med. 2011 Aug 31;3(98):98ra82.

Epithelioid sarcoma-like hemangioendothelioma

Epithelioid sarcoma-like hemangioendothelioma is a rare vascular tumor of apparent intermediate malignancy. It typically occurs on the extremities of adults.

Microscopically it is composed of sheets and nodules of eosinophilic epithelioid to spindled endothelial cells. Usually vascular differentiation is inapparent. There are no well-formed vessels and only rare intracytoplasmic lumens. It is strongly keratin positive causing the diagnosis consideration of epithelioid sarcoma. However, it is immunoreactive for vascular markers. In the original series all cases were positive for cytokeratin and Fli-1 and 5/6 were positive for CD31. In a more recent series described under a different name (pseudomyogenic hemangioendothelioma), all were positive for cytokeratin and Fli-1, but somewhat less than 50% were positive for CD31. In our experience immunoreactivity for CD31 is more common than that. The tumors also show immunoreactivity for ERG, a nuclear stain that is positive in vascular tumors. Importantly, all cases reported to date have been negative for CD34. The tumor has retained INI-1 expression.

The behavior of epithelioid sarcoma-like hemangioendothelioma appears to be that of a vascular tumor of intermediate malignancy, with a risk of local recurrence but no risk of distant metastasis. In the original description of this tumor, a case with presumed regional soft tissue metastasis was reported. The metastatic potential of this entity has recently be questioned. In a recent large series, it was proposed that this tumor is frequently multicentric. It is possible that the tumor with presumed soft tissue metastasis reported in the original series represented multicentric growth rather than metastasis. There is relatively little experience with this entity and the risk of lymph node or pulmonary metastasis may be uncovered over time.

The primary differential diagnosis in this case is epithelioid sarcoma. This differential is quite difficult. Epithelioid sarcoma tends to be more smaller nodules rather than a sheet-like growth. Epithelioid sarcoma usually has more nuclear atypia. By immunohistochemistry epithelioid sarcoma is positive for cytokeratin, positive for CD34 in 50-60% of cases, and is negative for CD31, Fli-1, and ERG. Epithelioid sarcoma demonstrates loss of INI-1.

Selected References

1. Billings SD, Folpe AL, Weiss SW. Epithelioid sarcoma-like hemangioendothelioma. Am J Surg Pathol. 2003 Jan;27(1):48-57. 2. Hornick JL, Fletcher CD. Pseudomyogenic hemangioendothelioma: a distinctive, often multicentric tumor with indolent behavior. Am J Surg Pathol. 2011 Feb;35(2):190-201.

Post-radiation cutaneous vascular neoplasms With the emphasis on the use of breast conserving surgery with adjunctive radiation for the treatment of early stage breast cancer, there has been an increase in the incidence of postradiation vascular tumors. Postradiation vascular tumors are an increasingly common diagnostic dilemma for pathologists and dermatopathologists. The two tumors that fall into this category are postradiation cutaneous angiosarcoma, malignant vascular neoplasms with significant morbidity and mortality, and atypical vascular lesions (AVL), vascular tumors that typically behave in a benign manner. Post-radiation cutaneous angiosarcoma

Clinical features Post-radiation cutaneous angiosarcoma usually presents as a large erythematous to violaceous cutaneous patch, plaque or nodule on the affected breast. Rare cases present as small papules or ill-defined erythema. Unlike other forms of post-radiation sarcoma, including post-radiation angiosarcoma, angiosarcoma in this clinical setting typically has a short latency. The average latency from completion of radiation to the development of angiosarcoma is approximately 5-6 years. Importantly, cases can present as early as one-year after radiation therapy.

Microscopic features Post-radiation angiosarcoma, like other forms of angiosarcoma can have a variety of patterns. They can be composed of well differentiated tumors consisting of complex interanastomosing vessels lined by hyperchromatic endothelial cells. Some tumors are more cellular, often with a sieve-like pattern of back to back neoplastic vessels. Some cases may have a spindled pattern with fascicles of spindled tumor cells. High grade tumors may have solid sheets of endothelial cells with vascular lumens apparent only on high power examination. Usually more solid tumors have some evidence of vasoformative channels at the periphery of the neoplasm. Nuclear atypia and pleomorphism tends to increase with increased cellularity.

Occasional tumors have deceptive growth patterns. Some tumors have a lobular growth pattern somewhat reminiscent of capillary hemangiomas. This pattern has been termed the capillary nodule pattern by Rosai and colleagues. Tumors with this pattern still demonstrate diffuse involvement of the dermis and more nuclear atypia than a conventional cutaneous hemangioma. Another pattern is what we have termed the “radiation dermatitis” pattern. In this pattern the tumor demonstrates significant nuclear atypia but architectural complexity is either inapparent or absent. At low power this pattern may be mistaken for radiation dermatitis. Immunostains for CD31 can unveil the more complex vessels than are appreciated on routine H&E stained sections and highlight the nature of the tumor cells.

Treatment and prognosis Mastectomy is the mainstay of treatment. There is no defined role for adjuvant therapy at this time. Post-radiation angiosarcoma like other angiosarcomas is a potentially aggressive tumor. There is a high rate of local recurrence (~50%) even with histologically negative margins. Post-radiation angiosarcomas frequently develop satellite nodules contributing to the increased risk of local recurrence. The reported metastatic rate varies from 17-40%. As expected the most common site of metastatic disease is the lung, followed, interestingly, by the contralateral breast. The overall mortality rate approaches 40% in some series. Brenn and Fletcher suggested that well-differentiated tumors may be less aggressive, but this was based on limited follow-up. Other studies have had opposite conclusions. More studies with longer follow-up are needed to determine if histologic grade impacts prognosis. As a general rule grade is independent of behavior in angiosarcoma, and this tumor is not formally graded by the FNCLCC grading scheme.

Differential Diagnosis The primary differential diagnosis is atypical vascular lesion (AVL). This will be discussed in more detail in the next section. Epithelioid angiosarcoma can sometimes be confused with recurrent breast cancer. Recognition of vasoformative channels and expression of vascular markers allows distinction. Other forms of post-radiation sarcoma could be considered. Usually, other forms of post-radiation sarcoma is a pleomorphic sarcoma (so-called malignant fibrous histiocytoma). Pleomorphic sarcoma has greater pleomorphism, often a storiform growth pattern and lacks expression of vascular markers.

Atypical Vascular Lesion (AVL) AVL was first described by Fineberg and Rosen in 1994 with the term atypical vascular lesions of the skin after radiation therapy. The essentially same entity has also been described in the literature as benign lymphangiomatous papules, acquired (progressive) lymphangiomas, acquired lymphangiectasia, and lymphangioma circumscriptum. The term atypical vascular lesion (AVL) is the more widely accepted term currently. AVL lesion has been defined as a vascular proliferation that develops within the radiation field relatively shortly after completion of radiotherapy and follows a benign clinical course. Although the incidence of AVL is not well- documented, they appear less common than postradiation cutaneous angiosarcomas.

AVLs present in women with a median age of 57 years (range 38-73 years), with a postradiation latency period on average of 3.5 years (range 1-17 years), after a similar dose of radiation (40-50 Gy) as seen in postradiation cutaneous angiosarcomas. AVLs have only been described in women, most commonly in the setting of breast conserving surgery and adjuvant radiation; however, they can occur in any postradiation field regardless of site. They typically presents as small solitary or multiple (0.1-6 cm) papules or nodules within the field of radiation. Lesions can present synchronously or metachronously.

Microscopic Features Microscopically, AVLs are relatively well circumscribed symmetrical vaguely wedge- shaped proliferation of well-formed vascular structures usually confined to the papillary and superficial to mid reticular dermis. In rare cases, the lesion may extend to the subcutis. The superficial portion of AVL typically shows more dilated vascular spaces than the deeper aspect. In the deeper aspects the vessels are compressed and can show a dissecting pattern between collagen bundles. The vessels are lined by a single layer of endothelial cells. Some cases have papillary projections formed by the underlying stroma, but multilayering of endothelial cells or mitotic figures are not present.

AVLs have been divided into two histologic patterns: 1) superficial lymphangioma- like (benign lymphangioma circumscriptum / lymphangioma) and 2) hobnail hemangioma-like / benign lymphangioendothelioma-like. The former histologically mimics a lymphangioma circumscriptum with large ecstatic vessels, while the latter presents as well formed small slit-like vessels lined by hobnailed nuclei. Patton et al, divided the two patterns into lymphatic and vascular types, with the former accounting for approximately two thirds of the cases of AVL. The vascular type of AVL were described as a circumscribed proliferation of ecstatic thin-walled vessels confined to the dermis and lined by hobnailed endothelial cells.

Both patterns tend to lack the multi-layering seen in postradiation cutaneous angiosarcoma, while showing only mild, if any, atypia in the form of hyperchromasia. Rarely, cases can develop significant cytologic atypia akin to angiosarcoma, yet still retain the architectural morphology of AVL with a pericyte layer investing the neoplastic vessels, as demonstrated by immunostains for SMA. These may represent transitional forms that are evolving into angiosarcoma.

Treatment and prognosis If possible AVLs should be completely excised. This is primarily a practical issue. In order for a case to be comfortably diagnosed, the entire lesion needs to be visualized.

By definition AVLs behave in a benign fashion. Rare cases recur locally, but most recurrences are actually considered new lesions in the same radiation field. Although AVL is considered a benign tumor, there are documented cases of AVL that have progressed to angiosarcoma. There are also cases of angiosarcoma that were initially diagnosed as AVL based on limited sampling. Areas of well differentiated angiosarcoma can be indistinguishable from AVL. For this reason, complete excision is the treatment of choice for AVL.

Differential diagnosis Differentiating AVL from angiosarcoma can be difficult as alluded to above. Key features are the circumscribed nature of AVL, lack of atypia, lack of multilayering, and absence of mitotic activity. A summary of histologic features comparing AVL and post-radiation angiosarcoma is provided below.

In selected cases ancillary studies may be helpful. Angiosarcoma usually lacks a pericyte layer and the neoplastic vessels are not highlighted by SMA stains. Nuclear labeling for Ki-67 has been suggested as an adjunct to the diagnosis, but in my experience this is less reliable. Recently, it has been reported that a significant proportion of post-radiation and lymphedema-associated angiosarcomas have amplification of c-MYC, ranging from ~50-100% whereas AVL are consistently negative for c-MYC amplification. In our experience at the Cleveland Clinic, FISH c- MYC amplification is consistently seen in post-radiation angiosarcomas. Of the cases we have tested, all have demonstrated c-MYC amplification. The presence of c-MYC amplification also corresponds to nuclear expression as demonstrated by immunohistochemistry. Immunostains for c-MYC are an effective surrogate marker for gene amplification, with post-radiation angiosarcomas consistently positive and AVLs consistently negative.

Amplification of c-MYC appears to be an obligate molecular event for the great majority of post-radiation angiosarcomas. Likely in the cases arising in AVLs, there is some second genetic “hit “ resulting in gene amplification that leads to the development of angiosarcoma in these cases.

Table: Comparison of AVL and post-radiation angiosarcoma Feature AVL AS Infiltration into subcutis - +++ Papillary endothelial hyperplasia - +++ Prominent nucleoli - +++ Mitotic figures - +++ Significant cytologic atypia - +++ “Blood lakes” - ++ Dissection of dermal collagen +/- +++ Anastomotic vessels ++ +++ Hyperchromatic endothelial cells +++ ++ Chronic inflammation* +++ + Relative circumscription +++ - Projection of stroma into lumen +++ -

Selected References:

1. Billings SD, McKenney JK, Folpe AL, et al: Cutaneous angiosarcoma following breastconserving surgery and radiation: an analysis of 27 cases. Am J Surg Pathol 28:781-788,2004 2. Brenn T, Fletcher CD: Radiation-associated cutaneous atypical vascular lesions and angiosarcoma: clinicopathologic analysis of 42 cases. Am J Surg Pathol 29:983-996, 2005 3. Di Tommaso L, Rosai J: The capillary lobule: a deceptively benign feature of postradiation angiosarcoma of the skin: report of three cases. Am J Dermatopathol 27:301-305, 2005 4. Diaz-Cascajo C, Borghi S, Weyers W, et al: Benign lymphangiomatous papules of the skin following radiotherapy: a report of five new cases and review of the literature. Histopathology 35:319-327, 1999 5. Fineberg S, Rosen PP: Cutaneous angiosarcoma and atypical vascular lesions of the skin and breast after radiation therapy for breast carcinoma. Am J Clin Pathol 102:757-763, 1994 6. Gengler C, Coindre JM, Leroux A, et al: Vascular proliferations of the skin after radiation therapy for breast cancer: clinicopathologic analysis of a series in favor of a benign process: a study from the French Sarcoma Group. Cancer 109:1584-1598, 2007 7. Mattoch IW, Robbins JB, Kempson RL, et al: Post-radiotherapy vascular proliferations in mammary skin: a clinicopathologic study of 11 cases. J Am Acad Dermatol 57:126-133, 2007 8. Patton KT, Deyrup AT, Weiss SW: Atypical vascular lesions after surgery and radiation of the breast: a clinicopathologic study of 32 cases analyzing histologic heterogeneity and association with angiosarcoma. Am J Surg Pathol 32:943-950, 2008 9. Requena L, Kutzner H, Mentzel T, et al: Benign vascular proliferations in irradiated skin. Am J Surg Pathol 26:328-337, 2002 10. Weaver J, Billings SD. Post-radiation cutaneous vascular tumors of the breast: A review. Seminars in Diagnostic Pathology.26:141-149, 2009. 11. Manner J, Radlwimmer B, Hohenberger P, et al. MYC high level gene amplification is a distinctive feature of angiosarcomas after irradiation or chronic lymphedema. Am J Pathol. 176:34-9, 2010. 12. Guo T, Zhang L, Chang NE, et al. Consistent MYC and FLT4 gene amplification in radiation-induced angiosarcoma but not in other radiation-

associated atypical vascular lesions. Genes Chromosomes Cancer. 50:25-33, 2011. 13. Fernandez AP, Sun Y, Tubbs RR, Goldblum JR, Billings SD. FISH for MYC amplification and anti-MYC immunohistochemistry: useful diagnostic tools in the assessment of secondary angiosarcoma and atypical vascular proliferations. J Cutan Pathol. 2012 Feb;39(2):234-42. 14. Mentzel T, Schildhaus HU, Palmedo G, Büttner R, Kutzner H. Postradiation cutaneous angiosarcoma after treatment of breast carcinoma is characterized by MYC amplification in contrast to atypical vascular lesions after radiotherapy and control cases: clinicopathological, immunohistochemical and molecular analysis of 66 cases. Mod Pathol. 2012 Jan;25(1):75-85.

IBD and IBD-Related Dysplasia

John R. Goldblum, MD

Chairman, Department of Anatomic Pathology Cleveland Clinic Professor of Pathology Cleveland Clinic Lerner College of Medicine [email protected]

- VI -

Patterns of Injury in Colorectal Biopsy Specimens

Pathologists are asked to evaluate colorectal biopsy specimens for a variety of reasons, but one of the more common reasons is to identify a pattern of injury with the hope that a specific diagnosis can be rendered. There are some forms of colitis that have specific histologic features, including lymphocytic colitis, collagenous colitis, ischemic colitis (in many cases) and the mucosal prolapse syndromes. In addition, there are a host of features that are fairly specific for chronic inflammatory bowel disease, although it is often difficult or impossible to distinguish between ulcerative colitis and Crohn's disease on the basis of colorectal biopsy specimens. The focus of this discussion will be on the evaluation of colorectal biopsy specimens in terms of patterns of injury and the role of the pathologist in arriving at a specific diagnosis. In my opinion, the most important aspect of evaluating colorectal biopsy specimens is having a keen sense of what is normal. Very often, it is difficult to decide whether a biopsy specimen is at the "upper limit of normal" or at the "lower limit of abnormal," and in many respects, this is a subjective evaluation. However, the most common diagnosis is "normal colon," and very often, this is exactly what the gastroenterologist is looking for. Thus, the pathologist should not be afraid to make a diagnosis of "normal colon" and terms such as "nonspecific chronic inflammation," "nonspecific colitis" and "acute and chronic inflammation" are inappropriate pathologic diagnoses that will render the treating gastroenterologist helpless. We review many biopsy specimens from other hospitals for patients referred to our gastroenterologists or colorectal surgeons for the evaluation of chronic inflammatory bowel disease based upon some of the above diagnoses, when in fact a referral would not have been necessary had the pathologist called these biopsy specimens "normal colon." There are important caveats to be aware of when evaluating colorectal biopsy specimens in terms on what constitutes "normal colon for site." For example, plasma cells are always present in the lamina propria of the colonic mucosa, regardless of where the biopsy specimen is taken. However, the lamina propria cellularity varies among different sites in the colon. In general, the cecum/right colon is quite a bit more cellular than other portions of the colon with a progressive decrease in the lamina propria cellularity as one moves from the right to left colon.1 If one were to compare a rectal and cecal biopsy side by side, one would have a keen appreciation of just how less cellular the rectum is when compared to the cecum. The crypts throughout the colon are generally straight, evenly distributed and come down to touch the superficial aspect of the muscularis mucosae. Even when a colorectal biopsy specimen is cut tangentially, the even distribution and appearance of the crypts can be appreciated. However, rectal biopsies do show more crypt architectural distortion than biopsy specimens from other portions of the colon. Thus, I accept a small degree of architectural irregularity in biopsies from this site, and I do not necessarily take this as evidence of a chronic colitis. Finally, there are normally lymphocytes found in the surface epithelium of the colorectal mucosa. There are approximately five lymphocytes per 100 epithelial cells in the colon.2,3 However, there are normally more surface epithelial lymphocytes found in the cecum/right colon than in the remainder of the colon, although these data are not particularly well known. In addition, one must be careful in counting surface epithelial lymphocytes in epithelium overlying lymphoid follicles, given the increased numbers of intraepithelial lymphocytes normally found in this location.4 Given all of the above caveats, it is very important to know the precise site of a colorectal biopsy specimen in order to determine if the changes present are normal for site or pathologic. This has become increasingly more difficult as gastroenterologists have tended to group biopsy specimens from different sites in the colon into one specimen container. The evaluation of patterns of injury in colorectal biopsy specimens is best performed at low magnification. In fact, virtually all of the important information that one can gather from such specimens can be obtained from low magnification. For example, I always evaluate the cellularity of the lamina propria and the architecture of the crypts at low magnification and compare these changes between different fragments within the same specimen container, as well as between specimen containers in patients who have had biopsies obtained from multiple sites. In my mind, the first distinction that the pathologist must make is to determine whether colitis is present or not and, if so, whether the colitis is acute or chronic. The most reliable markers of a chronic colitis are crypt architectural distortion and a basal plasmacytosis. Many colitides, including acute colitides, result in an expanded lamina propria by plasma cells; however, it has been shown that a basal plasmacytosis filling the space between the base of the crypts and the superficial aspect of the muscularis mucosae is probably the best single marker of a chronic colitis.5,6 This has also been found to be true in separating acute colitis from acute-onset inflammatory bowel disease, since these features of chronicity are virtually always present even in the initial biopsy specimen from such patients.7 Paneth's cells are also a useful marker of chronic colitis, although it must be kept in mind that Paneth cells may normally be present in the cecum/right colon. The term "activity" means that there is active epithelial injury that is mediated by either neutrophils or eosinophils. By definition, activity is always present in an acute colitis; activity may or may not be present in a chronic colitis. Thus, when activity is not present in a biopsy showing features of chronic colitis, the colitis can be termed "quiescent." One can also gain an appreciation at low magnification as to whether the colitis is focal or diffuse. In a focal colitis, there are areas of normal colonic mucosa among areas of colitis. In a diffuse colitis, there is no intervening normal mucosa among areas of colitis. The distribution of colitis can be determined in several ways – that is, by inspection of a resected gross specimen, by looking directly at the colonic mucosa through an endoscope, or by histologic evaluation. This discussion will focus on the latter (since after all we are all surgical pathologists). Focal active colitis (FAC) is a commonly encountered pattern of injury seen in colorectal biopsy specimens. It is characterized by a patchy distribution of acute inflammation within the colonic mucosa, and thus, by definition, there are some areas of the specimen that exhibit normal histology. There have been two relatively large studies evaluating the clinical significance of this pattern of injury. Greenson et al. evaluated 49 colorectal biopsy specimens with FAC, excluding patients with an established diagnosis of inflammatory bowel disease.8 Of these 49 patients, 42 were found to have adequate clinical follow-up (median 22 months) to allow for a confident clinical diagnosis. Interestingly, none of these 42 patients developed Crohn's disease within the follow-up period. Nineteen patients were diagnosed with an acute self-limited-like diarrheal illness (an acute, self-limiting, infectious-like diarrheal illness without an identifiable pathogen), six with irritable bowel syndrome, four with antibiotic-associated colitis and two with ischemic colitis. Six asymptomatic patients who underwent screening colonoscopy for neoplasia were also found to have FAC; this was interpreted as an incidental finding. In addition, these authors also evaluated another cohort of 31 patients with resection-proved Crohn's disease, none of which had FAC in their initial colorectal biopsy specimen. Thus, they concluded that FAC does not seem to be a presenting biopsy finding in patients with Crohn's disease and that, in most cases, it is an indicator of an infectious-type colitis with or without an identifiable pathogen. It had been our anecdotal experience at the Cleveland Clinic that FAC was in fact a presenting biopsy finding in a minority of patients with Crohn's disease. Thus, we undertook a similar study as that of Greenson et al in order to determine the clinical significance of FAC.9 We evaluated 31 cases of FAC with a mean follow-up period of 26 months. Clinical diagnoses included infectious colitis (15 cases, 48%), incidental FAC (9 cases, 29%), ischemic colitis (3 cases, 10%) and Crohn's disease (4 cases, 13%). The most striking difference between the findings of Greenson et al and our study is the absence of patients in the formal study in whom Crohn's disease developed. In our study, all biopsy specimens that demonstrated even minimal histologic features suggestive of a chronic colitis were excluded, in contrast to the study by Greenson et al., in which biopsy specimens demonstrating "mild crypt distortion or slight basal plasmacytosis" were included. Thus, our inclusion criteria for FAC were even more stringent than those of Greenson et al, and this cannot be used to explain our higher incidence of Crohn's disease. Furthermore, the relatively high percentage of patients (13%) in our study in whom Crohn's disease developed is not explained by an institutional bias, because we excluded cases referred to our institution for a possible or probable diagnosis of inflammatory bowel disease. None of the patients in our cohort in whom Crohn's disease developed had a biopsy specimen demonstrating an epithelioid granuloma since such cases were excluded from evaluation. It has been estimated that only 5% to 28% of colorectal biopsy specimens from patients with Crohn's disease demonstrate such granulomas.10 Caution must be exercised to avoid confusion of a true epithelioid granuloma with a crypt-rupture granuloma, which occurs as a response to mucin extravasation adjacent to crypt injury of any cause. Although in our study none of the histologic features correlated with a specific clinical diagnostic category for FAC, the presence of an epithelioid granuloma in a background of FAC is strongly suggestive of Crohn's disease, once a crypt-rupture granuloma is excluded and appropriate special stains are performed to exclude an infectious etiology. In summary, FAC most commonly correlates with an infectious etiology. It is not uncommonly seen as an incidental finding in an asymptomatic patient (discussed below), and on occasion, this pattern of injury is a harbinger of Crohn's disease. Aside from the presence of epithelioid granulomas, histologic features are not specific for any particular etiology. As mentioned above, the studies of Greenson et al and Volk et al found that FAC was not uncommonly seen as an incidental finding in asymptomatic patients who were often undergoing colonoscopic screening for neoplasia. There are several possible explanations for this pattern of injury in such patients. Over the past several years, it has become increasingly apparent that oral sodium phosphate bowel preparation may be a cause of FAC.11-13 Oral sodium phosphate is an increasingly used form of bowel preparation for patients undergoing colonoscopic examination. Taken orally, it has been shown to have a good patient tolerance as well as effective bowel cleansing properties. However, several recent studies have shown endoscopic and histologic mucosal alterations that are likely secondary to this bowel preparation. Although these alterations appear to be of no clinical significance, the changes that result from this preparation may mislead gastroenterologists and pathologists to ascribe a disease (such as Crohn's disease) in a patient who does not have any disease at all. Thus, it has been advised that an alternate form of bowel preparation be utilized in patients with suspected inflammatory bowel disease, given the confusion that may result from oral sodium phosphate-associated FAC. Finally, numerous drugs can also be responsible for a FAC pattern of injury. For example, NSAIDs have been associated with a range of colonic abnormalities, including FAC.14 Given the incredibly high frequency of use of such agents, this should always be an etiologic consideration for this pattern of injury, although the pathologist certainly would be unable to arrive at this specific diagnosis. A host of other drugs that are too numerous to name can also cause a FAC pattern of injury. Thus, the gastroenterologist must obtain a complete drug history in these patients and always consider the possibility that the endoscopic and/or histologic alterations found in the colonic mucosa could be drug-induced. Inflammatory Bowel Disease and IBD-related Dysplasia It is well known that patients with ulcerative colitis (UC) are at an increased risk of developing dysplasia and carcinoma, although the exact magnitude of this risk is unknown. It has been estimated that the risk approaches 8% by 20 years and 18% by 30 years.15 Various studies that have investigated this problem have suffered from small numbers of patients, relatively short follow-up intervals and referral center bias.16 Risk factors for colorectal cancer in IBD include a longer duration of colitis, greater extent of colonic involvement, family history of colorectal cancer, primary sclerosing cholangitis, young age of IBD onset and possibly even backwash ileitis.17 More recently, it has been suggested that increased severity of inflammation identified endoscopically and histologically is also associated with increased cancer risk.18,19 Patients with severe symptoms of colitis (fulminant colitis) may have a colectomy before their risk of dysplasia/cancer is significantly increased, and thus patients who are at highest risk are often those with long-standing well-controlled asymptomatic disease. Currently, most gastroenterologists enroll their patients with long- standing extensive colitis into colonoscopic surveillance programs to detect dysplasia or at worst, an early-stage curable carcinoma. Although the benefits of such a program are controversial (discussed below),20 it is important to understand the concept of dysplasia, how to recognize it histologically, and understand the pitfalls of using dysplasia as a marker of increased cancer risk. Dysplasia: A Marker and Precursor of Carcinoma Dysplasia is an unequivocal neoplastic change that is intraepithelial and confined by the basement membrane surrounding the gland within which it arises. The observation that dysplasia is often seen adjacent to an invasive carcinoma is supportive evidence that this lesion is a precursor to the invasive stage.21 An important issue that must be resolved is how often dysplasia is seen distant from an inflammatory bowel disease (IBD)-related carcinoma, as this would be useful in determining its utility as a marker of carcinoma. For example, Connell et al found dysplasia distant to an IBD-related carcinoma in 37 of 50 (74%) proctocolectomy specimens, with an average of 27 blocks evaluated.22 The Histologic Recognition of Dysplasia The recognition of dysplasia can be extremely challenging, particularly in biopsy specimens with active disease, as it can be difficult or impossible to separate reparative from true dysplastic changes. In an attempt to standardize the nomenclature and criteria for dysplasia, the Inflammatory Bowel Disease-Dysplasia Morphology Study Group proposed a three-tiered classification scheme for the evaluation of dysplasia in biopsy specimens from patients with IBD including (1) negative for dysplasia; (2) positive for dysplasia, either low- grade or high-grade, and (3) changes indefinite for dysplasia.21 The low-magnification evaluation of a biopsy specimen is probably the most useful way to recognize the presence or absence of dysplasia, as true dysplastic changes are typically recognizable at low magnification. Most dysplastic lesions in UC resemble sporadic adenomas seen in non-colitic patients and are therefore easily recognizable. Other dysplastic lesions do not resemble adenomas, but more closely resemble the form of dysplasia seen in BE. In low-grade dysplasia, there is nuclear crowding, often with some stratification of nuclei, nuclear pleomorphism and probably most important, nuclear hyperchromasia. The glandular architecture is usually not markedly abnormal. The distinction between low-grade and high-grade dysplasia is made on the degree of cytologic abnormality that is present, with more marked nuclear pleomorphism, hyperchromasia and stratification seen in high-grade dysplasia. In many cases, high-grade dysplastic lesions will be accompanied by marked architectural abnormalities, often with a villiform surface. Certainly in some cases, the distinction between low-grade and high-grade dysplasia is somewhat arbitrary. This distinction has become less important with the trend to perform colectomy for multifocal flat low-grade dysplasia. A chronic colitis that is active can show marked reparative changes that may be difficult or impossible to separate from true dysplasia. Under such circumstances, a diagnosis of indefinite for dysplasia is warranted. Although the Inflammatory Bowel Disease-Dysplasia Morphology Study Group subdivides indefinite for dysplasia into three subcategories (probably inflammatory, probably dysplastic, and unknown), we do not subgroup the indefinite category at the Cleveland Clinic, although we do attempt to offer an explanation in a comment as to why we believe the biopsy is indefinite for dysplasia. Although it is certainly possible to make a diagnosis of dysplasia in an active colitis, one must be cautious in this situation, particularly when there is an absence of cytologic atypia on the surface epithelium. Sporadic adenoma versus IBD-Associated Dysplasia (Polypoid Dysplasia) Given the frequency of sporadic adenomas, it is not surprising that adenomas can also arise in patients with UC. However, as previously mentioned, since most IBD-associated dysplasias resemble sporadic adenomas, there is no way to reliably distinguish these lesions histologically. The risk of synchronous or metachronous invasive carcinoma differs markedly in these two settings, as patients with IBD- associated dysplasia have a much greater risk of having or developing carcinoma. One can be confident that one is dealing with a sporadic adenoma if it occurs in an area not affected by IBD (for example, a right-sided adenoma in a patient with left-sided colitis). However, if the adenoma-like dysplastic lesion arises within an area involved by IBD, then this problem becomes much more difficult. In an attempt to distinguish sporadic adenomas from adenoma-like dysplastic lesions arising in IBD, studies have evaluated whether histologic features, molecular genetic features or the location of the lesion in question relative to the IBD can be used to reliably distinguish between these lesions. In 1993, Schneider and Stolte proposed histologic criteria that would allow for the separation of sporadic adenomas from IBD-related dysplastic lesions.23 However, this separation was based on predetermined, somewhat arbitrary criteria, and follow-up data was not provided. Although the morphologic observations that these authors made may be valid, follow-up information in this cohort would be most useful to validate their observations. In 1998, Torres et al attempted to evaluate this problem in another way.24 They evaluated 89 polypoid adenomatous lesions from 59 patients with IBD, including 51 patients with UC and 8 patients with Crohn's disease, and correlated the morphologic features of these adenoma-like lesions with clinical, endoscopic and follow-up data. Lesions were classified as probable sporadic adenomas if the lesion was not located within an area of histologically proven colitis. A lesion was classified as a probable IBD-associated dysplastic lesion if it developed within an area of colitis and there was associated flat dysplasia or adenocarcinoma that was detected during follow-up evaluation. Finally, lesions were classified as indeterminate type if they arose in an area involved by IBD, but neither flat dysplasia nor adenocarcinoma was detected within the follow-up evaluation. Of the patients with probable sporadic adenomas, follow-up information was available in nine of these patients, and none developed subsequent dysplasia or adenocarcinoma. In contrast, of the 31 patients with either indeterminate polyps or probable IBD-associated dysplastic lesions with follow-up information, 13 patients either had or developed flat dysplastic lesions or adenocarcinoma in adjacent mucosa (including five patients with probable IBD-associated dysplastic lesions who had flat dysplasia in adjacent mucosa at the onset). Using these categories, patients with probable IBD-associated dysplastic lesions were younger (median age 48 years) when compared to those with probable sporadic adenomas (median age 63.5 years), were more likely to have active disease (85% vs. 50%) and were more likely to have a longer duration of disease (median 11 years vs. 5 years). Histologically, patients with probable IBD-associated dysplastic lesions more commonly had lamina propria mononuclear inflammation (60% vs. 16%), lamina propria neutrophilia (60% vs. 36%) and most importantly, were significantly more likely to have an admixture of dysplastic and non-dysplastic crypts on the surface of the polyp (60% vs. 16%). In summary, although several studies have attempted to distinguish between sporadic adenomas and adenoma-like IBD-associated dysplastic lesions, in my opinion, these lesions cannot be distinguished on histologic grounds alone. More recently, there has been a trend to evaluate molecular genetic markers that might be useful in distinguishing between IBD- related dysplastic lesions and sporadic adenomas. Several recent studies suggest that the adenomatous polyposis coli (APC) gene is less significant in the dysplasia-carcinoma sequence of UC when compared to sporadic colorectal carcinomas.25 For example, Fogt et al. found loss of heterozygosity of the APC gene (on 5q) in 4 of 19 (21%) patients with sporadic adenomas, when compared to none of 15 cases of IBD-related high-grade-dysplasia without associated carcinoma.26 Furthermore, loss of heterozygosity of p16 (on 9p21) was found in 7 of 14 (50%) patients with IBD-related high grade dysplasia without associated carcinoma, compared to only 1 of 22 (4.5%) patients with sporadic adenomas. Fogt et al27 found loss of heterozygosity at the von Hippel Lindau gene locus (on 3p25) in 35% of either flat or polypoid dysplastic lesions arising in UC, whereas none of the adenomas tested had loss of heterozygosity at this locus. Studies by Mueller et al28 and Walsh et al29 have evaluated the utility of immunohistochemical markers for p53, bcl-2 and catenin in distinguishing between these lesions, but there is significant overlap in the immunophenotype of these lesions such that in the individual case, they are not particularly useful. In a study by Odze et al,30 there were no significant differences in the frequencies of LOH (loss of heterozygosity) of 3p, APC or p16 in the UC-associated adenoma-like lesions that occurred within areas of colitis compared to those that occurred in areas not involved by colitis. These data suggest that UC-associated flat dysplastic (non-adenoma-like) DALMs have a different molecular genotype than UC-associated adenoma-like lesions as well as non- UC sporadic adenomas. Furthermore, the latter two lesions may be pathogenetically similar, if not identical, to one another.31 More recent data suggest that a large proportion of these patients can be safely treated by endoscopic polypectomy, provided that strict criteria are followed. Two articles were published back-to-back in Gastroenterology supporting the role of endoscopic polypectomy for adenoma-like dysplastic lesions in UC.32,33 Authors from the Brigham and Women's Hospital evaluated 24 patients with UC who had an adenoma-like polypoid dysplastic lesion arising in an area involved by UC.32 In all of these patients, there was histologic documentation of UC in the surrounding non-polypoid mucosa. In addition, all lesions were either sessile or pedunculated polyps that were endoscopically and histologically indistinguishable from sporadic adenomas. All patients had numerous other biopsy specimens taken, and there was no evidence of flat dysplasia or carcinoma in any other area of the colon at the time of discovery of the adenoma-like polypoid lesion. All patients were treated by complete polypectomy and had follow-up surveillance colonoscopy with biopsy (mean follow-up period: 42.4 months). Ten patients did not develop any further dysplastic lesions, whereas 14 patients (58%) subsequently developed dysplastic lesions. In two patients, the lesions arose outside of areas involved by the colitis, and thus it can be assumed that these represent sporadic adenomas. Of the 12 patients who subsequently developed an adenoma- like polypoid lesion in an area involved by IBD, one patient had low-grade dysplasia detected in flat mucosa in a colectomy specimen that was performed six months later, but no patient developed adenocarcinoma within the follow-up period. Similar results were obtained from the Mt. Sinai Medical Center.33 Thus, both studies advocate endoscopic polypectomy for adenoma-like dysplastic lesions arising in areas involved by IBD, provided strict criteria are met. A more recent study from Brigham and Women's Hospital, with a much longer mean follow-up period (mean: 82.1 months), reaffirmed these conclusions.34 The Problem with Dysplasia The first problem encountered by the pathologist, gastroenterologist, and ultimately the patient, is the histologic recognition of dysplasia, and its separation from a reparative process. As noted in several studies, there is some degree of intra- and interobserver variability in recognizing dysplasia, particularly in the separation of negative versus indefinite for dysplasia versus low-grade dysplasia.35,36 The utility of dysplasia as a marker of cancer is limited by sampling. Given the huge surface area of the colon, the fact that many foci of dysplasia are not grossly recognizable by the endoscopist, and the current atmosphere of cost containment, the problem of sampling error is obvious. Mapping studies have shown that dysplasia may be localized to a small area within the colon, and may require extensive sampling beyond what is clinically or economically feasible.37 Furthermore, there are limited data that address the frequency and interval of progression from low-grade to high-grade dysplasia to invasive carcinoma. All of these problems (and more) have led to a search for other markers of carcinoma risk in UC. Ideally, such a marker would be more objective than the recognition of dysplasia and provide a high predictive value for the development of carcinoma, while still being associated with a low rate of synchronous carcinoma. Such a "magic bullet" marker has not yet been found, and thus the morphologic recognition of dysplasia still remains the best marker of increased cancer risk for patients with IBD. 1. Paski S, Wightman R, Robert ME, et al. The importance of recognizing increased cecal inflammation in health and avoiding the misdiagnosis of nonspecific colitis. Am J Gastroenterol 103:2294, 2007.

2. Lazenby AJ. Collagenous and lymphocytic colitis. Semin Diagn Pathol 22:295, 2005.

3. Lamps LW, Lazenby AJ. Colonic epithelial lymphocytosis and lymphocytic colitis: Descriptive histopathology versus distinct clinicopathologic entities. Adv Anat Pathol 7:210, 2000.

4. Goldman H, Antonioli DA. Mucosal biopsy of the rectum, colon, and distal ileum. Hum Pathol 13:981, 1982.

5. Kumar NB, Nostrant TT, Appelman HD. The histopathologic spectrum of acute self- limited colitis (acute infectious-type colitis). Am J Surg Pathol 6:523, 1982.

6. Nostrant TT, Kumar NB, Appelman HD. Histopathology differentiates acute self-limited colitis from ulcerative colitis. Gastroenterology 92:318, 1987.

7. Surawicz CM, Haggitt RC, Husseman M, et al. Mucosal biopsy diagnosis of colitis: Acute self-limited colitis and idiopathic inflammatory bowel disease. Gastroenterology 107:755, 1994.

8. Greenson JK, Stern RA, Carpenter SL, et al. The clinical significance of focal active colitis. Hum Pathol 28:729, 1997.

9. Volk EE, Shapiro BD, Easley KA, et al. The clinical significance of a biopsy-based diagnosis of focal active colitis: A clinicopathologic study of 31 cases. Mod Pathol 11:789, 1998.

10. Jenkins D, Balsitis M, Gallivan S, et al. Guidelines for the initial biopsy diagnosis of suspected chronic idiopathic inflammatory bowel disease. the British Society of Gastroenterology Initiative. J Clin Pathol 50:93, 1997.

11. Driman DK, Preiksaitis HG. Colorectal inflammation and increased cell proliferation associated with oral sodium phosphate bowel preparation solution. Hum Pathol 29:972, 1998.

12. Zwas FR, Cirillo NW, el-Serag HB, et al. Colonic mucosal abnormalities associated with oral sodium phosphate solution. Gastrointest Endosc 43:463, 1996.

13. Wong NA, Penman ID, Campbell S, et al. Microscopic focal cryptitis associated with sodium phosphate bowel preparation. Histopathology 36:476, 2000.

14. Goldstein NS, Cinenza AN. The histopathology of nonsteroidal anti-inflammatory drug- associated colitis. Am J Clin Pathol 110:622, 1998.

15. Eaden JA, Mayberry JF. Colorectal cancer complicating ulcerative colitis: A review. Am J Gastroenterol 95:2710, 2000. 16. Collins RH,Jr, Feldman M, Fordtran JS. Colon cancer, dysplasia, and surveillance in patients with ulcerative colitis. A critical review. N Engl J Med 316:1654, 1987.

17. Itzkowitz SH, Harpaz N. Diagnosis and management of dysplasia in patients with inflammatory bowel diseases. Gastroenterology 126:1634, 2004.

18. Gupta RB, Harpaz N, Itzkowitz S, et al. Histologic inflammation is a risk factor for progression to colorectal neoplasia in ulcerative colitis: a cohort study. Gastroenterology 133:1099, 2007.

19. Rutter M, Saunders B, Wilkinson K, et al. Severity of inflammation is a risk factor for colorectal neoplasia in ulcerative colitis. Gastroenterology 126:451, 2004.

20. Iztkowitz SH, Present DH. Consensus Conference: colorectal cancer screening and surveillance in inflammatory bowel disease. Inflamm Bowel Dis 11:314, 2005.

21. Riddell RH, Goldman H, Ransohoff DF, et al. Dysplasia in inflammatory bowel disease: Standardized classification with provisional clinical applications. Hum Pathol 14:931, 1983.

22. Connell WR, Talbot IC, Harpaz N, et al. Clinicopathological characteristics of colorectal carcinoma complicating ulcerative colitis. Gut 35:1419, 1994.

23. Schneider A, Stolte M. Differential diagnosis of adenomas and dysplastic lesions in patients with ulcerative colitis. Z Gastroenterol 31:653, 1993.

24. Torres C, Antonioli D, Odze RD. Polypoid dysplasia and adenomas in inflammatory bowel disease: A clinical, pathologic, and follow-up study of 89 polyps from 59 patients. Am J Surg Pathol 22:275, 1998.

25. Tarmin L, Yin J, Harpaz N, et al. Adenomatous polyposis coli gene mutations in ulcerative colitis-associated dysplasias and cancers versus sporadic colon neoplasms. Cancer Res 55:2035, 1995.

26. Fogt F, Vortmeyer AO, Goldman H, et al. Comparison of genetic alterations in colonic adenoma and ulcerative colitis-associated dysplasia and carcinoma. Hum Pathol 29:131, 1998.

27. Fogt F, Vortmeyer AO, Stolte M, et al. Loss of heterozygosity of the von Hippel Lindau gene locus in polypoid dysplasia but not flat dysplasia in ulcerative colitis or sporadic adenomas. Hum Pathol 29:961, 1998.

28. Mueller E, Vieth M, Stolte M, et al. The differentiation of true adenomas from colitis- associated dysplasia in ulcerative colitis: A comparative immunohistochemical study. Hum Pathol 30:898, 1999.

29. Walsh SV, Loda M, Torres CM, et al. P53 and beta catenin expression in chronic ulcerative colitis--associated polypoid dysplasia and sporadic adenomas: An immunohistochemical study. Am J Surg Pathol 23:963, 1999. 30. Odze RD, Brown CA, Hartmann CJ, et al. Genetic alterations in chronic ulcerative colitis-associated adenoma-like DALMs are similar to non-colitic sporadic adenomas. Am J Surg Pathol 24:1209, 2000.

31. Fogt F, Urbanski SJ, Sanders ME, et al. Distinction between dysplasia-associated lesion or mass (DALM) and adenoma in patients with ulcerative colitis. Hum Pathol 31:288, 2000.

32. Engelsgjerd M, Farraye FA, Odze RD. Polypectomy may be adequate treatment for adenoma-like dysplastic lesions in chronic ulcerative colitis. Gastroenterology 117:1288, 1999.

33. Rubin PH, Friedman S, Harpaz N, et al. Colonoscopic polypectomy in chronic colitis: Conservative management after endoscopic resection of dysplastic polyps. Gastroenterology 117:1295, 1999.

34. Odze RD, Farraye FA, Hecht JL, et al. Long-term follow-up after polypectomy treatment for adenoma-like dysplastic lesions in ulcerative colitis. Clin Gastroenterol Hepatol 2:534, 2004.

35. Melville DM, Jass JR, Morson BC, et al. Observer study of the grading of dysplasia in ulcerative colitis: Comparison with clinical outcome. Hum Pathol 20:1008, 1989.

36. Dixon MF, Brown LJ, Gilmour HM, et al. Observer variation in the assessment of dysplasia in ulcerative colitis. Histopathology 13:385, 1988.

37. Levine DS, Rabinovitch PS, Haggitt RC, et al. Distribution of aneuploid cell populations in ulcerative colitis with dysplasia or cancer. Gastroenterology 101:1198, 1991.

Pleomorphic Cutaneous Spindle Cell Tumors

Steven D. Billings, MD

Co-Section Head, Section of Dermatopathology Anatomic Pathology, Cleveland Clinic OH, USA Associate Professor in the Cleveland Clinic Lerner College of Medicine [email protected]

- VII -

Differential Diagnosis Most pleomorphic cutaneous spindle cell tumors occur in adults, and many involve sun-damaged skin. Thus, the typical differential diagnosis for these lesions is actually quite limited. Far and away the most common tumors encountered in this setting are sarcomatoid squamous cell carcinoma, desmoplastic/spindle cell melanoma and atypical fibroxanthoma/ superficial pleomorphic dermal sarcoma (malignant fibrous histiocytoma). Less common tumors that also need to be included in the differential diagnosis include leiomyosarcoma, spindle cell variants of angiosarcoma, and very rarely unusual tumors such as malignant peripheral nerve sheath tumor, rhabdomyosarcoma, pleomorphic liposarcoma, etc.

Immunohistochemistry Immunohistochemistry is critical in the evaluation of the classic differential diagnosis of sarcomatoid squamous cell carcinoma, desmoplastic/spindle cell melanoma, and atypical fibroxanthoma (AFX). There are a variety of potential pitfalls in the interpretation of the immunostains in this setting. Fortunately, the skin affords a great number of internal positive and negative controls that aid in the interpretation (Table 1). Internal positive and negative controls completely “trump” external controls, whether those external controls are placed on the same glass slide or not. Inappropriate staining of negative controls usually is the result of inappropriate antibody concentration and/ or excessive epitope retrieval.

Table 1: Commonly evaluated antigens and their normal positive internal controls in the skin

Antigen Normal Positive Internal Control in Skin

Pan-cytokeratin and low molecular weight Suprabasal keratinocytes and adnexae cytokeratins

High molecular weight cytokeratins Basal keratinocytes, myoepithelial cells p63 Nuclear stain that highlights keratinocytes and myoepithelial cells

S100 protein Melanocytes, Langerhans cells, nerves, myoepithelial cells

Melanocytic markers (e.g., HMB-45, Dermal melanocytes (may be HMB-45 Melan-A, MiTF, tyrosinase) negative)

Smooth muscle actins Pilar smooth muscle, vascular smooth muscle

Desmin Pilar smooth muscle, vascular smooth muscle

Endothelial markers (e.g., CD31, CD34, Endothelial cells; FLI-1 and ERG are nuclear FLI-1 protein, ERG, von Willebrand stains protein)

Cytokeratins

Background: The cytokeratins are a family of 20 intermediate filament proteins, the expression of which is largely, but by no means entirely restricted to epithelial cells. There are many different ways to think about the cytokeratins, as individual cytokeratins (e.g., CK 7, CK20), as acidic and basic cytokeratins pairs (e.g., CK 8/18) or as “low” and “high” molecular weight cytokeratins. With regard to the last category, it is important to realize that the division of the cytokeratins into low and high molecular weights is entirely arbitrary. It is more useful to think of the low molecular weight cytokeratins as those of simple epithelia, such as simple ductules, and the high molecular weight cytokeratins as those of complex epithelia, such as urothelium or skin.

For practical purposes, most of us tend to think about cytokeratins in terms of the antibodies we use to identify them. In our experience, the most widely used pan- keratin antibody is the AE1/AE3 cocktail. AE1 recognizes the acidic cytokeratins 10, 14, 15, 16, and 19, whereas AE3 recognizes the basic cytokeratins 1, 2, 3, 4, 5, 6, and 8. A recently developed wide spectrum cytokeratin antibody, OSCAR, appears to be a particularly sensitive marker, and is roughly equivalent to AE1/AE3 as a screening cytokeratin. The most widely used low molecular weight cytokeratin antibodies are Cam 5.2 (CK 8, 18, 19) and 35BH11 (CK 8 and 18). Almost all laboratories used the high molecular weight cytokeratin antibody 34BE12 (CK 1, 5, 10, 14/15), also known as “cytokeratin 903”, or CK5/6 which recognizes keratins 5 and 6. MNF-116, another commonly used antibody, recognizes keratins 5, 6,8, and 17. Many other antibodies exist and one should always carefully read the product insert and any relevant literature before using them in one’s own laboratory.

Cytokeratin expression in normal skin: The skin serves as a superb control tissue for the evaluation of cytokeratin expression, as each epithelial cell type present within the skin expresses unique cytokeratins, as detailed below.

Table 2: Cytokeratin expression in normal skin

Cell Type Cytokeratins Expressed

Suprabasal keratinocytes 1, 10, (11)

Basal keratinocytes 5 and 14

Hair shaft 5, 6, (14), 15, 16, 17

Eccrine ductular cells 7, 8, 17, 18, 19

Eccrine myoepithelial cells 5, 6, 14

Pitfalls in the evaluation of cytokeratin expression in pleomorphic cutaneous spindle cell tumors:

Falsely negative basal keratinocyte staining with broad spectrum cytokeratin antibodies3: Although these antibodies should react with basal keratinocytes (CK 5 and 14), these cells are frequently negative in many laboratories, for unknown reasons. This indicates that these high molecular weight keratins are not being identified. It is critical to realize that sarcomatoid squamous cell carcinomas may express almost exclusively high molecular weight cytokeratins such as CK 5 and CK 14. Therefore, the failure to appreciate negative basal cell staining with broad spectrum cytokeratin antibodies may result in the misdiagnosis of sarcomatoid squamous cell carcinoma.

Extremely focal or absent cytokeratin expression in sarcomatoid squamous cell carcinoma: Sarcomatoid squamous cell carcinomas may show only extremely focal cytokeratin expression, may entirely lose cytokeratin expression in spindled areas, or may express only high molecular weight cytokeratins. The most sensitive markers of sarcomatoid SCC are 34BE12 and CK 5/6. It is also important to recognize that essentially all sarcomatoid SCC will express vimentin (see below). p63

Immunostains for p63 have emerged as an effective adjunct in the diagnosis of sarcomatoids squamous cell carcinoma. It serves as a transcription factor that serves to maintain proliferative capacity of epidermal stem cells. It is positive in basal keratinocytes and myoepithelial cells. It is an effective marker in discriminating sarcomatoid squamous cell carcinoma from AFX. It should be used in conjunction with keratin stains. Immunoreactivity for p63 by itself should e interpreted with caution.

Vimentin

Vimentin is present in almost all embryonic cells and most cells of any lineage re- express vimentin when grown in culture. This latter situation may be analogized to spindled change in neoplasms, inasmuch as essentially any spindled cell will express vimentin. Vimentin is expressed by essentially all sarcomas, all melanomas, all spindled squamous cell carcinomas, all spindled angiosarcomas, etc. It is also expressed in almost all mesotheliomas, most gliomas, and a subset of non-spindled carcinomas. Therefore it has essentially no role in the diagnosis of cutaneous spindle cell tumors.

S100 protein

S100 protein is a 20 kd acidic calcium binding protein, which gets its name from its solubility in 100% ammonium sulfate. S100 protein is composed of 2 subunits, with 3 isotypes- αα (expressed in muscle), αβ (expressed in melanocytes, glia, chondrocytes, skin adnexae), ββ (expressed in Langerhans cells, Schwann cells). The commonly used polyclonal antibodies recognize all of these isotypes.

S100 protein is present in essentially 100% of normal melanocytes and nevi, and in approximately 97-98% of melanomas. S100 protein is thus the single best screening marker for melanocytic neoplasms. It is important to realize that very rare melanomas are S100 protein negative. S100 protein is not, however, completely specific for melanocytic tumors, and can be expressed in smooth muscle tumors (weakly), some carcinomas (especially breast cancer but not squamous cell carcinoma), myoepithelial tumors, peripheral nerve sheath tumors and others. Melanocytic markers

HMB-45

Background: HMB-45 is a monoclonal antibody that identifies a premelanosomal protein, gp100. HMB-45 is often negative in resting melanocytes and nevi, but may be expressed in “reactive” melanocytes. HMB-45 is organelle specific but not lineage specific, meaning that non-melanocytic tumors that contain pre-melanosomes (e.g., Bednar tumor, angiomyolipoma) will be HMB-45 positive. HMB-45 should not be used as a screening marker for melanoma- use S100 protein instead.

Diagnostic uses: Numerous studies have shown the sensitivity of HMB-45 in epithelioid melanomas to be approximately 85%. Sarcomatoid melanomas are less often positive (30-50%) and true desmoplastic melanomas are essentially never HMB- 45 positive in the spindled/desmoplastic zones.

Melan-A

Melan-A, the product of the MART-1 gene, is a 20-22 kd component of the premelanosomal membrane, of unknown function. Unlike HMB-45, Melan-A is expressed in resting melanocytes and nevi. As with HMB-45, Melan-A is organelle specific but not lineage specific. Melan-A should also not be used as a screening marker for melanoma. Essentially the same sensitivity and specificity as HMB-45 .Present in some HMB-45 negative melanomas, and vice versa.

Microphthalmia Transcription Factor (MiTF)

MiTF is a nuclear regulatory protein critical for melanocyte development and for the expression of tyrosinase. It is expressed in essentially all resting melanocytes and nevi. It is the only nuclear marker of melanocytic differentiation. There are both melanocyte-specific and non-specific isoforms of this protein- unfortunately, the commercially available antibodies are not specific for the melanocytic isoforms. MiTF was initially described as highly sensitive and specific marker of melanoma, and is expressed in well over 90% of epithelioid melanomas. Sarcomatoid melanomas are less often positive (40%) and true desmoplastic melanomas are very infrequently positive (<5%). Because the commercially available antibodies are not specific for the melanocytic MiTF isoforms, MiTF expression is not limited to melanomas, and may be seen in leiomyosarcomas, atypical fibroxanthomas, atypical lipomatous neoplasms, and very rare carcinomas. Thus, MiTF is best used for the confirmation of S100 protein-positive, HMB-45/Melan-A/ tyrosinase-negative tumors suspected of being melanomas. MiTF expression in the absence of S100 protein expression is not diagnostic of melanoma. MiTF is also expressed in “myomelanocytic tumors” such as angiomyolipoma and clear cell sugar tumor of the lung.

Tyrosinase

Tyrosinase is an enzyme critical for melanin synthesis and is therefore melanosome- specific, but not melanoma specific. Tyrosinase is expressed in resting melanocytes and nevi. The sensitivity of tyrosinase for melanoma is equal to or slightly better than that of HMB-45 or Melan-A. Its specificity is identical to that of HMB-45. CD68 : CD68 (KP1) is a relatively non-specific marker of lysosomes, rather than a lineage specific histiocytic marker. CD68-positive melanomas may be mistaken for atypical fibroxanthomas, if one is using this marker in an attempt to confirm “fibrohistiocytic differentiation”. CD68 has a very limited role in the diagnosis of cutaneous spindle cell tumors, as it may also be expressed in angiosarcomas, carcinomas, and leiomyosarcomas, among others.

CD163: CD163 is a relatively specific histiocyte marker that oddly enough suffers from being too sensitive. When tumors are stained with CD163, it highlights numerous entrapped histiocytes that are not readily appreciated in routine H&E stains. Therefore, tumors may be incorrectly interpreted as being positive for this marker.

CD10: CD10 is a widely distributed antigen that functions as a cell membrane metallopeptidase. It is expressed in early B-cells, but is widely distributed in other tissue types. Immunostains for CD10 have been touted as a good positive marker for the diagnosis of atypical fibroxanthoma, but this has not stood the test of time as approximately 50% of sarcomatoid squamous cell carcinomas also express this marker.

Markers of endothelial differentiation

CD31 (Platelet endothelial cell adhesion molecule-1)

CD31 is a130 kd transmembrane glycoprotein that is expressed on all endothelial cells, including lymphatic endothelial cells. CD31 is also routinely expressed by tissue macrophages and platelets. CD31 is the single most sensitive and specific marker of endothelial differentiation, and is the single best endothelial marker to have in your own laboratory, if you can have just one. CD31 is expressed by essentially all benign vascular tumors and by over 90% of hemangioendotheliomas and angiosarcomas. CD31 expression in carcinomas is extraordinarily unusual, with only case reports in breast and thyroid carcinomas. Non-endothelial sarcomas, in particular epithelioid sarcoma, are CD31-negative. CD31-positive macrophages within non-endothelial tumors can be confused with positive tumor cells, leading to an erroneous diagnosis of an endothelial tumor.

CD34

CD34, a transmembrane glycoprotein of unknown function, is widely expressed on a number of normal tissues, including hematopoietic stem cells, interstitial cells of Cajal, endothelial cells, and dendritic interstitial cells in the skin and around nerves. CD34 is thus not a specific endothelial marker, but is a useful marker in the appropriate histologic context. CD34 is a highly sensitive marker of benign, borderline and malignant endothelial tumors, being expressed by > 90% of cases. CD34 has also generally been regarded as the most sensitive marker for Kaposi sarcoma, although the KSHV-associated LANA protein is probably the current marker of choice for Kaposi sarcoma. In the differential diagnosis of epithelioid angiosarcomas, it is important to remember that CD34 is expressed by up to 60% of epithelioid sarcomas, but not by carcinomas. It is also expressed by dermatofibrosarcoma protuberans, gastrointestinal stromal tumors, leiomyosarcomas, solitary fibrous tumors, and often in reactive dermal fibroblasts. FLI-1 protein

FLI-1 protein is a member of ETS family of nuclear transcription factors and is expressed in all mature endothelia and during the earliest stages of endothelial differentiation. FLI-1 is best known as the partner of EWS in the Ewing sarcoma- specific t (11;22) (q12; q24) (EWS-FLI-1) fusion gene. It is also routinely expressed by small lymphocytes (probably T-cells). FLI-1 is the only nuclear marker of endothelium. FLI-1 is a highly sensitive marker of endothelial neoplasms. ,It is important not to mistake intratumoral FLI-1 positive endothelial cells and lymphocytes for positive tumor cells.

ERG

In our laboratory ERG has largely replaced FLI-1 as a nuclear marker of endothelial cells. It is also a member of the ETS family of transcription factors. It is consistently expressed in normal endothelial cells and in >95% of vascular neoplasms. It is also a good marker for Ewing sarcomas that harbor the less common EWSR!-ERG translocation.

Von Willebrand Factor (Factor VIII- related protein) vWF is a clotting factor that is theoretically present only in Weibel-Palade body of endothelium and in platelets, and which should be in theory the most specific marker of endothelial differentiation. Unfortunately, vWF is secreted into the serum, and the inevitable high background staining seen with this marker greatly reduces its “real world” utility. There is no real role for vWF stains in the diagnosis of vascular tumors.

Muscle Markers

Smooth muscle actin

The actins are a ubiquitously distributed family of intracellular proteins that may be broadly divided into muscle and non-muscle isoforms. The muscle-specific isoforms can be divided again into smooth, skeletal and cardiac isoforms. From a practical perspective, the single most useful actin antibody is mAb 1A4, which recognizes only smooth muscle isoforms. In the skin, smooth muscle actin is routinely expressed by pilar and vascular smooth muscle, reactive myofibroblasts, a subset of pericytes, and myoepithelial cells. Smooth muscle actin is a highly sensitive marker of smooth muscle and myofibroblastic tumors in the skin.

Desmin

Desmin is an intermediate filament protein that is expressed by muscle cells of all types, as well as submesothelial fibroblasts, a subset of lymph node dendritic cells, and endometrial stromal cells. Desmin expression is uncommon in myofibroblasts. Desmin is best used as a screening marker for tumors of skeletal muscle differentiation, which are extraordinarily rare in the skin. Although desmin expression is almost always present in pilar smooth muscle tumors, it may be absent in smooth muscle tumors of vascular smooth muscle origin (vascular smooth muscle less commonly expresses desmin, for unknown reasons). Smooth muscle actin is a much better screening marker for smooth muscle tumors. When the differential diagnosis includes true smooth muscle and myofibroblastic tumors, strong desmin expression supports true smooth muscle differentiation. Anomalous desmin expression may be seen in melanoma, schwannoma, tenosynovial giant cell tumor, Ewing sarcoma, and angiomatoid (malignant) fibrous histiocytoma.

Atypical fibroxanthoma/ superficial pleomorphic dermal sarcoma

Diagnostic criteria: AFX typically present as a rapidly growing mass in a sun- exposed region of an older adult. Histologically, most AFX are undifferentiated, pleomorphic spindle cells tumors, although relatively monomorphic variants do exist. The diagnosis of AFX should be reserved for small (<1-1.5 cm) lesions that are confined to the dermis and which are completely visualize. Larger lesions with an infiltrative growth pattern should be designated “superficial pleomorphic dermal sarcoma or superficial malignant fibrous histiocytoma”.

Immunohistochemistry: AFX is a histologic and immunohistochemical diagnosis of exclusion. There are no markers or combinations of markers that establish the diagnosis of AFX/MFH. The lesional cells of AFX must be negative for cytokeratins and S100 protein; a small amount of actin expression, indicative of myofibroblastic differentiation is acceptable. AFX commonly contains S100 protein-positive Langerhans cells, CD31-positive endothelial cells and macrophages and Factor XIIIa- positive dendritic cells, and these must be rigorously identified as non-lesional cells. CD68 or CD10 expression does not support or exclude the diagnosis of AFX. CD99 (MIC2 glycoprotein, O13, HBA71) expression is not at all specific for the diagnosis of AFX, and in our opinion has no role in this diagnosis.

Selected References

Desmoplastic (neurotropic) melanoma Clinical and Histopathological Features Desmoplastic melanomas (DM) comprise up to 4% of all melanomas. DM usually presents in older patients, generally in the sixth to seventh decade, but a wide age range may be affected, with rare reported cases in children. In most series men are more commonly affected than women. The most common site of presentation is the head and neck, accounting for up to 75% of all cases. However, a variety of sites may be affected, and in some series cases in women occurred most commonly in the lower extremities. Approximately half to two-thirds of cases are clinically amelanotic and present as flesh-colored papules or nodules, often resulting in a mistaken clinical impression of a benign cutaneous lesion, such as an epidermoid cyst, intradermal nevus or a non-melanoma skin cancers such as basal cell carcinoma. The diagnosis of melanoma is suspected in only about one-third of cases of cases. This innocuous clinical presentation frequently results in delayed diagnoses, and tumors may be present from months to years prior to biopsy. Desmoplastic melanoma can be a difficult diagnosis. A helpful clue to the diagnosis of DM is the presence of an overlying atypical junctional melanocytic proliferation, which can range from an atypical proliferation of single melanocytes to full-blown lentigo maligna-type melanoma in-situ. It is important to realize that the majority of DM will not have a clear-cut overlying in-situ melanoma, although some type of atypical junctional melanocytic proliferation is present in 40 to over 90% of cases. Less frequently DM is associated with superficial spreading melanoma or nodular melanomas. The diagnosis of DM should be considered for any unusual spindle cell proliferation occurring in sun-damaged skin. In our experience, the presence of nodular lymphoid aggregates, often deep in the dermis, is often the first clue that one is dealing with a DM, particularly when the junctional component is subtle or not represented. Another highly characteristic feature of DM is its distinctly “packeted” growth pattern, with aggregates of hyperchromatic spindled cells separated by dense, hyalinized collagen. In most cases, the spindled cells of DM will be large and hyperchromatic (put not pleomorphic), although some cases may consist of deceptively bland cells with wavy or buckled nuclei, simulating neurofibroma. Fortunately, most cases have at least moderate nuclear atypia with enlarged vesicular to hyperchromatic nuclei and prominent nucleoli. Melanin pigment is almost never present within the lesional cells, in contrast to blue nevus, a differential diagnostic consideration in some cases. Neurotropism, characterized by perineural or intraneural involvement by tumor cells, is present in approximately 1/3 of cases but is by no means required for diagnosis. Neurotropism is increasingly seen with deeper tumors, not surprisingly, considering that larger nerve twigs are generally present in the deeper reticular dermis. When neurotropism is present, the tumor may be referred to as a “desmoplastic neurotropic melanoma”. Mitotic figures are present in variable numbers. Most DM are locally advanced tumors at the time of diagnosis, with the majority extending to the level of the reticular dermis (Clark’s level IV) or subcutis (Clark’s level V) at the time of diagnosis. Occasional tumors extend into skeletal muscle or the parotid gland. Only a distinct minority are limited to the papillary dermis (Clark’s level II or III). The Breslow level in the majority of the DM ranges from 2.5 to >4mm. Epidermal ulceration is present in up to 25% of cases and evidence of regression may be seen in up to 15%. Immunohistochemistry is invaluable in the diagnosis of DM. Essentially all DM show strong, diffuse immunoreactivity for S-100 protein, often highlighting the very large size of the neoplastic cells and their long cytoplasmic processes. The exact number of S100-negative DM is for obvious reasons extremely difficult to measure, however, such tumors must exist. In contrast, fewer than 10% of DM are positive for any of the more specific markers of melanocytic differentiation, including HMB45, Melan A, tyrosinase and microphthalmia transcription factor (MiTF). DM may also express a variety of putative “nerve sheath” markers, including p75 neurotrophin receptor, CD57 and PGP 9.5. As discussed below, this may result in confusion with malignant peripheral nerve sheath tumors. Expression of p75 neurotrophin receptor may have a possible role in the neurotropism seen in DM. Many DM also show considerable smooth muscle actin positivity, reflective of a large number of intimately associated myofibroblasts; this apparent actin expression may result in confusion with a variety of myofibroblastic and myoid tumors. Focal anomalous cytokeratin and/ or desmin expression may also be seen, potentially resulting in misdiagnosis as a spindled carcinoma or myoid tumor, respectively. Differential Diagnosis: The differential diagnosis of DM is broad and includes a wide range of both benign and malignant entities. The primary differential diagnosis centers on other malignant spindle cell tumors of the dermis, especially spindle cell squamous cell carcinoma. Like DM both of these tumors occur in elderly patients on sun-exposed skin. In spindle cell squamous cell carcinomas, areas of more conventional squamous cell carcinoma or actinic keratosis are often present. By immunohistochemistry, most spindle cell squamous cell carcinomas will show cytokeratin expression, although this can be extremely focal or even absent. Some spindled squamous cell carcinomas may express only high molecular weight cytokeratin isoforms, identified by antibodies such as 34-beta-E12 or antibodies to cytokeratins 5/6. Spindled squamous cell carcinomas are negative for S-100 protein, in contrast to DM, although they may contain entrapped S100 protein-positive dendritic cells. Atypical fibroxanthoma and DM share a common clinical presentation, as amelanotic lesions on sun-exposed skin of the head and neck of elderly patients. The more common storiform, pleomorphic examples of atypical fibroxanthoma generally show more random nuclear atypia with bizarre multinucleated tumor giant cells than commonly seen in DM. The more spindled examples of atypical fibroxanthoma can bear a striking resemblance to DM. The associated junctional melanocytic proliferation see in most examples of DM can help in the diagnosis. Immunoreactivity for S-100 protein, except for some entrapped dendritic cells, is not a feature of atypical fibroxanthoma. Cutaneous leiomyosarcoma can sometimes be mistaken for DM. Clinically, cutaneous leiomyosarcoma is less common on the head and neck. Microscopically, the fascicles in leiomyosarcoma have a more ordered intersecting pattern and have more elongated, blunt ended nuclei than DM. Immunoreactivity for smooth muscle actin and desmin and the absence of immunoreactivity for S100 protein are helpful adjuncts in this diagnosis. Rare examples of spindle cell angiosarcoma can sometimes be relatively “bloodless” and mimic DM. Immunostains for vascular markers CD31 and/or CD34 are helpful in this differential diagnosis. Malignant peripheral nerve sheath tumor is often mentioned in the differential diagnosis of DM. However, in most cases this is not a practical problem. Examples of malignant peripheral nerve sheath tumor arising in the skin are quite rare and are often (~40%) in the setting of previously documented neurofibromatosis. Malignant peripheral nerve sheath tumors usually are a more cellular tumor without the interposing collagenous stroma or the admixed lymphocytes of DM. A co-existing precursor neurofibroma is often present in cases of cutaneous malignant peripheral nerve sheath tumor. Cutaneous malignant peripheral nerve sheath tumors often have an epithelioid morphology as well. The pattern of S100 protein expression is may be less helpful in distinguishing DM from cutaneous malignant peripheral nerve sheath tumors, as some cutaneous malignant peripheral nerve sheath tumors show strong S100 protein expression. Benign entities in the differential diagnosis include melanocytic nevi. Desmoplastic nevi and amelanotic blue nevi are the most difficult benign melanocytic tumors to differentiate from DM. Desmoplastic nevi occur on average in a slightly younger population, most frequently in middle-aged patients, and more commonly arise on the trunk. As in DM, desmoplastic nevi may show a junctional melanocytic proliferation, although it should not show nuclear atypia or a predominant single cell growth pattern. Desmoplastic nevi may show some nuclear variability, but should not show the distinct hyperchromatism seen at least focally in DM. Desmoplastic nevi (and blue nevi) may show tracking of cells along the epineurium of cutaneous nerves, which should be rigorously distinguished from the true perineurial invasion seen in DM. Mitotic activity is very sparse to absent in desmoplastic nevi, with a correspondingly much lower KI-67 labeling index than is seen in DM. Finally, desmoplastic nevi generally do not show involvement of the subcutaneous fat, a common finding in DM. Immunohistochemical stains are of limited utility in this differential diagnosis as both typically show strong immunoreactivity for S-100 protein. Desmoplastic nevi are, however, much more likely to express specific melanocytic markers, such as HMB45. Amelanotic blue nevi may also be confused with DM. Amelanotic blue nevi do not have an associated junctional melanocytic proliferation, lack cytologic atypia and do not show mitotic activity. Additionally, blue nevi generally show diffuse immunoreactivity for HMB45, in contrast to DM. In terms of benign fibroblastic entities, DM needs to be differentiated from scars. On the surface this seems are trivial matter. However, this can be somewhat problematic in re-excision specimens. Well formed scars are fairly easy to distinguish from DM because of the horizontally oriented fibroblasts and vertically oriented blood vessels. However, more recent scars are more cellular and less organized and may have mild cellular atypia, potentially causing confusion with recurrent or residual DM. In the setting of recurrent or residual DM the volume of tumor cells may be relatively low and more intimately admixed with bland fibroblastic cells of the scar. Therefore careful examination may be necessary to recognize the presence of occult residual tumor. It has been suggested that S-100 protein immunostains may be helpful in this setting to highlight the tumor cells hiding in the background of a scar. Unfortunately, S-100 positive spindle cells are a normal constituent of dermal scars, seen in both recent and well organized scars. These S-100 positive cells most likely represent entrapped dendritic cells such as Langerhans cells. Because CD1a expression on Langerhans cells is lost during migration through the dermis, this is not helpful for discriminating these cells from the tumor cells of DM. Thus, the differential diagnosis of recurrent/residual DM vs. a scar ultimately rests of routine histologic evaluation. In most cases this is not too problematic. There is generally sufficient nuclear atypia to allow distinction between DM and scars. However, in cases which cytologic features are bland, this is a potential diagnostic problem. In such cases, more vertical orientation of tumors, the clustering of tumor cells and the presence of histologic features such as intranuclear cytoplasmic inclusions can be helpful clues to the presence of residual DM. Benign fibrous histiocytomas are conceivably in the differential diagnosis, especially the cellular and atypical variants . Desmoplastic melanoma does not have the peripheral collagen trapping or overlying epidermal hyperplasia seen in fibrous histiocytomas. Fibrous histiocytomas are essentially negative for S100 protein.

Prognosis/Outcome: Our understanding of the clinical behavior of DM continues to evolve. DM was initially believed to be a particularly aggressive variant of melanoma, with 4 of 7 patients in Conley et al’s seminal description dying of metastatic disease. As more experience has been gained with DM, a somewhat different picture has emerged, and it appears less aggressive than was originally thought. In fact, stage I and II DM appear to have a generally better five year survival rate than comparable conventional melanomas. The local recurrence rate of DM does appear to be significantly higher than that of conventional melanoma, presumably due to its growth pattern, and varies from approximately 10-50%, in contrast to the less than 5% local recurrence rate seen with conventional melanoma. Neurotropism is associated with a higher rate of local recurrence, and can also lead to localized neuropathies and as a portal for brain invasion. DM appears to be less likely to involve regional lymph nodes than conventional melanomas, but has a comparable rate of visceral metastasis. Metastatic DM may assume a more epithelioid morphology. Prognostic factors predictive of a worse outcome across various studies include tumor depth (Clark’s level V, Breslow depth >4mm), high mitotic rate (>4/mm2), narrow margins of excision (<1 cm), ulceration, location on the head and neck, increased cellularity and prior misdiagnosis. Surgery is the mainstay of DM therapy. As narrow margins are associated with an increased risk of local recurrence, a minimum of 1 cm margins are required and when possible 2 cm margins may be preferable. Adjuvant radiation therapy may play a role in managing recurrent lesion or inoperable tumors.

Differential diagnosis of desmoplastic melanoma, atypical fibroxanthoma, and sarcomatoid squamous cell carcinoma

DM AFX SSCC

Clinical Head and neck Head and neck Variable

Growth Packeted, Storiform, fascicles Storiform, fascicles Pattern short fascicles

Cytology Pleomorphic to Pleomorphic to Spindled spindled spindled

Immuno S100+, CK- S100-, CK-, SMA -/+ S100-, CK+, SMA -/+

Atypical junctional Clues No epidermal melanocytes; AK or SCC component Lymphoid aggregates

Leiomyosarcoma

Most superficial leiomyosarcomas arise in the dermis and are derived from pilar arrector muscles. They typically present in young adults, most often on the extremities and may be painful. Microscopically they are like leiomyosarcomas elsewhere. They are composed of intersecting fascicles of eosinophilic spindled cells with varying degrees of cytologic atypia and mitotic activity. Dermal based leiomyosarcomas have an excellent prognosis. They may recur locally, but essentially never metastasize. This behavior has prompted the reconsideration of the tumor. Some have advocated referring to this group of tumors as “atypical intradermal smooth muscle neoplasm”. The same cannot be said of leiomyosarcomas arising in the subcutis. These are derived from vascular smooth muscle and have a much more aggressive behavior. They frequently metastasize and have a high incidence of mortality. The more aggressive behavior is attributed to its close association with underlying vessels and early access to the blood supply.

References Folpe AL, Gown AM. Immunohistochemistry for Analysis of Soft Tissue Tumors. In: Goldblum JR, Weiss SW. Soft Tissue Tumors, 5th ed. St. Louis: CV Mosby. 2008.

Miettinen, M. Immunohistochemistry of Soft Tissue Tumors. In: Miettinen, M. Diagnostic Soft Tissue Pathology, 1st ed. Philadelphia: Churchill Livingstone. 2003.

Miller K, Goodlad JR, Brenn T. Pleomorphic dermal sarcoma: adverse histologic features predict aggressive behavior and allow distinction from atypical fibroxanthoma. Am J Surg Pathol. 2012 Sep;36(9):1317-26.

Conley J, Lattes R, Orr W. Desmoplastic malignant melanoma (a rare variant of spindle cell melanoma). Cancer 28: 914-936, 1971. Chen LL, Jaimes N, Barker CA, Busam KJ, Marghoob AA. Desmoplastic melanoma: A review. J Am Acad Dermatol. 2012 Dec 22. doi:pii: S0190-9622(12)01177-2.

Mohebati A, Ganly I, Busam KJ, Coit D, Kraus DH, Shah JP, Patel SG. The role of sentinel lymph node biopsy in the management of head and neck desmoplastic melanoma. Ann Surg Oncol. 2012 Dec;19(13):4307-13.

Busam KJ. Desmoplastic melanoma. Clin Lab Med. 2011 Jun;31(2):321-30.

Busam KJ. Cutaneous desmoplastic melanoma. Adv Anat Pathol. 2005 Mar;12(2):92-102. Kaneishi NK. Cockerell CJ. Histologic differentiation of DM from cicatrices. Am J Dermatopathol. 20:128-34, 1998. Harris GR. Shea CR. Horenstein MG. Reed JA. Burchette JL Jr. Prieto VG. Desmoplastic (sclerotic) nevus: an underrecognized entity that resembles dermatofibroma and DM. Am J Surg Pathol. 23:786-94, 1999. Chorny JA. Barr RJ. S100-positive spindle cells in scars: a diagnostic pitfall in the re- excision of DM. Am J Dermatopathol. 24:309-12, 2002. Busam KJ, Mujumdar U, Hummer AJ, Nobrega J, Hawkins WG, Coit DG, Brady MS. Cutaneous DM. Reappraisal of morphologic heterogeneity and prognostic factors. Am J Surg Pathol. 28:1518-25, 2004. Allison KH. Patel RM. Goldblum JR. Rubin BP. Superficial malignant peripheral nerve sheath tumor: a rare and challenging diagnosis. American Journal of Clinical Pathology. 124:685-92, 2005. Thomas C, Somani N, Owen LG, Malone JC, Billings SD. Cutaneous malignant peripheral nerve sheath tumors. J Cutan Pathol. 2009 Aug;36(8):896-900.

Kraft S, Fletcher CD. Atypical intradermal smooth muscle neoplasms: clinicopathologic analysis of 84 cases and a reappraisal of cutaneous "leiomyosarcoma". Am J Surg Pathol. 2011 Apr;35(4):599-607.

Weiss SW. Smooth muscle tumors of soft tissue. Adv Anat Pathol. 2002 Nov;9(6):351-9. Farshid G, Pradhan M, Goldblum J, Weiss SW. Leiomyosarcoma of somatic soft tissues: a tumor of vascular origin with multivariate analysis of outcome in 42 cases. Am J Surg Pathol. 2002 Jan;26(1):14-24.

Cutaneous Soft Tissue Tumor of Uncertain Histogenesis

Steven D. Billings, MD

Co-Section Head, Section of Dermatopathology Anatomic Pathology, Cleveland Clinic OH, USA Associate Professor in the Cleveland Clinic Lerner College of Medicine [email protected]

- VIII -

This lecture will focus on a variety of tumors that for the mos part do not fit into a well defined category, with two exceptions: angiomatoid fibrous histiocytoma, a fibrohistiocytic tumor of intermediate malignancy, and low-grade fibromyxoid sarcoma, a fibroblastic tumor.

Epithelioid sarcoma

Epithelioid sarcoma (ES) usually presents in young adults, most commonly on the distal extremities, but epithelioid sarcoma can occur in any age and almost any location. I have personally seen cases presenting on the face of young children and also in elderly patients. The clinical presentation is often as a nondescript subcutaneous mass. Some lesions present as ulcerated masses.

Epithelioid sarcoma usually presents as nodules of epithelioid cells with densely eosinophilic cytoplasm that resemble histiocytes. The nodules can have a pseudogranulomatous pattern, especially in nodules that have central necrosis. The tumor nuclei are enlarged but generally bland without prominent nucleoli. Occasional cases can have a predominantly or exclusively spindled morphology. This morphology can be extraordinarily deceptive. The spindled cells are often embedded within a hyalinized collagenous stroma. Mitotic figures are usually present but the mitotic rate is usually relatively low.

By immunohistochemistry, ES is positive for cytokeratins and will be positive for antibodies AE1/3, CAM5.2, and CK7. They are negative for CK20, and most are negative for CK5/6. Epithelioid sarcoma is positive for EMA, vimentin, and approximately 50% are positive for CD34. ES is negative for CD31. It should be noted that ES might be Factor XIIIa-positive. More recently, negativity for INI-1 has been demonstrated a useful finding in the diagnosis of ES. Non-neoplastic tissue is essentially always positive for this nuclear marker, as are most tumors in the differential diagnosis.

Differential Diagnosis

The differential diagnosis of ES is varied. In more epithelioid tumors, the differential diagnosis includes granulomatous disease such as infection and sarcoidosis. Other epithelioid tumors such as metastatic carcinoma or epithelioid vascular tumors could be considered in the differential diagnosis.

Epithelioid hemangioendotheliomas can be confused with ES, especially in cases with a more hyalinized stroma. In such cases the key feature of the cord-like arrangement of the tumor cells may be more difficult to recognize. Evidence of intravascular tumor can be helpful, as approximately 50% of epithelioid hemangioendotheliomas may show histologic evidence of arising within a vessel. The presence of vacuolated tumor cells, so-called blister cells, is also helpful. By immunohistochemistry, epithelioid hemangioendotheliomas can show significant overlap. Both may show immunoreactivity for cytokeratins and CD34. Immunoreactivity for CD31 is helpful, as ES does not express this marker. Epithelioid sarcoma-like hemangioendothelioma can bear a striking resemblance to ES. The former tends to have more of a sheet-like growth pattern rather than the

nodular pattern of ES. Epithelioid sarcoma-like hemangioendothelioma is positive for CD31 and ERG and has retained INI-1 expression.

Exclusively or largely spindled ES are often deceptively bland appearing and may contain hyalinized collagen, reminiscent of cellular fibrous histiocytoma. For this reason, we are in the practice of routinely immunostaining all suspected CFH on the extremities with pan-cytokeratin antibodies, particularly in younger patients. ES lacks the circumscribed nature, epidermal hyperplasia, and peripheral collagen trapping. As a word of warning ES can express Factor XIIIa, and keratin markers should always be performed when considering this differential diagnosis.

ES has a high rate of local recurrence, between 50-85% of patients and a metastatic rate of 30-50%, often with late metastases. Favored metastatic sites are the lung and regional lymph nodes. Sentinel lymph node biopsies should be considered in cases of ES.

References:

1. Halling AC. Wollan PC. Pritchard DJ. Vlasak R. Nascimento AG. Epithelioid sarcoma: a clinicopathologic review of 55 cases. Mayo Clinic Proc. 71:636-42, 1996. 2. Billings SD. Hood AF. Epithelioid sarcoma arising on the nose of a child: a case report and review of the literature. J Cutan Pathol. 27:186-90, 2000. 3. Orrock JM. Abbott JJ. Gibson LE. Folpe AL. INI1 and GLUT-1 expression in epithelioid sarcoma and its cutaneous neoplastic and nonneoplastic mimics. American Journal of Dermatopathology. 31(2):152-6, 2009 Apr. 4. Maduekwe UN. Hornicek FJ. Springfield DS. Raskin KA. Harmon DC. Choy E. Rosenberg AE. Nielsen GP. DeLaney TF. Chen YL. Ott MJ. Yoon SS. Role of sentinel lymph node biopsy in the staging of synovial, epithelioid, and clear cell sarcomas. Annals of Surgical Oncology. 16(5):1356-63, 2009 May. 5. Hornick JL. Dal Cin P. Fletcher CD. Loss of INI1 expression is characteristic of both conventional and proximal-type epithelioid sarcoma. American Journal of Surgical Pathology. 33(4):542-50, 2009 Apr. 6. Quante M. Patel NK. Hill S. Merchant W. Courtauld E. Newman P. McKee PH. Epithelioid hemangioendothelioma presenting in the skin: a clinicopathologic study of eight cases. American Journal of Dermatopathology. 20:541-6, 1998.

Cellular neurothekeoma (CNT)

Once considered a nerve sheath tumor, it is now recognized CNT is most likely another tumor in the so-called fibrohistiocytic family of neoplasms, as there has been no convincing demonstration of nerve sheath differentiation. CNT are typically tumors of children and young adults. They usually present as small, asymptomatic cutaneous nodules on the upper extremities or head and neck area.

CNT is a dermal tumor with a distinctive nested to fascicular pattern of growth. Within the tumor nests, the cells may have a subtle, whorled arrangement or be more sheet-like. The tumor cells are predominantly epithelioid in appearance with well-defined cytoplasmic membranes and abundant pale, eosinophilic cytoplasm. Spindled cells comprise a minority of the cellular constituents. The surrounding stroma ranges from hyalinized, to myxoid. The nuclei are round to slightly oval and have a vesicular chromatin pattern, but nuclear hyperchromasia and moderate nuclear atypia are common focal features. Some mitotic figures are usually present, but the

mitotic rate is usually less than 5/10HPF. Rare cases with atypical mitotic figures have been reported. Associated osteoclast-like giant cells are relatively common.

Rarely, CNT may show more worrisome features including large size (>5 cm), a sheet-like growth pattern, a more infiltrative growth pattern, vascular invasion, prominent nuclear atypia and a high mitotic rate (>10mf/10 HPF). These atypical features do not appear to be associated with a more aggressive clinical course. CNT are consistently immunoreactive for NKI/C3 and CD10, with close to 100% of cases expressing this combination of markers. CNT is also positive for S100A6 and MiTF. CNT is positive for PGP9.5, but this marker has little diagnostic use because of non- specificity. It is important to know that CNT may be positive for SMA, CD68 and Factor XIIIa. CNT is negative for S-100 protein, HMB45, Melan-A, cytokeratins, desmin, EMA and CD34.

CNT behaves in a benign fashion with only very rare documented cases of local recurrence. No metastases have been reported.

Differential Diagnosis:

CNT may be confused with myxoid neurothekeoma (MNT) (dermal nerve sheath myxoma). MNT has a similar nested to lobular growth pattern, but with more prominent fibrous septae separating the tumor nodules. Although MNT may show foci with epithelioid cells similar to cellular neurothekeoma, it is predominantly a spindled cell tumor. Myxoid stroma is invariably abundant in this neurothekeoma variant, whereas myxoid stroma is more variable in CNT. MNT also shows strong and consistent immunoreactivity for S-100 protein.

The tumor nodules of CNT can resemble the nodules of histiocytic cells in plexiform fibrohistiocytic tumor (PFT). PFT usually occurs in children and young adults as a dermal or subcutaneous mass of the extremities. PFT has a distinctly plexiform growth pattern and usually has a biphasic appearance of nodules of histiocytic cells often with osteoclasts and fascicles of fibromatosis-like spindled cells. Some cases may be essentially monophasic with predominantly histiocytic nodules or fibromatosis-like fascicles. The morphologic similarity of the histiocytic nodules of PFH to CNT has lead some to suggest a relationship between them, but the differing behavior suggests that the resemblance is merely coincidental. PFH recurs in ~20-30% of cases and rare cases of metastasis have been reported in contrast to the benign behavior of CNT. PFH also do not express MiTF.

CNT arising on the head can be confused with extracranial meningiomas. Meningiomas may also have a nodular growth pattern that may have a subtle whorling arrangement of tumor cells with relatively abundant cytoplasm. Extracranial cutaneous meningiomas typical occur along cranial closure lines. The presence of psammoma bodies may provide a clue to the diagnosis of cutaneous meningioma. Meningiomas are also positive for EMA.

Epithelioid sarcoma may have a nodular growth pattern that could be confused with CNT, but the tumor nodules within the epithelioid sarcoma frequently show central tumor necrosis, a feature not seen in CNT. By immunohistochemistry, epithelioid sarcoma shows strong immunoreactivity for cytokeratins, EMA and

immunoreactivity for CD34 in roughly half of cases. Epithelioid sarcoma is discussed in more detail below.

References

1. Fetsch JF, Laskin WB, Hallman JR, Lupton GP, Miettinen M. Neurothekeoma: an analysis of 178 tumors with detailed immunohistochemical data and long-term patient follow-up information. Am J Surg Pathol. 31: 1103-1114, 2007. 2. Barnhill RL. Mihm MC Jr. Cellular neurothekeoma. A distinctive variant of neurothekeoma mimicking nevomelanocytic tumors. Am J Surg Pathol. 14:113-20, 1990. 3. Husain S. Silvers DN. Halperin AJ. McNutt NS. Histologic spectrum of neurothekeoma and the value of immunoperoxidase staining for S-100 protein in distinguishing it from melanoma. Am J Dermatopathol. 16:496-503, 1994. 4. Busam KJ. Mentzel T. Colpaert C. Barnhill RL. Fletcher CD. Atypical or worrisome features in cellular neurothekeoma: a study of 10 cases. Am J Surg Pathol. 22:1067-72, 1998. 5. Laskin WB. Fetsch JF. Miettinen M. The "neurothekeoma": immunohistochemical analysis distinguishes the true nerve sheath myxoma from its mimics.Human Pathol. 31:1230-41, 2000. 6. Fullen DR. Lowe L. Su LD. Antibody to S100a6 protein is a sensitive immunohistochemical marker for neurothekeoma. J Cutan Pathol. 30:118-22, 2003. 7. Page RN. King R. Mihm MC Jr. Googe PB. Microphthalmia transcription factor and NKI/C3 expression in cellular neurothekeoma. Mod Pathol. 17:230-4, 2004. 8. Campbell LK. Thomas JR. Lamps LW. Smoller BR. Folpe AL. Protein gene product 9.5 (PGP 9.5) is not a specific marker of neural and nerve sheath tumors: an immunohistochemical study of 95 mesenchymal neoplasms. Mod Pathol. 16:963-9, 2003. 9. Enzinger FM, Zhang RY. Plexiform fibrohistiocytic tumor presenting in children and young adults. An analysis of 65 cases. American Journal of Surgical Pathology 1988;12(11):818-26. 10. Remstein ED, Arndt CA, Nascimento AG. Plexiform fibrohistiocytic tumor: clinicopathologic analysis of 22 cases. Am J Surg Pathol 1999;23(6):662-70. 11. Salomao DR, Nascimento AG. Plexiform fibrohistiocytic tumor with systemic metastases: a case report. Am J Surg Pathol 1997;21(4):469-76. 12. Billings SD, Folpe AL. Cutaneous and subcutaneous fibrohistiocytic tumors of intermediate malignancy: an update. Am J Dermatopathol. 2004;26:141-55.

Low-grade fibromyxoid sarcoma Low-grade fibromyxoid sarcoma (LGFMS) is uncommon sarcoma. It was originally described in 1987 by Harry Evans who described two cases of a deceptively bland sarcoma with paradoxically aggressive behavior. This initial description was followed a report of 12 additional cases that seem to confirm this unusually aggressively behavior despite relatively low-grade histologic features. In this report, 7 out of 12 patients developed metastases and 4 died of their disease. Also described in the late 1990s was the hyalinizing spindle cell tumor with giant collagen rosettes. It was noted that this tumor has similar clinical and histologic features to LGFMS and suggested that it was possibly a variant of LGFMS. This relationship was confirmed with a identification of a novel translocation t(7;16) involving effusion of the FUS/CREB3L2 genes shared by both tumors, supporting that the concept that they were ends of a spectrum within the same tumor.

LGFMS usually occurs in young to middle-aged adults. It predominantly presents as a deep soft tissue mass most commonly in the lower extremity, but they may be seen in almost any location including the trunk, shoulder area, inguinal region, buttocks and head and neck. In an initial series of LGFMS occasional cases were described as being located as superficial tumors of the subcutis or dermis. A subsequent review of LGFMS from the AFIP and Indiana University demonstrated at up to 20% of LGFMS

present as superficial neoplasms. In this series, a disproportionate number of the superficial tumors were seen in children. Normally, pediatric cases account for roughly 10-20% of all cases of LGFMS. In this series almost 40% of the superficial tumors were seen in a pediatric population. Similar to their deep counterparts, superficial LGFMS most commonly presents in the lower extremity but a wide variety of locations was reported.

Both the clinical and histologic diagnoses considered in most of the cases of superficial LGFMS were benign entities. The clinical presentation is that of a nonspecific nodule or mass and entities considered in the clinical diagnosis included a benign lymph node node, sebaceous cyst, schwannoma or lipoma. The histologic differential diagnoses included benign fibrous histiocytoma, nodular fasciitis, infantile fibromatosis, myxoma, neurofibroma, fibromatosis, and dermatofibrosarcoma protuberans. Only one submitting diagnosis considered the possibility of a LGFMS, likely in part due to the under recognized presentation of LGFMS as a superficial tumor.

Microscopic Features: LGFMS has subtle somewhat protean histologic features. It is characterized by a proliferation of hyperchromatic spindled to stellate tumor cells embedded in a variably collagenous and myxoid stroma. The tumor usually shows transitions from fibrous to myxoid zones that may be gradual or quite abrupt with distinct myxoid nodules. In the majority of cases, the tumor cells are arranged in a subtle swirling growth pattern. Focal areas with a fascicular growth pattern may be seen in fibrous zones. In the myxoid zones a prominent curvilinear vasculature is often seen. The tumor cells usually demonstrate subtle nuclear atypia. Mitotic figures are usually rare. Collagen rosettes, characterized by variably sized mats of collagen surrounded by oval to somewhat round tumor cells, are present in some cases and may occasionally be prominent. Areas of intermediate grade sarcoma may be focally present in a minority of cases.

Immunohistochemical stains for superficial LGFMS are generally of little use in the diagnosis of LGFMS with one recently described exception. LGFMS is immunoreactive for vimentin and may occasionally show focal immunoreactivity for EMA, CD34 and actin. LGFMS is typically negative for S100 protein, desmin and cytokeratin. Recently, immunoreactivity for MUC4 has been demonstrated to be a very sensitive and specific marker for LGFMS.

FISH utilizing breakaway probes for FUS are a useful adjunct to the diagnosis of LGFMS, positive in at least 70-90% of cases.

Differential Diagnosis:

Nodular Fasciitis Nodular fasciitis may be histologically mistaken for LGFMS. Nodular fasciitis may show alternating fibrous and myxoid zones, but the myxoid zones typically represents cystic breakdown. The tumor cells of nodular fasciitis are arranged in more defined tissue culture-like fascicles rather than the more gently swirling arrangement seen in LGFMS. The tumor cells are hyperchromatic and typically show an open chromatin pattern with small nucleoli and wispy cytoplasm. There is usually associated

hemorrhage. By immunohistochemistry, nodular fasciitis will often express some smooth muscle actin.

Soft Tissue Perineurioma Soft tissue perineurioma may have a similar swirling growth pattern with a variably collagenous and myxoid stroma. However, the tumor cells tend to be smaller with less atypia and hyperchromasia. Soft tissue perineurioma universally express EMA and also express claudin-1 and Glut-1, markers not expressed by LGFMS. Perineuriomas do not have collagen rosettes.

Neuroblastoma-like Schwannoma This variant of schwannoma may be confused with the hyalinizing spindle cell tumor giant collagen rosette variant of LGFMS. It may show similar mat-like collagen surrounded by round to oval tumor cells. However, schwannoma has a thick fibrous capsule and it strongly expresses S100 protein.

Neurofibroma Neurofibroma has random rather than swirling arrangement of spindle cells. Spindle cells typically have comma-shaped nuclei and diffuse express S100 protein.

Cutaneous Myxoma/Superficial Angiomyxoma Cutaneous myxoma is a less cellular and has a more uniform myxoid stromal background in comparison to LGFMS. Myxomas often have acellular myxoid pools. A sparse perivascular infiltrate of neutrophils is frequently seen in cutaneous myxoma, a features not seen in LGFMS.

Fibromatosis Focally, fascicular areas of LGFMS may resemble fibromatosis. This differential diagnosis is more relevant for deep soft tissue tumors. Superficial fibromatosis is generally restricted to hands, feet and the penis. It has a characteristic clinical and histologic presentation in those locations. In selected cases, nuclear immunoreactivity for beta catenin can be helpful in establishing the diagnosis of fibromatosis rather than LGFMS.

Myxoid Dermatofibrosarcoma Protuberans DFSP may show areas with striking similarity to the myxoid zones of LGFMS. However, the tumor cells of myxoid DFSP tend to have a more random arrangement rather than the swirling or fascicular arrangement seen in LGFMS. The vasculature is often more delicate in myxoid DFSP having a plexiform capillary pattern similar to myxoid liposarcoma. Myxoid DFSP invariably shows some areas with the characteristic infiltrative pattern resulting in a honeycomb pattern of tumor cell infiltration around subcutaneous fat. Also, other forms of DFSP are usually seen in association with myxoid DFSP. Conventional DFSP is seen in greater than 60% of cases when adequately sampled. Other forms of DFSP such as giant cell fibroblastoma, Bednar tumor and fibrosarcomatous DFSP is seen in approximately 10% of cases of DFSP. DFSP is diffusely reactive for CD34 and harbors a different cytogenetic abnormality (t1722).

Myxofibrosarcoma Myxofibrosarcoma may show some areas quite similar to LGFMS. Myxofibrosarcoma typically presents in an older patient population. Myxofibrosarcoma often has a more diffusely optically clear myxoid stromal background and is more infiltrative than LGFMS. Myxofibrosarcoma usually shows at least some moderate cytologic atypia and higher mitotic rate. Pseudolipoblasts are also a clue to the diagnosis of myxofibrosarcoma.

References 1. Evans HL. Low-grade fibromyxoid sarcoma. A report of two metastasizing neoplasms having a deceptively benign appearance. American Journal of Clinical Pathology. 88(5):615-9, 1987 Nov. 2. Evans HL. Low-grade fibromyxoid sarcoma. A report of 12 cases. American Journal of Surgical Pathology. 17(6):595-600, 1993 Jun. 3. Goodlad et al. Low grade fibromyxoid sarcoma: clinicopathological analysis of eleven new cases in support of a distinct entity. Histopathology. 26(3):229-37, 1995 Mar. 4. Lane et al. Hyalinizing spindle cell tumor with giant rosettes: a distinctive tumor closely resembling low-grade fibromyxoid sarcoma. American Journal of Surgical Pathology. 21(12):1481-8, 1997 Dec. 5. Folpe et al. Low-grade fibromyxoid sarcoma and hyalinizing spindle cell tumor with giant rosettes: a clinicopathologic study of 73 cases supporting their identity and assessing the impact of high-grade areas. American Journal of Surgical Pathology. 24(10):1353-60, 2000 Oct. 6. Storlazzi et al. Fusion of the FUS and BBF2H7 genes in low grade fibromyxoid sarcoma. Human Molecular Genetics. 12(18):2349-58, 2003 Sep 15. 7. Reid et al. Low-grade fibromyxoid sarcoma and hyalinizing spindle cell tumor with giant rosettes share a common t(7;16)(q34;p11) translocation. American Journal of Surgical Pathology. 27(9):1229-36, 2003 Sep. 8. Panagopoulos et al. The chimeric FUS/CREB3l2 gene is specific for low-grade fibromyxoid sarcoma. Genes, Chromosomes & Cancer. 40(3):218-28, 2004 Jul. 9. Oda et al. Low-grade fibromyxoid sarcoma versus low-grade myxofibrosarcoma in the extremities and trunk. A comparison of clinicopathological and immunohistochemical features. Histopathology. 45(1):29-38, 2004 Jul. 10. Billings, et al. Superficial low-grade fibromyxoid sarcoma (Evans tumor): a clinicopathologic analysis of 19 cases with a unique observation in the pediatric population. American Journal of Surgical Pathology. 29(2):204-10, 2005 Feb. 11. Mertens et al. Clinicopathologic and molecular genetic characterization of low-grade fibromyxoid sarcoma, and cloning of a novel FUS/CREB3L1 fusion gene. Laboratory Investigation. 85(3):408-15, 2005 Mar. 12. Doyle LA, Möller E, Dal Cin P, Fletcher CD, Mertens F, Hornick JL. MUC4 is a highly sensitive and specific marker for low-grade fibromyxoid sarcoma. Am J Surg Pathol. 2011 May;35(5):733-41. 13. Hornick JL. Fletcher CD. Intestinal perineuriomas: clinicopathologic definition of a new anatomic subset in a series of 10 cases. American Journal of Surgical Pathology. 29(7):859- 65, 2005 Jul. 14. Bhattachary, et al. Nuclear beta-catenin expression distinguishes deep fibromatosis from other benign and malignant fibroblastic and myofibroblastic lesions. American Journal of Surgical Pathology. 29(5):653-9, 2005 May. 15. Reimann JD. Fletcher CD. Myxoid dermatofibrosarcoma protuberans: a rare variant analyzed in a series of 23 cases. American Journal of Surgical Pathology. 31:1371-7, 2007. 16. Mentzel T. Scharer L. Kazakov DV. Michal M. Myxoid dermatofibrosarcoma protuberans: clinicopathologic, immunohistochemical, and molecular analysis of eight cases. American Journal of Dermatopathology. 29:443-8, 2007. 17. Calonje E. Guerin D. McCormick D. Fletcher CD. Superficial angiomyxoma: clinicopathologic analysis of a series of distinctive but poorly recognized cutaneous tumors with tendency for recurrence. American Journal of Surgical Pathology. 23:910-7, 1999.

Angiomatoid fibrous histiocytoma

Angiomatoid fibrous histiocytoma (AFH) is a fibrohistiocytic tumor of intermediate malignancy. It presents most commonly as a dermal or subcutaneous nodular, multinodular or cystic mass of the extremities in children and young adults, but may present at any age. It is usually painless but may be associated with systemic symptoms such as fever and night sweats, possibly related to cytokine production by the tumor.

Histologically, AFH is quite distinctive and is characterized by the presence of a dense fibrous pseudocapsule and a surrounding lymphocytic infiltrate that may also have plasma cells and germinal center formation. In the central portion of the tumor, there is a proliferation of histiocytoid or spindled tumor cells that may be arranged in a variety of growth patterns, including sheets, meningioma-like whorls, and short-fascicles. Frequently there are blood-filled cystic spaces that are lined by flattened tumor cells rather than true endothelium. Although the cells of angiomatoid (malignant) fibrous histiocytoma are typically cytologically bland, occasional cases may show striking pleomorphism, which does not appear to have clinical significance. The mitotic rate is typically low. Cases without the large cystic spaces show at least some evidence of hemorrhage, such as intracellular hemosiderin or focal hemorrhage. In some cases this hemorrhage may be extensive and elicit a striking desmoplastic reaction, obscuring the underlying neoplasm.

AFH has a unique immunophenotype, with expression of desmin, epithelial membrane antigen and CD68 in about 50-60% of cases. Cytogenetically, AFH conations t(2;22)EWS-CREB1 (72%), t(12;22)EWS-ATF1 (21%), or t(12;22)FUS- ATF1(7%). FISH to detect the translocation is a very valuable diagnostic tool. It should be noted that the same translocations are seen in clear cell sarcoma of tendons and aponeuroses, emphasizing the need to integrate molecular findings with morphology and immunohistochemistry.

Differential Diagnosis

Due to the fibrous capsule and lymphoid aggregates, AFH is frequently mistaken for metastatic tumor involving a lymph node. The absence of a true subcapsular sinus helps with this differential diagnosis. Furthermore, the population usually affected by AFH is unlikely to have underlying carcinomas.

AFH can be confused with dermatofibromas with areas of hemorrhage, so- called aneurysmal dermatofibromas. Most of the confusion is due to the similarity of nomenclature. Aneurysmal dermatofibromas are circumscribed but lack the fibrous capsule. Other features of conventional dermatofibroma such as collagen trapping and epidermal hyperplasia are helpful in this differential diagnosis.

Because of the hemorrhage and angiomatoid spaces, AFH is frequently confused with a vascular neoplasm. However, AFH is negative for both CD31 and CD34.

There are occasional cases where results of immunohistochemical stains cause erroneous interpretation. As this tumor often presents in children, the possibility of

rhabdomyosarcoma is often considered and resulting immunoreactivity for desmin can result in interpretation of the lesion as AFH. The individual tumor cells of AFH do not really resemble the high-grade, primitive round cells of rhabdomyosarcoma. AFH lacks immunoreactivity for muscle specific markers such as myogenin or Myo- D1.

AFH may locally recur and may rarely develop regional lymph node metastasis. Distant metastasis is not a feature. Infiltrating margins, skeletal muscle involvement and head and neck location increase the risk of local recurrence. Surgical excision is the treatment of choice.

References:

1. Antonescu CR. Dal Cin P. Nafa K. Teot LA. Surti U. Fletcher CD. Ladanyi M. EWSR1- CREB1 is the predominant gene fusion in angiomatoid fibrous histiocytoma. Genes, Chromosomes & Cancer. 46:1051-60, 2007. 2. Weinreb I. Rubin BP. Goldblum JR. Pleomorphic angiomatoid fibrous histiocytoma: a case confirmed by fluorescence in situ hybridization analysis for EWSR1 rearrangement. Journal of Cutaneous Pathology. 35:855-60, 2008. 3. Thway K. Angiomatoid fibrous histiocytoma: a review with recent genetic findings. Archives of Pathology & Laboratory Medicine. 132:273-7, 2008. 4. Tanas MR, Rubin BP, Montgomery EA, Turner SL, Cook JR, Tubbs RR, Billings SD, Goldblum JR. Utility of FISH in the diagnosis of angiomatoid fibrous histiocytoma: a series of 18 cases. Modern Pathology. 23:93-97, 2010.

PEComa

PEComa (perivascular epithelioid cell neoplasm) is an unusual group of neoplasms derived from immunophenotypically unique perivascular epithelioid cells that demonstrate smooth muscle and melanocytic differentiation. Included in this group of tumors are angiomyolipoma (AML), lymphangiomyomatosis (LAM), pulmonary clear cell sugar tumor (CCST) and PEComa. PEComa is the term currently applied to tumors that are not AML, LAM or CCST. PEComa may present in any location including the skin. There is a female predominance with a F;M ratio of 7:1.

Histologically, PEComa is composed of epithelioid cells with abundant clear to lightly eosinophilic cytoplasm arranged in nests or fascicles arranged around thin blood vessels. By immunohistochemistry the tumor cells are positive for HMB45 (96%), SMA (80%), Melan-A (72%), MiTF (50%), desmin (36%), S100 (0-30%). Co-expression of a melanocyte specific marker and SMA is seen in 84% of cases and is very specific. Differentiating PEComa from melanocytic lesions can be very difficult. The lack of a junctional component, the abundant clear cytoplasm and expression of SMA is helpful. Most PEComas lack S100 protein expression in contrast to melanocytic lesions. The histologic appearance can mimic metastatic renal cell carcinoma. Renal cell carcinoma lacks expression of melanocytic markers, and is positive for epithelial markers (e.g. EMA). Most PEComas behave in an indolent fashion. Tumors with two or more of the following features are considered malignant: size > 5 cm, infiltrative growth pattern, high nuclear grade, high cellularity, mitotic activity > 1 MF/50 HPF, necrosis, or vascular invasion. Tumors with one of these features are considered to be of uncertain malignant potential. Tumors with none of these features are considered benign.

Reference 1. Folpe AL, Mentzel T, Lehr HA, et al. Perivascular epithelioid neoplasm of soft tissue and gynecologic origin: A clinicopathologic study of 26 cases and review of the literature. Am J Surg Pathol 2005;29:1558. 2. Walsh SN, Sangüeza OP. PEComas: a review with emphasis on cutaneous lesions. Semin Diagn Pathol. 2009;3:123-30.

Pleomorphic hyalinizing angiectatic tumor of soft parts Pleomorphic hyalinizing angiectatic tumor of soft parts (PHAT) is a rare tumor of intermediate malignancy that often presents as a superficial subcutaneous mass. It usually presents in adult patients, with a female predominance. It is most commonly seen involving the lower extremity but about any body site can be involved.

Microscopically, PHAT has a lobulated appearance, They can be relatively circumscribed but most cases have infiltrative borders. One of the distinctive features of PHAT is the thin-walled ectatic, hyalinized vessels. The tumor cells themselves vary from plump spindled cells to markedly pleomorphic tumor cells that often have a random to sheet-like arrangement but can also be organized in fascicles. Intranuclear cytoplasmic inclusions are a common findings. Intracytoplasmic hemosiderin is also a common finding, especially in tumor cells adjacent to the hyalinized vessels. Despite the striking pleomorphism, the mitotic rate is low, generally <1 MF/50 HPFs. By immunohistochemistry, about 70-80% are positive for CD34. Importantly, they are consistently negative for S100 protein and cytokeratins.

PHATs are considered a tumor of intermediate malignancy. They may recur locally in up to 50% of cases, but metastasis has not been reported. There have been rare recurrent cases that have shown histologic progression to high grade sarcoma.

The differential diagnosis includes schwannoma with ancient change and pleomorphic undifferentiated sarcoma (malignant fibrous histiocytoma). Schwannomas can have a similar vasculature and can have atypical cells with intracytoplasmic inclusions. In distinction, schwannomas have a well formed capsule, usually less atypia and are S100 protein positive. Pleomorphic undifferentiated sarcoma is usually more cellular, has more diffuse atypia, a higher mitotic rate, and lacks the hyalinized vasculature of PHAT.

References 1. Lewin MR, Montgomery EA, Barrett TL. New or unusual dermatopathology tumors: a review. J Cutan Pathol. 2011 Sep;38(9):689-96. 2. Kazakov DV, Pavlovsky M, Mukensnabl P, Michal M. Pleomorphic hyalinizing angiectatic tumor with a sarcomatous component recurring as high-grade myxofibrosarcoma. Pathol Int. 2007 May;57(5):281-4. 3. Capovilla M, Birembaut P. Primary cutaneous myxofibrosarcoma mimicking pleomorphic hyalinizing angiectatic tumor (PHAT): a potential diagnostic pitfall. Am J Dermatopathol. 2006 Jun;28(3):276-7; author reply 277-8. 4. Folpe AL, Weiss SW. Pleomorphic hyalinizing angiectatic tumor: analysis of 41 cases supporting evolution from a distinctive precursor lesion. Am J Surg Pathol. 2004 Nov;28(11):1417-25. 5. Silverman JS, Dana MM. Pleomorphic hyalinizing angiectatic tumor of soft parts: immunohistochemical case study shows cellular composition by CD34+ fibroblasts and factor XIIIa+ dendrophages. J Cutan Pathol. 1997 Jul;24(6):377-83.

6. Smith ME, Fisher C, Weiss SW. Pleomorphic hyalinizing angiectatic tumor of soft parts. A low-grade neoplasm resembling neurilemoma. Am J Surg Pathol. 1996 Jan;20(1):21-9.

Ossifying fibromyxoid tumor

Ossifying fibromyxoid tumor (OFMT) typically presents in the subcutis and a significant proportion in my experience are clinically encountered by dermatologists. It usually presents in middle aged adults as a subcutaneous mass of the extremities, but any location can be affected. The tumor classically has a peripheral shell of lamellar bone, but nonossifying variants account for at least 20% or more of cases. The tumor cells tend to be quite uniform with a characteristic nested to cord-like arrangement in a myxoid to collagenous stroma. Usually the cytologic features are bland. Chondroid areas and calcification can be encountered. The diagnosis primarily rests on recognition of the histologic features, but immunohistochemistry can be helpful in supporting the diagnosis, as about 70% are positive for S100 protein. Desmin immunoreactivity is seen in about 20%. They are negative for cytokeratins and melanocytic markers. Recently a rearrangement of the PHF1 gene on chromosome 6 has been described in OFMT, which will likely result in the ability to use FISH to support the diagnosis in the future.

OFMT is considered a tumor of intermediate malignancy with a risk of local recurrence but without a significant risk of metastasis. That said, there are fully malignant forms with the capacity to metastasize. Tumors with high nuclear grade or high cellularity with a mitotic rate of > 2 MFs/50 HPFs should be regarded as malignant.

Differential Diagnosis

In ossifying tumors the differential diagnosis, the differential diagnosis includes an extraskeletal osteosarcoma and myositis ossificans. The former has more diffuse high grade atypia with intermingled malignant osteoid rather than a peripheral arrangement of bone. It should be pointed out that malignant OFMT may have areas resembling the malignant osteoid of osteosarcoma. The nested to cord-like growth pattern and immunoreactivity for S100 can help discriminate a malignant OFMT from extraskeletal osteosarcoma. Myositis ossificans has a peripheral shell of bone but presents in the skeletal muscle rather than the subcutis is helpful. The growth pattern is also different. The cellular proliferation of myositis ossificans has the same loose fascicular (tissue culture-like) growth pattern of nodular fasciitis.

Nonossifying variants may be confused with myoepithelial tumors and schwannoma. Both tend to have a cord-like growth pattern. Myoepithelial tumors frequently show glandular differentiation (cutaneous mixed tumors) and will be positive for actins and epithelial markers (keratin, p63) in addition to S100 protein. Schwannomas are usually more spindled cells with nuclear palisading and Antoni A and B areas.

References

1. Gebre-Medhin S, Nord KH, Möller E, Mandahl N, Magnusson L, Nilsson J, Jo VY, Vult von Steyern F, Brosjö O, Larsson O, Domanski HA, Sciot R, Debiec-Rychter M, Fletcher CD,

Mertens F. Recurrent Rearrangement of the PHF1 Gene in Ossifying Fibromyxoid Tumors. Am J Pathol. 2012 Sep;181(3):1069-77. 2. Graham RP, Dry S, Li X, Binder S, Bahrami A, Raimondi SC, Dogan A, Chakraborty S, Souchek JJ, Folpe AL. Ossifying fibromyxoid tumor of soft parts: a clinicopathologic, proteomic, and genomic study. Am J Surg Pathol. 2011 Nov;35(11):1615-25. 3. Miettinen M, Finnell V, Fetsch JF. Ossifying fibromyxoid tumor of soft parts--a clinicopathologic and immunohistochemical study of 104 cases with long-term follow-up and a critical review of the literature. Am J Surg Pathol. 2008 Jul;32(7):996-1005. 4. Cibull TL, Billings SD. Cutaneous malignant ossifying fibromyxoid tumor. Am J Dermatopathol. 2007 Apr;29(2):156-9. 5. Soldano AC, Vazquez-Martul E, Romero JA, Prieto VG. Subcutaneous ossifying fibromyxoid tumor. J Cutan Pathol. 2006 Nov;33(11):749-53. 6. Saadat P, Pullarkat S, Kelly L, Vadmal M. Ossifying fibromyxoid tumor of the skin: a report of 2 cases with light microscopic, immunohistochemical, and electron microscopic characterization. J Am Acad Dermatol. 2005 Apr;52(4):644-7. 7. Folpe AL, Weiss SW. Ossifying fibromyxoid tumor of soft parts: a clinicopathologic study of 70 cases with emphasis on atypical and malignant variants. Am J Surg Pathol. 2003 Apr;27(4):421-31. 8. Enzinger FM, Weiss SW, Liang CY. Ossifying fibromyxoid tumor of soft parts. A clinicopathological analysis of 59 cases. Am J Surg Pathol. 1989 Oct;13(10):817-27.